Papers: Technical Proceedings

Technical ProceedingsTH
Papers
6 CUSTODY MEASUREMENT TECHNICAL EXCHANGE MEETING (CMTEM 2011)
November 21-23, 2011 Dammam Office Building Saudi Aramco
THE CHANGING ROLE AND RESPONSIBILITY OF THE FISCAL FLOW COMPUTER

Roger A. Burnison
FocQus (Burnison Engineering LLC)

10635 Shadow Wood Drive, Houston, Texas, USA In cooperation with Mark Godfrey OMNI Flow Computers Inc 12620 West Airport Blvd Sugar Land, Texas, USA
Introduction For many years, the flow computer has been a stand-alone black box receiving signals from meters and other instrumentation in order to compute actual volumes of oil and gas as well as converting the same data to standard conditions Regulations and Standards for the performance, security, uncertainty and storage of data and audit information have been mandated, in some cases by regulators such as the ERCB, Canada and in others by standards bodies such as API, OIML etc In more recent years with the advent of faster, more compact and less expensive hard- and software, companies have been demanding more open access with the result that there is, and has been, a gradual transfer of responsibility for the data produced by the flow computer. Multiple access through supervisory systems is becoming increasingly popular and operators through their Engineers and Technicians will need to develop different skills set to those of the past. Management will need to understand the changing panorama and be prepared for the consequences. The paper traces some of the history of the flow computer, the traditional functions of the same and the requirements for interface between the Operator / Technician, the computer and how additional functionality is altering its role. Flow Computers A Quick History Traditionally a 1970s, flow computer would have been standalone device, based solely on analog electronics, one per meter run counting and totalizing meter pulses or integrating analog signals. Output would be either an analog signal most likely sent to a chart recorder or a numeric indicator. In the 1980s as microprocessors evolved from military use to industrial applications and their complementing external technologies, it was gradually permitted to print hardcopy reports, store limited amounts of historical data and communicate to a supervisory system or a station controller / totalizer. See figure 1.0
Page 1/11
Figure 1.0 The traditional flow computer (1990s) By the 1990s, flow computers had evolved to process signals from multi meter runs, plus prover control and station capability into a single computer with multiple communication ports servicing multiple external systems. o o o o o o o Print multiple reports; hourly, daily, monthly, batch, snapshot, alarms, logs Control Meter Proving Control valve control Block valve control Recalculate batches retrospectively Run diagnostics Archive hourly and daily data for 30 or more days (mandated by API, etc)
By 1990s standards, the ability to store or archive multiple days of raw data in a few Kbytes of memory with the ability to be retrieved at a later date was a huge cost saving innovation. During the 1990s other industrial computing platforms, such as PLCs emerged to threaten the future of flow computers on dedicated hardware platforms. Due to several failures in programming, insufficient knowledge of the implementation of flow measurement standards, poor real time clocks for time period measurement together with the eventual realization that hardware and software must always be developed together, PLC platforms largely failed to deliver their promise in flow computing. See Figure 2.0
Page 2/11
Figure 2.0 Typical Flow Computer application ~ year 2000 to date For the next 20 years, nothing much changed until the widespread deployment of microprocessor based flow meters such as those utilising the principles of Coriolis and Ultrasound and the adoption of Ethernet communications in industrial computing systems. With these meter types, the traditional pulse outputs were modulated by the meters own internal processor on a non-real-time basis. However, the advantage of these meters is that the huge amount of diagnostic data available provides an ability to infer, from analysis of the data, information as to what is occurring with the fluid process within in the meter tube. This triggered the need for flow computers to have considerable amounts of onboard memory and faster processing speeds. Ethernet based communications permitted relatively large amounts of data to be exchanged over shorter periods of time and introduced the flow computer to the world of connectivity, Local area networks, the Internet and most importantly, security issues.
Flow Computer User Demands - 2011 To calculate flow rate and totalize from a wide rage of flow meter types and be expected to be forward and backwards compatible with traditional and future technologies. An Open Programming Ability similar to that seen in the 1990s from the PLC o o o o User controlled software/firmware revision ability User Defined Register lists for ease of connectivity between devices Increased security; User Name and Password log in, SSLetc Remote and unique Access to multiple users o SCADA & DCS (Operations) o Production Accounting o PLIS o Customers o Office Systems, LAN printers etc o Maintenance / Diagnostics Page 3/11
Flow Computer Vendor Challenges - 2011 The electronics Industry is moving to exclusively surface mount components as opposed to Through Hole components. On the face of it, this does not sound such a big deal. However, the reality is that surface mount electronic assemblies are very cost effective for mass market commodity products such as consumer electronics such as DVD players, cellular telephones and so on. Flow Computers, on the other hand, are not produced in the millions or hundreds of thousands per batch, but in significantly smaller quantities of, perhaps, one thousand or less. This puts the manufacturing cost up to small volume production products. A cost that, of course, has to be passed on to the ultimate buyer/user. Additionally, the repair of circuit boards containing mostly surface mount components is extremely difficult tending towards impossible. Therefore, the flow computer circuit boards themselves become like the cell phone, basically disposable, and therefore repair is by replacement. Electronic components become obsolescent at alarming rates. Integrated Circuit (IC) manufacturers do not allow dust to gather on components on shelves. Once they see a declining trend in demand for an IC, they very quickly announce a last-buy opportunity before they terminate production. The days of the venerable Motorola 69000 processor in production for 30 years is over. A ten (10) year production life is now an eternity. Additionally, IC deliveries are now measured in months not weeks. Therefore a flow computer manufacturer has to continually accommodate component changes not all of which are drop-in replacements. Producing the same product, unchanged for 10 or 20 years is a thing of the past. Competition in the market means that a vendor needs to offer latest hardware specs (Specmanship), generally this would include:o o o o o o Faster processors Multiple processors Massive memory capability Improving ADC accuracy over a wide temperature stability operating range Low power consumption Ergonomics (& attractive appearance)
Not all the above have a benefit to the user of a flow computer. The most complex, iterative flow calculations can now implemented in just a few micro-seconds but with measurement standards calling for complete cycle calculations in less than one second, there is little to no benefit to the user. Economic compromise for all the above can mean that a flow computer vendor has to make compromises in his hardware platform and will have to offer one hardware platform that must be used across multiple and often unrelated applications. This results in Control Platforms such as those for PLC, DCS or control specifications being used for measurement. The vendor must make compromises in the hardware performance resulting in compromise in measurement performance often going unnoticed by users!
Open End User Flow Computer Programming Increasingly we have seen movement away from the constraints or protection of vendor supplied and controlled firmware, towards flow computing platforms offering the user an open programming ability. That is, the user takes control of the software or firmware revision history and can implement or change calculations and routines as they require. The attractions from the users perspective is that they are no-longer constrained by what a computer vendor offers and can implement or modify at their desire without cost or recourse back to the original vendor. Effectively, the vendor supplies to the user a hardware platform and a software tool kit to implement whatever is desired.
Page 4/11
Of course, this also comes with some downsides. o The End user is fully accountable for all his actions implemented in the flow computer both the Good and the Bad alike. Maximum flexibility - and liability - are on the end user, not the vendor. Computer vendor can only support his software tools, not necessarily the implementation by users of programming. The end user must therefore maintain very detailed records of all changes made in the same way that a vendor must produce a revision control system together with a history. Skill sets necessary to monitor and maintain an open or flexible system must reside with end user, otherwise there is no cost saving.
o o
As a result, a single (1) FAT type test during a systems lifetime will not suffice. Each time a revision is made to the software, the user must demonstrate to any and all interested parties the validity of numbers processed and generated by the computer. In particular, for offshore installations, this may imply significant additional and hidden costs of ownership of open systems. Finally, security of data must be taken considerably more seriously than current practice. Most flow computers have multi level passwords for data protection. Sadly, in spite of users claims to take security seriously, all too often these passwords are the factory default settings or the equipment TAG number written above the computer. User Name and unique Password can be implement at local levels. Restricting access by MAC address for external data access should be considered for remote access. So accepting that more control will be put into the hands of users, the conclusion to how to have more flexibility with control is that the User will be forced to invest, deploy and rely on more system management tools and infrastructure to ensure system / measurement integrity.
Metering Management Tools Traditional data gathering associated with custody transfer measurement based solely on SCADA or DCS systems will not yield the controls necessary for future systems management. A review of what is necessary to monitor and manage future measurement systems is diagrammatically represented below.
Page 5/11
A suite of applications is needed, combining remote or central storage and reproduction of measurement data along with the ability to control and/or log events, whether automatically or manually. The key to Management Systems are the right people having access to data and other system information, that they need to do their respective jobs; Operators, Technicians, Engineers and Managers alike all play their part. Operators need to be able to implement control actions such as valve control, alarm management, batching, proving and meter factor control. Technicians need to have access to calibration data and task lists, product manuals, maintenance procedures, etc. Engineering needs the ability to review events that can impact the integrity of measurement and plan interventions that minimize the potential for mis-measurement. Additionally, they need to have ready access to data and the necessary tools to make the retrospective corrective adjustments to previously reported data. Management needs an overview and executive summary of the entire measurement operations process. See Figure 3.0
Page 6/11
Figure 3.0 Supervisory and Measurement connection to multiple Flow Computers
Measurement Data Management System A Measurement Data Management system should be accessible by management and operators in order to view data received from the Remote Terminal Units (RTUs) principally the Flow Computers. The Flow Computers provide audit trail data which is directly logged in an Event & Task logbook. The system is also connected to an Equipment Data Base with the capability to generate and store calibration records as well as related documentation. Additionally, the Measurement Data Management System (MDMS) should to incorporate the procedures as held in the Quality Manual. For more detailed information, see addendum 1. With increasing centralization of data and control of the processes involved in oil & gas production many corporate IT departments have established extremely rigid controls around access to the data gathered and stored by the SCADA, DCS, or other supervisory systems. In addition the data requirements of the majority of (non-measurement) specialists in Operations are very different from those required by the Measurement Management Team, Focal Point(s) and Metering Technicians. As a result, there is now a tendency to utilise the increased connectivity of the flow computer to access the data generated separately from the central corporate system. There are now a number of instances where the original supervisory system is becoming overloaded. With a separate secure connection to the flow computer(s) the MDMS can access all of the data, maintain a details log as to who has access, what change(s), if any, are made and when. Auditors are better able to view all of the composite data rather than just the output from the flow computer such as batch or daily reports.
Page 7/11
Additionally, there is an increasing requirement for adequate well monitoring of production, measurement allocation and check on balances across facilities for loss control and other purposes. The move towards separate data access and reporting of measurement data can now and will be done in such a way that the primary function of the flow computer is unaffected. Remote access to the flow computer may include configuration checks as well upload to the flow computer of such changes as may be determined, duly logged and reported. With the correct controls in place the Measurement Management Team are able with direct access to validate data generated, recalculate totals as required and view system logs to ensure timely correction of any errors or mis-measurements which may occur, see Figure 4.0 below
Figure 4.0 Measurement Data Management System
Page 8/11
Conclusions The expectation of the ability of the flow computers to do more that just calculate and totalise fluid flow will continue as technology develops. Users seek less reliance on the flow computer vendor and look towards themselves for centralized support or their chosen local integration partner. Once a system is put into operation, a modern flow computers ability to better control data access to people or systems authorized to do so must be implemented. Management of the vast amounts of data that a modern flow computer can send or receive must be controlled by competent personnel in the right place at the right time with access to the right tools. To minimize Engineer overload, intelligent automation of alarming and data trending allows the minimum of personnel to monitor and successfully manage compliant measurement systems. The move away from vendor supplied closed flow computer firmware to an open structure implies a significant increase in the need for adequate controls by the user. As the responsibility for the access to and wider understanding of the increased features of the latest generations of flow computers a structured Measurement Data Management System will materially assist Operations to maintain control and provide validated data. In some cases the software to provide the data management function may, in future, reside in the Flow Computer.
Guide to Abbreviations (used in text) ADC API DCS ERCB FAT IC LAN OIML PLC PLIS SAT SCADA SSL TAG Analog to Digital Control American Petroleum Institute Distributed (&/or Digital) Control System Energy Resources Conservation Board Factory Acceptance Test Integrated Circuits Local Area Network Organisation Programmable Logic Controller Pipeline Integrity System (Leak Detection) Site Acceptance Test Supervisory Control and Data Acquisition Single socket layer (security methodology) A label uniquely identifying an equipment or instrument or part thereof
Page 9/11
Addendum 1 Measurement Data Management System (MDMS) The MDMS may be accessed with suitable permissions by: o o o o o o o o Management Engineering Operations Technicians Internal Auditors Production Accounting External Auditors Licence Partners
The system may produce Reports such as: o o o o o o o o o o Daily Totals Batch Alarms and Events Configuration(s) Communication Log System Function Log Meter Factor Control Charts Uncertainty for o Analysis o Trending o Statistics o Calibration o Work plan
The RTUs connected to the system may include: o o o o o o o o Flow Computers Tank Gauging Controllers Pump Controls Compressors Valves Well Monitors etc.
The Equipment Data Base should include the information on each piece of equipment, such as the following: o o o o o o Serial Number Maker Model Size Range Calibration requirements such as procedures, ranges, limits, periodicity, last date, next date, etc. as well as the current Certificate Number, if applicable
Ideally the system should store associated documentation such as: o o o o o o o Drawings Manuals Data Sheets Reports Certificates Photographs Calculations, etc.
Page 10/11
The Event and Task Log should cover o o o o Locations, including stores and off-sites Facilities such as Meter Skids, Laboratories, Test Equipment, etc Processes such as Meter Streams, Quality Loops, Provers, Pumps, Well Monitors,etc Instrumentation including: o o o o o o o o Flow Pressure Temperature Sediment and Water Density Chromatograph (gas) Sampler Valves and Actuators
The Measurement Quality Manual incorporates the following, developed and written specifically for the systems being operated: o o o o o o o o o o o Operations Maintenance Calibration Alarm Management Event Recording Sampling Flow Computer Verification Laboratory Processes Emergencies Contingencies Errors and Mis-measurement
NB. there may be more than one manual or parts of the manual for an extended operation.
Page 11/11
Natural Gas Metering With Ultrasound A New Dimension of Metering
Dr. Achim Zajc, RMG by Honeywell, Messtechnik GmbH, Butzbach, Germany
Metering Overview of Natural Gas with Ultrasound
Table of Contents
Abstract ...................................................................................................................................................................................... 4 Introduction ................................................................................................................................................................................ 4 Operating Principle .................................................................................................................................................................... 5 Application Considerations ...................................................................................................................................................... 6 Noise ....................................................................................................................................................................................... 6 Dirt .......................................................................................................................................................................................... 7 Profile Distortion ...................................................................................................................................................................... 8 Installation Requirements ........................................................................................................................................................ 9 Design Choices ........................................................................................................................................................................ 11 In General ............................................................................................................................................................................. 11 Number of Paths .............................................................................................................................................................. 12 Path Orientation ............................................................................................................................................................... 12 Transducer Design ........................................................................................................................................................... 12 Pulse Detection Algorithms .............................................................................................................................................. 12 Meter Bore........................................................................................................................................................................ 13 RMG by Honeywell Makes the Difference............................................................................................................................. 13 Number and Orientation Paths ......................................................................................................................................... 13 Transducer Design ........................................................................................................................................................... 17 Detection Algorithms ........................................................................................................................................................ 17 Meter Bore........................................................................................................................................................................ 18 Diagnostic Output: Meter Validation ...................................................................................................................................... 19 Transducer Gain Level .......................................................................................................................................................... 19 Transducer Signal to Noise Ratio.......................................................................................................................................... 19 Speed of Sound .................................................................................................................................................................... 20 SoS Comparison to Calculated Value .............................................................................................................................. 20 Path Speed of Sound Comparison ................................................................................................................................... 21 Profile Distortion .................................................................................................................................................................... 21 Conditioning Based Monitoring (CBM): The RMG by Honeywell Way .................................................................................. 21 CBM SoS Comparison of each Path ............................................................................................................................. 22 CBM Signal-to-Noise Ratio............................................................................................................................................ 22 CBM Live Comparison of SoS due to AGA 10 [10] ....................................................................................................... 23 CBM Flow Profile Factors .............................................................................................................................................. 24 CBM Swirl Angle ........................................................................................................................................................... 30 CBM Live - RMG Precision Adjustment (Patented [11]) .............................................................................................. 30 Conclusion and Outlook ......................................................................................................................................................... 33 Literature .................................................................................................................................................................................. 34
Table of Figures
Figure 1: Principal schematic of a transit time ultrasonic meter ................................................................................................... 6 Figure 2: Laminar and turbulent flow profiles ............................................................................................................................... 8 Figure 3: End view representations of swirl, asymmetry and cross flow ...................................................................................... 9 Figure 4: Typical installation requirements for a unidirectional operation .................................................................................. 10 Figure 5: Typical installation requirements for a bidirectional operation ................................................................................... 11 Figure 6: Typical ultrasonic pulse (raw signal) of a USZ 08 ....................................................................................................... 13 Figure 7: RMG by Honeywell path orientation ........................................................................................................................... 14 Figure 8: Path configuration of the RMG by Honeywell 6-path ultrasonic flow meter according to Gauss-Chebyshev.............. 15 Figure 9: Vector analysis of the gas velocity in a single level ................................................................................................... 16 Figure 10: Vector analysis of the gas velocity in all three levels to cover the complete flow profile ........................................... 16 Figure 11: Completely metal-encapsulated titanium sensor ...................................................................................................... 17 Figure 12: USZ 08 of RMG by Honeywell installed in a tandem arrangement in a gas metering station for an underground storage facility in Germany (picture displayed with the permission of ENECO-Epe) ................................................................ 18 Figure 13: Meter bore complete open inner diameter, no intruding sensors or reduction of the meter bore .......................... 18 Figure 14: RMGView USZ 08 diagnostics and operation software live page ........................................................................... 21 Figure 15: Signal-to-Noise Ratio (SNR) in relation to the gas velocity ....................................................................................... 23 Figure 16: Live comparison of the SoS theory with SoS measured due to AGA 10 [10] ............................................................ 24 Figure 17: X- and Y-profile factors in a cross path arrangement ................................................................................................ 25 Figure 18: Flow profile factor analysis done separately for the X- and Y-profile factors ............................................................ 25 Figure 19: X-profile factor separated in two sub-profile factors: X1 and X2 profile factors. ........................................................ 26 Figure 20: Y-profile factor separated in two sub-profile factors: Y1 and Y2 profile factors ......................................................... 27 Figure 21: Picture showing how the Y2-profile factor changes if swirl is in the gas flow profile ................................................. 27 Figure 22: Standard flow profile for turbulent flow ..................................................................................................................... 28 Figure 23: The inner surface of the meter or the entire pipe will change over years of operation, due to deposition of dirt or liquid in the button ...................................................................................................................................................................... 28 Figure 24: Demonstration of how the flow profile and X-profile factor will change ..................................................................... 29 Figure 25: Example of a contaminated USM with inlet spool. This contamination could be detected with the X-profile factor analysis. ..................................................................................................................................................................................... 29 Figure 26: Graphical explanation of the swirl angle ................................................................................................................... 30 Figure 27: Pie meters for the three levels showing live swirl angles in the RMGView software package ............................... 30 Figure 28: Direct USM measurement of signal .......................................................................................................................... 32 Figure 29: Reflective measurement (echo measurement) ......................................................................................................... 32 Figure 30: Effect of the RMG by Honeywell precision adjustment ............................................................................................ 33
Abstract
Introductory concepts for the measurement of natural gas using the principle of pulse transit time measurement are discussed in this white paper. Diagnostics typical to multipath transit time meters using ultrasonic pulses are also described, along with multipath arrangements and transducer types. Guidance is provided for the proper application, installation and condition-based monitoring of ultrasonic meters with respect to RMG by Honeywell technology to assure continued measurement integrity.
Introduction
Over the past 20 years, gas ultrasonic meters have transitioned from the engineering lab to wide commercial usage as the primary device-of-choice to measure gas volume for fiscal accounting. Acceptance by gas pipeline companies has occurred during this time due to the devices: Reliability Accuracy Repeatability Capacity Rangeability Low maintenance Adoption of industry standards for fiscal measurement applications
Briefly, the historical development of fluid velocity measurement in closed conduits with sonic pulses was first considered in the 1920s with the discovery that transmission and reception of repetitive sound bursts could be used to describe the location and speed of moving objects; this principle was soon used to build sonar and radar arrays. Attempts were made over the years to apply this principle to measurement of fluid velocity in conduits, but it wasnt until the development of eco nomical high-speed electronics and digital signal processing in the late 1970s that a repeatable instrument with sufficient resolution for gas applications was devised. Over the following decade and a half, the practical challenges of making the technology commercially viable as a flow measurement device were described and addressed through innovation and development that resulted in the production of a gas ultrasonic meter that utilizes: Robust transducers generating repeatable pulses (amplitude and frequency) Multiple paths to average axial velocity components over the cross-section of a closed conduit (i.e., pipe) High-speed electronics complete with an accurate clock to detect, resolve and time transmission/reception of sonic pulses with sufficient time domain resolution Integrated transducer and electronics to permit high pulse transmission rates. Their transit time measurement allows rapid integration of fluid flow velocity so accurately measured values can be reported once per second. Virtually all ultrasonic meters used for fiscal measurement are flow calibrated at meteorologically traceable test labs. Flow tests are conducted at multiple points over the meters operating range to characterize its proof curve. Meter factor(s) are then calculated and applied to correct the meters output to the labs reference standards. An advantage of modern ultrasonic meters is that once a meter is flow calibrated, diagnostic assessments can describe proof (i.e., meter factor shifts due to a fault in the meters operating elements, such as transducers and/or processing electronics) so that re-calibration generally isnt required (although some regulatory authorities mandate re -certification at set intervals, these mandates vary by jurisdiction).
Operating Principle
Knowledge regarding the measurement principle of ultrasonic meters lays a foundation for optimal field application as well as providing the basis for understanding whether the meter continues to accurately and reliably measure gas volume. Multi-path ultrasonic meters are typically used for gas custody transfer to calculate gas flow rate from velocity measurements made over a pipes cross-section. This is accomplished using the following process: Transducer pairs are installed in a meter body and used to make transit time measurements of ultrasonic pulses, which each transducer transmits and receives. Pulses shot in the downstream direction are accelerated, while those shot upstream are decelerated by the gas flow. (At zero flow, transit times in the up and downstream directions are equal.) Velocities are calculated for each transducer pair, or path, from the measured transit time difference between pulses shot in the up- and down-stream directions Multiple path velocities are averaged into the bulk velocity usi ng a weighting scheme that depends on the paths location in the pipe cross-section for which velocity is sampled Bulk velocity is multiplied by the meter bodys cross -sectional area to calculate uncorrected flow rate
Velocity measurements are made along multiple paths using transducer pairs arrayed in a known position in the meter body. Since the absolute digital travel time measurement method is employed (firing pulses in rapid succession in opposite directions across the same flight path in the pipe), fluctuations in pressure, temperature and gas composition do not affect velocity measurement due to the nearly instantaneous sonic pulse emissions by individual transducer pairs. Below are the basic equations [1,2] used for transit time measurement (Equation 1, 2), bulk velocity (Equation 3), speed of sound (Equation 4), flow calculation (Equation 5) and a schematic (Figure 1) of a transducer pairs geometry that puts the vector sums into context. Note that the only thermodynamic term in any of these equations is c, the speed of sound. This is a fluid property, and the only term in these equations that varies with gas composition, pressure and temperature (i.e., fluid density).
Figure 1: Principal schematic of a transit time ultrasonic meter
Solving for velocity, , in equations 1 and 2, and combining their terms, results in the solution of interest: fluid velocity (Equation 3). Note that the density-dependent term, c (speed of sound) cancels and drops out of equation 3. This is possible because fluid density (i.e., composition, pressure and temperature) is assumed to be constant at the time when the up- and downstream pulses are fired (a reliable assumption given that up- and downstream pulses are fired within milliseconds of one another). The mutual effect, and therefore cancellation of c on the up/down pulses, gives rise to the description of the meters operating principle as the absolute digital transit time measurement method. Application of the absolute digital transit time method provides the technique needed to render accurate gas measurement, and its integrity is dependent on: Knowledge of path length (distance traversed by ultrasonic pulses) Confirmation of clock accuracy (accurate transit time measurement) Validation of pipe cross sectional area (accuracy in flow calculation, and transit time measurement) Validation of flow profile (accuracy of calculation of bulk velocity from individual path velocity measurements)
Diagnostic outputs monitored must verify the constancy of the bolded items in these bullets, in whole or part, to provide an assessment of the meters operating condition.
Application Considerations
All gas measurement technologies, including ultrasonic meters, have limitations. It is important for engineers and technicians that use any flow measurement technology to consider the limitations of the primary device proposed for use at a particular meter station prior to installation. Careful consideration before construction and installation can avoid costly re-work should the chosen technology prove less than optimal for the particular measuring stations operating scenario.
Noise
As noted in the operating principle discussion, ultrasonic flow measurement depends on accurate transit time measurement of sonic pulses. Sound is characterized by its tone (frequency) and loudness (amplitude). In the case of an ultrasonic meter, the tone or frequency of the pulses is above the range of human hearing (20 kHz), hence the modifier ultra.
Noise inside the pipe work can interfere with detection of sonic pulses if it is of coincident frequency with the meters transducers, and drowns out the pulse if it is sufficiently high in amplitude. If pulses are drowned out, detection, and therefore pulse transit time measurement, become impossible and flow measurement stops. Designers should always be concerned with the possibility of noise interfering with an ultrasonic meters function, and should avoid installation near a noise source. Thats easy (and obvious) to state, but in practice hard to accomplish since t he most common noise sources are flow and pressure control valves, which are almost always co-located with meters at gas transfer stations. It is also notable that the noise offensive to an ultrasonic meter is inaudible to human hearing, so the valve sets that commonly cause interference are quiet or whisper trim style valves. These valves achieve their inaudible noise characterist ic with trim designs that push the noise out of the range of human hearing, but into the ultrasonic range where UMs operate. Therefore, designers should install ultrasonic meters where they will be least affected by the noise generated by such valves. Most UM manufacturers can provide additional guidance particular to their product. In general: Meter installation: Install ultrasonic metering upstream of regulating devices. Noise reducers: Locate noise attenuating elements between the meter and the noise source (e.g., tees, separators, etc.). Consult the meter manufacturer: They may have transducers of alternate frequency less susceptible to noise interference and/or the knowledge-base with respect to their meters response. This way a proposed installation can be analyzed for possible impact based on the valve type, flow and pressure drop, followed by a recommendation for attenuators that militate against possible interference.
Dirt
Dirt and liquids can impact the performance and accuracy of ultrasonic meters, as they do all other flow measurement technologies. The effects vary depending upon the nature of the technology. For example, dirt gathering on the interior of an orifice tube or plate will have impact on the dimensional characteristics upon which orifice meters rely. In the case of turbine meters, it may cause the meter to run slow due to an increased bearing friction. In ultrasonic meters, it concerns exits with respect to transducer blockage and dimensional integrity (diametral and path length changes). In terms of diametral changes, recall that in an ultrasonic meter, velocity is measured and uncorrected flow is calculated from the product of cross sectional area and the measured velocity. A percent change in area equates to a percent change in calculated flow. Therefore, there is a 1:1 relationship between diameter error and measurement error. Path length changes due to trash build-up on transducer faces also cause measurement errors, but these dirt-induced errors are easily detected using speed of sound comparisons (see Diagnostics below). Dirt build-up in the meter causes larger measurement errors, but is harder to detect than a dirt-induced path length change. Subtle diagnostic indicators can be monitored, but regardless of the suspicion these indicators might flag, it is usually necessary to make a visual inspection and then clean the meter internals to eliminate the diametral reduction. Operators should: Make meter installations with a site assessment of the possibility for liquids/dirt contamination in mind and consider the addition of inlet separators, filters and drains on the meter run to either prevent contamination from occurring, or to provide a mechanism to drain liquids from the meter run. Consider installing the meter run with a mild slope to discourage accumulation of liquids down the length of the meter run and through the measuring section. Liquids will tend to accumulate at the lower end of the meter run, which would also be the ideal location for a drain. Install inspection ports or even tees with uni-bolt closures to allow for visual inspection with a bore-scope and, in the case of inspection tees/caps, to permit for meter run cleaning.
Adopt a regular program of diagnostic and visual inspection to detect build-up and avoid measurement errors due to diametral reduction.
Profile Distortion
As noted in the operating principle discussion, resolution of path velocities into a representative bulk, or average, fluid velocity, is essential to accurate calculation of uncorrected flow at line conditions. As such, it is necessary to ensure the flow profile is consistent with that found during flow calibration and, additionally, that the profile is symmetrical (figure 2) in shape so that the individual path weighting factors applied by the meter manufacturer retain their validity.
Figure 2: Laminar and turbulent flow profiles
Therefore, path velocities should be mapped during flow calibration and subsequent field inspections should compare the asfound flow profile to that documented when the meter factor was established during calibration in the flow lab. Meter manufacturers usually employ different path configurations from one another in the quest to characterize the flow profile and accurately calculate bulk velocity; some designs are more effective in this regard than others, although arguments can be made that high-performance flow conditioners, designed to generate consistent flow profiles, render this distinction moot. Regardless of whether or not a flow conditioner is used, it is still critical that a multi-path meter provide indication when the asfound profile deviates from that expected (i.e., from flow calibration). A good meters path design will enable calculation and reporting of swirl and asymmetry. Profile distortions can occur due to meter run obstructions, debris accumulation or surface roughness changes on the pipe wall, as well as protrusions installed upstream of the meter (e.g., sampling probes or thermo-wells). However, the most common source of profile disturbance is standard pipe work such as tees, elbows and headers. These elements generate swirl, asymmetry or a combination of the two (Figure 3).
Figure 3: End view representations of swirl, asymmetry and cross flow
General practices to ensure a consistent flow profile include: Design meter runs that minimize profile distortions (long runs of straight pipe of the meter), or include elements such as flow conditioners, that normalize them . (Note: if users elect to use flow conditioners, the meter must be calibrated with the flow conditioner installed in the same position in the meter run, relative to the meter, at both the flow lab and field installations.) Select a meter design employing a multi-path design that properly characterizes the flow profile and can report via its diagnostics whether swirl and/or asymmetry are present. Upon initial meter start-up, map the flow profile once again to verify that the translation of the flow calibrated meter system (meter run, meter and flow conditioner if used) to the field is valid; if the profile in the field is different than that measured at the lab, there is a shift in meter factor!
An area of continuing manufacturer research is development of paths designs that permit repeatable disturbance descriptions and measure their coincidental impact on meter factor shifts. With this design and information on meter factor shift, it will then be possible to correct meter output for the disturbance. This is a steep challenge since the degree and type of disturbances varies, and characterizing meter response to these many influences demands a large and statistically reliable body of accurate data.
Installation Requirements
As discussed in the previous chapters, for a correct measurement it is very important that the flow profile isstable and repeatible. Typical upstream piping elements such as bends, headers, T-joints, flow conditioners, filtration equipment, diameter changes (steps, expanders or reducers) and valves will create swirl and asymmetry to the flow profile. As described in ISO 17089-1 [2], asymmetric profiles may require an inlet spool of 50 DN without a flow conditioner, and swirl may require 200 DN straight pipe without a flow conditioner before a fully developed flow profile can be assumed. Obviously, 50 DN or 200 DN straight inlet pipes are not suitable for a standard meter run installation, but the ISO Standard 17089-1(2) and AGA Report No. 9 apply different solutions to making shorter meter runs for custody use . ISO 17089-1 [2] recommends a straight inlet pipe between 30 DN and 50 DN without flow conditioner and an outlet spool of 3 DN as a minimum. With the flow conditioner, the recommendation is to have 10 DN straight pipe between the flow conditioner and the ultrasonic meter. This contrasts to what the AGA 9 Report suggests as a default configuration of a [1]. flow conditioner should be placed between two 10 DN straight pipes, the latter pipe leading towards the ultrasonic meter and then followed by a 5 DN straight outlet spool. Most North American users employ a 5D, flow conditioner and 10D spool upstream of the meter inlet, and 5 D downstream. Some users prefer honed upstream sections to millitate against dirt build-up, but this adds to cost.
10
Unfortunately, ISO 17089-1 and the AGA 9 Report do not clearly state installation conditions or requirements for uni- and bidirectional operation. And what is stated is in contrast to each other and may cause confusion and even prevent more sophisticated technology with the need for smaller straight inlet spools to be used. Many end users follow Standards and Reports to be on the safe side, but could end up losing money (higher investment) and space if they are not following new technology trends. Different approaches between AGA and ISO also exist for the allowable protrusions (for e.g., thermowells) and diameter steps [1, 2]. Both ISO 17089-1 and the AGA 9 Report state that the inlet and outlet spools should be straight and have the same diameter as the ultrasonic meter. The AGA 9 Report says the inner diameter of the inlet and the meter have to be within 1% of each other. ISO 17089-1 states within 1% as preferable, but within 3% as maximum. Figure 4 shows the installation requirements of the USZ 08 from RMG by Honeywell for unidirectional operation. Figure 5 shows the installation requirements of the USZ 08 from RMG by Honeywell for bidirectional operation. Both installation requirements align with the type approvals for custody transfer measurements according PTB and MID [3, 4]. It is recognized that flow conditioners are typically used in the Americas at 10D upstream of the meter and the RMG by Honeywell meter performs beyond the limits of AGA 9. First, the inlet spool requirements are significantly smaller than the ISO and AGA Guidelines requested (10 DN without flow conditioner and 5 DN with flow conditioner) for both operation modes and this is confirmed in the type approval certificate (although the flow perturbation tests due to MID, TRG 13 and OIML 137-1 are fulfilled by the USZ 08 with these much shorter inlet spools) [5, 6 and 7]. This helps to reduce investment expenses and costs for a much smaller installation. Secondly, the diameter set up is allowed to vary from -2% to +5% (7%) and the ultrasonic meter still measures the requested accuracy in the ISO Standard and AGA 9 Report. This is much less strict than the AGA Report or the ISO Standard and provides a great deal of installation flexibility. This test has been done on third-party-approved test rigs and the results are documented in the type approval for custody transfer measurement [3, 4].
Figure 4: Typical installation requirements for a unidirectional operation
11
For the installation of temperature measurement, both the ISO Standard and AGA 9 Report suggest the installation of 2 DN to 5 DN downstream of the ultrasonic meter for unidirectional operation. For bidirectional use the temperature measurement should be installed from the ultrasonic meter flange 3 DN to max 5 DN. In general, the temperature measurement has to be installed in such a way that it is representing the gas temperature and is not affected by the ambient temperature.
Figure 5: Typical installation requirements for a bidirectional operation
Design Choices
In General
In the development of gas ultrasonic meters, manufacturers necessarily make design choices that balance the desire to build the perfect meter with the need to provide an affordable solution that meets customer needs and existing measurement standards (e.g., AGA TMC Report No. 9 [1] and ISO 17089 [2]). Although the absolute digital pulse transit time method is a measuring principle common to all ultrasonic meter manufacturers, the design choices theyve made for their products set each apart: Transducer design Path orientation/geometry Number of paths Pulse detection algorithms
12
Number of Paths
From the previous discussion of profile distortion and the need for profile repeatability, it is evident that the more samples of the velocity stream (that is, more paths in a meter), the better the measurement. However, its also obvious that more paths means higher cost and results in the Signal Processing Unit (SPU) having to digest more data, which might slow down volume output reporting because of a higher calculation overhead. Additionally, in smaller meter sizes, it is not possible to install a large number of transducers due to physical limits on available space.
Path Orientation
Limits on the number of paths and a desire to maximize sampling/characterizing of the flow profile can be stretched (if not overcome) with careful positioning, or orienting, of the transducer paths. Once again, there are trade-offs between getting more slices of the measuring sections cross section versus axial orientation that can measure swirl, cross-flow and asymmetry. Some manufacturers use a path orientation that bounces ultrasonic signals off pipe walls to increase axial sampling, while others use point-to-point transmission to avoid pulse attenuation and warping that can compromise signal detection.
Transducer Design
Ultrasonic transducers are engineered to deliver maximum amplitude at given frequencies while maintaining transmission of a uniform pulse shape that is repeatable so the pulse can be reliably detected. This is a challenge thats been answered in various ways, once again involving trade-offs between benefits, cost and reliability. Some manufacturers offer transducers with exposed elements so the signal isnt attenuated by a cap that would protect the device from foreign object damage and also inhibit dirt build-up. Others cap the elements but pressure balance to achieve better signal integrity.
Pulse Detection Algorithms

An SPU that can consistently detect pulses in a challenging operating environment (i.e., noisy, dirty, etc.) is also a key element in devising a reliable meter. Ultrasonic pulses are affected by noise interference, and sometimes noise other than pulses may be detected by a SPU and erroneously interpreted as an ultrasonic pulse. This can result in large measurement errors. Pulse mis-detection, also known as a cycle jump or peak skip, might also occur, resulting in systemic transit time measurement errors (refer to Figure 6 below for a depiction of a typical ultrasonic wave packet). Individual peaks in the pulse envelope are often used in manufacturer pulse detection algorithms as start/stop points for transit time measurements to achieve the resolution needed for accurate flow measurement. A peak skip is manifest when the SPU selects the wrong peak on which to start or stop transit time measurement, a fault that can be found using the path speed of sound correlation suggested in Chapter Diagnostic Output : Meter Validation below. Therefore, manufacturers have applied various criteria to pulse validation (counting peaks or measuring individual peak amplitude, for example, and selecting a specific peak to use for transit time measurements). Another technique is to sample repetitive pulses and compare them to one another, which is often called stacking. Again, trade-offs drive design choices; stacking can be an excellent way to validate pulses, however, it slows down processing, reduces the number of validated pulses, and as such, may provide statistically small time samples that reduce sensitivity and reliability in these measurements. Sometimes processing may be slow enough to cause data refresh rates that exceed one second.
13
Figure 6: Typical ultrasonic pulse (raw signal) of a USZ 08
Meter Bore
A meter bored to match pipe schedule is generally preferred to maximize the devices rangeability. Some manufacturers have found that a reduced or tapered meter bore can improve the velocity profile at low flow rates to compensate for use of fewer paths.. A reduced bore also jets the flow so the velocity field pulls away from the pipe wall in the measuring section and doesnt generate near-wall turbulence that can impact signal detection with lower amplitude (i.e., lower energy) pulses. A reduced bore may also skirt the need to offer meter bodies in various pipe schedules. Once again trade-offs are involved; in this case, some rangeability is sacrificed, and the opportunity exists for liquid and debris to collect at the low points at the entrance and exit of the meter bores to compensate for low energy transducers.
RMG by Honeywell Makes the Difference

With all these performance trade-offs in mind, RMG by Honeywell has selected the following to optimize its meter design [8, 9]:
Number and orientation paths

Six paths arrayed in an X pattern in three horizontal planes: a central plane, and two geometrically similar planes. This orientation permits measurement of swirl, cross-flow and asymmetry, as well as transparent path velocity weighting per the Gauss-Chebyshev profile model for compressible fluids (Figures 7 and 8). This path design was introduced by RMG by Honeywell at the end of 1998. So this is already more than 13 years ago and there is no reason up to now to change this path arrangement because of a couple of hundred successfully installed ultrasonic meters in the field, and with this design, it is possible to detect or measure the asymmetry, swirl, and cross-flow.
14
Figure 7: RMG by Honeywell path orientation
There is no compensation needed because the flow profile distortion is actually measured. Measurements with crossed paths ensure an optimal analysis of the velocity components v1 to v6 even in the case of asymmetries, swirl, and cross flow. The flow rate Q results from multiplying the weighted mean flow velocity by the pipe diameter (Equation 6). An additional Reynolds
Number correction is not required. Where: Q = Uncorrected flow (acfs or m3/s) A = Pipe diameter (ft or m) wi = Weighting factor = Path velocity (fps or m/s)
RMG by Honeywell has also opted to utilize a point-to-point pulse path to avoid problems with signal attenuation or warping that can occur with bounce path technology. Pulse warping can be reduced by the use of reflectors (flats attached to the interior of the measuring section where pulses are bounced), however, reflectors compromise a full bore design and themselves generate turbulence. Note: previous RMG by Honeywell designs have utilized bounce paths with reflectors for DN 100 (4) and DN 150 (6), but development of smaller transducers has allowed RMG by Honeywell to array 12 on a meter body to DN 100 (4) and DN 150 (6) diameter s so that use of a point-to-point path construct is possible. As such, this path arrangement of six paths is now available from the smallest DN 100 (4) to the largest DN 1000 (40) ultrasonic meter available.
15
As previously discussed, the six paths arrayed in an X pattern in three horizontal planes allow direct measuring of the asymmetry of the flow profile. Figure 9 shows this technique. The figure is separated into three sections: Section 1: Axial flow velocity Section 2: Tangential flow velocity Section 3: Total flow velocity
The velocity in general is a vector. For further discussion, we must analyze the axial flow velocity in more detail. The axial flow velocity is the main flow direction of the gas in the ultrasonic meter (Z direction). The results of the ultrasonic meter measurement are the two blue vectors. The addition of the two blue vectors in a vector parallelogram results in the black vector (Vz) which is the gas velocity at that level in the direction of Z.
Figure 8: Path configuration of the RMG by Honeywell 6-path ultrasonic flow meter according to Gauss-Chebyshev
The same consideration is valid for the tangential flow velocity. Here, the addition the two smaller blue vectors results in the green vector, the gas velocity in the direction of X (Vx), or in other words, the asymmetry of the flow profile. Now we have two resulting vectors: one is Vz for axial flow and another is Vx for tangential flow. By adding both vectors, we get the total flow velocity vector (the red vector in Figure 9 in the total velocity section). The angle between the total flow velocity vector and the Vz vector is the so called swirl angle. In ideal conditions, the angle between both equals zero.
16
Figure 9: Vector analysis of the gas velocity in a single level
Figure 10: Vector analysis of the gas velocity in all three levels to cover the complete flow profile
17
The USZ 08 ultrasonic meter from RMG by Honeywell is designed with three levels of measurement (2 paths on each of 3 horizontal planes), and the straight-forward vector analysis can be done in all three levels (upper plane, center plane and lower plane) as indicated in Figure 10. This path arrangement makes possible the best coverage of the flow profile with a minimum number of paths. Therefore, an additional Reynolds Number correction is not required because the USZ 09 is measuring the flow profile.
Transducer Design
RMG by Honeywell has developed compact, Titanium-encapsulated, high-energy transducers in 120 and 200 kHz models, making the unit resistant to dirt. Alternate frequency designs are available to help customers cope with noisy environments. The high amplitude capacity of the piezo-ceramic sensor permits the use of a dirt-resistant cap (which must still be thin shell Titanium to avoid attenuation) without the need to pressure balance the unit. Figure 11 shows the 120 kHz transducer used in the USZ 08.
Figure 11: Completely metal-encapsulated titanium sensor
RMG by Honeywell transducers are EExd approved for hazardous areas, but are not intrinsically safe. Their detailed rating is Ex II 2G Ex de IIC T5/T6 and the transducers can be used at pressures of up to 250 bar (ANSI 1500). Furthermore, wide measuring ranges (1:100 and above) with correspondingly high flow velocities of more than 40 m/s (131.2 ft/s) are possible with these robust transducers.
Detection Algorithms
RMG by Honeywell utilizes numerous criteria to validate pulses without compromising high firing rates (10 pulses per second). One of the criteria common to many manufacturers, including RMG by Honeywell, is peak identification and quantization in regards to position and amplitude in the pulse envelope. However, use of comparative analysis of pulses, or stacking, has been avoided since it was found a burden for signal processing in challenging environments (i.e., noisy, turbulent, etc.), resulting in either data refresh rates exceeding one second, or a reduction of evaluated samples falling below statistical acceptability. As a result, RMG by Honeywell has implemented additional qualitative analysis to evaluate the pulse envelope and identify ultrasonic pulses, while still maintaining high firing rates. Figure 12 shows an example of an USZ 08 ultrasonic meter installed in a gas station.
18
Figure 12: USZ 08 of RMG by Honeywell installed in a tandem arrangement in a gas metering station for an underground storage facility in Germany (picture displayed with the permission of ENECO-Epe)
Meter Bore
The ISO 17089-1 Standard distinguishes between a full bore and reduced bore ultrasonic meter. A full bore ultrasonic meter has the same inside diameter as the flange diameter. A reduced bore ultrasonic meter has a smaller inside diameter than the flange diameter. ISO 17089-1 also recommends that changes in the inside diameters and protrusion should be avoided in order to minimize disturbances of the velocity profile. For this reason, RMG by Honeywell decided during the market introduction of the RMG USZ 08 to design it as a full bore ultrasonic meter without transducers intruding into the meter pipe (Figure 13).
Figure 13: Meter bore complete open inner diameter, no intruding sensors or reduction of the meter bore
19
The classification of full bore versus reduced bore sounds marginal, but Figure 12 shows why it is not. This is a tandem bidirectional operation of the USZ 08 for a natural gas underground application. If one of the two meters was reduced, it would create a flow profile disturbance for the other meter. So, in this kind of installation, it would not be a good design choice to use a combination of full bore and reduced bore ultrasonic meters. Rather, it is preferable to go ahead with a full bore ultrasonic meter for a very compact and cost-reduced installation. If two reduced bore ultrasonic meters are the design choice then there has to be a straight pipe of 10 DN between the two units. In other words, the installation needs more space and the investment costs are higher compared to the previously discussed solution.
Diagnostic Output: Meter Validation

All commercially viable gas ultrasonic meters offer diagnostic outputs that indicate meter operating condition, up to and including the ability to judge whether or no t volumetric output is accurate. The nature of the meters operating principle helps define these outputs and also their interpretation. As noted, ultrasonic meters depend on transmission and recognition of sonic pulses using precise timing measurements and known geometry (path length and angle) to accurately measure gas velocity. Manufacturers have incorporated signal (pulse) recognition and processing algorithms as well as highly accurate clocks to make timing measurements. Therefore, signal strength, signal-to-noise ratio and clock accuracy are fundamental to accurate and reliable meter performance. General (and rather generic) descriptions of meter diagnostics follow, but detailed descriptions are manufacturer-specific and beyond the scope of this paper.
Transducer Gain Level

Transducer gain level is a measure of the signal strength, or amplitude, at which each transducer is excited by the meter electronics to generate ultrasonic pulses. Gains are automatically adjusted by the meter electronics so that sufficient pulse amplitude is applied to enable pulse detection. Gains vary depending upon fluid density (i.e., flowing pressure, composition and temperature), and pulse reception quality. Transducer gains should be considered as pairs (gain for the A and B side units rationed to one another is the most common method of treatment). If a transducer pair gains ratio breaks pattern with its previous footprint, or with the pattern of other units, it may indicate: Transducer fault Dirt accumulation on a transducer Other issue(s)
Transducer inspection will be required if a fault is detected in its replacement; otherwise, meter cleaning might be needed as a remedial action.
Transducer Signal-to-Noise Ratio

Evaluation of an individual transducers Signal -to-Noise Ratio (SNR) provides feedback on whether noise is impacting meter function. For example, when the SNR falls to 1:1, the signal is overcome by noise and measurement stops. It is important to note that variation in SNR itself is not an indication that the meters accuracy is in question, but rather that pulse recognition, (i.e., detection) is threatened. If pulses cannot be detected, measurement ceases. The SNR for transducers facing a noise source is typically lower than for transducers facing away from the source. There is no remedial action that can be taken in regards to the meter if SNR falls to the point that measurement ceases. The only
20
remedies available are to eliminate the noise source, change that sources frequency , or alter the meters operating frequency by changing transducers. Alternatively, attenuating elements can be installed between the meter and noise source, but this is time-, labor- and cost-intensive, unless these elements are planned for and installed during the construction phase.
Speed of Sound
Speed of Sound (SoS) is a critical and powerful diagnostic tool available in ultrasonic meters from which users can determine if a meters operating performance has shifted. Two tests can be conducted using meter measured values for SoS: An absolute comparison of meter corrected SoS versus that calculated from the gas thermodynamic properties A per path comparison to determine if an outlier on a particular path suggests its path length has changed (path length changes are either due to meter configuration input errors or debris build-up on transducer faces)
SoS Comparison to Calculated Value

Recall Equations 3 and 4 from the operating principle discussion: Simple inspection of these equations reveals that the fluid velocity, v, and speed of sound, c, are both directly dependent on the meters path length and transit time measurements. Essentially, meter pulse transit time measurements and path length data permit calculation of both v and c.
The SoS in natural gas can also be calculated from its fluid properties of composition, pressure and temperature, with calculation standards adopted by the AGA 10 [10] as applicable. Therefore, it is possible to compare the meter measured value of SoS to that calculated with the AGA 10 equations. The measured versus calculated values should agree closely (a limit of +/- 0.25% is typically used, but may need relaxing depending on the quality of compositional data). Should a significant offset between the measured and calculated values be found, it indicates one or more of the following: The meters path length(s) is in error The meters clock has shifted, causing transit time errors (or there is pulse misdetection, which is also an SPU problem) The data used to make the SoS calculation per AGA 10 is incorrect o o Compositional data is in error suggesting a GC issue One or both of the pressure and temperature transducers is incorrect
These conclusions can be made because meter clock/pulse detection (or SPU function), path length and fluid data are the only variables that can cause disagreement between the measured and calculated value of SoS. Furthermore, it can be stated that good correlation of measured and calculated SoS proves that clock/SPU functions and path length are valid, and it can be concluded, therefore, that the meter factor has not changed!
21
Path Speed of Sound Comparison

Per path SoS footprints can be used to evaluate individual path issues related to path length, and possibly transducer function. Should an individual paths as found SoS deviate from the established foot print (once again, the footprint established during flow calibration and/or meter start-up should be used for reference), it can be concluded that this paths function (either path length or pulse detection on one of its transducers) is at issue. While its true that a disparity could also be caused by a clock issue, the same clock is used for timing measurement on all paths, and were there a clock problem, all paths would probably shift in like fashion. Nonetheless, if the clock were at fault, the SoS comparison between meter corrected and AGA 10 calculated values would indicate the fault.
Profile Distortion
Profile distortions can be detected with comparison of a given manufacturers diagnostic output for Swirl Angle and/or Asymmetry using a footprint technique similar to that suggested for Transducer Gains and per path SoS evaluation. Users should review the available diagnostics for each of these quantities for a particular manufacturers product, and consi der the basis of calculation that each provides, since the variation in path geometries between manufacturers means that differences in the quality and sensitivity of these outputs also exists. Due to these sensitivity differences, it is not possible to provide a generic description for alarm treatment of swirl and asymmetry outputs. However, any multi-path meter affords users the opportunity to also footprint the per path velocity patterns for its given geometry, which can then be compared with the as found velocity pattern. As a cautionary note, flow profiles vary with fluid velocity so any as found to footprint comparison needs to be made at roughly the equivalent velocity/Reynolds Number. To face this complication,it is better to select a meter manufacturer that provides ready outputs for swirl and asymmetry. It is recommended that users request manufacturers to specify the measurements and calculations made to generate swirl and asymmetry values to ensure theyre understood. Thus the various path designs/geometries offered by manufacturers necessarily means the treatment, and therefore sensitivity, of these outputs also varies.
Conditioning Based Monitoring (CBM): The RMG by Honeywell Way

One of the advantages of ultrasonic meters, in comparision with all other flow measurement technologies, is the availability of a lot of additional information diagnostics beyond just delivering pulses or signals proportional to the gas volume. All this additional information and diagnostics is mostly handled through separate Windows TM-based software. RMG by Honeywell has two ways of using additional diagnostics; one is the flow computer ERZ 2000 and the other is the Windows TM-based software RMGView (Figure 14).
Figure 14: RMGView USZ 08 diagnostics and operation software live page
22
The RMGView parameterization and diagnostics software can handle the following CBM parameters live in parallel with the standard operational features: Monitoring of the AGC levels Comparison of the Speed of Sound (SoS) of each path Signal quality: o o Signal-to-noise ratio (SNR) in dB Valid samples in %
Comparison of the Speed of Sound (SoS) due to AGA 10: o o Estimated velocity of sound from the composition of the natural gas Measured velocity of sound from the ultrasonic meter
Evaluation of the flow profile: o Comparison of flow profile factors
Monitoring the swirl angle Live - RMG Precision Adjustment
CBM SoS Comparison of Each Path

The RMG by Honeywell path design uses 6 paths, and for each of the paths there is a measured Speed of Sound (SoS). If everything is okay (including the ultrasonic meter and the flow profile), then the ratio of the SoS of the single paths should be equal to one. Perhaps the following example makes this clearer. Here is the single SoS (C1 to C6) listed, and for example, the ratios C1/C2 and C1/C6 are calculated and they are both 1 so the measurement point is in standard conditions. Example: Single SoS per path: C1 = 341.91 m/s; C2 = 341.89 m/s; C3 = 341.93 m/s C4 = 341.77 m/s; C5 = 342.08 m/s; C6 = 342.09 m/s Ratio of SoS of different path: C1/C2 = 1.00006 C1/C6 = 0.99947 etc.
CBM Signal-to-Noise Ratio

Signal to Noise ratio, or SNR, is a parameter indicating the degree to which ambient noise my be present in the pipe and can be used to pinpoint why a meter fails to report data at specific operating conditions. USMs depend on detection of sound pulses, but as noted previously, should the noise in the flowing stream, perhaps caused by a throttled control valve, be of sufficient amplitude in a frequency band corresponding to that of the meter transducers, interference can occur making signal detection impossible. RMG by Honeywells USM operates best (i.e. has optimal signal detection) when the SNR is above 20 db. Below that level, pulses may be rejected until noise overwhelms the meters ability to detect them at all and measu rement output stops. Honeywell is certified to operate for fiscal measurement purposes by PTB with accepted pulses to 40% .This provides
23
significant operating band-width even in the Non-Normal range of the graph above. Mesurement will theoretically continue until the SNR reaches 1, but somewhere between the span of 20 db to 1 db, accepted pulses will likely fall below 40%, putting meter operation outside the PTB certified range. Ambient noise that interferes with a USMs operation is usually generated by a throttled control valve installed in near proximity to the meter. A noise evaluation should be jointly conducted with the operator during the stations design phase if they propose to install such a valve, particularly a quiet trim valve, near a me ter so that measures to isolate the meter from the noise source can be taken in during pipe work design (placing attenuating elements between valve and meter such as filters or blind tees). A finding of noise interference after station build has occurred usually means an expensive piping re-work, but alternate frequency transducers are offered by Honeywell that might provide relief by moving the meters operating frequency away from that of the offending noise, or if the operator can change the valves tri m cage or operating profile, these too may be solutions.
CBM Live Comparison of SoS due to AGA 10 [10]

In larger gas stations, such as a gas border station, underground gas storage facility or a crossover station, a process gas chromatograph (PGC) is usually installed. A PGC separates the natural gas in its 11 main components, which are: Nitrogen, Methane, Ethane, Propane, i/n-Butane, i/n-Pentane, neo-Pentane and the sum of the higher boiling hydrocarbons called C6+.
Figure 15: Signal-to-Noise Ratio (SNR) in relation to the gas velocity
With this data and the formula published in the AGA 10 Report it is possible to calculate the Speed of Sound out of the composition [10]. This is the so-called theoretical Speed of Sound (SoS theory). On the other hand, the ultrasonic meter itself is measuring the SoS out of the differences in the transient travel time from up and down stream pulses. This is the measured Speed of Sound (SoS measured). How can users recognize this comparison in a live and on-line environment? The answer is shown in Figure 16. The data of the PGC is transmitted to the Flow Computer (ERZ 2000) and also the measured data from the ultrasonic meter (SoS measured) is transferred to the Flow Computer (ERZ 2000). It is a standard feature of the ERZ 2000 to calculate the Speed of Sound (SoS theory) according to the AGA 10, out of the gas quality data of the PGC. Now the Flow Computer has both data available: SoS measured and SoS theory. Finally, it is very easy to compare both values. In a normal case the difference of the two is <0.25%.
24
In cases with larger differences, what is wrong? It can be the PGC, USM, temperature or pressure measurement. In other words, a more detailed investigation is necessary. In more than 80% of these cases it helps to start a manual calibration of the PGC and the operation conditions back to normal. So, based on this experience, the question is: how is checking done? Does the PGC control the USM or the other way around? Regardless, this comparison is a very simple and helpful tool to check the complete metering run.
Figure 16: Live comparison of the SoS theory with SoS measured due to AGA 10 [10]
CBM Flow Profile Factors

Up to now, our CBM discussion has addressed application of the RMG by Honeywell USMs sophisticated diagnostic outputs to indicate whether a functional issue has occurred with the meter. But what of an operational issue that might not be readily apparent from the functional diagnostics of SoS or SNR? A change in flow profile induced by blockage, protrusions or pipe contamination can affect fiscal output with little noticeable affect in the diagnostics discussed til now. So we must also address operational diagnostics to determine whether such a profile change has occurred and then evaluate its pote ntial impact on meter accuracy. This is best accomplished using individual path velocities to diagnose profile changes from baseline, and in that evaluation, a manufacturers path orientation design choice has great impact. Depending on the design choice, the flow paths and their arrangement define the type and number of profile factors in an ultrasonic meter. With RMG by Honeywell, six paths arrayed in an X pattern in three horizontal planes results in two main categories of profile factors (Figure 17): X-profile factor o Flow profile in the horizontal direction using the Sensor-Cross-Planes
Y-profile factor o Indicates the vertical portion of the flow profile
In normal conditions the X-profile factor equals the Y-profile factor (X = Y). But in a disturbed flow profile the profile factors are different from one another (X Y). This indicates the integral, not closed for the flow profile, is not completely covered. In other words, some parts of the profile are missing. To clarify this situation, further analysis of the sub-profile factors is necessary (Figure 18). All profile factors mentioned in this chapter can be tracked and displayed in real-time with the RMGView software.
25
Figure 17: X- and Y-profile factors in a cross path arrangement
The X-profile factor itself comprises two sub-profile factors (Figure 19): X1 profile factor X2 profile factor
The X1 profile factor is created by the comparison of the center plane with the upper plane, and the X2 profile factor is built from the comparison of the center plane with the lower plane. Under normal conditions X1 = X2.
Figure 18: Flow profile factor analysis done separately for the X- and Y-profile factors
26
Figure 19: X-profile factor separated in two sub-profile factors: X1 and X2 profile factors
27
In comparison to the X-profile factors, the Y-profile factors are deducted in a similar way. The Y-profile factors are shown in Figure 20. The Y-profile factor is divided into two sub-profile factors: Y1 profile factor and Y2 profile factor. The Y1 profile factor consists of the Y3,1 and Y3,5 profile factors, and the Y2 profile factor consists of the sub-profile factors Y4,2 and Y4,6. Y3,1 and Y4,2 profile factors are created by the comparison of the center plane with the upper plane. Y3,5 and Y4,6 profile factors are built from the comparison of the center plane with the lower plane.
Figure 20: Y-profile factor separated in two sub-profile factors: Y1 and Y2 profile factors
From Figure 20 it is also obvious that if there are no crossed paths in one plane it is impossible to build Y-profile factors. The Y-profile factors indicate the swirl in the flow profile as shown in Figure 21. Here is shown the area of the Y2 profile factor. If there is swirl in the flow, the areas are indicated with +/- for faster and lower gas velocities.
Figure 21: Picture showing how the Y2-profile factor changes if swirl is in the gas flow profile
28
The above consideration was somehow theoretical. So it comes to the question: how does tracking help in daily live metering systems? For example, an ultrasonic meter is installed in a metering skid for years and is in operation 24 hour a day, 365 days a year. During this time the inner surface of the meter may change due to deposition of dirt or liquid, or as a result of roughness created rust. Figure 22 shows a standard turbulent flow profile shortly after the start up of the ultrasonic meter. Figure 23 shows graphically what can happen to the inner surface of the meter after years of operation. This figure also explains in a very demonstrative way why reflective ultrasonic meters may have problems after years of operation, and why they are limited in gas velocity. How does this change to the inner surface influence the flow profile? The answer is shown in Figure 24. The gas velocity vectors at the outer planes will be reduced and the gas velocity at the center plane will be fast compared to the ideal condition. This is also reflected in the profile factors X1 and X2, which will increase.
Figure 22: Standard flow profile for turbulent flow
Figure 23: The inner surface of the meter or the entire pipe will change over years of operation, due to deposition of dirt or liquid in the button
29
Figure 24: Demonstration of how the flow profile and X-profile factor will change
As described earlier, the profile factors can be tracked and displayed live. Therefore, the USZ 08 is able to detect these kinds of disturbances on-line as a standard feature. This can be very clearly demonstrated as it happened in one gas station recently. The status of the diagnostics is as follows: Comparison of SoS due to AGA 10 Result: OK Signal Qualitiy (SNR) Result: OK Profile factors X1/X2 show a significant difference WHY ?
The answer is indicated by Figure 25. Over time rust formed and there was more rust on the top than the bottom. This effect is evidenced by the change of the profile factors X1/X2 and is directly indicated by the USZ 08 of RMG by Honeywell.
Figure 25: Example of a contaminated USM with inlet spool. This contamination could be detected with the X-profile factor analysis.
30
CBM Swirl Angle

The swirl angle is the difference of the gas velocity vector from the axial direction (Figure 26). How the swirl angle is measured is explained earlier in detail. The swirl angle can (Page 22) be directly monitored on the live page of the RMGView software for all three levels (Figure 27). At standard conditions the sum over all three swirl angles equals zero. In cases where the sum of all swirl angles is zero, the flow profile is not 100% captured and the measured flow can be lower or higher than the real flow.
Figure 26: Graphical explanation of the swirl angle
Figure 27: Pie meters for the three levels showing live swirl angles in the RMGView software package
CBM Live - RMG Precision Adjustment (Patented [11])

Up until now, it has been state-of-the-art to perform a Zero Flow Verification Test to adjust an ultrasonic gas meter. This test is described in AGA 9 (6.3) [1]. This adjustment is necessary due to the fact that besides the time of flight of the ultrasonic pulses, delay times also occur within the system, which are caused by the signal processing electronics, properties of the transducers, and the calculation algorithms. As these delay times cannot be determined directly, they must be determined by a costly measurement.
31
Assuming there is no flow through the meter, the time of flight of a sound pulse is given by the following equation: To determine the system delay time tw all other measured values of this equation must be determined exactly. The time of flight is directly measured by the ultrasonic gas meter. The path length L can be measured exactly, at least for all meters with face-to-face arrangement of the transducers (working without reflections). More challenging is the determination of the theoretical Speed of Sound Cth. It can be calculated by the use of state-of-the-art algorithms (AGA8/AGA10), taking into account the gas composition and the actual gas temperature and pressure. To minimize measurement uncertainty, it is recommended that the meter be filled with a gas having a known Speed of Sound (e.g., N2). Pressure and temperature must be kept stable during the procedure and measured precisely. Most critical is the measurement of temperature, as levels of differing temperatures may occur inside of the meter. Obviously, this method includes various possible sources of errors, which contribute to and increase the measurement uncertainty. As such, it is not a live verification of the delay time. The RMG by Honeywell ultrasonic meter USZ 08 and its electronics USE 09 allow for a precise adjustment of the delay time by a new method, which avoids all disadvantages of the classical method described above. For this adjustment, two measurements have to be done per shot: Time of flight between S1 and S2: t1 (direct measurement, figure 28)
First echo on the receive sensor: t2 (reflective measurement, figure 29)
32
Figure 28: Direct USM measurement of signal
Figure 29: Reflective measurement (echo measurement)
The fundamental equations are: Instead of the time of flight t1 for direct distance between sender and receiver, the time of flight t 2 for the first echo, reflected on the receiver and sender, is measured. According to Figure 29, the path length in this case has tripled. Both measurements provide a measured value for the speed of sound (C1 and C2). Out of these measurements the delay time can be determined precisely and shot by shot! That means LIVE! This new method provides the following unique advantages: The composition of the gas inside of the meter must not be known The measurement is independent of the theoretical value of the Speed of Sound As the absolute value of the Speed of Sound is not needed, pressure and temperature do not have to be measured The determination of the delay time is done automatically Higher accuracy in the determination of Speed of Sound Live monitoring of the altering process of the transducers Temperature, pressure, moisture, aging of sensors, and electronics have no influence on the calibration result A verification of the meter can be performed in the field under operating conditions
33
Figure 30: Effect of the RMG by Honeywell precision adjustment
Figure 30 shows the influence of the live dry calibration compared to the standard modus without echo measurements. As we have explained, this echo measurement allows for a much more accurate determination of the Speed of Sound. The determination of the Transit Time is also more accurate, and this implies that the flow measurement overall is higher than conventional ultrasonic meters without echo measurements.
Conclusion and Outlook

Today, ultrasonic meters are widely accepted for custody transfer and allocation metering because of their technical advantages over other flow metering technologies like turbine meters and vortex meters. This situation is also due, in part, to advancements in ultrasonic meter technology and establishment of the ISO standard [2] in 2010. Ultrasonic meters are now the overwhelming technology-of-choice for large capacity gas measurement stations because of their reliability and rangeability. Every year, more and more ultrasonic meters are sold (with greater pressure on pricing), resulting in the development of smaller size meters (< DN 100 [4]) for distribution networks and downstream applications. For these applications, the installation requirements for ultrasonic meters have to be simplified a challenge to be met in the near future by improvements to the technology. Another reason for the on-going success of ultrasonic meters is the potential to provide operators with simple diagnostic techniques to validate meter integrity in the field, such as RMG by Honeywells precision adjustment measurement and the comparison of Speed of Sound diagnostics. The coming years will bring additional diagnostic advancements within ultrasonic meters, which will simplify installation, operation, and meter validation. There is also a clear trend towards ultrasonic meters in larger sizes. Unfortunately, there are no test rigs capable of serving these applications in a proper way. This situation is coupled with the challenge of obtaining a time slot on a test rig for highpressure calibrations.
34
Literature
1. 2. AGA Report No. 9, Measurement of Gas by Multipath Ultrasonic Meters, American Gas Association, 2nd Edition, 2007, 113 pages ISO / FDIS 17089-1, International Standard, Measurement of fluid flow in closed conduits -Ultrasonic meters for gas Part 1: Meters for custody transfer and allocation measurement, International Organization for Standardization, 2010, 100 pages Type-approval certificate of the USZ08 ultrasonic flowmeter, Physikalisch- Technische Bundesanstalt (PTB), Braunschweig, 7.241 / 01.04 MID Type-approval certificate of the USZ08 ultrasonic flowmeter, Physikalisch- Technische Bundesanstalt (PTB), Braunschweig, DE-09-MI002-PTB003 Letter, Confirmation of Flow Tests , Physikalisch-Technische Bundesanstalt (PTB), Braunschweig, 2010 Technical Directive G13, Einbau und Betrieb von Turbinenradgaszhlern, Physikalisch -Technische Bundesanstalt (PTB), 2005 OIML R 137-1, Gas meters Part 1: Requirements, International Recommendation International Organization of Legal Metrology, Issue. 2006, 40 pages Hans-Jrgen Kastner, Andreas Weber and Jrn Weber, 6-path ultrasonic gas meters - Redundant, high-precision measurements without pressure drop, GWF 2006 James W. Bowen, Achim Zajc, Natural Gas Metering with Ultrasound A New Dimension of Metering, GWF International Issue 2009, 14-21
3. 4. 5. 6. 7. 8. 9.
10. AGA Report No. 10, Speed of Sound in Natural Gas and Other Related Hydrocarbon Gases , American Gas Association, 2003, 177 pages 11. German Patent No. DE 10 2008 026 620 A1 2009.12.10. Verfahren zu Kalibrierung eines Ultraschall Durchflussmessers, RMG Messtechnik, 2009, 5 pages
More Information For more information on Natural Gas Metering with Ultrasonic Technology, visit our website at www.honeywellprocess.com or contact your Honeywell account manager. Automation & Control Solutions Process Solutions Honeywell 1250 West Sam Houston Parkway South Houston, TX 77042 Honeywell House, Arlington Business Park, Bracknell, Berkshire, England RG12 1EB UK Shanghai City Centre, 100 Junyi Road Shanghai, China 20051
WP-11-16-ENG September 2011 Printed in USA 2011 Honeywell International Inc.
2011 Custody Measurement Technical Exchange Meeting November 21 23, 2011
An Analysis of Gas Ultrasonic Meter Recalibration Intervals

Thomas Kegel [tkegel@ceesi.com] Colorado Engineering Experiment Station, Inc (CEESI) Nunn, Colorado, USA Abstract This paper describes data, discusses analytical results and presents a mathematical model that relates recalibration shift, meter size, velocity, and recalibration time interval. The results can be applied as a tool to assist in determining an appropriate recalibration interval for an ultrasonic meter. The database supporting this project is a result of twelve years of history in the operation of an ultrasonic gas flow calibration facility. The database includes 95 recalibration events, recalibration time intervals from less than one year to nine years, meter sizes from DN100 to DN500, and gas velocities between 3 and 30 m/s. Introduction The application of ultrasonic gas meters has been steadily increasing following the publication of the first edition of AGA Report 9 in 1998. Neither the first nor second editions of AGA 9 specify a recalibration interval. While custody transfer meters used in Canada require a five year recalibration interval, most other regulatory agencies have no specification. In the United States, recalibration time intervals would be included in a contract, but most contracts are silent on this topic. The absence of clear guidance is due in part to the lack of significant recalibration data with accompanying analyses. The project summarized in this paper represents a contribution to the industrys understanding of the factors that contribute to ultrasonic meter recalibration shifts. Previous Publications The topic of meter recalibration has begun to appear in various publications, this section provides a brief survey of the available literature. Reference 1 discusses several topics from the perspective of a calibration laboratory. Calibration data analyses are based on long term data from ultrasonic check standards used in the laboratory. The check standards indicate random effects characterizing repeatability and reproducibility that increase as velocity decreases. The observed random effects are separated based on velocity and time, the resulting analysis quantifies long term variation. Reference 2 provides a discussion from the perspective of an ultrasonic meter user. Several case studies are described where meters are removed from service and recalibrated. Data presentation and discussion include the effect of meter cleaning and component replacement as well as shifts observed upon recalibration.
Reference 3 describes data from 35 meters re-calibrated in a flowlab. The objective was to investigate the effects of recalibration interval on the performance of ultrasonic meters. Results indicated that ultrasonic flowmeter performance changes over with time, data were presented as a function of recalibration time interval and velocity. The present study is a continuation if this work.
0.8 0.4
Meter Error [%]
0.0 -0.4
2005
-0.8 -1.2
2001 Shi
0 10 20 30 Reference 4 summarizes the results of a Velocity [m/s] two year study. The recalibration data of 34 meters were reviewed, 22 of the total had Figure 1: Sample Ultrasonic Meter Recalibration been in service at least six years when they were recalibrated. The details of selected calibrations are discussed to illustrate project conclusions. Related topics covered in the paper include diagnostics, component replacement and cleaning.
Analysis Method The analysis method of the present study is described based on the sample calibration data shown in Figure 1. The meter error is defined as the percent difference between the flowrates indicated by the meter and calibration standard. This particular meter was first calibrated in 2001, the data represent the performance of the meter as received. In 2005 the meter was recalibrated, the data represent the performance of the meter with correction coefficients restored to the factory settings. The shift curve which characterizes the difference between the two calibrations is defined by a second order polynomial. It represents the average value of the meter error fitted over the velocity range. The shift curve does not account for the random effects observed in association with the two calibration curves. Numerical values of calibration shift are calculated at 3.28 m/s (10 ft/s) intervals, they are symbolized by closed circles in Figure 1. These numerical values, called velocity points, become data points that are used in the analysis to categorize velocity based effects. This process described above was repeated for multiple ultrasonic meter recalibrations. Care was taken to properly interpret the data sets. The same set of coefficients was confirmed to be present during both calibrations. If a meter was cleaned, only the clean data were compared. The status of any component replacements was noted.
Table 1: Database Scope Nominal Diameter 100 150 200 250 300 400 500 Recalibration Events 3 6 31 13 18 14 10 Velocity Points 24 44 213 80 127 97 62
Database Scope The database used in the present analysis represents most of the ultrasonic meter recalibrations completed at the CEESI facility. The analysis comprises 95 recalibration events where an event is defined as one meter returned for one recalibration. The same meter can be recalibrated several times each recalibration represents a different recalibration event. In the current study nineteen meters were recalibrated at least twice. Bidirectional meters result in two recalibration events, one each for the forward and reverse directions. The current study includes four bidirectional meters. The distribution of recalibration events and velocity points are contained in Table 1. A different view of the database scope is shown in Figure 2. The abscissa represents the recalibration interval expressed in years and the ordinate represents meter inside diameter expressed in millimeters. The recalibration intervals ranged to nine years, the data are reasonably evenly distributed based on meter size. The entire database consists of 646 data points.
600 500
Inside Diameter [mm]
400 300 200 100 0 0 2 4 6 Recalibraon Interval [years] 8
Figure 2: Database Scope
2.0 1.0
Shi [%]
0.0 -1.0 -2.0 0 10 20 Velocity [m/s] 30
A majority of the meters were from a single manufacturer (Daniel) while a few were Figure 3: Recalibration Shift as a Function of Velocity from a second manufacturer (Instromet). This is a reflection of the calibration business rather than the result of a selection process. Those manufacturers that entered the market more recently are not represented, likely as a result of the fact that the meters have not yet been returned for recalibration in significant quantities. Once again, this was not the result of a selection process. Summary of Results It is proposed that the recalibration shift is dependent on the meter size, velocity and recalibration interval. In this section the relationships between these variables are explored. Results summarizing the effect of velocity are contained in Figure 3. The individual symbols represent the 646 velocity points that make up the database. For each of the ten velocities a mean value is calculated,
Three data points were obtained at a velocity of 33.5 m/s. They appear on the graph of Figure 3, but are not included in subsequent analyses. The standard deviation associated with the data points corresponding to each velocity is also calculated. The calculated values are used to define a statistical interval that contains 95% of the data. The two solid lines in Figure 3 represent that statistical interval centered about the dashed mean line. The interval width is consistent through the velocity range, increasing slightly at higher velocity. The consistency in the interval width suggests that the recalibration shift does not vary significantly with velocity. An ultrasonic meter user might not need to consider velocity as a variable in making recalibration decisions.
Shi [%]
the ten values are identified by the dashed line. The mean values are within 0.06%, meaning that ultrasonic meter recalibration data are equally likely to indicate a positive or negative shift.
0.2
0.1
150 mm
0.0
100 mm
-0.1 0 10 20 Velocity [m/s] 30
Figure 4: Average Recalibration Shift as a Function of Line SIze and Velocity
3.0
Shi [%]
2.0
0.6
m/s
1.2
2.4
1.0
4.9 19.5
9.8
30.5
0.0 0 2 4 6 8
The data points in Figure 3 are not sepaRecalibraon Interval [years] rated by meter size. It is possible that the data for one meter size may be shifted in one direction, while those representing a Figure 5: Summary of Results From Reference 3 different meter size are shifted in the opposite direction; behavior that would not be apparent in the graph. With this possibility in mind the analysis shifted to data sets sorted by line size. For each meter size a linear fit is determined that relates recalibration shift and velocity, the fitted lines are contained in Figure 4. The five solid lines correspond to meters sizes between DN200 and DN500. Meters of these sizes exhibit similar behavior characterized by a gradual increase in recalibration shift with gas velocity. The two smallest meter sizes exhibit a gradual decrease in recalibration shift with velocity. All of the data fall within 0.1% except for velocities less than 15 m/s measured with the DN150 meters. It is concluded that the recalibration shift does not vary significantly with line size. Some results from Reference 3 are summarized in Figure 5. The ordinate represents the absolute value of the recalibration shift, the abscissa represents the recalibration interval, and the solid lines represent different velocities. Clearly the recalibration shift increases as the velocity decreases. It appears as if Figures
3 and 5 are contradictory in regards to the velocity effect. The reason for the differences lies in the analytical methodologies. The previous study included the repeatability of the calibration results, it is well known that the random effects increase in magnitude as the velocity decreases. The present study only considers the recalibration shift of the average values; the random effects resulting from repeatability are not included. The second variable that might affect recalibration shift is meter size, summarized results are contained in Figure 6. Once again the individual symbols represent the 646 velocity points that make up the database. The analysis is similar to that applied to the data of Figure 3; mean and standard deviation values are calculated for the data corresponding to each of the seven nominal diameters. The solid lines identify a 95% confidence interval. The lines are represented by:
2.0 1.0
Shi [%]
0.0 -1.0 -2.0 100 200 300 400 500 Internal Diameter [mm]
Figure 6: Recalibration Shift as a Function of Meter Size
1.0 0.5
Shi [%]
0.0
600 [Eq. 1] S1 = 0.25 + 1.3 d -0.5 where S1 represents the recalibration shift -1.0 and d represents the inside diameter in mm. 0.2 1 10 Clearly the recalibration shift magnitude inRecalibraon Interval [years] creases as the meter size decreases. This apparent trend is useful to an ultrasonic meter Figure 7: Recalibration Shift as a Function of Recalibration user that may be evaluating the recalibration Time Interval schedule for meters of several sizes. For example, they may elect to recalibrate smaller meters more frequently than larger meters.
The calculated mean values are shown in Figure 6 as a dashed line, the values are all within 0.1%. As described above, the analysis continues by separating the 646 data points based on velocity to identify asymmetry within the database. The mean values corresponding to the nine 3.1 - 27.5 m/s velocity values are each within 0.08% while the mean values corresponding to the 30.5 m/s mean values are within 0.18%. From the analysis thus far it is concluded that the recalibration shift is symmetric to within 0.1% for most values of velocity and diameter. The DN150 meter and 30.5 m/s data are symmetric to within 0.2%.
The third variable that might affect recalibration shift is recalibration time interval. Results summarizing this effect are contained in Figure 7. The individual symbols represent 628 velocity points from the database. Not shown are data from two of the DN100 meters as well as the two lowest velocity data points from one of the DN200 meters. These data fall points outside the ordinate scale, they are readily identified in Figures 5 and 6. As shown in previous graphs, the dashed line represents the mean of the entire data set as the shift varies with time interval. The mean values remains within 0.1%, thus further re-affirming the symmetry of the shift data. The solid lines represent a manually developed estimate of the 95% confidence interval. The lines are represented by: 0.43 S2 = 1.0 0.5 [Eq. 2] t where S2 represents the recalibration shift and t represents the recalibration time interval in years. The shape of the confidence interval indicates a fairly rapid initial increase in recalibration shift that gradually decreases over time.
0.3 0.2 0.1

Shi [%]
m 27.4
/s
0 -0.1 -0.2 -0.3 0 2 4 6 Recalibraon Interval [years] 8
30. 5m
/s
Figure 8: Average Recalibration Shift as a Function of Velocity and Recalibration Time Interval
0.4 0.2
Shi [%]
150 mm
500 mm
0.0 -0.2 -0.4 0 2
200 mm
250 mm
400 m
300 m m
4 6 Recalibraon Interval [years]
Figure 9: Average Recalibration Shift as a Function of Meter As discussed above, asymmetries associatSize and Recalibration Time Interval ed with velocity and meter size are not evident in Figure 7. To identify asymmetry the data were first organized by velocity, and then mean values were determined as a function of recalibration interval. The results are shown in Figure 8. Eight of the ten linear fits are similar, a slight downward slope is observed. The recalibration shift trends slightly negative, the average slope is 0.013% per year, a ten year recalibration interval might result in average shift of -0.13%. The two highest velocities exhibit much larger amplitude slopes. This observation may also be noted in Figure 3 where the highest velocities are accompanied by an increase in statistical interval width. All the data of Figure 8 fall within 0.2% except for intervals greater than 6.3 years for operation at 30.5 m/s.
The data are then organized by meter size and mean values determined as a function of recalibration interval. The results are shown in Figure 9. The available data for the DN100 size are not shown because
they are based on only three meters subject to similar recalibration intervals. The data of Figure 9 indicate more variation than previous graphs. The two smallest meter sizes indicate that the recalibration shift trends slightly positive. The average slope is 0.030% per year, a ten year recalibration interval might result in average shift of +0.30%. The four largest meter sizes indicate that the recalibration shift trends slightly negative. The average slope is -0.033% per year, a ten year recalibration interval might result in average shift of -0.33%. All the data of Figure 9 fall within 0.2% except for intervals greater than 5.25 years for DN150 meters and 6.65 years for DN400 meters. The discussion continues by writing a general form of Equation 2:
0.3 0.2 Shi [%] 0.1 0 -0.1 -0.2 -0.3 0 10 20 Velocity [m/s] 30
wn e U nkno ics le ctron
New electronics
No ne w ele ctronics
Figure 10: Average Recalibration Shift as a Function of Electronics Status and Velocity
0.2
Shi [%]
0.1 0.43 Si = K i 1.0 0.5 [Eq. 3] 0 t

-0.1 -0.2 -0.3 100
U nk n e le c own tron ics
N o ne w s e le ctronic
Ne w e le
where the subscript i refers to the data from a particular meter size, Si represents recalibration shift and Ki is a constant. It was observed that as the meter size increases, the constant Ki decrease to maintain the 95% confidence interval. In other words, a plot similar to Figure 7 that contains data for only one meter size will show the solid lines closer together. As the meter size increases from DN200 to DN500, Ki decreases from 1.0 to 0.6. This is the same behavior that leads to the interval width reduction with meter size observed in Figure 6. In summary, the meter user might choose to exercise additional caution when measuring the higher velocities indicated in Figure 8. Meanwhile, Figure 9 shows that all meter sizes will exhibit recalibration shifts over time that are similar in magnitude. Over the years the electronic components of ultrasonic meters have been improved. For a variety of reasons some users will upgrade the electronics on existing meters while some will chose not to. The present analysis includes the effect on recalibration shift of replacing electronic components. The status
ctronics
200
300 400 Inside Diameter [mm]
500
Figure 11: Average Recalibration Shift as a Function of Electronics Status and Meter Size
0.2 0.0 -0.2
No ne w e le
Shi [%]
wn e le ctronic s ctronics
U nkno
Ne
-0.4 0
on ctr ele
ics
2 4 6 Recalibraon Interval [years]
Figure 12: Average Recalibration Shift as a Function of Electronics Status and Recalibration Time Interval
of electronic components was not always known based on the information available for the present study. Out of the total of 646 data points, 296 did not involve new electronics, 115 included new electronics, and the status of 235 were unknown. It is noted that data points in the unknown category might include field replacement. It is further noted that with many older calibrations the serial number of the electronics was not recorded and therefore the status would be unknown. To investigate the effect of electronic component replacement the data were divided into three categories based on knowledge (yes, no, unknown) regarding the replacement of electronic components. Linear fits were determined of recalibration shift as it varied with velocity, internal diameter, and recalibration interval; the results are contained in Figures 10-12. Note that the linear fits cover different recalibration interval ranges, this is a result of grouping the data points. There appears to be an effect associated with whether or not the electronics have been replaced. In all three graphs the no new curve is centered at the zero shift position; all the data fall within 0.08%. The new curves fall consistently low, the overall average for all three graphs is -0.15% while the range goes from -0.34% to +0.08%. The trends of the lines show no particular pattern between the graphs. Overall the data all lie within 0.2% except for the case where new electronics are installed in conjunction with a recalibration time interval of less that 5.1 years. Predictive Model The entire database was fit to a curve of the form:
Sp = c0 + c1v + c 2 d + c3 t [Eq. 4]
where: Sp = average predicted shift [%] v = velocity [m/s] d = inside diameter [mm] t = recalibration time interval [years] c0 = 5.308424x10-2 c1 = 4.255990x10-4 c2 = -8.919045x10-5 c3 = -1.099365x10-2 It is noted that Equation 4 only predicts the average trends. Any one meter can exhibit recalibration shifts within the confidence intervals shown in Figures 3, 6 and 7. The Sp values all lie within 0.08% when applied to the entire database.
Uncertainty Considerations Much of the discussion concerns observations of small effects, either recalibration shifts or trends. The significance of an observed effect must be judged within the context of the measurement uncertainty. In particular the present study has compiled data of numerous comparisons of the form: S = A - B [Eq. 5]
where S is the recalibration shift and A and B are calibration events. The uncertainty of S can be expressed as:
2 2 2 [Eq. 6] uS = u A + uB uC
where: uA = uncertainty associated with calibration A uB = uncertainty associated with calibration B uC = correlated effects between calibrations A and B Correlated effects represent uncertainty components that remain unchanged between A and B. A simple example would be the equation of state used to calculate the natural gas compressibility. The equation will not change between A and B and the uncertainty of S will be reduced by the uncertainty in the equation of state. The present study is based on calibrations completed in the CEESI Iowa calibration facility with an estimated uncertainty of uA= uB= 0.23%. The process of estimating uC is complex because it will vary with recalibration time interval, velocity and meter size. The details of this process are beyond the current scope of the project, but will be considered for Curve 3 future work. Cu
Failure Rate
rve 1
In the absence of a detailed uncertainty analysis, it is likely that the uncertainty in S will likely exceed uS = 0.2% and thus many of the observations discussed above will fall within the uncertainty. General Observations
e Curv
Time
Figure 13: The Bathtub Curve
Reliability engineering models can be based on a bathtub curve that might be relevant to the current analysis. A simple example is shown in Figure 13. Curve 1 corresponds infant mortality failures while Curve 2 corresponds to wear out failures. Curve 3, the bathtub, results from combining Curves 1 and 2. The shape of the statistical interval width suggests Curve 1 in Figure 13. Perhaps components of the ultrasonic meter change in time in the manner of Curve 1. Perhaps similarly to the wear in period for a set of new bearings.
Supposing that the bathtub curve describes ultrasonic flowmeter performance, the current study has not identified any trends that indicate the presence of a Curve 2 type wear our behavior. It is possible that sufficient recalibrations with long time intervals have not yet been recorded. The authors of Reference 4 came to a similar conclusion. The statistical interval of Figure 6 shows the variation is recalibration shift decreasing as the line size increases. A fixed time measurement shift, a change in the clock, might be responsible for the observed trend. The fixed time shift becomes a smaller percentage of the time measurement that increases with meter size. Further analysis based on this observation has not currently been completed. In general the two highest velocities seem to exhibit more drift than the lower velocities. Users that operate meters at higher velocities (over 27 m/s) might want to consider more frequent calibration. Larger meters seem to have lower recalibration shifts that smaller meters. The present study included data from DN100 meters that was often removed from the analyses. The DN100 data was limited in scope (3 meters) and range (of recalibration time interval).
8 6 4 2 0 Sep-00
Recalibraons per month
The authors of Reference 4 concludes the absence of an effect due to electronics replacement. The current study seems to indicate an effect, though it might be small enough to fall within the uncertainty. The analysis of Reference 3 identifies an increase in amplitude corresponding to year three when recalibration shift is plotted against recalibration time interval. The present study also shows a similar increase in magnitude between years 3-4 with lower shift values between years 4-7. It is noted that all the data from Reference 3 is include in the present study, some similarities can therefore be expected. Future Work
Sep-02
Sep-04
Sep-06
Sep-08
Sep-10
Figure 14: Monthly Recalibrations
On a daily basis calibrations continue in the Iowa facility. Figure 14 shows that recalibrations have been gradually increasing. These data represent increased knowledge and will be added to the database. While the present database is quite large, some recalibration events have not been included. Future plans include filling in these gaps. The authors of Reference 4 propose that diagnostic parameters represent a powerful tool to predict the need to recalibrate ultrasonic meters. Inclusion of available diagnostic parameters is planned for the future. The details of how correlated effects influence the uncertainty of the recalibration shift is considered for future work.
References 1. Kegel, T. and Britton, R., Characterizing Ultrasonic Meter Performance Using A Very Large Database, 26th International North Sea Flow Measurement Workshop, 2008. 2. English, S. and Clancy, J., Ultrasonic Meter Recalibration Program, Seventh International Symposium on Fluid Flow Measurement, 2009. 3. Trostel, B., Kegel, T. and Clancy, J., Ultrasonic Flowmeter Calibration Intervals, Flomeko, 2010. 4. Hall, J., Zanker, K. and Kelner, E., When Should a Gas Ultrasonic Flow Meter be Recalibrated? 28th International North Sea Flow Measurement Workshop, 2010. 5. http://en.wikipedia.org/wiki/Bathtub_curve
Maintain Measurement Accuracy from Design to Operation

Authors: Peter Jakubenas Installed Product Services Manager, FMC Technologies Inc Bryan Partridge After Market Manager, FMC Technologies Inc Jeffery MacKenzie Global Aftermarket Manager, FMC Technologies Inc
ABSTRACT
Custody Transfer Metering is unique. Unlike general flow metering applications this small but very important component of the total installed equipment can consume some major costs from design, through construction and then in on-going maintenance. Efforts to keep providing high-quality overall solutions (design, Testing, Calibration, field service and technical support) to this unique subset of installed equipment is often costly. Further as it becomes an ever decreasing component of what we do, it is more and more difficult to maintain a critical mass of knowledge and capabilities on-site. The Design, Operation and Maintenance of these complex Measurement Systems, once a key to their success, now provide a challenge to major Oil and Gas corporations. Is it a core business or noncore business? For many Oil and Gas companies there are many questions being asked in boardrooms and at operational levels At that level the questions are: Why do we need to implement this? Why is it so expensive? Can we do without Custody Transfer equipment? At other levels in these corporations there are more questions, such as: How do we design that? Where do we site it? How big will it be? What technology do I need? As the equipment is placed in service the questions turn to: It needs what in terms of maintenance? Is it necessary to have a specialist Team of people? How often do we have to do that? The questions are all real, all relevant and not unique. At risk with inept operation and maintenance is your measurement certainty. Even a small increase in uncertainty can cost millions in a short time. This paper will look into Caring for Custody Measurement. Initially the paper will try to identify the system design criteria that ensure simplified best performance and increased support services over the life of the equipment. Also the support services necessary to ensure high performance of your installed measurement equipment over the short, medium and long term. The second component of the paper will be to take a microscope to the identified essential Services for Custody Metering Systems and look into the critical components and the options Operators have for these services. Advanced software features are now able to determine need for meter proving, evaluate proving results, and track system performance in terms of uncertainty. Condition of the primary elements can be analyzed and maintenance intervals determined based on need rather than just passage of time. Finally, the paper overviews a Case Study of an Innovative approach to operation and maintenance. It shows how a well conceived approach can lead to increased reliability, improved integrity and enhanced performance for an operator. It will also look at how that operator has gained many such benefits while also lowering their costs and exposure to the risk of uncertainty of measurement throughout the life of the project.
DESIGN
The value of crude oil, natural gas, and refined petroleum products continues to be significant, clearly high accuracy is tantamount to good measurement. True! High accuracy is impossible however unless
Page | 1
there is a yardstick by which to compare fact versus measurement. This cannot be done continuously so the industry target is low uncertainty of measurement. Whatever adjectives we apply to the measurement, high accuracy or low uncertainty they will result in short payback time if designed into the overall metering package. In this paper we will focus our discussion on low uncertainty rather that high accuracy. Design for lowest possible uncertainty has to include high performance measuring systems that offer state of the art solutions to high precision liquid measurement requirements; however, when it is necessary to achieve ultimate performance, it is important that the primary measurement elements be surrounded by carefully selected ancillary components. Then all these components need to be seamlessly integrated into a system which will incorporate the operational needs; the accuracy (low uncertainty); and the long term maintainability necessary for the life of the field. Measurements of this complexity necessitate a systems approach to the procurement of meter stations, usually with a turnkey contract to an experienced component and system supplier. Overall uncertainty could be as low as + 0.1%, where circumstances permit, generally an uncertainty in the order of + 0.25% or better is achievable with properly designed systems. System Economic Justification Obviously, metering is only one type of Custody transfer system. Tanks with level metering, usually servo or radar gauges, are the other well used methodology. While a metering system initial purchase price and operation costs may seem to be somewhat more than that for tank gauging equipment, there are issues of calibration (strapping) of tanks, installation of datum points, etc that are often not included in the tank level systems. There is also the additional demerit of decreased uncertainty. The differential in uncertainty is in the order of 0.5% (0.7% for tank gauging vs. 0.2% for metering). For oil that is $100/bbl; there is a potential risk of $0.50 per barrel. For a 10 inch pipeline flowing at 10,000 BPH, $3.6 million is at risk every month. Keep in mind that if the difference is in the favor of the buyer, the seller is depriving himself of deserved revenue. If the difference is in the favor of the seller, he may be opening himself up to potential lawsuits and retroactive charges. For this reason the measurement needs to reflect the truth and it is essential to include traceability to world recognized standards. It is therefore not only desirable keep uncertainty low but also to obtain traceability. These two desires combine to ensure that an in-situ proving method is integrated into the solution. Control over the capital costs, with an eye on recurring expenses is the means to reducing the overall development costs. One additional benefit of metering is that custody transfer tanks (tanks that have been and will need future calibration) can be reduced and live tanks (Live tanks allow flow into them and out of them at the same time) can be used instead. The net result is fewer, possibly up to 50% fewer, tanks will be needed. What does Designing systems for low total cost of ownership actually mean? This is a broad field and has numerous components. Ongoing costs for a measurement system include the following costs: energy consumed, operation, calibration and, maintenance.
Over the life of an asset or, in this case the life of a project, purchase price can be a small component of the life cycle costs. So, purchasing equipment with the lowest initial cost is certainly an oft used selection it may not, over the life of the development, be the best choice. An overall understanding of all of the components of the metering asset ownership, a design to lower ongoing costs, and the experience and capability to maintain, in a word expertise is very important.
Page | 2
Energy Consumption Considerations Energy. Why is that a consideration? Three reasons 1. Costs to run pumps 2. Stress on components at higher pressures 3. Increased prices for larger pumps and equipment The ongoing cost of energy is mainly due to pressure drop losses (pumping costs) associated with flow through the system. Keeping pressure drop low needs understanding and experience in metering technology, header sizing and (multiple) meter runs design, strainer design, mixing method for sampling, along with block and flow control valve selection. Even when the developments energy cost on site is low, as it can be when the produced natural gas drives turbine generators, stress on the components leads to increased maintenance and operation of the energy producing and most of the other equipment. Focusing on lowering the overall energy costs ensures that energy, hence cost is minimized. Selecting the optimum meter run size and quantity for a total station flow rate becomes important from not only an initial cost standpoint, but also from an operation and maintenance cost standpoint. A large number of small runs entail higher maintenance and instrument calibration costs for the meter runs, but less for the prover. Fewer larger meter runs result in less maintenance and calibration costs but higher prover maintenance costs. Those meters that typically require more cycles for proving may contribute to higher prover maintenance costs. Larger diameter equipment is more costly to maintain because of larger and higher quantity of flange bolts and heavier components requiring larger lifting equipment. Ultrasonic meters offer the lowest pressure drop per unit of flow rate of all the technologies available, but may not be the best choice due to fluid properties. The next lowest pressure drop per unit of flow rate is the PD meter, followed by the helical turbine meter and the conventional turbine meter. The Coriolis meter usually has the highest pressure drop per unit of flow rate measured. Keeping velocity in headers at 12fps (3.6m/s) to 15 ft/sec (4.5m/s) keeps pressure losses low. Special design considerations or larger size piping also can reduce the losses at inlet and outlet tee branches. Strainers for inferential meters need to be designed with tapered nozzles and oversize bodies to keep pressure losses low. To keep pressure drop low it is important to have proper basket design and mesh selection to only eliminate large debris. It should be noted that meter run strainers are not necessary for ultrasonic meters; however a single strainer either on the meter run or on the prover inlet is recommended to protect the prover and the downstream equipment (such as control and block valves). For large systems with turn down ratios of 20 or 30 to 1 and sampling in the inlet or outlet header, positive mixing of the stream ahead of the sampling device is usually required. A static mixer is often used to accomplish this, however the static mixer introduces pressure drop and in some cases a jet mixer pump is used. Both of these alternatives consume energy and selection of the method is best made by quantitative comparison of the two. Maintenance costs for the additional pump should also be a consideration in this analysis. If a jet mixer is used, it may be desirable to program its operation only when the header velocity is below what is required for good mixing. If the turn down ratio is low and the sampling device is located downstream of the flow meter/s and/or pumping stations, mixing equipment with the resultant pressure drop may not be necessary. This of course assumes that the meters actions result in more mixing, this would not be true for ultrasonic meters. To keep pressure drop low block valves upstream of inferential meters should be full bore and be one or two pipe sizes larger than the meter. Ball valves or slab gate valves are quite often used for this application.
Page | 3
Valves downstream of the meters involved in diverting flow for proving must be high integrity double block and bleed valves. To keep pressure drop and maintenance costs low they should be sized with velocities in the 15fps (4.5m/s) to 20 fps (6.1m/s) range. These valves are subjected to many open and close cycles and should not have seals that rub against the seats. For this reason ball valves are not recommended for this application. To keep pressure drop low control or flow balancing valves should be selected so that they are fully open during normal metering operations. These are usually high performance butterfly valves and are used primarily to balance flow when proving meters.
OPERATION
Keeping operating manpower costs low can be tricky and hard to ascertain initially. The key here is making things convenient and easier for the operator. Primarily this means keeping the operator in the control room and giving him access to an extensive array of tools. This is accomplished through the incorporation of the latest HMI software along with the integration of flow computers and metering supervisory computers. Meter skid design should ensure that the meter skid hardware is easy to access, has proper, clear tagging of all equipment in the field and has computerized manuals to facilitate operation. Extensive graphic interfaces are now available for guiding the operator through some of the more complex metering tasks. Flow computers with touch screen graphic panels provide a view of the metering operation in real time. Add the important digital and analog operating parameters, essential for the operators are clearly represented under all circumstances. While the debate rages about the best type of prover to use, the need for meter proving still exists. For the operator, proving can be based on various criteria programmed into the control system. Added to these criteria there can be additional criteria for accepting a series of proving runs. The acceptability of the resultant meter factor can also be judged based on customer preferred comparison methods with past meter data. Statistical analysis of meter performance based on API recommended methods and graphic tracking of the data is an essential tool for determining the health of the meters and other components.
MAINTENANCE
There are many, many influencing factors when discussing maintenance. Often some of the factors that will influence the long term life of equipment are unknown at the time of design because many developments are untried commercially, so the documentation is based on our understanding, rather than fact. Designing in this environment is often based on experience of previous problems and similar experiences. Decreased maintenance costs can be achieved by selecting equipment suitable for the total environment that is both outside and inside of the equipment. This could include providing long life protective coatings, providing convenient access to equipment and lengthening the interval between required activities such as instrument calibration based on condition monitoring and historical information. Initial cost of redundant pressure, differential pressure, and temperature transmitters with condition based monitoring can often be justified by less ongoing cost of calibrations.
TESTING
While the System Vendors would like to cookie cut every system they build, generally every i ntegrated measurement system is a one off design. This is as a result of the differences in each development, product characteristics, flows, pressures temperatures. Additionally there is often a need to verify or
Page | 4
ensure that all the equipment interacts as it is intended before they get to site. It is for this reason that these critical systems are usually pre assembled and tested at the manufacturing facility. For the best integrators the equipment is tested prior to and during factory acceptance tests (FAT). During these tests the equipment is checked for proper operation, ease of operation and ease of maintenance. It is clearly much easier and far less costly to make even the most minor of changes at the design and/or manufacturing facility than it is to do this on a remote site. Suppliers with many years of system integration experience can accomplish pre-FAT and FAT in a relatively short time. During and after the FAT there can be formal classroom and hands on training of operation and maintenance personnel if required. In the past this was quite often done, with follow up courses annually to ensure compliance. With the constant pressure on manpower and the drive for efficiency and core competency this training has far fewer students and is rarely repeated.
FIELD INSTALLATION
In the 40+ years of systems supply experience the field installation process has changed significantly. In the early years once the FAT was complete the system would be shipped to its permanent installation site. After arrival on-site and system installation, the client would request commissioning. Soon after the request, a team of technical personnel from the designer/supplier would be sent to commission the system. This commissioning was comprehensive and structured but at the time, most users of these types of systems had their own teams of dedicated technical staff to expand on any perceived shortcuts. In more recent times the large dedicated technical teams that end users employed have either been reduced to a shell of their previous glory or no longer exist at all. This means that there are many more tasks that contract specialists are asked to provide. It is now commonplace for systems integrators to have technical and service engineers on-site to supervise the many important functions that are necessary. Tasks that are now commonly performed by the System supplier are: unpacking & checking, re-assembly (if any) and installation of the equipment. critical calibrations, commissioning
The last two (2) have always been an important part of the commissioning process. Critical calibration of provers, meters, transmitters, analysers and then commissioning of the whole system is the precursor for the final site acceptance tests (SAT). After the SAT some fine tuning of the equipment can be done. This fine tuning will be painstakingly done to ensure the reliability in measurement certainty, operability and maintainability.
TECHNICAL SUPPORT FOR OPERATION AND MAINTENANCE

Custody transfer measurement is a very specialized field. The design operation and maintenance of custody transfer systems should be trusted to only properly trained and experienced people. This core competency was once developed by end users for their own use and for peace of mind. However, equipment providers, some system suppliers and service suppliers provide the base expertise and the critical mass to ensure ongoing technical standards. Various educational institutes, universities and international renowned testing authorities now offer formal and informal classes, seminars, and hands on training on some of the individual components of these systems. Additionally, those that provide service on this type of equipment 24/7, have the ongoing expertise to be up-to-date continuously.
Page | 5
Since the measurement equipment at a production facility, offshore platform, refinery or terminal may be a small part of the total equipment, dedicating fully trained experts to operate and maintain the equipment is sometimes difficult for end users. Often operation and maintenance for lowest measurement uncertainty and optimum use of the assets employed falls by the wayside. This is not the best way to ensure low uncertainty where a little often is much, much better than a lot too late. Over the last 2 decades it has become common for specialized measurement teams to provide operational and maintenance services, under contract to the user. Several vendors now have services available in several formats varying in magnitude from crews on site 24/7, to monthly, quarterly or annual visits coupled with various internet and phone support options. These are normally combined with periodic training programs that give end users personnel essential knowledge rather than bringing the up to speed which was the case in the early years.
CASE STUDY
Almost a decade ago a large oil company decided that they had issues with their operation and maintenance of critical infrastructure. They had an impression that the integrity of their metering systems was not at the high level of excellence that they were renown for and had internally expected. This impression lead to an invitation to FMC Technologies and others to discuss alternatives for the operation and maintenance of around 200 measurement assets (159 onshore and 39 offshore) in the North Sea and Western Europe. For these discussions to take place the company developed the following Vision as the basis for the basis for this endeavour: To ensure that the technical integrity and quality of the systems associated with product movements, including measurement of oil, gas, water and environmental discharges consistently meets internal, partner and regulatory requirements while: maintaining outstanding competence, providing innovative solutions, and delivering superb execution / operational excellence. As a result of these discussions FMC Technologies obtained a multiyear contract to provide an extensive array of services on these assets. Following on from the company Vision for this contract a Mission Statement for the endeavor was derived. That Mission Statement is: To protect the companys reputation in terms of measurement through ensuring that that the companys fiscal, allocation and environmental metering systems are fully compliant with statutory and contractual requirements, while maintaining the highest HSE standards and helping to achieve the companies roadmaps deliverables. Implementation required an array of tasks to be immediately undertaken as the timelines and expectations were both tight and extensive. Initially the task required mobilization of a planning, scheduling, and field operations team; a consultancy and engineering team and a management team. The broad scope of sites and the breadth of the vision and mission meant that there was a large array of equipment that needed to be supported. The list of equipment includes the following: Flow Computers: S500s, S600s, Micro 5000s, Omni, Solartron 7915, 7951, 7925, 7922, 7955, 869M, Instromet FC500, FC2000, NetOil Computers, Daniel S103, Promac, Contrec Primary and Secondary Devices: magnetic flow meters, mulitphase meters, turbine meters, ultrasonic meters, orifice meters, v-cone meters, averaging pitot tubes, coriolis meters, positive displacement meters, differential
Page | 6
pressure and pressure transmitters, temperature transmitters, Multi-Variable transmitters, gas and oil density transducers Micrometer, Yokogawa, Solartron, Foxboro, E+H, ABB, Krohne, Roxar, Framo, Faure Herman, Brooks, Daniel, Rosemount, Peco, Heco, FMC Technologies, Micro Motion Sizes 1, 2, 3, 4, 5, 6, 8, 10, 12. Analysers: Oil In Produced Water analysers, Water In Oil analysers, Relative Density analysers (Densitometers), pycnometers, chromatographs, dew point analysers, H2S analysers Roxar, Daniel, Solartron, E+H, Siemens, Michell, Ametek, ABB, Hobre Provers and Master Meters (10 36): Sampling Systems: Welker, Jiskoot
Key to the contract was the Consultancy and Engineering components. Here there were also many issues to assimilate from the vision and mission statements. The overall scope of these two (2) components included checking and verifying calibration data, such as: Meter factors / linearity checks and uncertainty calculations Technical improvements / developing plans for system upgrades Implementation of Condition Based Maintenance Maintenance optimization (e.g. SAP) Audit support / Securing reputation Procedural efforts including, but not limited to such things as re-writing work instructions Protection of financial interest on fiscal custody transactions
Contracts such as this must have Key Performance Indicators (KPIs) to enable both short and long term analysis of the agreement. KPIs were set up for Health and Safety, Behavioural, Performance, Innovations and Savings. From these KPIs a collection of results are obtained and tracked over near term, mid-term and protracted time frames of the contract. This cascade from the Vision, Mission, KPIs and results lead to a pool of c ontract benefits. In this instance these benefits included a complete bespoke management system and infrastructure that could be continuously developed and changes simply implemented. These benefits included: o o o o o o o Operational infrastructure with greater up time and lower uncertainty Operating systems and procedures Planning and scheduling tools Work instructions Calibration database Measurement management (alternatively called MIS-measurement management) Audit Management
Page | 7
Case Study Results: This program is an ongoing contract and, as such, is still covered by commercial confidence. While the results of the program are proprietary and it is not possible to show the exact quantitative metrics, we are able to report below achievements: Decrease of mismeasurement Optimization of maintenance frequencies with resultant manpower and downtime reductions Increased cost transparency for operation and maintenance activities Introduction of condition based maintenance with resultant manpower and downtime reductions Production of charts and reports analysing corrective maintenance vs. planned maintenance Value added by understanding and supplying the customer with important information regarding installation integrity Successful 5 year track record with 100% score on maintenance compliance FPI Fundamental Process Indicators due to comprehensive scheduling methods utilized by FMC Technologies. More than 2500 days without any Lost Time Incidents (LTIs) or recordable incident.
CONCLUSIONS
The benefits and results from the case study show that employing a team to operate and maintain expensive custody transfer has definite payback. Staying ahead of the game, maintaining knowledgeable, focused and motivated people ensure consistent improvements in the use, operation and performance of the infrastructure contribute to better outcomes for both buyer and seller. Additional conclusions from over 40 years of Systems Integration experiences are broad. They include these: The purchase price of a Custody Transfer Metering System, while seemingly high, is only a small component of the Total Life Costs. Ongoing costs of operation and maintenance over a 20 year project life of petroleum custody transfer measurement systems, can predominate over the initial equipment investment. Good design should include methods to decrease the Life-cycle costs. In implementing life-cycle design it may mean that initial cost may be higher. Sophisticated diagnostic systems and equipment will reduce ongoing costs. In the case study example, condition based maintenance clearly reduced costs. Operation and maintenance of these systems requires specialists. Sustaining a group trained on the large variety of components comprising the systems is often a difficult task even if the metering is seen as core business. Contract operation and maintenance, where technicians are overseen by engineers, is an economical methodology in custody transfer metering. Systematic scheduling of preventive maintenance and corrective maintenance can lead to lower maintenance costs. A sophisticated reporting system enables greater transparency of total costs and allows decisions to optimize costs.
Page | 8
2011 CUSTODY TRANSFER MEASUREMENT TECHNICAL EXCHANGE MEETING 21st 23rd November 2011 Flow meter Validation and the use of Smart Diagnostics
Mr Jim McCabe, SGC Metering
1.
INTRODUCTION
Numerous types of flow meters have evolved over many years. These include traditional meters such as: Differential (Head), Positive Displacement (P.D) and, Velocity (e.g Turbine) flow meters. There is now a vast array of types of flow meters currently available for selection, including newer technology meters such as Coriolis and Ultrasonic flow meters. All flow meters, with the possible exception of the orifice meter, currently require a flow calibration to be carried out prior to installation to determine the meters initial accuracy. The performance of a flow meter however may change over time and regular meter inspections, and/or recalibrations may therefore be required to ensure continued meter function and optimum accuracy. This paper will list the various types of methods used for the initial calibration, subsequent validation and verification of a number of different types of flow meters. Over the last decade or so there has been a large increase in the use of new technology flow meters. The specific problems associated with the new technology and smart flow meters will be discussed, some possible solutions, and ways to overcome these problems will be presented.
2.
METER CALIBRATION
All flow meters must be calibrated at a certified test facility to determine the meter accuracy before they are put into service. Historically all flow meters had to meet national metrology requirements, or international requirements such as OIML R117. In Europe there is now a new EU directive, called the Measurement Instruments Directive (M.I.D) that came into force in 2009 [1]. All new gas and liquid flow meters, and new liquid metering systems, which include other system components, to be put into use in the EU for custody transfer or fiscal (tax) purposes now have to meet the requirements of this directive. Calibration is "The set of operations that establish, under specified conditions, the relationship between values of quantities indicated by a measuring instrument or measuring system and the corresponding values realised by standards" [2] A primary standard calibration is one that is based on measurements of natural physical parameters (i.e., mass, distance, and time). A secondary standard calibration is not based on natural, physical measurements. It often involves calibrating the user's flow meter against another flow meter, known as a "master meter," that has been calibrated itself with a primary standard. Flow calibration is the process of quantifying the accuracy of a flow meter, where accuracy is defined as the deviation from a known reference value. Calibration of a flow meter is typically performed in a laboratory at several different flow rates, densities, or temperatures. When a calibration is performed, the meters calibration factors are determined. The actual method of flow calibration that is used depends on the type of flow meter, the fluid being measured and also the metering application. The procedures used for flow calibration are different for gas and liquid flow meters and it is generally much easier to calibrate liquid flow meters than gas flow meters. Calibration methods, frequency intervals and accuracy requirements are defined in various international standards. Government standards also state specific calibration and meter proving requirements [3]. 2.1 Liquid Flow Meter
The calibration of turbine meters and positive displacement meters can be performed by testing against a reference device and this can be done by several methods: o Gravimetric - smaller flow meters can be calibrated gravimetrically by weighing the quantity of liquid collected in a vessel. The vessel is weighed and the weight (in air) of the fluid collected is noted.
2011 CUSTODY TRANSFER MEASUREMENT TECHNICAL EXCHANGE MEETING 21st 23rd November 2011
o o o Volumetric this can also be used to measure the level of fluid discharged through the meter. Conventional pipe prover or small volume prover (SVP) can be used to calibrate liquid volume flow meters. Master meter - turbine meters can be used as a liquid Master meter, or Transfer standard/s, to scale up capacity to allow testing of larger meter sizes.
The output of the meter under test is monitored while a known volume of liquid flows through the meter by using the reference device. The calibration can then also be subsequently validated by "proving" [4]. Liquid flow meters can be tested in an open system and traditional flow meters such as positive displacement and turbine meters can be calibrated using water. Such a calibration however cannot then be transferred to operation when the meter is used with other fluids or operating conditions. It is thus required to validate the liquid meter with the actual product at the conditions that the meter will be operating at in the field. 2.2 Gas Flow Meter
Gas meters have to be tested in a closed loop, pressurised system. Gas calibration laboratories typically use turbine meters as transfer standard reference meters where the transfer meters have themselves been validated using a primary standard such as a gravimetric system, gas piston prover or sonic nozzles etc. It is preferable that the same test facility is used due to possible differences between test facilities uncertainties, due to the type of reference meters, fluid, operating conditions and test procedures used etc. A fundamental metrological principle is to calibrate a meter against a device which is traceable to national standards, which has an uncertainty lower by an order of magnitude than the meter being calibrated. Example: If a meter has a stated accuracy of 0.1%, the test device used to test it should have at least a stated accuracy of 0.05%. This is often not possible for gas flow meters such as gas turbine or ultrasonic meters. Gas meter flow calibrations typically have more individual sources of uncertainty than a liquid flow meter calibration, resulting in a much higher typical uncertainty ranging from 0.3 to 0.5%. This is an order of magnitude greater than that achievable with the water gravimetric system. More than one reference meter and other associated equipment may have to be used operating in parallel to achieve sufficient flow capacity. The uncertainties of these meters and the associated equipment used during the flow calibration start to add up giving increased flow calibration uncertainty. A six, or eight-point flow calibration is normally used for a gas turbine meter [5]. 2.2.1 Orifice Meter
Fluid flow using an orifice plate meter can be accurately determined without the need for actual flow calibration of the meter. Orifice meter calculations use empirically derived coefficients and these are directly based on accurately measurable dimensions of the orifice plate and pipe diameters. Orifice meters are subject to periodic physical inspection of the meter tube and plate as detailed in standards such as AGA3, API 2530, ISO 5167 etc. An orifice meter system can be flow calibrated in special cases to determine the actual discharge co-efficient however it is not standard practice to do this. 2.2.2 Gas Ultrasonic Meter
Most gas ultrasonic flow meter manufacturers perform a Dry calibration on the meter prior to shipment. This test involves performing a Zero-flow calibration on the meter, where blind flanges are fitted to the meter and the meter pressurized with a known pure gas such as high-purity, 99.99% (white-spot) nitrogen. A check is then carried out by comparing the observed meter Velocity of Sound against a value calculated using AGA 10 or other calculation using inputs of gas temperature, pressure and gas composition [6]. Some manufacturers may then adjust the meter settings at this time to correct, or optimise, the initial meter settings. Some manufacturers are able to perform a flow calibration with air at ambient pressure, whilst others may require a minimum of several bar pressure to function correctly.
For fiscal metering applications it is normally required to perform a high-pressure flow calibration at a calibration facility as close to the intended meter operating conditions as possible. Calibration is generally conducted under the conditions available at the flow calibration facility. Since almost all of these facilities utilize natural gas flowing in a pipeline bypass arrangement, it is normally not possible to vary parameters such as line temperature, pressure and gas composition etc. When the ultrasonic meters are then used in applications where the operating conditions are different from their calibration values, questions are often asked if the calibration remains valid. Most manufacturers and organizations such as AGA, ISO, GERG etc have carried out tests to determine these effects. Calibration of the flow meter also normally includes the complete meter run upstream and downstream piping. This is the procedure that is also currently carried out for any re-calibration of the gas ultrasonic flow meter. A six-point flow calibration is normally used for a gas ultrasonic meter. The flow through the reference meter/s is measured, by the test facility, over the test period which is typically at least 100 secs, at each flow point. This test period may need to be extended, particularly at low flow rates. The flow through the ultrasonic meter is determined by counting the output pulses from the ultrasonic meter during a test period, or alternatively by periodically polling the meter over a serial communication link. The pulses from the reference meter are also counted over the same time period. At the end of each test, the flowrate value obtained from the ultrasonic meter and the reference meter are then compared. Three repeat measurements are made at each flow point and the acceptance criteria would as per AGA9 [7]. 2.3 Calibration Facilities
Many of the flow meter manufacturers have their own flow calibration facilities. These however are usually only suitable for testing of smaller meter sizes and generally use water or air as a test medium. These facilities generally have limited use for flow calibration of meters large meters with hydrocarbon fluid at actual process conditions and can only really be used to test that a meter is functioning correctly. There are a number of independent flow calibration laboratories throughout the world which can be used to provide a certified flow calibration of a meter using real product and at representative operating conditions. These calibration laboratories specialise in testing and flow calibrating a variety of types of flow meters including Ultrasonic, Coriolis, Positive Displacement, Turbine and even Differential pressure meters. Each calibration laboratory has its own capabilities and limitations.
3. METER VALIDATION
As fluids pass through flow meters some of these may be abrasive, erosive, or corrosive. Over time this can lead to wear of meter components and can subsequently cause a shift in meter calibration. Traditional flow meters such as turbine and positive-displacement meters need to be inspected and/ recalibrated periodically to determine the extent and effects of wear on the meter bearings, rotors and other moving parts, blade damage and possible contamination. Validation is the process of checking a flow meter by some means to confirm that it is still functioning correctly and to also determine that the meter accuracy is still within the required limits. Validation can simply involve a physical inspection of the meter, testing against a prover, testing against a calibrated meter or by physically removing the meter for re-calibration at a suitable calibration test facility with a primary flow standard.Most traditional liquid flow meters can be validated by proving in-situ to determine meter accuracy. These proving systems are specifically designed to count pulses from meters such as turbine or PD meter. These meters generate pulses directly as a result of the rotation of the rotor, or shaft, caused by the fluid flow through the meter. There is no simple rule to determine when to validate or re-calibrate a flow meter, since calibration factors and susceptibility are highly relative to the specific flow meter technology and application of the meter. The re-calibration frequency of a flow meter depends on a number of factors, including: o o o o 3.1 The type of flow meter The fluid being measured Operating conditions How often the flow meter is used time or volume throughput Gas Turbine Meter
It is difficult to validate gas turbine meters on site, although master meters or check meters can be used. See gas ultrasonic meter for further details. Standard practice is periodically to physically remove them from service and recalibrate them at a certified test facility on a prescribed basis. The calibration of gas meters transfers more readily from the calibration laboratory to site as unlike liquid turbine meters there are no product viscosity issues. 3.2 Liquid Turbine Meter
Once installed liquid turbine meters should be validated. This is normally accomplished using a prover of certified volume or with a master meter. It is very important that the proving be done under conditions (flow rate, pressure, temperature, and liquid characteristics) that are as close to those of normal operation as possible. Recommended test procedures are covered in the API Manual of Petroleum Measurement Standards, Chapter 4, [4]. 3.3 Gas Ultrasonic Meter
Gas ultrasonic meters generally have to be physically removed from service and periodically recalibrated at a certified test facility on a prescribed basis. There is no international standard or regulation that defines the recalibration period of ultrasonic meters for custody transfer. In the UK gas ultrasonic meters must be recalibrated after 6 months in service and again after an additional 6 months. If the flow weighted mean shift between each calibration has been less than 0.3% on two subsequent recalibrations then the meter would generally qualify for a relaxation in calibration frequency to 12 months. In the Netherlands large ultrasonic meters (20) only have to be recalibrated every 5 years, whilst in Germany the recalibration period is 8 years if the meter has not been installed in series with another meter.
3.4 Liquid Ultrasonic Meter
Liquid ultrasonic flow meters can generally be Proved in -situ to confirm meter accuracy. This new technology meter however presents some distinct problems for meter calibration and validation and this will be discussed in detail later in this paper.
4.
METER PROVING
Proving differs from calibration in that it is performed in the field under operating conditions. Liquid flow meters can generally be Proved in-situ to confirm meter accuracy. Proving is necessary to determine whether variations in fluid properties, such as viscosity, and process conditions cause a shift in the meters calibration under actual operating conditions. The proving process allows the user to correct the reading from the meter to provide the true quantity of fluid that went through the meter. When a meter is proved in the field, a meter correction factor is determined. The correction factor is multiplied by the reading from the meter to offset the meter measurement. If the operating conditions vary significantly during operation, the meter should also be proved at the different operating conditions, to determine if different meter factors are needed for each set of conditions. Proving is accomplished by comparing the reading from the meter to a reference flow volume to validate the meters flow measurement accuracy. Proving not only validates the meter, but also validates all of the equipment used to prove the meter (detector switches, valves etc.). Meters are proved at a frequency which has been agreed by the responsible government body and the Pipeline Operator. 4.1 Prover
A pipe prover is a precision length of pipe fitted with a detector switch at each end. The volume between the detectors is calibrated very precisely on a periodic basis, normally at least annually, by an independent certified company. During proving the fluid flow is directed through the meter under test and through the prover. This propels a sphere through the prover which is driven along the length of the pipe from one detector to the other. The pulses from the meter under test are counted between the detector switches, which are used to gate the pulses. 4.2 Master Meter
A master meter is a secondary standard that has itself been calibrated against a primary device and is then used to prove another meter. A master meter is normally either a turbine or positive displacement meter that has been proved at various flow rates on a similar fluid to that which the line meter will be metering, however this is typically calibrated in a laboratory setting. A master meters meter factor has not been established under actual operating conditions. The master meter itself is proved using a volumetric prover and is proved at least four flowrates through the range of the meter. Each proving requires 5 consecutive runs has to be repeatable to within 0.05%. The master meter can then be taken to site and used to prove another meter [3]. In recent years newer technology meters such as coriolis and ultrasonic have been regularly used as master meters. There is no agreement on master meter validation requirements. Some countries state that a master meter should be validated every 2 years. The international recommendation OIML R-117 states no less than every 4 years [8]. 4.3 Transfer Meter
A transfer meter is normally either a turbine or positive displacement meter that has been proved at various flow rates on a similar fluid to that which the line meter will be metering, on site and at operating conditions. The transfer meter is normally proved at site using a volumetric prover and can then be used to prove another meter. A transfer meters meter factor has been established under actual operating conditions.
4.4 Check Meter
An alternative method used is to install two meter in series and then continually compare the volume readings from each. An acceptable deviation would be agreed with the regulatory authorities and as long as the two meters are found to be within this limit then there is no need for removal or re-calibration. Operators in Norway, the Netherlands and in Germany have been successfully using this method for many years. An issue with this method is when there an unacceptable deviation is found it is often difficult to determine which meter is correct and which meter is faulty. Other techniques, such as the comparison of speed of sound or density or other meter diagnostics, can be used for arbitration in order to determine which meter is most likely to be at fault. In most cases in Europe it is common to use two different meter technologies such as an ultrasonic meter and turbine meter. The reason this is done is to avoid commonmode errors which may occur if two meters with the same measuring principle are installed in series. The two meters may be permanently installed in series. Alternatively a master meter, or transfer meter, can be installed in the line or configured to be in series by valving etc so that the master meter can be used periodically to check the duty meter calibration.
5.
NEW TECHNOLOGY FLOW METERS
Although classed as new technology, flow meters, such as coriolis and ultrasonic meters have been around for at least 30 years. These meters are now increasingly used to measure liquids and they also need to be calibrated, validated (proved) and verified. Theoretically as these types of meter have no moving parts they should not require re-calibration. Unlike PD meters and turbine meters, the coriolis and Ultrasonic meter has no elements that wear or change with usage. This presents the possibility of calibrating the UFM in a laboratory under conditions similar to where it is to be installed, thereby eliminating the need for an onsite prover. This however is still not currently normal practice. In practice however coriolis meters however can suffer from wear and contamination of the flow tubes. Ultrasonic meters can also suffer from contamination of the meter causing a shift in effective meter area. Coriolis meters are rare in that they can be calibrated using a water based gravimetric system then used on gas and the same calibration remains valid [9, 10]. There are significant advantages to calibrating a flow meter on an "easy" fluid such as water. A gravimetric calibration is relatively simple to accomplish, and minimizes the system uncertainty. A tank is filled with fluid and then weighed. The fluid is then discharged through the meter under test and the mass reading compared to the mass discharged from the tank. A typical water-based gravimetric system used to calibrate a coriolis meter is capable of 0.04% uncertainty (by mass) or better, with the only significant contribution to the uncertainty attributed to the linearity and repeatability of the scales used for weighing the water. This allows for a calibrated coriolis mass flow meter to achieve +/- 0.1% of reading [9, 10]. Larger meter sizes may still need to be sent to an independent accredited test facility for flow calibration using hydrocarbon fluid. Potential problems: When a turbine meter is proved, each pulse counted represents an incremental volume that is passing through the meter at that instant, the pulse is effectively occurring in real time. If there were a step change in flowrate, the pulse frequency from the turbine would almost instantly reflect that change. API Chapter 5 mandates that a liquid flow meter should achieve a run-to-run repeatability of 0.05% for a group of 5 proof runs [4, 11]. Traditional flow meters can achieve this but new technology flow meters have generally not been able to meet these requirements. There are some fundamental differences to consider between Turbine, Positive Displacement meter and new technology meters such as Coriolis and Ultrasonic. The key differences are as follows:
i) With a turbine meter each pulse counted represents an incremental volume that is passing through the meter at that instant in time, the pulse is effectively occurring in real time. If there were a step change in flowrate, the pulse frequency from the turbine would almost instantly reflect that change. A turbine meter produces a uniform pulse train output, see Fig 1. A good turbine meter will have a pulse-to-pulse standard deviation of better than 1-2%.
Fig. 1 - Uniform pulse train A turbine meter sees all of the flow through the meter. Ultrasonic meters however take periodic samples of fluid velocity along one or more sample paths. The number of samples taken is equal in number to the sample frequency for the sample period. Coriolis and ultrasonic meters do not output a continuous steady frequency output. Instead these meters generated a non-uniform pulse train output, see Fig 2. This variation can be seen as poor repeatability when trying to prove these meters. Due to this frequency variation consideration must be given in selection of proving methodology
Fig .2 - Non-Uniform pulse train ii) Pulses from new technology meters are not direct output pulses. These are manufactured pulses generated by the flow meter control electronics. These micro-processor generated pulses are therefore not real time and the output pulse lags the actual flow in the pipe by some fixed delay. This delay is caused by the interrupt driven processes, calculations and conversions performed in the meter electronics and is directly related to the processing speed and efficiency of the meter electronics. If there is a change in flow during a proof run, even within accepted API limits, the derived K-factor will have a bias [12]. A typical ultrasonic meter will have a pulse-to-pulse standard deviation of 2-3%. iii) Turbulent eddies caused by shear stresses, and other small variations in flow are normally present in the flow stream [13]. There are also natural high and low frequency fluctuations in the flow, see Fig 3 & Fig 4.
Fig. 3 High frequency fluctuations
Fig. 4 Low frequency fluctuations
This turbulence and the natural flow fluctuation are averaged by the turbine rotor [13]. The turbine meter sees a much reduced average. This produces very little data scatter due to inherent inertial dampening of the measurement element, see Fig 5.
Fig. 5 Turbine meter observed flow This turbulence produces random variations in the observed axial velocity along each ultrasonic path. An ultrasonic meter takes multiple reading per second this causes data scatter due to the dynamic ability of the meter to measure these minute variations in velocity [13]. The result of this is poor short term repeatability which is a function of this flow turbulence. This allows the meter to see all of the fluctuations in the flow and as a result the output from an ultrasonic meter will produce a greater degree of data scatter, see Fig 6. To get a true measure of the velocity, many samples over a period of time must be obtained. It may be necessary to use longer proof runs, with more samples, to obtain acceptable results for meter repeatability.
Fig. 6 Ultrasonic meter observed flow This problem does not affect coriolis meters so they tend to have a much better repeatability than liquid ultrasonic meters. iv) New technology meters use filtering, damping and response factors. These are used to smooth the output from the meter and to filter noise etc. The direct effect of using these is to slow the response of the meter and also leads to a further delay in the output pulse.
There is no standard method used by manufacturers that specifies what this delay is. Some manufacturers state that the flowmeters time constant as Short, Medium, or Long. Others give a time constant setting of 0 through 4, and it isnt linear [12]. The baseline K-Factor change when the flowmeter time constant settings are changed (1:5) was less than 0.01% for Coriolis meters. This figure however is generally much greater however for an ultrasonic meter [12]. Proving run repeatability is used as an indication of whether the proving results are valid. Proving run repeatability for coriolis and ultrasonic meters may not fall within the typical 5 run, 0.05% span of repeatability. API Chapter 4.8, Table B-2 gives the recommended prover volumes to achieve a 0.027% uncertainty of meter factor. The prover size required is in 5 runs for a 16 meter is 521 bbl. This is not practical so an alternative solution needs to be investigated [4]. The API Chapter 4.8, Table A-1 recognizes that an increase in proving runs may be needed in order to achieve a 0.027% uncertainty of meter factor for meter such as the ultrasonic meter [4]. The number of pulses generated during a proof is not the issue and increasing the number of pulses does not necessarily improve repeatability. It is the proving volume is the critical factor in achieving a successful prove and increasing proving volume improves LUSM repeatability. Conventional pipe provers with insufficient capacity or small volume provers are not the ideal type of prover to be used when proving new technology flow meters but many new users of coriolis and ultrasonic flow meters already have these proving systems in place. The critical thing is the proving process that is performed not the prover itself. The recommendation for proving of new technology meters is to use a suitably sized conventional pipe prover, although this may prove to be impractical and also too costly to implement. The preferred method is to use a Master Meter/Small Volume Prover (S.V.P) combination as a transfer standard. The Master Meter can be calibrated against the prover in -situ, using real fluid and under actual site operating conditions. This can then be used to prove the meter under test. Master Meter proving is accepted and recognized by API and described in the standard API MPMS Chapter 4.5 as a valid method to prove Liquid Ultrasonic Meters [11]. This proving method overcomes the problems associated with direct proving of new technology flow meters. By using this methodology sufficiently long proving cycle times using the master meter can be used to improve the meter repeatability and it can also be used in-situ [13]. This solution can be used on site and eliminates the uncertainties due to laboratory calibration of the master meter on different fluids at different operating conditions. Custody Transfer measurement applications using new technology meters need to be proven like traditional technology meter. Whilst the output from a new technology meter has been made to look like a traditional turbine or P.D meter they are not the same thing, therefore proving a Liquid Ultrasonic Meter is not the same as proving a turbine meter It is essential to have flow pulse signal processing respond quickly to minimize potential meter factor bias errors. Any change made to speed of response requires the meter to be re-proven, although this may not be necessary for coriolis meters. Practical steps that can be taken by users of new technology flowmeters to minimize the effects of meter response factors etc are [12]: o o o o o o Optimize the flowmeter response by reducing the response factor to the minimum recommended by the manufacturer. Set any filtering, stacking or any other damping factors to zero or minimum. Take steps to minimize any flowrate variations during the prove process. APIs MPMS Chapter 4 recommends upto 5% flowrate variation during a proof however this may not be large enough for new technology flow meters [4]. Use as large a prover as is practical. The bias error introduced by the delayed manufactured pulse is inversely proportional to the prove time between the switches. If a large prover is not an option, consider proving the flowmeter using the master meter transfer method described here, using a turbine meter as a master meter.
6. METER VERIFICATION
Meter verification establishes confidence in performance by analysis of the secondary variables associated with flow. Such verification techniques give a qualitative indication of whether the meter is working correctly. Some verification techniques include spin- down tests for turbine meters and speed of sound and transducer gain checks for ultrasonic meters. Sometimes calibration or validation is carried out on meters when only verification is needed and many times these terms are used interchangeably. 6.1 Orifice Meter Verification
Orifice plates can wear, become damaged, warped, or contaminated etc. Differential pressure transmitters can also fail. Traditionally orifice flow meters would need to be checked periodically to ensure continued accuracy and are subject to removal and periodic physical inspection of the meter tube and plate to verify continued conformance with the requirements of AGA3, API 2530 or ISO 5167. The inspection frequency could be once a year, but it depends on the specific application, the type of fluid being measured and whether the fluid is clean or dirty etc [14]. Orifice meter inspection criteria include: o Orifice Plate Edge Sharpness If the sharp edge of the orifice plate becomes rounded the orifice discharge coefficient may increase and the orifice meter significantly under-measures the flow rate. o Notches or Grooves on the Sharp Edge Notches or grooves on the edge of the plate can give errors in the range of -0.02% to +0.51%, depending on how deep the groove is. o Bent or Warped Orifice Plate For small deflections, the effect of plate bending may reduce the plate bore causing an over measurement error. However, for large deflections, the orifice bore is increased significantly and under measurement of flow rate is the result. American Gas Association, Report 3 and API 2530 A.G.A. reference that states that any test equipment should have twice the stated accuracy as the instrument it is being utilized to test. If a transducer has a stated accuracy of 0.1%, the test device used to test it should have at least a stated accuracy of 0.05%. A typical rule is the higher the accuracy, the higher the cost. At each testing interval, the orifice plate should be removed and inspected for: o Dirt, grease, any other build up o Orifice bore direction o Bore diameter using a round-tipped micrometer o Verification of the bore diameter with the stated record, the stamped size on the plate, and the bore diameter entered in the flow computer program o Bore condition, looking for nicks, gouges, dullness, or sharpness o Flatness o Thickness o Eccentricity o Seal ring condition
10
7. METER DIAGNOSTICS
New technology flow meters, such as coriolis and ultrasonic meters, all produce lots of data that can be used to evaluate the meter status and performance. This data however first needs to be converted into useful information that is easy to understand and interpret by the user. Meter diagnostics are very powerful tools and they can be used for maintenance purposes or meter verification on site. They can be used to determine if the meter is operating correctly, is still accurate or signal to the user if the meter needs repair or recalibration [3]. Diagnostics can also be used to provide supplementary information when investigating a problem with a meter. Meter diagnostics can now be used to detect the following conditions: o o o o o o o o o o 7.1 Contamination Corrosion Erosion Wall roughness changes Process upsets Bad installation Distorted flow profile Liquid Carry Over Gas breakout Stratification Coriolis Meter
Historically secondary parameters such as drive gain (drive power) were used to verify coriolis meter performance. This method however is not directly correlated with the flow measurement and is now considered as a poor meter diagnostic. A method of verifying Coriolis meters by using a known density fluid has been used successfully. There is a correlation between density of a pure fluid (water or air) and the meters specific flow calibration factor, but this method of verification can be difficult to implement. Coating and corrosion can impact the stiffness of the meter tubes. The meter stiffness can be compared to a factory baseline. Since the stiffness is directly related to the Flow Calibration Factor (FCF), this verification technique confirms that the meter is accurately measuring flow [15]. Smart Meter Verification provides real-time verification of the coriolis meter sensor and electronics and can also be used to identify any physical damage or degradation in the performance of the meter. There is currently no independent third party testing available to prove the smart meter verification technique. There is a significant amount of internal testing that has been done to prove that Smart Meter Verification can detect changes to the flow meter, and the manufacturer are planning to work with third parties to provide independent analysis. 7.2 Ultrasonic Meter
There are several types of ultrasonic meters available with different path configurations, number of paths, signal processing techniques and associated meter diagnostics. The path configurations available are reflection and chordal, or a combination of both. Most ultrasonic meters use chordal path arrangements with four, eight, or as much as 32 paths. With a fully developed flow profile a small number of paths will adequately cover the flow profile. If however the flow profile is not ideal then more paths may be required. There are many diagnostic features inherent in the design of ultrasonic meters that can be used to provide confidence that the meter is functioning correctly [16]. Diagnostics have historically been used only as a health-check and maintenance tool. The diagnostics cannot provide absolute quantitative proof of the meter calibration accuracy. Diagnostics however have the potential to allow calibration intervals to be extended. There are a number of different types of meter diagnostics but there are three main common groupings than are now being used [17].
11
NOTE: The typical values given for diagnostic parameters in this section are specific to a multi-path, chordal, design of meter and should be used for information purposes only. Appropriate values for diagnostic limits are dependant on many factors and should be determined empirically for each meter type and physical installation. Manufacturers should be able to provide suitable guidance for their meter and proposed installation. 7.2.1 Functional Diagnostics
These diagnostics provide information on the physical operation of the meter and can be used to determine if the meter is operating correctly [16] [18]. o Signal gain (AGC)
The signal gain indicates the level of signal attenuation. The meter uses automatic gain control (AGC) to ensure that the received signal amplitude is consistent and is the same, irrespective of operating conditions. The gain will vary depending on gas composition, pressure, velocity, and path length. The gain however should be fairly constant and unexpected changes, either gradual or sudden, can indicate transducer contamination or possible transducer failure. A high gain is an indication that the signal strength may be too low for proper measurement. o Signal to Noise Ratio (SNR)
The signal to noise ratio provides a measure of the quality of the ultrasonic signals received. A low S/N ratio can be an indication that the signal is being attenuated by the product. When this ratio drops this can indicate either low signal levels or high noise levels due to some process issue. A typical minimum value for a multi-path, chordal, gas ultrasonic meter is 10 [16]. Differences between the upstream and downstream SNR could suggest the presence of ultrasonic noise, possibly from a nearby control valve with a large pressure drop. o Performance
The performance is the simple arithmetic ratio of pulses received to pulses transmitted. This value describes how many of the ultrasonic signals are acceptable to be used for custody transfer flow measurement. The value is displayed as a percentage indicating how many of the transmitted signals are being used. When this ratio below a critical value, this can indicate that there is something wrong either with the meter or the process. A typical minimum value for gas ultrasonic meters is 25% [16]. 7.2.2 Process Diagnostics
These can be used to determine if stream conditions are stable and therefore suitable for custody transfer measurement. In an ideal situation the process condition diagnostics can be used to confirm that the installation provides a fully developed, symmetrical, non-swirling flow profile, and also confirm that there are low turbulence levels [16]. o Turbulence
Turbulent eddies caused by shear stresses, and other small variations in flow are normally present in the flow stream at he pipe wall, see Fig 7. A typical value, for a multi-path, chordal, gas ultrasonic meter is 2 to 4% outer chords (near pipe wall) & 1 to 2% for inner chords. These values will increase by a factor of three or more when a partial blockage of the flow conditioner occurs [19].
12
Fig. 7 - Pipe Wall Induced Turbulence When an inverted profile is observed this is often indicative that a large degree of swirl is present in the metering stream [16]. Note: This is applicable to a multi-path chordal ultrasonic meter only and other meter path configurations may produce a different result. o Profile factor
The Profile Factor is a dimensionless ratio of the inner chord velocities to the outer chord velocities. The profile factor provides a measure of the flatness of the profile which is a function of the entire metering assembly including meter, piping, and flow conditioning components as well as their location. A typical value for a multi-path chordal, gas ultrasonic meter is 1.17, +/- 0.5% [16]. A change in the flow profile may indicate a change in the fluid viscosity, and/or a change in the pipe wall roughness. Cross-checking with other diagnostic features may enable the operator to determine the source of the change and to estimate its significance [17]. A change in profile factor can indicate a change in the meter accuracy. o Symmetry
Symmetry is a dimensionless measure of the flow symmetry and compares the flow in the top half of the pipe or meter, with that in the bottom half of the pipe or meter. A typical value for a multi-path chordal, gas ultrasonic meter would be 1, +/- 1% [16]. o Cross flow
Cross-Flow is a dimensionless measure of the flow symmetry comparing the chords on one side of the meter to the chords on the other side. A typical value for a multi-path, chordal, gas ultrasonic meter would be 1, +/- 1% [16]. o Swirl angle
The Swirl Angle is an estimate of the swirl (to the nearest degree) [20]. A value of 0 degrees indicates no appreciable swirl. The ISO 5167 standard for orifice meters allows +/- 2 deg of Swirl and is an acceptable limit [21]. 7.2.3 Measurement Integrity Diagnostics
These can be used to determine if the measurement system is operating accurately and is within design specifications [16]. o Speed of Sound (SOS)
The gas speed of sound, or velocity of sound, is a measure of the meter average transit time. This can be used to verify that the absolute time measurement is working correctly. A check of the transducer health on each path can be made by comparing the velocity of sound measured on the different paths. Since each path has an individual VOS measurement and they are measuring the same liquid, the VOS should be the approximately the same for each path. The measured SOS may also be continually compared with the theoretical calculated from gas composition via an on-line gas chromatograph (GC) together with an equation of state, such as AGA 10, and the measured temperature and pressure of the gas [AGA 10]. For a typical gas pipeline, deviation between the measured and calculated SOS of greater than 0.21% may indicate errors in the 13
measurement of temperature and/or pressure, in the operation of the on-line gas chromatograph, or in the operation of the USM [22]. o Molecular weight
One ultrasonic meter manufacturer uses a patented method used to derive the gas molecular weight from the SOS, temperature and pressure. This could be used to determine the meter density. The algorithm was invented in the early 1980s and was enhanced with Virial Equations of State algorithms in the early 2000s. 7.3 Liquid Ultrasonic Meter
Similar to gas ultrasonic meters, liquid ultrasonic meters provide a variety of diagnostic information that can be used to verify accurate measurement. Diagnostics include fluid density, fluid viscosity, velocity of sound in the fluid, Reynolds Number of the flow stream, individual path velocities, standard deviation of path velocities, flow profile, signal-to-noise ratios, and signal gain [23]. 7.4 Turbine Meter
Proving turbine meters on a periodic basis against known volume allows us to determine the repeatability and meter factor at that instant. This however cannot detect problems that occur in real time when the meter is operational. Proving is an important process to guarantee meter accuracy but certain turbine meter problems will only be detectable after they have occurred. Turbine meters are influenced by many factors which alter their accuracy and repeatability. Common factors are debris collecting on blades or shroud, bent or broken blades, worn bearings, dry bearings, and hydraulic problems such as flashing, pulsation, swirl, etc. Subtle problems may only be found by monitoring the results of the successive proofs and then analysing these. It is common practice to use pulse integrity (verification) with turbine meters. This is a form of meter diagnostics as the two pulse trains are compared and are checked for timing, regularity, and phase. The pulse trains are checked for missing pulses and some versions also detect added pulses. There are very few actual diagnostic tools available for turbine meters. One such tool creates a baseline signature for the meter and then compares the dynamic reading in real-time against this during operation of the meter. The turbine meter signature is controlled by the following factors: the blade's angle and spacing between each other (linearity), bearing resistance in relation to rotational momentum (meter factor), and hydraulic flow characteristics such as excessive pulsation, flashing and swirl (repeatability). A pulsation factor is determined by monitoring the time for one revolution and comparing it to the previous revolution. The turbine meter diagnostic package monitors turbine meter irregularities on-line in a real time mode. It monitors the pulse period, the phase angle, the standard deviation, and changes of blade signature and rotational speed. This procedure detects turbine meter problems such as missing blades, clogged shroud, turbine wobble, bent blades, and normal wear and tear [24]. 7.5 Orifice Meter
Previously there were no DP meter diagnostic capabilities available. There is now however a patented self- diagnostic technique that can also be used to determine DP base meter status and performance [24]. Figure 8 shows an orifice meter with instrumentation sketch and figure 9 shows the (simplified) pressure fluctuation through the meter body. A third pressure tap (d) is added further downstream of the plate. This allows the measurement of two extra DPs [25]. These extra DPs can be used to determine if there are any issues with the orifice meter.
14
Fig.8 Pressure tappings
Fig.9 Pressure fluctuation
7.6
Process Instruments
Many process instruments, such as pressure transmitters, are of a digital design and also have in-built diagnostics that provide a means for early detection of abnormal situations in the process. Some instrument diagnostics work on the fact virtually all processes have a unique noise or variation signature under normal operation. Changes in these signatures may signal that a significant change in the process, process equipment, or transmitter installation will occur or has occurred. The transmitter creates a baseline of the process noise or signature and then uses this to detect changes in the dynamic process noise. The transmitter can also provide early warning of process, equipment and installation problems related to process pulsation. Some monitor slight pressure fluctuations of the fluid to detect a blockage in the impulse lines.
15
8. CONDITION BASED MONITORING
As specified previously, the standard practice for gas meters is to remove them from service on a periodic basis for a flow re-calibration. There are significant impacts of this process, such as the logistics and costs to remove and re-install the meter, shipping and calibration. There is also the downtime, witness costs and normally the cost of a spare meter required when the other is out of service for flow calibration. This does not include any safety or environmental issues related to the venting and release of hydrocarbon gases to the atmosphere. The UK Department of Energy and Climate Change (DECC) measurement guidelines have recognised the possibility of using diagnostics data as a means of extending the interval between re-calibrations [17]. Extract: Para. 4.8.2 Multi-path ultrasonic meters offers a number of inherent diagnostic capabilities that can be used to give at least a qualitative indication that the meter has not shown any drift in its operating characteristic. Should these self-diagnostic facilities become sufficiently well understood, it may be possible to extend the interval between meter calibrations beyond the current 12 month horizon; it may even ultimately be possible to abandon the strategy of removal and recalibration altogether, in favour of a continuous health-checking regime. This is in fact the ultimate goal of Government, Operators and meter manufacturers. [3] Historically, monitoring of diagnostics was by periodic interrogation of the meter to gather diagnostic data and the results were subject to individual interpretation. Each manufacturer has their proprietary version of diagnostic software package which could be used to log data from a meter intermittently and provide snapshots of the meter performance. Diagnostics are only useful if they are easy to understand and quick to interpret. Interpretation of the diagnostics often used to require analysis and diagnosis by an expert from the manufacturer. Condition Based Monitoring (CBM), sometimes called Condition Based Maintenance is a new concept for flow metering, although this has been used for many years to monitor vibration in rotating machinery. A CBM software package would operate automatically and in real-time and would allow multiple diagnostics to be monitored, and logged simultaneously, on a real-time basis. The diagnostics could be used to track key physical characteristics of the flow meter throughout operation. CBM would continuously monitor, compare and alarm when any of the key diagnostics parameters shifts by more than a preset limit. This could then be used to alert operators of future maintenance requirements, potential device failures, or unwanted process drifts before they even happen. If properly implemented then CBM could be used as a powerful tool for periodic verification, to decide if a recalibration is necessary or as a credible means of extending re-calibration intervals. CBM can also replace a calendar based re-calibration and maintenance philosophy, with a condition based methodology [16]. At this time there is only one fiscal metering application in the UK where a condition-based maintenance strategy has been adopted [16]. Independent studies carried out on behalf of the Pipeline Research Council Institute (PRCI) concluded that ultrasonic meters should not be recalibrated on a calendar basis. Meter diagnostics should be used to determine if and when meter recalibration is required [24]. Most of the meter manufacturers suggest the following approach for a CBM package: o o o o o Calibrate the meter Footprint the meter during the calibration a. Gather key diagnostics at each flowrate b. Store the calibration diagnostic data and use this as a reference Install the meter Validate the meter installation Footprint the meter after installation a. Gather key diagnostics at each flowrate b. Store the installation diagnostic data 16
c. Compare the installation diagnostic data against the reference diagnostic data d. Confirm compliance with established limits e. Alarm if out of limits Continually gather key diagnostics during meter operation a. Store all the operation diagnostic data b. Compare the operation diagnostic data against the installation data c. Confirm compliance with established limits d. Alarm if out of limits
This approach follows the current API recommendations, which encourages users to monitor operational meter diagnostic data and compare this to reference, or baseline data. The DTI Measurement Guidelines [3], state: ...for USM manufacturers to quantitatively relate changes in meter diagnostics to meter performance . If enough operators start to employ the CBM diagnostic concept then there may be a time in the not too distance future when there is enough empirical data available to correlate shifts in meter diagnostics with quantitative shifts in meter accuracy.
17
9. SUMMARY
Calibration, validation and subsequent verification of flow meters can sometimes be difficult to achieve in practice. This paper has discussed the various ways that flow meters can be flow calibrated and validated including proving and re-calibration. Proving can be used for validation and re-calibration of traditional flow meters. The same procedures however cannot always be used with new technology meters. This paper has also outlined the specific problems associated with these meters and presented possible ways to overcome these problems. The paper has introduced flow meter diagnostics and then finally discussed how these could be used within a CBM package to verify the meter performance onsite without the need for meter re-calibration.
18
REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Measurement Instrument Directive, Directive 2004/22/EC on measuring instruments, Official Journal of the European Union L135/1, 30 April 2004. International Vocabulary of Basic and General Terms in Metrology (VIM), BS PD6461, 1995. BERR, Guidance notes for petroleum measurement, Issue 7, 2003. API, Manual of Petroleum Measurement Standards, Chapter 4 - Proving systems. ISO 9951: Measurement of gas flow in closed conduits: Turbine meters, 1994. Zanker, K, J: The Calibration, Proving and Validation of Ultrasonic Flow Meters, ISFFM, 2006. AGA Report No. 9, (1998): Measurement of Gas by Multipath Ultrasonic Meters OIML Recommendation, R117: Measuring systems for liquids other than water, 1995. AGA Report No 11, API MPMS 14.9, Measurement Of Natural Gas by Coriolis Meter, 2003. Ravenswaaij, B, van: AGA Report No 11 Expands Market for Proven Metering Concept, NSFMW, 2004. API, Manual of Petroleum Measurement Standards, Chapter 5 Metering, Section 8 Measurement of Liquid Hydrocarbons by Ultrasonic Flow Meters Using Transit Time Technology, 2005. Elliot, K: Proving liquid meters with microprocessor based pulse outputs, ISHM, 2005. Cousins, T. Augenstein, D: Proving of Multi-Path Ultrasonic Flowmeters, NSFMW, 2002. Morrow, T: Effects of Abnormal Conditions on Accuracy of Orifice Measurement, ISHM, 2004. Cunningham, T, J: An in-situ verification technology for coriolis flow meters, ISSFM, 2009. Peterson, S. Lightbody, C. Trail, J. Coughlan, L: On Line Condition Based Monitoring of Gas USMs, NSFMW, 2008. Griffin, D (DECC): Guidance document: Condition-Based Maintenance on Gas Ultrasonic Meters, 2011. ISO/FDIS 17089-1 2010 'Measurement of fluid flow in closed conduits - Ultrasonic meters for gas - Part 1: Meters for custody transfer and allocation measurement. Proving Multi Channel Ultrasonic Flowmeters in Custody Transfer of Liquid Hydrocarbons, ISA 2005. Zanker, K, J: Diagnostic Ability of the Daniel Four-Path Ultrasonic Flow Meter, SE Asia HFMW, 2003. Zanker, K. Advanced USM Diagnostics, NSFMW 2004. Letton, W. Pettigrew, D.J. Renwick, B. Watson, J: An Ultrasonic Gas Flow Measurement System with Integral Self Checking, NSFMW, 1998. Laird, C, B: Accuracy Diagnostics of Liquid Ultrasonic Flow Meters, ISHM, 2009. Smith Meter, AccuLERT II, Installation/Operation Bulletin, MN02008. Skelton, M. Barrons, S. Ayre, J. Steven, R: Developments in the Self-Diagnostic Capabilities of Orifice Plate Meters, NSFMW 2010. Hall, J: When Should a Gas Ultrasonic Flow Meter Be Recalibrated, NSFMW, 2010.
[12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26]
BIBLIOGRAPHY [1] [2] [3] ISO 7278-2 Liquid Hydrocarbons - Dynamic Measurement - Proving Systems for Volumetric Meters - Methods for Design, Installation and Calibration of Pipe Provers. ISO 7278-3, Liquid Hydrocarbons - Dynamic Measurement - Proving Systems for Volumetric Meters - Pulse Interpolation Techniques. ISO 7278-4, Liquid Hydrocarbons - Dynamic Measurement - Proving Systems for Volumetric Meters - Guide for Operators of Pipe Provers.
19
CUSTODY MEASUREMENT STANDARDS GLOBALIZATION INITIATIVE

BY KHALID A. AL-FADHL Chairman, Custody Measurement Standards Committee Custody Measurement Unit Process Automation Systems Division Process & Control Systems Department Saudi Arabian Oil Company (Saudi Aramco)
1. INTRODUCTION: Saudi Aramco Engineering Standards are at the center of engineering intellectual know-how and serve as the safety net for building safe, reliable and cost effective facilities. These are the documents where Saudi Aramco 75 years of engineering experience are captured and maintained. Recently, a dialogue started on whether Saudi Aramco Engineering Standards are considered in alignment with International Standards or not. Some people have indicated that these standards, in their current state, are the main reason behind the exponentially increasing capital projects cost. In their point of view, abandoning the company standards and fully adopting International Standards is the easy and quick solution. Another party indicated that the company standards are well in alignment with international standards and the company only needs to optimize them through Globalization. This paper will shed a light on the Saudi Aramco's position on the Globalization of Saudi Aramco Engineering Standards issue and the efforts exerted by the Custody Measurement Unit and the Standards Committee to fulfill this initiative and achieve the desired goals.
2. WHAT IS STANDARDS GLOBALIZATION? Globalization is a process to enhance the alignment Saudi Aramco Engineering Standards with the international or Industry standards
This globalization initiative was a result of a benchmarking exercise of Saudi Aramco standards management practices conducted in 2007 by Standards Coordination Unit in collaboration with Consulting Services Department. The study was done against regional and international oil and gas companies such as: Gulf Standards Technical Committee for Oil & Gas Industry (GSO) Regional standards committees, the International Association of Oil & Gas Producers (OGP). European Standards Organizations (CEN), Technical Committee for Materials, Equipment and Offshore Structures for Petroleum, Petrochemical and Natural Gas Industries Standards Engineering Society International Standards Organization (ISO)
1 of 11
The result of this exercise revealed that Saudi Aramco is in the leading edge in terms of engineering standards management practices and systems. Saudi Aramco engineering standards are based on a mixture of industry and international standards with various additions and exceptions. Specific emphasis has been noticed on the value engineering exercise and the process of handling standards internally. However, it has been brought to attention in almost all mentioned above venues that Saudi Aramco has not been involved enough in the development process of international standards. Chairmen of several technical committees in regional and international standards development bodies have expressed their eagerness to have Saudi Aramco as an active member. Subject matter experts of Saudi Aramco are being recognized as of high caliber. It has also been concluded from the benchmarking exercise that some additional requirements may not add value to the design but rather increase equipment cost. Manufacturers add an additional cost for designing non-standard equipment. To be able to reduce company specific requirements, subject matter experts have to work on injecting the company legitimate additional requirements into international standards. By doing so, cost will be dramatically reduced as manufacturers will be designing and producing standard equipment with minor modifications to internationally agreed upon specifications. Company requirements represent a source of practical field experience and provide valuable input to the continuing revision of the industry or international standards. Companies are encouraged to make their input directly into the working groups developing international standards.
3. BENEFITS OF GLOBALIZATION Some of the benefits that can be gained from this are: Eliminate special requirements that may increase project costs Increase the number of potential suppliers of material Reduce manpower required to maintain company standards Help build local industry around international standards to better compete in the global market Develop cost estimates more economically.
4.
SAUDI ARAMCO ENGINEERING STANDARDS STRUCTURE Saudi Aramco Engineering Standards are the center of engineering intellectual knowledge and serve as the safety net for building safe, reliable and cost effective facilities. These are the documents where Saudi Aramco 75 years of engineering experience are captured and
2 of 11
maintained. As part of its functions and responsibilities, the Board of Engineers chaired by Engineering Services Chief Engineer bears the responsibility of the review and approval of Saudi Aramco Engineering Standards and Specifications. Standards Coordination Unit of the Engineering Knowledge and Resources Division (EK&RD) is the company liaison office for Saudi Aramco Engineering Standards review, approval and waiver processes coordination. Saudi Aramco standards system manages more than 2300 standards; from which, more than 1200 are considered mandatory. Subject matter experts from various engineering disciplines are participating in the development and maintenance of these standards. There are 37 standards committees with members of more than 500 people from 17 different organizations including engineering departments, project management, operations, maintenance and operating plants (see Figure 2). Several procedures have been developed and maintained to manage the review and approval process of the engineering standards. Corporate management systems have been established in both Saudi Aramco SAP system and the intranet; i.e. e-Standards Website, SAP e-Standards Review and Approval, SAP e-Waivers Process, and SAP RVL/RSA Maintenance. Saudi Aramco standards management practices have been recognized nationally, regionally and internationally to be in the leading edge as a result of a recent benchmarking study. The study was done on a pilot group in cooperation with Consulting Services Department.
Saudi Aramco standards undergo a systematic approach in the development and revision process. The standard review cycle or model ensures that Saudi Aramco standards are continuously being revised and benchmarked. Through this review cycle, standard revisions are triggered by events at any time. These events could be feedback from the capital program, waivers or feedback from the standard users through our corporate standards web site.
In addition, revisions to standards are time-based. Standard surveys are conducted on an annual basis to receive feedback on standards and trigger revisions as necessary. In addition to that, standards are systematically subject to planned major revisions and value engineering studies every 3-5 years (See figure 1).
3 of 11
Figure 1: Standard Review Model
5. GLOBALIZATION APPROACH AND STRATEGY Saudi Aramco has started to get onboard all these mentioned above venues. Commitment and strategic planning is required to achieve the desired goals. The following approach will ensure the success of this globalization initiative. Develop a snapshot of the current state of Saudi Aramco standards Gap analysis is proposed to be conducted by Saudi Aramco Subject Matter Experts each in his discipline. The analysis should be done through the active participation in the different venues mentioned earlier. Performing Value Engineering studies for the engineering standards to improve Document VALUE through the use of Value Engineering Methodology in order to ensure clarity and ease of use, cost effectiveness and conformity to industry standards and guidelines. Also to ensure that cost are not increased by applying codes and standards that exceed the actual needs of SA Facilities Involvement in International Standards Development To achieve the goals and gain the benefits mentioned above, active involvement in the development of international standards has to take place. There are two different approaches to achieve this goal which is based on the type of external standards
4 of 11
development organizations. Industry Standards Development Organizations which are mainly developed by national legislations are considered to be easier to tackle. Saudi Aramco can be easily a member of any of those organizations. Subject matter experts can directly interact with these organizations and represent Saudi Aramco. Members are usually oil and gas companies, equipment manufacturers, national standards organizations, etc. Each member has the rights to vote on standards as long as they are actively participating in the development and review process. Examples of these organizations are API, ASME, ASTM, PIP, etc. However, international standards development organizations such as ISO and IEC have a slightly different approach. As ISO and IEC are considered an international standards development organizations, countries are the members. Each country is counted as one vote in the standards development process. Therefore, Saudi Arabia is considered one member and counted as one vote. Subject matter experts can participate in the ISO/IEC committees, however, they have to be designated as delegates by the national standards organization, i.e. the Saudi Arabian Standards Organization (SASO). To have a greater influence, Saudi Aramco has to participate not only through SASO but also through other venues. The Gulf Standards Technical Committee for Oil & Gas Industry in the GCC Countries is one of the very important venues. Companies in the GCC countries are aligning their activities and efforts in the area of standardization to have a greater influence in the international standards development organizations. Actively participating in the GSO TC7 standards committee and working groups will guarantee all GCC votes in the international standards development organizations technical deliberations. Another venue, which will also increase Saudi Aramco footprint in the area of standardizations, is the International Association of Oil & Gas producers (OGP). OGP has more than 50 members from the oil and gas industry around the globe. The OGP mission is to create value added standards for the oil and natural gas industry. The OGP strongly supports the internationalization of key standards used by the petroleum and natural gas industries. Its goal is to assist in the development of international standards that can be utilized for the broadest possible application. Success stories are publicly available and quantifiable (See Figure 2).
5 of 11
Figure 2: Involvement in International Standards Development
6. CUSTODY MEASUREMENT STANDARDS GLOBALIZATION INITIATIVE
A. Custody Measurement Engineering Documents:

Document No. SAEP-21 SAEP-22 SAEP-28 SAEP-50 SAES-Y-100 SAES-Y-101 SAES-Y-103 SAES-Y-301 SAES-Y-501 Document Title Project Execution Requirements for Saudi Aramco Royalty/Custody Metering Systems Tank Calibration Requirements Radar Tank Gauging Equipment Verification Requirements Project Execution Requirements for Third Party Royalty/Custody Metering Systems Regulated Vendors List for Custody Measurement Equipment Custody Metering of Hydrocarbon Gases Royalty/Custody Metering of Hydrocarbon Liquids Royalty/Custody Measurement of Hydrocarbon Liquids using Radar Tank Gauging Systems Dynamic Sampling of Hydrocarbon Liquids for Royalty/Custody Transfers
6 of 11
Document No. 34-SAMSS-112 34-SAMSS-114 34-SAMSS-119 34-SAMSS-120 34-SAMSS-121 34-SAMSS-122 34-SAMSS-167 34-SAMSS-168 34-SAMSS-321 34-SAMSS-525 34-SAMSS-841 34-SAMSS-846 34-SAMSS-847 34-SAMSS-851 Orifice Meter for Gas Custody Measurement
Document Title
Ultrasonic Flow Meter for Gas Custody Measurement Bi-Directional Meter Prover Uni-Directional Meter Prover Small Volume Prover Atmospheric Portable Tank Prover Truck Loading Metering Systems Truck Unloading Metering Systems Radar Tank Gauging Systems for Royalty/Custody Measurement of Hydrocarbon Liquids Automatic Sampling Systems for Crude Oil and Refined Products Flow Computer for Custody Transfer Measurement of Hydrocarbon Liquids Flow Computers for Custody Transfer Measurement of Hydrocarbon Gases Preset Controller for Truck Loading and unloading Systems Metering Supervisory Computer for Custody Transfer Measurement of Hydrocarbon Liquids
The above engineering documents are either developed primary in accordance with the requirements of American Petroleum Institute (API) Petroleum measurement standards. Some of standards are developed in-house because no reference document exists to meet the purpose and the scope of the document.
B. Development and Revision Process of Custody Measurement Engineering documents: The following procedure is followed to develop a new document: a) The engineering Unit, responsible for the discipline, will submit a request with the proposed scope to the Chairman of the Standards Committee to plan for developing a new engineering document. b) The request will be evaluated by the Chairman and inform the Standards Committee Members, nominated by the concerned engineering and operational organizations, of the new document so they plan for participating in the review of the document. c) The Chairman assigns a qualified engineer or Subject Matter Expert to develop the document per the appropriate procedure depending on the type of the engineering document. d) The author shall explore the existing international and/or Industrial standards documents that have similar scope and can be used as reference documents and the proper format of the document whether Exception or Narrative format. e) After developing the initial draft, the author will conduct peer review and solicit comments from the approved suppliers of the document subject. f) The author resolves the comments received from the peer reviewers and the vendors
7 of 11
and produce the document for committee members review. g) The author proposed a resolution for the committee members comments. If the comments resolutions were approved by the reviewer, the document will be revised to incorporate the resolution of the comments. If the comments resolutions require more discussion, the Chairman calls for a committee meeting to finalize the outstanding comments. h) The revised document with the committee comments resolutions will be submitted for BOE Review. i) The BOE comments will be resolved by the author and incorporates the approved resolutions in the document. j) The final document is sent to Standards Coordination to obtain the Management approval.
The above process ensures that no unnecessary requirement is mandated in the document.
C. Gap Analysis This analysis was made very easy because of the Exception format used for the original development of the documents such as the material specification documents of the proving systems. The other documents were developed using the Narrative format because there were no industrial of international standard documents addresses the scope of the Aramco engineering documents. The requirements of the document scope were extracted from several documents. Performing this Gap analysis requires conducting technical comparison between the document and the reference documents. The document shall be submitted to the concerned reference standard organization for review for possible adaptation.
D. Value Engineering Studies: In addition to the process followed for developing new document, all documents are planned for Value Engineering Study. Per the plan prepared for this phase, Custody Measurement Unit performed three value engineering studies for number of its engineering standards and Material system specification documents. Experts in the measurement discipline from internationally recognized system integration companies, material suppliers Saudi Aramco operating organizations in addition to subject matter experts from the engineering participated actively in these studies. The studies were highly successful and generated many proposals and recommendations that enhanced the compliance with the international standards and introduced cost avoidance as shown in the table below:
8 of 11
Document No.
Document Title
Date
No. of VE Proposals
No. of VE Recommendations 30
Estimated Cost Avoidance ($)
SAES-Y-101
Custody Metering of Hydrocarbon Gases 5/9-13/2009 Royalty/Custody Metering of Hydrocarbon Liquids
14
7,400,000
SAES-Y-103
18
33
30,000,000
34-SAMSS-119 Bi-directional prover 34-SAMSS-120 Uni-directional meter prover 8/2-4/2010 34-SAMSS-121 Small Volume Prover 34-SAMSS-122 Atmospheric Portable Tank Prover 34-SAMSS-112 Orifice Meter for Gas Custody Measurement 10/4-5/2011 34-SAMSS-114 Ultrasonic Flow Meter for Gas Custody Measurement
4 4 2 0 1
19 6 33 9 8
816,200 816,200 880,000 0 TBD
11
TBD
Per the requirements of the Value Engineering study criteria, the cost avoidance of the implementation of approved proposals are estimated based on five years period
Custody Measurement Unit is planning to perform two set of its engineering documents as shown in the table below:
Document No. SAES-Y-501 34-SAMSS-525 34-SAMSS-526 34-SAMSS-841
Document Title Dynamic Sampling of Hydrocarbon Liquids for Royalty/Custody Transfers Automatic Sampling Systems for Crude Oil and Refined Products Automatic Sampling Systems for Hydrocarbon Gases Flow Computer for Custody Transfer Measurement of Hydrocarbon Liquids Flow Computers for Custody Transfer Measurement of Hydrocarbon Gases Preset Controller for Truck Loading and Unloading Systems Metering Supervisory Computer for Custody Transfer Measurement of Hydrocarbon Liquids
Planned date
July 2012
34-SAMSS-846
October 2013
34-SAMSS-847 34-SAMSS-851
E. Participation in the International Standards Organizations Custody Measurement Unit participates effectively in the measurement related activities in
9 of 11
the international standards organizations. It assumes the voting power in the following standards organizations a) American Petroleum Institute (API) Petroleum Measurement subcommittees: COGFM: Committee on Gas Fluid Measurement COMA: Committee on Measurement Accountability COLM: Committee on Liquid Measurement COMQ: Committee on Measurement Quality
b) American Society for Testing and Materials (ASTM) Committee D02 on Petroleum Products and Lubricants
c) International Organization for Standards (ISO) Saudi Standard, Metrology and Quality Organization (SASO) as the Saudi national standard organization, delegated Saudi Aramco represented by CMU to review and vote, as a Participation Country, on the draft international standards submitted by ISO Technical Committee # 28, Petroleum products and lubricants. CMU also, when requested, reviews documents submitted by TC-30, Measurement of fluid flow in closed conduits.
d) American Gas Association (AGA) CMU participates in AGA activities by reviewing the reports developed in the field of gas measurement.
7. CHALLENGES:
a) Implementation process of the initiative differs from one discipline to another. Some disciplines are matured enough and complete set of standards are exist covering requirements of all applications. Others are not, such as measurement, where the existing standards do not cover all aspects of measurement. SA needs to develop standards that do not exist. b) Involvement of the Subject Matter Experts in the increasing day-to-day activities. c) Standards Cross Referencing. Although SA have initiated a task to reduce the standards cross referencing, but it is still a challenge to optimize the number of documents being cross referenced. d) Shortages in the local and international subject matter experts e) Geographical diversity does not permit attending International standards organizations working group meetings. SA member in the working group contribution will be limited to
10 of 11
review the developed documents electronically and he will not be able to persuade the other working group members to approve injecting SA requirements in the document. f) Long time for developing and approving the reference standard document. The need for the standard mandates that SA develops or revises its own documents without waiting for the international standard. This step will increase the number of SA standards instead of reducing them g) Marketing SA standards requires referring to international standards not SA documents. The SA engineering standard recommended to be for adaptation by an international or industry standards organizations shall be revised to either refer to an international reference standard or state the requirement, narratively, from the SA reference standard.
8. CONCLUSION:
To be standards leader instead of standards follower, Saudi Aramco through globalization initiative, succeeded to optimize the alignment of its technical requirements with the international standards. Also, some of the standards were adopted by GSO and others were submitted to ISO for adaptation. Saudi Aramco will continue implementing of the standards Globalization initiative strategies for the rest of its engineering documents to further enhancement of the alignment of the company requirements with international standards and explore all venues to incorporate its legitimate requirements in the reference international and industry standards through effective and professional participation.
9. RECOMMENDATIONS
All the necessary resources should be allocated to implement the Globalization Initiative requirements such as Participation of Subject Matter Experts in the International Standard Organizations activities. Conducting Value Engineering Studies Standards. Standards Coordination shall provide a certified VE facilitator to manage the study and issue the final VE Report.
11 of 11
Dr. Edward Shelton

Oil and Gas Measurement Training and Education: Knowledge Transfer and Employee Competency September 17, 2011
Abstract
Oil and Gas Measurement System organizations recognize that their workers competencies are an important source of competitive advantage but struggle or encounter difficulties when they attempt to identify, transfer and assess those competencies. Specific competencies are often ambiguous to managers, and individual perceptions of a firm's competencies may vary significantly. This lack of specificity may mask significant misunderstanding and confusion about competencies. When this misunderstanding or confusion occurs among middle managers who are responsible for the day-today management of these competencies, the implications for an organization can be serious. This paper outlines Measurement System challenges that most organizations will face over the next 10 years, as a large part of the knowledgeable workforce will be retiring or leaving. These separations from the skilled workforce need to be identified, collected, organized and transferred into the main body of workers in order to maintain operational efficiency. Developing a program that can identify, assess and transfer competencies is the challenge for Measurement System managers. As part of the strategy, Measurement System professionals must develop a model that identify key skill competencies for their managers' that not only improve performance but is proven to be an effective knowledge transfer vehicle. A proposed model is presented as a preferred methodoly that identifies skills and developmental activities that result in knowledge transfer and skill gap reduction. In addition, we describe a method that any organization can use to assess its competencies, and we show how management can use the assessment to enhance the firm's competitive advantage through knowledge transfer and skill development.
Introduction The concept of competency-based development has gone from a new technique to a common practice in the [1] 35 years since Cohen first proposed them as a critical differentiator of performance. Today, almost every organization with more than 1000 people uses some form of competency-based management. Major consulting companies, such as Sorenson Capitol Group, Development Dimensions International, and Price Waterhouse Cooper and thousands of small consulting firms and independent consultants have become worldwide practitioners of competency assessment and development; yet there is very little academic and applied research literature to draw upon for application and development. This has resulted in continued skepticism on the part of many Test and Measurement (T&M) professionals, as to the usefulness of competency models. Some of this is due to the observation that many of the competency validation studies have been done by consultants who have little patience for the laborious process of documenting and getting the results published. This paper is devoted to our understanding of competencies, how they drive performance and how they are developed. In it, we show the relationship between T&M competencies and performance in the T&M area, from an industry perspective. We also show how these same competencies are developed, assessed and developmental activities applied. All of the information and data in this paper has been applied by T&M professionals and has been integrated into a T&M business or operation. Purpose The purpose of this paper is to outline a model f or development of competencies needed for todays Test and Measurement (T&M) managers and leaders requires program design and teaching methods focused on skill identification and development. Competencies are defined and an overview is provided using Competency Mapping (Competency) model for use by most T&M organizations. Properly identified skill competencies predict effectiveness in professional, management and leadership roles in many sectors of T&M. A properly developed T&M Competency Model lays the foundation for this paper regarding skill competency and knowledge transfer. Competencies needed to be effective in T&M organizations can be developed, however, despite widespread application and need for such development; there are few published studies of the empirical link between supply competencies and performance. There are even fewer published studies showing that these competencies and skills can be developed. Background For organizations to succeed in today's competitive and complex environment, employees at all levels need to develop and demonstrate "the personal capabilities, underlying characteristics and behaviors that drive superior performance at work." Since their introduction two decades ago, competency models and competency-based processes have become the 'best practice' foundation of high-quality business for thousands of international business and public sector organizations. Competencies have stood the test of time - and are now the required approach for organizations seeking to integrate with the strategic and operational needs of the organization. (Table 1) Knowledge Transfer and Competency Development Over the next 10 years, there will be a large part of the global skilled workforce retire or leaving. These separations are from an aging skilled workforce that has a reservoir of tacit and explicit knowledge or competencies that needs to captured so to pass them on to the emerging and upcoming talent pool. In the most general terms competencies are capabilities that must be demonstrated to be effective in a job (Table 5). Technical competencies, like other types of competencies are identified through the study of jobs and roles. They are groups of skills that are measurable and linked to business objectives. A competency can be subdivided into raw forms of ability, either it is tacit or explicit. Oran and Weick suggests that tacit knowledge is highly personal and hard to formalize while explicit knowledge is formal and systematic, therefore easier to train for. Competencies are sets of related skills and skill groups around an underlying construct. Skills can be either task related or behavior based. Tasks being the easier to train and
[2]
develop opposed to behaviors. The identification and construction of competencies require the identification of specific skills and a measurement method that will allow for assessment of current status of ability. A large portion of the research in this area has centered on the Competency Mapping Model). The Competency model is a systematic approach used for identifying and developing skills, which includes assessing learners skills and evaluation criteria, using appropriate methods to foster effective learning, ensuring transfer, and evaluating the effectiveness of the learning intervention. The model will review mapping methodology supported by an assessment and reporting database. Competencies can be developed Research suggests that competencies enable companies to execute future business strategy, identify employees' areas of weakness, and employ targeted interventions. The responsibility of identifying corporate competencies typically falls within the central HR function, with some companies relying on individual business-units to identify technical competencies. In addition, competencies are used in connection with training and development more than any other process. Though profiled companies do not incorporate competencies into 360 Degree reviews, research indicates that companies utilize competencies in conducting skills-gap analysis as well as business process reengineering. There are several advantages of the competency approach to talent management, first the learner plays an active role in indentifying his/her competency level and in identifying appropriate developmental activities that will speed his time to competency. For example, it is the learner not the instructor who bears the responsibility for identifying what should be learned, determining the most appropriate methods, and evaluating the extent to which knowledge and skill acquisition resulted from the learning activity. Moreover, over reliance on the Competency model to direct developmental activities is that it does not give enough consideration to learner engagement and the impact on learner motivation. For example, Aguinis and Kraiger (2009), Brown and Sitzmann (2010), and Chen and Klimoski (2007) provide unique valuable insights into current research on knowledge transfer and competency development, but none of them discusses the potential contribution of learner engagement and learner motivation (Table 3). However, according to Strinivasan, the "honeymoon effect" of traditional skill training programs might start with improvement immediately following the program, but within months it drops precipitously which are the ultimate purposes of development efforts (Table 2). But showing an impact on outcomes, while desired, may also blur how the change actually occurs. Furthermore, when a change has been noted, a question about the sustainability of the changes is raised because of the relatively short time periods studied. The few published studies examining improvement of more than one of these competencies show an overall improvement of [4] about 10 percent in emotional intelligence abilities three to eighteen months following training Videl. Developing others is a leadership competency of increasing importance; however, many managers don't know where to start. The first step in competency identification and development is to put in place an employee development framework that encompasses processes, resources, and a culture that values learning and development. Many Test and Measurement (T&M) organizations have accomplished this but have failed to follow through by equipping their managers with the skills and knowledge they need to develop their direct reports. A fairly small investment in improving managers' skills can produce significant improvements in the effectiveness of employee development programs, offering an attractive bargain to companies in these tough economic times. Aramco for example has identified differences between coaching and mentoring, in developing a hierarchal structure to support the individual user and has adopted a broader focus to do a good job developing their employees. The T&M culture, starting at the top, has to attach importance to developing people. Setting expectations through the performance management process and reinforcing them through rewards and recognition are critical for embedding these priorities in the culture, and managers are held accountable for supporting the development of their staff. T&M Managers recognize the importance of having a competency model both from a knowledge transfer value and for defining career paths for employees; both insure successful job performance to include development and career advancement. Skill inventories for jobs and career path or career mapping provide the foundation for competency based training and provide workers with a developmental framework of skills that identify and describe skill expectations. The organization can choose from an extensive library of skills that have been identified and developed internally by designated Subject Matter Experts (SME) who are senior level employees possessing the knowledge and experience needed for effective transfer of knowledge
[3]
and continual high performance impact for the organization. The Skill Inventory collection is designed to identify all critical job skills and specifies the required and recommended training, experiences, documents, and reading, for each of four levels: Awareness, basic, Skilled and Mastery. These skill competencies focus on specific job related skills that impact directly in getting the work done and generally do not emphasize soft skills. Soft skills are mapped separately in behavior and leadership maps separated from the technically skilled applied maps. Competency Program Competency development and effective job knowledge transfer has its beginnings by developing a list of valuable competencies or skill inventory in the organization (Table 4). The selection of Subject Matter Experts (SMEs) is members who have a diverse range of experience and tenure within the organization and the industry. SME and participant confidentiality is important to avoid attempts to give the right answer, that is, responses favored by the CEO or superiors. Competency identification begins with the identification and building of Skill Groups by Subject Matter Experts (SME), followed by employee self assessments and supervisor/mentor assessments. Reviewing the assessments by both the user and the mentor/supervisor together allows for meaningful dialogue that allows the user to buy in and participate directly in the identification of resources best suited for them individually. This act of inclusion and participation explains the high rate of acceptance for such programs that are usually seen as the companys program. During this review process, an Individual Development Plan (IDP) is drafted that outlines and schedules developmental programs for the year or what period of time is agreed upon. The IDP completes the process and begins a continuous cycle of review completes to be determined by management. The primary competency retention goals are: 1. To accelerate speed-to-competence for employee technical skills. 2. To provide a methodology for supervisors to know the technical expertise of their employees and maintain IDP continuity. 3. To align employee technical competencies with industry-established standards. 4. To empower employees to take more ownership of their development by providing competency maps linked to IDPs.
Access to resources for learning and development Each identified skill that has been identified by SMES has a cluster of developmental activities that contain learning and development opportunities such as e-learning, reading assignments, classroom courses, seminars, work assignments, conferences and seminars, all are an essential part of a learning framework. Further developmental initiatives can include advanced developmental constructs such as professional certifications that are growing in popularity and are a valuable source of employee development. For senior employees, attendance at professional society meetings and conferences allows them to stay current in their field and are usually associated with the mastery level of development. These activities provide a good source of networking and help with business development as well.
Coaching and Mentoring An important function that is integrated into the competency mapping process is the use of coaches and mentors. This position is a challenge for many organizations due to the time and manpower required vs. what is available. However, organizations that understand and can have a dedicated mentoring program tend to have a strong talent management strategy to grow talent internally versus hiring from the outside which in turn will have a greater effect on succession planning through the development and performance of their employees. A mentoring mindset that gives priority to succession planning and developing people will better be able to achieve sustainability in their competency development.
[5]
Wicker explains that mentoring and coaching feedback skills have long been recognized as fundamental to good management of people, and in the past several years we have seen a welcome resurgence of interest in coaching and mentoring in organizations, but to be effective in developing their direct reports, managers need to understand and participate in the assessment, development process and have direct contact with the
mentor if delegated away. Being aware of skill gaps, willing to run interference for better training and recognizing individual assessments by constant interaction with the mentor and u ser is absolute. The users supervisor/manager should be involved in the entire process and is particularly critical in providing support, feedback and opportunities to apply what has been learned. This is accomplished partly by using a skills assessment tool which generates an Individual development Plan (IDP) and is a starting point for finding information about the employee's development needs. The key elements of successful development at the level of the individual learner are well established by [6] Shapiro, who outlines the following elements must be available to the learner: Information about the individual learning needs; the ability to set developmental goals identifies resources needed to achieve them, and prepare an effective development plan; support for development and feedback on progress and opportunities to try out and apply new skills and knowledge and to improve existing skills through practice. Assessment Process Establishing competency baselines can best be accomplished through employee self assessment and a mentor/supervisor assessment to establish a knowledge base for all employees that includes the participation from both the employee and an assigned mentor or supervisor. Employee self-assessments and supervisor assessments are conducted to provide objective data - the assessments are for development purposes so employees and supervisors will benefit only if they are honest in the assessment process, otherwise employees will attend courses that are too simple or too complex. There is no performance ranking about assessing someone for development. It provides the avenue for sharing and working together to develop the workforce in a non-threatening way. The skills inventory assessment identifies the skill competency level using a skill gap assessment tool and aligns the skill gaps to development activities that are available to the close the competency gap. This skill gap assessment tool is the most accurate way to collect data on development needs, which is a competency assessment, based on a collection of real-life skill situations. This program is costly in terms of manpower and time commitment. However, the impact it has is unquestionably effective and has proven to pinpoint areas that need development. A well designed comprehensive competency program accelerates the speed to competence for employees by identifying skill strengths and weaknesses then targeting the skill gaps with balanced resources and development. Developmental activities include in house training courses, seminars, readings, certifications, work assignments along with job rotations as means for improving skill competency. A competency assessment provides top management with a wealth of information about middle And front line managers' perceptions of their organization's competencies. The competency assessment process provides managers with valuable information and allows management to provide accurate manpower projected needs based on gaps in skills, knowledge transfer and talent drain. The assessments allow for the prioritization of training & development needs of the organization by identifying the skill gaps at the organizations level. For the front line supervisor, they have an accurate assessment that can be used to know what employees know and what they do not know. As supervisors are moved, continuity continues based on the objective data from the assessments. Management will also be able to identify the best technical talent in their organizations. At a glance new supervisors will know more about their peoples capabilities than what was possible before. Finally, a comprehensive Talent Management Program allows employees to identify their strengths and identify their weaknesses based on a detailed description of the work they are actually doing. It provides a career path for them to focus on and help them develop in the direction the company needs. Those with initiative can take advantage of the numerous self development activities linked to the competency. Setting personal growth objectives and planning development is the purpose of the individual development plan (IDP), which includes: Clear linkage between the competencies being developed and job discipline; skills that specify in measurable terms what is to be learned or developed; list of learning and development activities that will be undertaken to develop competencies for the skills and the identification of resources and support to implement the plan successfully.
Having good developmental activities is the most important part of the development plan. The objectives should be specific, measurable, and actionable. Managers should help their direct reports prepare their development plans. They can arrange stretch assignments, temporary or rotational assignments, and other developmental activities such as shadowing opportunities or special projects. Development support and feedback is perhaps the most critical aspect of the manager's role in developing others. An inherent tension exists between performance and learning in the workplace. It is difficult to find time for learning, and planned training or development activities often are pushed aside by urgent business needs. Managers can help make sure that sufficient time is provided to the employee for learning and development activities. Another way to deal with time limitations is to build learning into work. This avoids the "down time" associated with formal training activities but, perhaps most important, creates the most effective learning experience. Again, the manager is key in facilitating a learning process grounded in real work. The learning cycle often is short-circuited by a lack of feedback or by inadequate feedback on performance. Practice should be followed by feedback. The task should be repeated until a satisfactory level of proficiency is achieved, with feedback provided to the learner each time. Sometimes feedback is fairly automatic; for example, the frequency and type of questions posed by the audience during a presentation can pinpoint problems and areas needing improvement. In many cases, however, the learner must rely on others to provide specific information about the nature of problems or errors, why they occurred, and how to avoid or correct them in future repetitions To properly implement any talent management initiative requires a paradigm shift from traditional thinking. The organization must commit resources starting with a champion who can advocate the initiative across the organization and who will communicate the importance of the initiative as well as provide support to implement the initiative. The manager must take full responsibility of this program and work closely with a champion such as HR to insure adequate resources is available and that program integrity is maintained. Opportunities to apply new knowledge and skills is a important role of the manager and are critical since he/she has the ability to assign work that will give the employee a chance to apply new skills and knowledge including decisions about which projects or tasks within a project will provide the best opportunity for practice and opportunities to apply new learning such as special assignments, briefings or presentations, temporary details, or "stretch" assignments.
Future Application Application of talent management tools such as the competency model has proven to be effective in knowledge transfer and skill development, there remains a great deal of unharnessed potential with this [7] program. Nonaka & Takeuchi (2005) criticizes current management practices that ignore the critical importance of tacit knowledge (Table 5). They claim that most organizations overemphasize explicit knowledge and, therefore, often miss valuable strategic opportunities. They argue that innovation emerges from a cycle of middle-up-down knowledge transformation where valuable knowledge often emerges among middle managers and is pushed up to top levels. The most valuable knowledge is often tacit, and according to this argument, effective firms undergo disciplined cycles of making the knowledge explicit as it is communicated throughout the organization. The efforts to transfer knowledge, in turn, create valuable tacit knowledge that continues in this virtual cycle. Although it may be difficult to capture in a strictly codified form, the tacit knowledge underlying many of the most valuable competencies is often communicated through shared experiences or stories and narratives. Senior managers, therefore, can use the information gathered from the competency model process to manage the virtual cycle of developing competencies. For example, middle managers may be encouraged to share stories regarding key tacit competencies to exploit their most valuable elements. In addition, these stories may help transfer competencies within the firm, building a new set of tacit experiences that add deeper value.
Conclusion Experience in administering the competency model shows that T&M consensus on competencies is associated with higher performance. If organizations are serious about effectively managing knowledge transfer and competency development, managers should be directly involved and support a dialogue regarding their units competencies. An ongoing conversation and review regarding competencies allows managers to monitor their organizations' competencies, and consensus about competencies, in a timely
fashion. Organizations should institute a process that seeks employee engagement and participation in the application of the process. Future and forward thinking processes should include feedback surveys or forums that allow T&M Managers to communicate their support about valuable competencies as well as solicit current effectiveness and feedback for improvements. In addition, managers can utilize information technology and corporate intranets to facilitate online communities of users and managers throughout an organization. These processes may help anticipate the competencies the firm will need to develop for future success, as well as contribute to building consensus. While many organizations have made impressive investments in establishing a framework for competency development and knowledge transfer, they may not be realizing the full benefits because managers lack the [8] encouragement or knowhow to develop their direct reports or don't view it as a priority (Arntzen). Successful employee development to competency requires both a framework and a management team that understands the development process and possesses the skills needed to develop others. By making managers more effective in developing their direct reports, organizations can help increase the return on what are often sizeable investments in employee development.
For all industrial Test and Measurement organizations and the policy makers will have and will continue to have the heavy responsibility of shaping the formal and informal training and development of their Human Capital. With competitive globalization and the demand for product quality and safety, there is an increased need for organizations to become more active in being a positive change agent in the business environments in which they operate. Insuring that tacit and explicit knowledge is retained and passed on to the incoming generation of workers to create a powerful environment of motivation, commitment and trust. social and regulatory bounds for the definition Genuine interest in an organizations front line leaders and managers to improve training and development as well as accept accountability and ethical responsibility of its business development need and to recognize the gravity of their position as the heads of one of the critical fields in todays world.
Tables
Table 1 Employee Engagement Scale 1. I am willing to work harder when my development needs are being addressed 2. I am willing to devote myself of performing my job duties 3. I get excited thinking about new ways to do my job more effectively 4. I am enthusiastic about providing a high quality product or service 5. I am willing to go the extra mile in order to do my job well 6. My job is a source of personal pride and training and development is a critical part of my attitude 7. I am ready to put my heart and soul into my work 8. I get the message that I am important to the organization
Table 2 Competency Based Relationships HR System Assessment Based Competency Mentor Classroom Based Brick and Mortar Commitment Result
Employment duration Employment Relationship Practice Configuration
18 yr avg Transformational Emphasis on Personal Consideration
10 yr avg Transactional Emphasis on technical and standardization of organizational needs Work structured with low Low supervisor engagement Low follow-through
Assessment Assessment Assessment
Work Design
Work Independence High supervisor engagment High follow-through Individual Based Supervisor recognition Competency recognition
Assessment
Rewards and appraisal Emphasis
Individual and Group Social group recognition Transactional recognition
Both
Table 3 Motivation and Sustenance of Positive Behavior Work Environment impact Competency Based with Mentor or Supervisor Outside of workplace Group Impact
Motivation Commitment Engagement Trust Skill Development Equity
High High Medium High High High
Medium Medium to High Medium Low Medium Low
High Medium Medium Medium low Medium High
10
Table 4 Hydrocarbon Measurement Engineer Skills Liquid Metering Density Meters Liquid Flow Instrument Verification Liquid Flow Meters Liquid Metering Control Systems Liquid Metering System Valves Liquid Metering Systems Liquid Temperature Measurement Measurement Reports and Documents Metering Calibration Online Gas Chromatographs in Liquid Pressure Measurement Devices Provers and Calibration Sampling of Hydrocarbon Liquids Standard Volume Calculations
Tank Measurement Automatic Tank Gauging Manual Tank Gauging Standard Volume Calculations for Tank Gauging Tank Calibration Tank Sampling Verification and Calibration of Tank Gauging Equipment Calculation of Gas Quantities (HME) Calibration and Inspection of Gas Flow Meters Documentation of Gas Metering Tickets
Gas Measurement Gas Meters Gas Metering Control Systems Gas Metering System Valves Gas Temperature Measurement Devices Online Gas Chromatographs in Gas Metering Applications Pressure and Differential Pressure Measurement Devices Sampling of Hydrocarbon Gases
Weigh Scales Calculation of Petroleum Quantities using Weight Scales Calibration of Weight Scales Standards and Practices Measurement Industry Standards and Practices Operating and maintenance instruction Manuals Home Organization Standards for measurement Product Sample Testing Calibration of Analytical Testing Equipment Testing of Hydrocarbon Gases and Liquids Measurement Fundamentals Basic Measurement Concepts Custody Measurement
11
Table 5 Major Finding in Failed Test and Measurement Operations (2010) Company A 14,500 + Employees Company B 511,000 Employees Company C 1000 + Employees Company D Below 1000 Employees
Calibration standards and policies were not followed Poor training and development Money and resources were not adequate Shortcut mentality existed Lack of accountability
Shortcut mentality existed Failed leadership and guidance Poor training and development Lack of accountability Falsified documentation
Calibration standards and policies were not followed Failed leadership and guidance Poor training and development
Lack of qualified and trained workers Lack of accountability Poorly maintained testing equipment Money and resources were not adequate
12
REFERENCES
1.
Cohen, M.A. and Lee, H.L., 1989. Resource deployment analysis of global manufacturing and distribution networks. Journal of Manufacturing Operations Management 2, pp. 81104.
2. Orton, J. D., & Weick, K. E. (1990). Loosely coupled systems: A reconceptualization. Academy of Management Review, 15, 203223. 3. Strinivasan, K. (2001). Designing optimal for a fast growing specialty retail chain. 4. Vidal, C.J., Goetschalckx, M., 1997. Strategic production-distribution models: A critical review with emphasis on global models. European Journal of Operational Research 98, 118 5. Wicker, D. (2004). Reorganizing for lean enterprise: A case study of selected divisions within General Motors. Dissertation Abstracts International, 3033. 6. Shapiro, J. F. (1999). Bottom up vs. top down approach to modeling: Qualitative modes for (p. 737). Boston: Kluwer Academic. 7. Nonaka & Takeuchi, 2004. The strategic middle manager. San Francisco: Jossey-Bass. 8. Arntzen, B. C., Brown, G. G., Harrison, T. P., & Trafton, L. L. (1995). Project management and digital equipment corporation. Interfaces, 25(1), 6993.
9.
13
Comparative Study between Metering and Tank Gauging

Dr. N. A. Sherief, Measurement Engineer Saudi Aramco Mobil Refinery (SAMREF) Yanbu, Saudi Arabia Abstract: The aim of the paper is to reassure that tank gauging is equally accurate measurement method compared to metering thus eliminating additional expenses of metering or tank gauging station that may run into millions of dollars. Normally the auditors recommend measurement on both sides. There is no need of additional measurement facility on the other side to verify or confirm the measurement. This paper presents summary of large data collected over 23 months of custody transfer of crude oil took place between two companies. The measurement methods involved are manual tank gauging by a third party inspection company and metering by turbine meters. The measurement was performed methodically on every batch transfer of 400 MBD. Result of study: For the total transfer of 256 million barrels, the deviation between these two methods observed was 0.003%, metering being higher which is far lesser than permitted 0.5%.
1. INTRODUCTION:
The question of comparison arose when the auditors recommended : Assess the cost versus benefit of installing custody meters to validate the accuracy of Saudi Aramco ((SA) Custody meters for crude oil receipts. SAMREF (Saudi Aramco Mobil Refinery) buys crude oil from Saudi Aramco on regular basis for its 400 MBD export refinery located in Yanbu, Saudi Arabia. Both companies share common boundary. Saudi Aramco has its meter station with two 16-inch Turbine Meters with an online prover and an automatic sampler. SAMREF has two dedicated tanks for the receipts and every time a tank is scheduled to be filled, low gauge is taken, crude is transferred, allowed to settle for 4 hours and manual gauge is taken by third party inspector. Auditors from share holder companies (Saudi Aramco and Exxon Mobil) conduct audits periodically on measurement and recommend improvements in this aspect.
2. FLOW PATH AND MEASUREMENT SEQUENCE: The route of crude oil flow with respect to measurement is as shown below: 1. Crude transfer to Crude Oil Terminal (COT) via Aramco East-West Pipeline 2. Settling around 2 hours at COT, drain free water 3. Drain residual water of Samref tank. 4. Samref Gauge tank opening. 5. S. Aramco Start pumping. 6. S. Aramco Batch gross transfer through 2 nos. 16-inch turbine meter as custody transfer. 7. S. Aramco auto sampling (AS) for Sediment and Water (S&W) 8. Samref auto sampling of crude 9. Batch transfer complete, settle for 4 hours. Tank close reading and report Samref gross crude receipts 10. Drain free water and commence crude feed to refinery
Boundary
5. Pump 1.E/W Pipeline 6. Meters
2. Saudi Aramco Tank 7. AS Drain Fig.1. Crude Transfer Journey in 10 steps 8. AS Boundary 4,9.SAMREF Tank 10. CDU 3. 10. Drain
3. SCOPE OF STUDY:
The flow of oil, metering and tank gauging are taking place ever since the refinery started in 1985. A recent data has been listed and compared the crude transfers. The transfer data of the following months has been analyzed:
Year 2009 2010 2011 Total
Months March to December Jan-Feb; June to Dec. Jan- April
No. of months 10 9 4 23
3.1 Two months typical data.

Typical two months data is presented from March to April 2011 to show how the comparison is moving. The chart shows the trend. The actual monthly summaries for March and April 2011 are shown below.
March 2011 Data and chart Batch seq. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Percent 0 0.023 0.003 -0.03 -0.06 -0.04 0.453 -0.062 0.032 -0.01 0.033 -0.022 0.063 -0.067 -0.084 -0.084 -0.011 0.066 0.021 0.049 0.046 0 Pctx100 0 2.3 0.3 -3 -6 -4
Percentagex100 30 40 50
Percentage Deviation from Aramco
45.3 -6.2 3.2 -1 3.3 -2.2 6.3 -6.7 -8.4 -8.4 -1.1 6.6 2.1 4.9 4.6 0
20
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
-10
-20
Batch Sequence No.
DATA Batch # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Pct
0.000 0.027 0.149 0.148 0.050 0.024 0.057 0.029 0.053 0.049 0.072 0.006 0.102 0.026 0.073 0.001 0.024 0.077 0.031 0.083 0.029 0.075 0.024 0.139 0.000
April 2011
100xPct 0 2.7 14.9 14.8 -5 2.4 -5.7 2.9 5.3 -4.9 -7.2 -0.6 -10.2 -2.6 7.3 0.1 -2.4 -7.7 -3.1 -8.3 2.9 7.5 -2.4 13.9 0 Positive20
SAMREF
Gain
Percentage Deviation from Aramco
15
Negative - SAMREF Loss

10
Percentagex100
Series1
0 1 -5 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
-10
-15
Batch Sequence No.
3.2 Comparative Tables: Aramco deliveries versus SAMREF receipts in barrels for 23 months are presented the following tables. The total amount of barrels is 256 million. The Overall difference is 8297 barrels i.e., percentage difference is 0.0032% which is negligible quantity compared to 0.5% allowed between two measurement methods.
Legend: STO- Single Tank Operation. Not useful for comparison Year 2009 Aramco Samref Difference % Diff. Year 2009 Aramco Samref Difference % Diff. Jul 11,470,280 11,471,536 -1,256 -0.0109 Aug 11,159,892 11,162,879 -2,987 -0.0268 Jan. STO Feb STO Mar 10,239,439 10,244,892 -5,453 -0.0532 Sept. 11,398,349 11,403,043 -4,694 -0.0412
T/A- Turnaround - Refinery shutdown Apr 11,100,364 11,110,203 -9,839 -0.0886 Oct 11,961,105 11,959,804 1,301 0.0109 May 11,411,689 11,413,628 -1,939 -0.0170 Nov 11,099,799 11,112,623 -12,824 -0.1154 Jun 11,459,840 11,447,404 12,436 0.1086 Dec 10,576,207 10,581,068 -4,861 -0.0459 Yearly Total 111,876,964 111,907,080 -30,116 -0.0269
Year 2010 Aramco Samref Difference % Diff.
Jan. 10,775,680 10,775,261 419 0.0039
Feb 7,482,965 7,479,615 3,350 0.0448
Mar
Apr
May
Jun 11,513,688
Turnaround
Turnaround
Turnaround
11,512,194 1,494 0.0130
Year 2010 Aramco Samref Difference % Diff.
Jul 12,462,017 12,457,840 4,177 0.0335
Aug 11,858,690 11,843,140 15,550 0.1313
Sept. 11,398,557 11,392,991 5,566 0.0489
Oct 12,021,303 12,013,609 7,694 0.0640
Nov 12,049,454 12,049,115 339 0.0028
Dec 12,199,466 12,199,647 100 0.0008
Yearly Total 101,761,820 101,723,412 38,408 0.0378
Year 2011 Aramco
Jan. 11,608,953
Feb 10,282,275
Mar 9,999,374
Apr 10,539,179
Total 42,429,781
Samref Difference % Diff.
11,612,608 -3,655 -0.0315
10,284,088 -1,813 -0.0176
9,993,564 5810 0.0581
10,539,516 -337 -0.0032
42,429,776 5 0.0000
4. RESULTS: Summary: Year Aramco Samref Difference % Diff. Result: 2009 111,876,964 111,907,080 -30,116 -0.0269 2010 101,761,820 101,723,412 38,408 0.0378 2011 42,429,781 42,429,776 5 0.0000 Total 256,068,565 256,060,268 8,297 0.0032
Overall difference for 23-month comparative data is found to be 0.0032%. Allowable 0.5%.
Discrepancy: 8297 Barrels in 256 MILLION Barrels, i.e., 0.0032% Difference, Meter is higher. i.e., @ $100/bbl, total discrepancy (loss) over 23 months is US$83000. i.e., $ 3600 per month, negligible compared to metering station installation cost, its operation and maintenance expenditure. 5. CONCLUSIONS:
1. Auditors recommendation is answered.
2.
Recommendation: Assess the cost versus benefit of installing custody meters SAMREF to validate the accuracy of Saudi Aramco (SA) Custody meters for crude oil receipts. There is no need to install additional measurement station on the other side of the business partner. This saves millions of dollars of cost and maintenance of measurement station. Tank gauging and metering produce comparable results.
2011 CMTEM - Custody Measurement Technical Exchange Meeting

November 21-23, 2011
Prover Loops: Replacement, Design, Installation and Commissioning Practical Project Management Issues
Author: Michael L. McGraw, Project Engineer I Saudi Aramco E&PM, S. Area Project Management KAUST & Thuwal Town Development Project michael.mcgraw@aramco.com +966.2.664-2280 x8130 (work) +966.55.393-7829 (mobile) +1.847.305.4063 (USA VOIP)
Pg 1
TABLE OF CONTENTS
Table of Contents Executive Summary Definitions Purpose of Presentation Paper Project Management Life Cycle Phase Perspectives: Business Goals Design Goals Detailed Design Procurement Logistics Construction/Installation Start-up & Commissioning Operations & Maintenance Conclusions Attachments References
page 2 page 3 page 4 page 6
page 7 page 8 page 9 page 10 page 11 page 12 page 13 page 14 page 15 page 16 page 22
Pg 2
Executive Summary In 2006, the author was charged with conducting a post mortem that involved evaluating what had occurred to-date with the installation of a bidirectional prover at a refinery based outside Chicago, IL USA. The author was most familiar with FCC units, specialty equipment and operations and had been involved in all phases of worldwide refinery projects since 1980. This was the authors first time experience with the nomenclature, semantics, design and operation of custody transfer equipment. His expertise was and is in project management. Much of the background of the reviewed prover project is relevant to established, if not old refineries outside of KSA. These refineries still have or had multiple methods of cracking and had multiple process units that took fractions (olefins or aromatics) and further processed them. When process assets were either bought or sold, if it did not exist, the need for custody transfer equipment then had to be inserted at the assets battery limits. As physical plot space is always at a premium, workarounds to this issue were the usage of mobile provers. Knowledge of this type of prover was limited to a handful of plant personnel. Unfortunately, the old mobile prover was destroyed in a fire. As the need for flow verification was and is perpetual, a rush project began to replace the mobile unit with a permanent prover. In order to save costs, the refiner chose to outsource the work and have a local engineering firm manage the project. Surprisingly, this firm chose to procure and retrofit a used 4-way valve diverter valve into the new fabricated equipment. The prover was water-draw calibrated; however, months later, the prover was successfully mechanically installed but the start-up and commissioning of it proved too difficult. Today, the prover sits unused after never being placed into service. This paper is a project management story; it is not about process technology lessons learned. Instead there are management practices that can be learned from what transpired.
Pg 3
Definitions Bidirectional Prover: a positive displacement type pipe prover with flow and displacer traveling in both directions by means of a 4-way diverter valve (a.k.a., liquid flow provers or known traceable volume provers). Device has field calibration connections, vents, instrument taps, launch chambers, flanges and mechanical means to access the moveable spheres. Please reference Attachment 2 for an animated illustration of how this device operates. Please reference Attachment 4 for an example picture of a working unit in the field. CMTRs: Certified Material Test Reports are documents issued by manufacturers attesting to the origination and/or elemental composition of the subject products (i.e., metallurgy CMTRs from foundries). Coriolis Meters: measures flow by analyzing the changes in the Coriolis force of a flowing substance. The force is generated in a mass moving within a rotating frame of reference. An angular, outward acceleration, factored with linear velocity is produced due to the rotation. With a fluid mass, the Coriolis force is proportional to the mass flow rate of that fluid. Detector switches: activated by the passing of a physical sphere within a prover pipe. Four way diverter valve: a ported valve that allows flow to enter one launch chamber and exit the opposite chamber; thereby cycling the valve and reversing the flow direction. Fractions: refineries and gas plants create raw materials (naphtha, methane, ethane propane, butane, isobutane and pentane) that are used by the petrochemicals industry. Downstream thermal and catalytic cracking creates the building blocks of the petrochemicals industry that can be classified into two groups: olefins (ethylene and propylene) and aromatics. Olefins are in-turn converted into intermediate and final products such as, adhesives, detergents, solvents, rubbers and elastomers films and fibers, polymers and resins. INCOTERMS: a series of pre-defined commercial terms published by the International Chamber of Commerce (ICC) widely used in international commercial transactions. A series of three-letter trade terms related to common sales practices, the Incoterms rules are intended primarily to clearly communicate the tasks, costs and risks associated with the transportation and delivery of goods, i.e., EXW or Ex-Works (named place of delivery). Launch Chambers: pipe section prior to the pre-run and calibrated section used to launch and receive spheres; normally are two pipe sizes larger to allow flow by-pass around the sphere and allow the sphere to decelerate; allow easy removal of sphere; can be mounted at any angle. NDE: Non-Destructive Examination is a reference to radiographic, ultrasonic, magnetic-particle, liquid penetrant, and remote visual inspection (RVI) quality testing methods. OEM: Original Equipment Manufacturer. Propylene: chemical formula C3H6 , an Olefin as described above; three types are commercially sold with a mixture of propane: PGP (Polymer Grade Propylene min. 97% to 99.5% purity), CGP (Chemical Grade Propylene min. 92% to 95% purity) or RGP (Refinery Grade Propylene min. 50% to 65% purity). Quick opening closure: typically hinged or davit type; however many refiners now opt for blind flanges. ROI: Return On Investment is a financial calculation that calculates how quickly a specific investment will be re paid by profits from the calculated project; calculations produce a specific time value (i.e., hours, days, months or years). Sphere: balls normally inflated from 2% to 4% larger than the prover pipe inside diameter; filled with a 50/50 mix of glycol and water.
Pg 4
T&M: Time & Materials is a commercial arrangement where a vendor supplier OEM engineer contractor is compensated for all actual manhours (generally contractually pre-determined) and reasonable travel and living expenses. Provers: comprised of Bidirectional (as defined above), Unidirectional or Volumetric Tank styles.
Pg 5
Purpose of Presentation Paper The author believes that every project, either successful or unsuccessful, has lessons that can be learned from it. The prover project that will be discussed was a failure; it was waste of time and money (please reference Attachment 1) and today it remains unused. We can benefit from it because our ability to objectively review past decisions is at the heart of continuous improvement within the refining and petrochemicals industry. Throughout this paper the reader will see one common theme, many unqualified personnel working on a complex project without asking for help. This project had passive refinery management and weak design personnel who could not fathom the projects complexities. The purpose of this paper is not to decide which type of bidirectional prover would have been better, mobile or permanent; either, if properly designed would have been useful. The purpose is to reflect on the management of the project; their strengths and weaknesses, and their execution decisions.
Pg 6
Project Management Life Cycle Phase Perspectives: Business Goals The reasons for adding a custody transfer point are contractual and pragmatically based. Contractually, if a refiner is shipping too much product to a client, the refiner is losing money. Conversely, if the shortage is willful and the refiner is under-supplying a client, they expose themselves to financial damages. Both reasons support the need for accurate volumetric metering. With reference to the project being discussed in this paper, the author learned that process analysts were long questioning the accuracy of the PGP-RGP (Polymer Grade Propylene Refinery Grade Propylene) pipeline metering instrumentation. Following the fire and destruction of the mobile prover, analysts realized that the ROI on a new prover project could easily justify the project moving forward.
NEED AN EXAMPLE ROI CALCULATION FOR PROVER (MEASUREMENT INACCURACIES vs. PROJECT COSTS). PREFER TO HAVE ONE THAT USES PROPYLENE.
Pg 7
Design Goals With reference to the project being discussed in this paper, the author believes the inexperience and the unfamiliarity of the plant personnel coordinating the scope of work is where problems began to occur. Design considerations that were considered appear to have been largely based on the way the mobile prover had operated in the refinery before the fire. The installation of a permanent prover is quite different as the permanent infrastructure requirements are substantially greater. Examples of permanent infrastructure design requirements that were initially missed were: - Oily water sewers - Fire protection monitors - Security vs. emergency egress - Operational process interfaces It is no small task to have on hand qualified, experienced engineering project management personnel who have the competence to start work on a new process unit design. Further complicating matters is when the project is deemed easy and has schedule/business unit pressure to complete as soon as possible. The failure of this refiner to select either a chemical engineer or an instrument/controls engineer with capabilities in the opposite skill to manage the project resulted in the dismal project performance and eventual outcome. In lieu of this, if the engineer had at least had knowledge or experience in custody transfer, they would have at least selected a stronger candidate at the detailed designer. If the refiners project coordinator was known to be weak, then the last resort would b e to completely subcontract the effort to a known outside specialist. Advocating a low cost, multiple subcontracts approach to perform a complex project that the refiner will eventually be left with is a predictable path to mismanagement of a petrochemical project.
Pg 8
Detailed Design With reference to the project being discussed in this paper, the author believes the inexperience and the unfamiliarity of the contract designer performing the scope of work is where problems continued. Although the engineer selected to manage the project had a controls background, the individual did not have related refinery experience tempered by a process-operational background. Provers have a high degree of complex mechanical, controls/instrumentation, piping, and process inter-relationships. Coriolis meters are in-line devices, provers are highly machined piping systems, and detector and pressure instruments are linked to computer controls. All of this equipment is affected by upstream and downstream operations. Appreciation of the causes of turbulence within the upstream piping and the measurement disruption caused by the presence of it is another factor that must be known in advance by the mechanical design department of the engineering firm. When the prover is opened, new requirements and/or obligations occur that have to have been known long before the project was detail engineered. Similar to how refineries worldwide have moved from open sample taps to closed sampling systems is how the operator and engineer must think about a prover. Issues that must be addressed are: - Personnel protection - Fire protection - Hydrocarbon leakage control - Emergency egress routes - Remote control room(s) coordination and control - Parking - Lighting - Computer controls in a closed environment with proper positive atmosphere or temporary sheltering with sunshade capabilities The contracting approach a detailed designer advocates can either help or hurt the project itself. A designer has the ability to: - Do everything themselves (i.e., develop data sheets, prepare specifications, assemble requisitions, tenderbidaward contracts for custom complex instruments and components, fabrication contracts, etc., in addition to performing all site activities such as field walk-downs, above and below ground interference checks, etc.). This approach exposes the engineering firm to the most risk. - Do some of the above (preparing the deliverables and providing them to a refiner to procure). This approach exposes the engineering firm to varying amounts of risk. - Supply talent to the refiner who are based on-site or offsite. This approach exposes the engineering firm to the least risk. - Completely subcontract the effort to a specialist. This exposes the engineering firm to the least risk, is also the most expensive and consequently the least profitable approach. Without an appreciation for the above, the results of the project are likely to be poor.
Pg 9
Procurement Provers can be completely sole sourced or put together through a series of purchases. It is typical to purchase a prover skid from one vendor/fabricator, which also has the ultimate responsibility to supply a calibrated unit Ex-Works (an INCOTERMS contractual delivery method). When an end user (refiner-operator) or a designer begins to insist on free issue supplying related components, the approach is introducing risk to the project. The subject project did award the design/fabrication/assembly to a qualified and well-known OEM. Surprisingly, both the designer and the refiner chose to look for cost savings and the OEM offered a used 4-way diverter valve that could be re-conditioned. Although this decision did not affect the prover to be waterdraw calibrated, it was one of targets of the post-mortem investigation that all parties spent a lot of time researching. Ultimately, the used 4-way diverter valve was further re-conditioned and tested. If the prover project were to ever begin anew, it would be no surprise if the used 4-way diverter valve were to be ultimately scrapped and replaced. The OEM supplied the designer and refiner acceptable CMTRs, NDE and testing documentation in accordance with the contract requirements. The OEM did manufacture a new bidirectional prover, and successfully water-draw calibrated the unit. The OEM was obligated to witness and support start-up and commissioning activities for a period not to exceed of two weeks. Normally, this would be quite sufficient in a well-managed project. When commissioning proved difficult, the refiner operator asked the OEM to stay onsite to which the OEM refused without a commercial change order on a T&M basis. When this occurred, the refiner began to physically work on the prover themselves. This was wrong and should have been stopped. Upon doing so, the calibration and warranty of the prover was potentially voided. The OEM continued to support the project from their manufacturing location and supplied replacement spheres, seals, detectors and miscellaneous parts. In many respects, the OEM showed great restraint at the behavior of the engineer and refiner as they tried in vain to commission and validate the prover. In retrospect, although the communication could very likely result in a lost business opportunity, it would have been in the refiners best interest that the specialty OEM either refuse the tender package or privately discuss the issue unqualified management working on the project. This could have only been done if the OEM had a back door into a specialty group of the refiner and they were a valued supplier. Why the OEM offered a retro-fitted 4-way diverter valve is an issue that cannot be further discussed.
Pg 10
Logistics The OEM shipped the bidirectional prover skid to the refiners jobsite. There was no indication of shipping or excessive vibration related damage. Loose components that were too sensitive to stay physically mounted to the prover were carefully boxed, crated and shipped with the primary unit. Spare parts were also shipped with the loose parts.
Pg 11
Construction / Installation A qualified union contractor was employed by the refiner to finish the construction of the prover project components. Before the prover arrived, the contractor prepared the jobsite, demolished impediments where necessary, revised existing large bore piping and constructed foundations. Other than security fencing, no other site improvements were performed. At this time, someone should have had a candid conversation with the refiner about the absence of underground utilities. Why this did not appear to have taken place is not the subject of this paper. The bidirectional prover skid was off loaded and set onto engineered foundations at the refiners jobsite. There was no indication of deflection or physical contact related damages. Tie-ins to the existing pipe racks were performed without regard for 90 degree elbows in proximity to the physical prover. An analyzer shed was constructed in close proximity nearby the physical prover to house a customized computer that performed the comparison computations. No remote prover valve or other instrument controls were installed in this housing. Furthermore, no similar controls to remote control rooms and/or the main control room were ever made. One can only surmise that the designer envisaged an operational approach that was very similar to the way the mobile prover was operated. Chain-link fencing was added before mechanical completion of the project. Only one way of ingress-egress was built. In retrospect, bollards should have been constructed to facilitate personnel emergency evacuation. No fire water deluge monitors were added. Regardless of the absence of the above utilities, the proximity issues and the remote control concerns, the contractor was not a factor in the failure of this project.
Pg 12
Start-up & Commissioning Initially, the team of people who worked on commissioning the bidirectional prover worked for either: the OEM, the refinery (plant operators), the outside engineering firm and the refinery project management group. Unknown to some or all was the fact that upstream process units had auto adjusting control valves that modified flow when pressure swings occurred. There was no consolidated set of simplified process P&ID type drawings that tried to highlight every related valve and/or pump set. There did not appear to be an attempt to pre-discuss the potential process upsets that the commissioning and testing of the prover could cause. After the two week period of OEM assistance ended and the OEM left, the decision to continue onsite activities was flawed. As noted earlier, anyone performing physical work on the prover should have been prevented from doing so. Doing so gave the OEM the right to correctly state that the calibration and warranty of the prover were potentially voided. The questioning of the size and composition of the prover sphere was a basic engineering, designer, OEM issue; not a start-up and commissioning issue. Replacement of these further violated the physical integrity of the prover giving the OEM another excuse or out . Replacement of the detectors was acceptable as these were from tapped connections. Piping design turbulence was never even a known concern until the post mortem was performed. Shortcomings regarding the used, re-conditioned 4-way diverter valve were also unknown until the post mortem was performed.
Pg 13
Operations & Maintenance With reference to the project being discussed in this paper, the author believes that if the subject bidirectional prover had gone into service, had there been an H/S/E or fire event, the refiner would have be exposed to significant risk.
Pg 14
Conclusions Why should a paper and discussion take place about a failed project such as this? Because it points out the fact that continuous improvement is an ongoing daily-monthly-yearly challenge. Obtaining qualified personnel to work on complex projects is an ongoing daily-monthly-yearly challenge. Multi-angle process project management engagement between a refiner and its subcontractors is essential; that means process unit operators and specialty groups meeting in the beginning of any project to thoroughly look at a projects complexities and pitfalls. Resisting quickly moving into an ill-defined project should be rewarded, not criticized. Strangely, the author feels the correct project management execution should have been the replacement of the mobile prover with another mobile prover. Please reference Attachment 3 to view a relatively modern unit. The ability to un-hitch the truck and related potential ignition source away from the prover trailer and prover area is considered very desirable. The fact that the refiner tried to shoe-horn in the permanent bidirectional prover into the same area where the former mobile unit had used was flawed. This decision was likely the reason why the new site did not have adequate utilities installed and/or contemplated.
Pg 15
Attachments Attachment 1 discussion paper subject refinery prover pictures Attachment 2 example animation of a Bidirectional Prover with Master Meter flow Attachment 3 example of a mobile bidirectional prover Attachment 4 example of a large crude unit bidirectional prover
Pg 16
Attachment 1
Front view of installed bidirectional prover
Rear view of installed bidirectional prover
Pg 17
Pg 18
Attachment 2
http://www.ltssales.com/Prover.aspx
Example animation of a Bidirectional Prover with Master Meter flow
Source: http://www.ltssales.com/Prover.aspx
Pg 19
Attachment 3
Example of a mobile bidirectional prover
Source: http://www.weamco.com/Portable.htm
Pg 20
Attachment 4
Example of large crude bidirectional prover
Pg 21
References
Att. 2 - obtained from web http://www.ltssales.com/Prover.aspx LTS Sales, 421 West 12th Street, Unit 63, Erie, PA 16501 USA: 1-877-454-1818, Phone: 1-814-454-1818, Fax: 814-454-6363, E-mail: sales@ltssales.com; Permission granted 20Oct11 by Charles R. Myers Product Marketing Manager, FMC Corp.; email: chuckr.myers@fmcti.com P:814-898-5235 M:814-440-4751. Att. 3 - obtained from web http://www.weamco.com/Portable.htm WeamcoMetric, P.O. Box 9555,Tulsa, OK, 74157-0555, 1-800-331-3790, Phone: 1-918-445-1141, Fax: 1-918-445-4617 Att. 4 obtained from web
Pg 22

Papers: Technical Proceedings

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Papers: Technical Proceedings

Uploaded by

Copyright:

Available Formats

Technical ProceedingsTH

6 CUSTODY MEASUREMENT TECHNICAL EXCHANGE MEETING (CMTEM 2011)

November 21-23, 2011 Dammam Office Building Saudi Aramco

THE CHANGING ROLE AND RESPONSIBILITY OF THE FISCAL FLOW COMPUTER

FocQus (Burnison Engineering LLC)

Figure 3.0 Supervisory and Measurement connection to multiple Flow Computers

Figure 4.0 Measurement Data Management System

Natural Gas Metering With Ultrasound A New Dimension of Metering

Dr. Achim Zajc, RMG by Honeywell, Messtechnik GmbH, Butzbach, Germany

Metering Overview of Natural Gas with Ultrasound

Metering Overview of Natural Gas with Ultrasound

Metering Overview of Natural Gas with Ultrasound

Metering Overview of Natural Gas with Ultrasound

Metering Overview of Natural Gas with Ultrasound

Figure 1: Principal schematic of a transit time ultrasonic meter

Metering Overview of Natural Gas with Ultrasound

Metering Overview of Natural Gas with Ultrasound

Figure 2: Laminar and turbulent flow profiles

Metering Overview of Natural Gas with Ultrasound

Figure 3: End view representations of swirl, asymmetry and cross flow

Metering Overview of Natural Gas with Ultrasound

Figure 4: Typical installation requirements for a unidirectional operation

Metering Overview of Natural Gas with Ultrasound

Figure 5: Typical installation requirements for a bidirectional operation

Metering Overview of Natural Gas with Ultrasound

Pulse Detection Algorithms

Metering Overview of Natural Gas with Ultrasound

Figure 6: Typical ultrasonic pulse (raw signal) of a USZ 08

RMG by Honeywell Makes the Difference

Number and orientation paths

Metering Overview of Natural Gas with Ultrasound

Figure 7: RMG by Honeywell path orientation

Metering Overview of Natural Gas with Ultrasound

Metering Overview of Natural Gas with Ultrasound

Figure 9: Vector analysis of the gas velocity in a single level

Metering Overview of Natural Gas with Ultrasound

Figure 11: Completely metal-encapsulated titanium sensor

Metering Overview of Natural Gas with Ultrasound

Metering Overview of Natural Gas with Ultrasound

Diagnostic Output: Meter Validation

Transducer Gain Level

Transducer Signal-to-Noise Ratio

Metering Overview of Natural Gas with Ultrasound

SoS Comparison to Calculated Value

Metering Overview of Natural Gas with Ultrasound

Path Speed of Sound Comparison

Conditioning Based Monitoring (CBM): The RMG by Honeywell Way

Metering Overview of Natural Gas with Ultrasound

Evaluation of the flow profile: o Comparison of flow profile factors

Monitoring the swirl angle Live - RMG Precision Adjustment

CBM SoS Comparison of Each Path

CBM Signal-to-Noise Ratio

Metering Overview of Natural Gas with Ultrasound

CBM Live Comparison of SoS due to AGA 10 [10]

Figure 15: Signal-to-Noise Ratio (SNR) in relation to the gas velocity

Metering Overview of Natural Gas with Ultrasound

CBM Flow Profile Factors

Y-profile factor o Indicates the vertical portion of the flow profile

Metering Overview of Natural Gas with Ultrasound

Figure 17: X- and Y-profile factors in a cross path arrangement

Metering Overview of Natural Gas with Ultrasound

Metering Overview of Natural Gas with Ultrasound