Gas Turbine Meter Measurement Wayland Sligh Daniel Measurement and Control, Inc- Houston Texas Introduction ane aes industry has a wide variety of ceatSurement equipment to satisty nearly ‘ery application need. Metering is no €xception. The bulk of gas measurement in the U.S. is handled by four basic types. of meters: diaphragm, rotary, turbine and orifice meters. In any specific application, there are trade-offs to be made in Selecting one type of meter over another Knowing the strengths and weaknesses of each type of meter is fundamental to the gas measurement engineer or Manager of measurement Meter Types The four types of meters menticned above fail into two general categories Diaphragm and rotary meters are referred to as positive displacement meters, and turbine and orifice meters are referred to as inferential meters or (velocity meters) A positive displacement meter contains measurement compartments that alternately fill and empty as the meter rotates. These compartment represent a know quantity of gas. ‘An imerential mete! has no measurement june ot compartments. We infer the vo gas Pinat has passed through the inferential meter by some physical action. in the case of the orifice meter, the flow ig related to the pressure drop across the orifice plate. In the turbine meter, the volume 's ial to the number of revolutions The rotational speed tional to the ses through proportiot of the turbine rotor. of the rotor is directly propor velocity of the gas as it pas: the rotor blades. Operation As the gas enters the meter inlet, it is forced to flow through a passage created by the inside diameter of the meter body and the outside diameter of the diffuser. This passage has an area equal to about one third the area of the upstream pipe. Therefore, in order for the flow through this restricted area to be equal to the flow tnrough the inlet pipe, the gas velocity must increase by a factor of three. This increase in velocity in turn boosts the kinetic energy of the gas, making more energy available to turn the turbine rotor. ‘As the gas passes through the turbine rotor, it impinges on the face of the blades, causing the rotor to turn. if the meter was completely frictionless, it would take an extremely small amount of kinetic eneray to cause the rotor to begin turning. Ina mechanical drive turbine the blade will begin tuming at approximately 1% of the meters maximum capacity. As the gas velocity increases to 10% of the meter's maximum capacity, the available energy to overcome friction and drive the rotor increases and the mieter accuracy reaches acceptable limits. In the Daniel electromagnetic turbine, the energy required to turn the rotor is somewhat less because of the lack of retarding forces from the4 prime charact Is its ré ii Sngeabllity. This is the ratio of the teristic of the turbine meter which the turbine meter will accuracy. The present day gas pressure is the table 1 conditions to its specified most it atmospheric hown in retain rangeability of turbine meters at 10 to 1 or greater as SI maximum: ee meter Capacity to the minimum below. r the stated operating TYPICAL TURBINE METER PERFORMANCE MereT eae FLOW MIN. FLOW _ Approx. Men ‘ate Rate Differential SCFH RANGE SCFH Inches W.C. Mini Gas 5900 10:1 590 1.0 ‘ 18,000 15.4 1,200 12 36,000 15.1 2,300 15 S 60,000 20:1 3,000 18 150,000 20:4 7,500 17 Table 4 Meter Capacity The maximum capacity is determined by such factors as rotor speed, pressure drop, and gas velocity. The limitation of the meter capacity due to rotor speed is a relative matter. The bearings used in the turbine meter have a give life under a given load and at a specific speed. As the speed of the rotor is varied, the life of the bearings varies. Although the turbine meter is capable of turning at three to four times its maximum speed without immediate damage, its life can be substantially reduced if this speed Is maintained for prolonged periods. The pressure drop across the rotor al: influences the maximum capacity oft be turbine meter. In some applicati oe where line pressures running from 7 to 1 inches of water above atmos; ce hy pressure loss caused by inweinesan: equipment must be at a minimum The gas velocit factor due tothe aoc ong characteristics of the rotor blades mn also the need to avoid the possibiite obtaining sonic velocity within the re Pr.f= Density Pressure, ter nperature, specific ang pPecific gravil) ane compressibility of the gas all effect Chavos Const. So as these parameters anes Cause an increase in density, pe moles fe rate that is achievable ji le meter is de: increasing the rangeability. Prat be density Of the gas does not affect the maximum rated capacity of ameter. The imum flow rate is determined by the based on the manufacturer. ror the meter er me selecte bearings that were Installation Where space permits, @ turbine meter installation should include ten eh of straight piping ahead of the meter wil 1 straightening vanes and five diameters ©! straight piping downstream. alternative to this is the short-coupled installation. See figures 1 and 2. Figure 1. SSSOURERETE STURT aTon or an neseUltrasonic Gas Flow Meters The Field User's Perspective Charles W. Derr DANIEL MEASUREMENT AND CONTROL Introduction ither case promote the ation and the move to Operational cost (New for the sake of new" and “old for the sake of familiarity” may in é waste of resources. It has now been a number of years since deregul open sales. transportation and “hubs” has resulted in fierce competition. Ope hnician” ie ands. where practical, are a necessary part of success and indeed, survival. “Techiicett 'S an overly generalized and many times unappreciated title. “Technicians” are the Professionals that really make systems for control, compression, dehydration, odorization, and measurement a success or bad venture and highly influence a gas company's prosperity This document focuses on a ‘highly proven before release” new technology that offers great Savings to gas companies and provides some new challenges to the field professional. Ultrasonic meters are easy to learn and they add some new dimension and value to the user's measurement experience. Gas ultrasonic meters, (USM's), are here to stay The Ultrasonic Meter « Applications ) Ultrasonic gas flow meters are used to measure hydrocarbon gases in applications which require wide range (tum-down ratio), very high or low flow, low restriction, high accuracy or any combination of these from low pressure to super-high pressure. The two basic types are ‘spooled meter bodies or hot-tapped (in-situ). The spooled types are either heavy wall cast or pipe/sleeve fabricated. The two types of spooled meter bodies may be path-configured as single path or multi-path. Multi-path meters may be path configured as center-line bounce (which requires a calculated profile) or parallel straight across paths (which section the profile and do not require assumed or calculated profile correction) ‘Single path spooled body meters are used for check meters, compressor control, system balance, leak detection and receipt or delivery control. Jd body meters are used forall the applications of the single path and aes are, condlions of controled flow condoning and ab. eallratin rove "OY custody service. spooled body meters (usually three or more paths) are capable of the highes ee eaoy of custody transfer in pipeline gas measurement history. Path technigose ree be chosen based upon the application. Heavy wall cast meters have extremely smait pressure coefficients compared to “pipe” wall meters. If pressure is constant this Would not matter. If pressure varies significantly it may be a majorconsideration.d into an existing on-line pipeline. The a are or several percent, depending upon the tapping, ests me ese need X-Y-2 axis precision tapping positioning during welding ia" rrors which will cost approximately a one percent error per one-hy RD Voi aignment errr, pipeline tapped will have a relatively thin wall-vs-bore ratio and feats rrors from growth as a function of temperature or pressure change are a very s he saeration. P & T compensation of the pipe growth and it's effect on the path ae on ‘can be online compensated on some versions. API 12.2, chapter 4 single wal pr iculations demonstrate the pipe geometry error magnitude if uncorrected. Additive to this th elongation or shortening as well as the path bounce angle change, the resultant pa ‘ ‘ ‘ Compensating onboard electronics avoid most of this lost accuracy. How Ultrasonic Meters Work Sound energy waves travel or “propagate ”, by definition, at Mach-1 or “the speed of sound’, (S.0.S.). Sound propagates at different rates in different gas compositions and also propagates at different rates for different temperatures or pressures for any given gas composition. RATE is, by definition, Distance over Time. We commonly express that relation in feet per second, (fps). Some practical examples are: 1,150 to 1,400 fps Production Gas @ 80-90+% Methane. 1,400 to 1,470 fps Gas Plant High Methane Residue... Measuring the iime that it takes for a burst of sound waves to wavei a fixed and known distance between a pair of sound transducers is referred to as measuring the transit time. The centerline of the sound energy waves as they travel from the transmitting to the receiving transducer is the path. Ultrasonic transducers both transmit and receive sound. They are solid state “transceivers”. If you number a pair of ultrasonic transducers as: 1A and 1B; then, they alternate their functions of send and receive in order to clock the transit time each direction along the path. The sequence of transmitting and receiving or “firing” the transducers would be: 1A (Tx)...1B (Rx), 1B (Tx)... 1A (RX).....tepetitively. If there was no flow and the transducers were placed diagonally across the pipe then it would take exactly the same transit time to go each direction. The difference in time, or “Delta”, would be zero regardless of the total transit time each direction. The actual time in this example would be a function of the path length and the S.O.S. As mentioned above any change in gas composition, temperature of pressure will change the speed of sound. By examining the difference in transit time for the two directions the S.O.S. cancels out. Calculation example: Flow Rate = 0: Path Length =12.0” $.0.S. = 1,000 fps. ‘At 1,000 feet per second the transit time would be 1/1,000 or 0.001 seconds; which is the same as one millisecond ( thousandth of a second ) and which is also the same as 1,000 microseconds ( millionths of a second ).king with long strings of decimals the transit time is usually expressed in Incroseconds. Once again, at zero flow, it would take the same 1,000 microseconds to go Gach way and the transit times, ( T1 and 72), for the 1A to 1B and 1B to 1A transducers fare equal to a “delta-T” of zero. No flow, no time difference To avoid wo Continuing the example: If we measured the transit time while the ‘sound was traveling the ‘one foot path length against the flow and it measured 1,010 microseconds and then we reversed the transducer pair's transmit/receive functions and measured 990 microseconds with the flow; we then observe: The delta-T (1,010-990 ) equals 20 microseconds. Ultrasonic Flow Meter The upstream “shot” was 10 microseconds slower. The downstream “shot” was 10 microseconds faster. The USM's onboard electronics will then go through several steps to convert that delta-T into flow rate. It will evaluate the velocity by solving: Note: Path Length “L” is used in feet. The “X” value for this example is: 0.7071 ft. (8.485 inches). The times are in seconds. These dimensions; (feet for lengths and seconds for times) should give the velocity in feet per second, (fps). Velocity (fps) = L? (11-12) or 4 (0.001010 0.000990) ‘ 2X = (T1*T2) 0.5 (0.001010 * 0.000990) This solves to: Velocity (fps) = 0.00002 _or 40 fps 0.0000005This would be the path average velocity between transducers. If the transducers were mounted at a 45 degree angle across the pipe the average axial velocity would be the result of the Cosine of the 45 degree angle (0.707) or, the path (at 45 degrees to the flow) would only be observing 70.7% of the axial velocity down the pipe which would thereby be 1.414 times the path velocity or: 40 fps * 1.414 = 56.56 fps. Each path represents a certain portion of the total area of the pipe. This portion has a “weight” in the final calculation of the average axial velocity for the whole cross -sectional area of the meter. For custody transfer it is beneficial to be able to reconstruct the final flow values from the primary time measurements. This equates to reconstructing a flow computer's flow rate by recording D.P., T., P., and gas composition and using independent calculation to prove the imbedded flow calculations. Unpublished weighting information may prohibit reconstruction in case of measurement troubleshooting needs or custody disputes: After the average path velocities for each path are measured and the path centerline to meter axial centerline angle calculations have been performed, the weighting factor is applied to each path and then all path data is summarized into average axial velocity times the meter area and the actual volume flow rate is calculated and transmitted. Data is transmitted via serial data port, frequency and analog and may additionally be used onboard the advanced ultrasonic meter’s transmitter to further calculate and transmit the standard rates and totalized volumes (relative to base conditions). The transmitter may be capable of onboard data logging for reconstruction in the event of loss of flow computer or RTU. Measured Data Accuracy (Percent of Value ) ‘The designers and field users should recognize that the accuracy statements of the very wide tum-down ultrasonic meters are in percent of flow value or “percent of reading”. It is easy to underestimate the performance of this generation of gas meters when you are very familiar with percent of full scale devices. Consider the following example: USM operating range: 1.0 to 100 fps. USM accuracy statement: s8 % of reading velocity: 5 fps Test apaureey: +/- 0.25% ( Ex. error = one half maximum allowed ) The error equals: 0.0025 * 5 = 0.0125 fps ‘This results in twelve and a half thousandth of one foot por second on a meter that works to 100 fps! Further: 0.0125 fps / 100 fps =_an error of 0.000125 of full scale or 0.0125% of full £000" of full scale! These are realistic numbers for the upper echelon of USMS 11 ere ‘allowance means that in this example the meter read the velocity to 1/ 4,000" of full scale.The following example will @ equivalent accuraci intial based / evaluate the eqi ‘acy scenario for differential meters: The assumption will be th; ; me ni rf ni P. error m ; n ill be that the meter is mecha ically per fect ra i DF and only the D.P. error Meter: Perfect differenti vy 2 al Differential transmitter accracs Transmitter range: 1 DLS RSD, Test differential O21 WAC, Tesora! Nr uIeY - 0.1% ror in * W.C. (#1-.0.001* 180) = 4/-0.15"W.c Positive error: (15 + 0.15 = 15.15 * W.C. ) Sq. root of 15.15 = 3.8923 The Sq. root of 15 = 3.8729 3.8923 — 3.8729 = 0.0194 / 3.8729 = 0.005 or +0.5 % of reading une eM at low flow (5 fps ) reads the flow to +/- 114,000" of full scale. differential device at low flow reads the flow to +/- 1/200" of full scale. The accuracy of the ultrasonic meter in this case at 5 % 1 0 1 of full scale flow equals the accuracy of the differential device at 31.6 % of full scale flow. Ultrasonic meters achieve accuracy’s at low flows that convention technologies reached at higher mid-range flows. Appreciation of the accuracy level helps impose accuracy guidelines for proving and field performance testing Calibration The ultrasonic meter is dry calibrated at the factory. This consists of measurements of the meter body housing and the delay times of the transducer pairs ‘and electronics. The meter is blind flanged and multiple cycle purged with ultra high purity nitrogen. After thermal stabilization, the meter is check for zero flow offset and path speeds of sound -vs- calculated speed of sound for pure nitrogen at a precisely measured temperatures and pressure. ‘Small residual biases may be flow calibrated and corrected by a meter factor. The quality and test procedure control of the supplier has a great bearing on the relation of dry calibration accuracy to flow calibration accuracy. Manufacturers achieve their stated accuracy’s with dry When seeking third party proof or attempting to better the dry calibration calibrations. f accuracy, flow calibrating is the user's option. Piping Considerations ndations! Manufacturers Follow the manufacturer's minimum piping requirement recomme! ul have large fatermining the minimum piping requirements spent lage Su se eats the premier test facilities of the world. ‘of meter run piping requirements is 40-20 diameters upstream and 35 reo downstream. mete tually aciated by the number of paths, Flow cencinong may be required in some installations. They are not required in ALL installations. Multi-path . meters are more tolerant to flow distortions than most meters. Well formed smooth flow is always better than asymmetric flow regardiess of the application or meter technology.ae Field Verification The speed of sound is calculated and available from the ultrasonic meter. This value is calculated from the same measured t1 and t2 values by observing the total found trip time (tt plus t2) and dividing this into two path lengths, (the total round trip distance). By knowing the gas composition from a gas chromatograph and measuring the pressure and temperature; the speed of sound can be precisely calculated and cross-checked. Excellent agreement (to approximately +/- 0.25 % of S.0.S. calculated value) means the meter is working properly. Poor agreement means that either a measurement or data base item is probably wrong. This is an excellent test and does not interfere with metering operation. The field user should also, at whatever interval that is required for data base check on the other energy system components such as flow computers, etc., check for proper constants and that no diagnostic errors exist. Ultrasonic meters are surprisingly easy to use. They have rapidly become a friendly measurement component. Take advantage of training programs offered and enjoy working with them.1 agement - Mero nea oe 8906A 43" gt, FO Motion, Ine Tulsa, Oklahoma 74146 Abstract Coriolis meters have gained world in Naud applications since the early teens eee installed base Of more than 400,000 units. Newer ns have increased low Pressure Gros eases lOvow sensitivity. lowered incre: enabling performance oh ereased,_ Noise immunity or better than tradition: Coriolis also has attri logy can achieve. Sor these attributes are the meters immunity to fey disturbances, fluid compositional change, and it contains no wearing parts. With more than 25,000 Meters measuring gas phase fluids around’ the world, many national and international measurement organizations are investigating and writing industry feports and measurement standards for the technology \n December of 2003 the American Gas Association and the American Petroleum Institute co-published AGA Report Number 11 and API Manual Petroleum Measurement Standards Chapter 14.9, Measurement of Natural Gas by Coriolis Meter. jerview of theory, selection, installation, A conanee: and benef of Goris meters wil be ‘Application details will be presented to Hustrate both the range of natural gas applications, Including producto, fuel flow conto gas Power turbines, master metering. city/in ial custody transfer, and pace oeinsoring Laboratories include the Colorado Experiment Station inc. (CEES!), s Research institute Gomori (SwRI), and Pigsar (Germany) Introduction Coriolis is one of the fastest growing technologies in the Oil and Gas market. Newer designs and technology developments since the early 1990's have enabled Coriolis to measure gases that are extremely light, heavy, dirty, clean, sweet, sour, hot, Cold, andlor in a partial two phase state. AGA Report Number 11 specifically concentrates on the Measurement of natural gas mixtures within the normal and expanded compositional ranges called out by AGA Report Number 8, Compressibility Factors for Natural Gas and Other Hydrocarbon Gases, The low flow sensitivity of Coriolis meters has been dramatically improved in recent years allowing the technology to easily achieve flow turdowns of 30:1 Or more at pressures of 300 psi with turndown increasing as pressure increases. Al in all, it can be argued that Corialis technology solves more problems and offers even more value for gas than liquid measurement. This is because gases are compressible, and with more traditional gas technologies (orifice, turbine, rotary and Ultrasonic) process pressure, temperature, and gas composition must be accurately measured or controlled, the devices regularly maintained (Orifice plates, flow tubes, and transmitters checked. Turbine bearings, flow tubes, transmitters, and gear ollers checked, rotary gears, particle jamming, and gear oilers; Ultrasonic flow tubes, flow conditioners, and transmitters checked) and adequate gas flow testing performed on the technologies that are sensitive to gas density and flow profile. Since Coriolis measures the flowing mass of fluids its accuracy is independent of fluid composition, flow pulsations and flow profile/swirl. The meter is more accurate over a wider range of operating condition: less costly to install and maintain and man;applications and especially in 300 ANS! applications and higher. Coriolis is a smaller line-size technology: the largest offering from any vendor for gas applications 1s 4 450mm (6") pipe diameter. The pressure drop and flow range of a Coriolis meter draws a direct relationship to the actual flow area through the meter when comparing it to other metering technologies; i.e, the flow area trough a turbine meter is area not displaced by the turbine internals and rotor, the flow area of an orifice meter Is that of the orifice diameter. Because of this relationship a Coriolis meter will typically be one pipe size smaller than a turbine meter and several sizes smaller than an orifice while having similar pressure drops at flowing pressures in the 300 ANSI class and above. Therefore it is typical to see a 150mm (6") Coriolis meters installed in up to 200mm (8") line sizes. ‘4 2°meter installed in a typical gas installation Coriolis meters are very cost competitive with other metering technologies on an installed cost basis, where installed cost includes: Instrument purchase price Instrument laboratory gas calibration Temperature and pressure compensation Flow conditioning and meter flow tube requirements = Engineering and Procurement of _ these instruments ~ Labor to install metering equipment ‘When. operating costs are included into the evaluation of Coriolis compared to traditional technalogies high turndown technologies, Coriolis is the undisputed fiscally responsible meter choice in the 300 to 00 ANSI class in line sizes of 200 mm (8°) and below, Application “Sweet Spots” + Gas. delivery locations/Pressure cut easurement teBations where high, “Xo Meas controller noise iS a concen "2 «Line sizes 150mm and smaller High turndown requirements (201 50:1 is common), eliminating vara fretering runs of differential head meter, having to change orifice plates. «Dirty or wet gas where maintenance ig an jgsue and the “wet” which is lqug hydrocarbons are considered high le. . aro for adequate straight-runs (re, Turbine, Orifice, and Ultrasonic) Changing gas composition and flowing density (Turbine) Critical phase fluids such as Ethylene (C2H,) or Carbon Dioxide (CO2), where fluid density in nearly impossible to determine accurately on-line. + Custody transfer, process control, or system balances where mass based measurement provides a lower uncertainty. Theory of Operation A Coriolis meter is comprised of two main components, a sensor (primary element) and a transmitter (Secondary). Coriolis meters infer the gas mass flow rate by sensing the Coriolis force on a vibrating tube or tubes. The conduit consists of one or more tubes which are vibrated at their resonant frequen..y. Seasiny calls located an the inlet and outlet sections of the tube(s) oscillate in proportion to the sinusoidal vibration. During flow the vibrating tube(s) and gas mass flow couple together due to the Coriolis force causing a phase shift in the signals produced by the sensing coils. The phase shift which is measured by the Coriolis meter transmitter is directly proportional to the mass flow rate.b Q 3 % a % 6 joie that the vibration frequency is proport Fequency is proportional the flowing density of the thd. For gee speenante For gas applications, he lowing or “tve" density is not used tor gas measurement, but can be used as an indicator to change in a Coriolis meter's flow factor For a more complete discussion of the Coriolis theory of operation, please contact the author. Standards work, approvals, and research Coriolis meters have long been used for process Control, and a number of worldwide approvals and Feports or recommended practices exist for fiscal (custody) transfer. These include: AGA Report Number 11 APIMPMS 14.9 ‘APIMPMS Ch. 56 NIST (USA) C.0.C German PTB Dutch NMi Numerous other countries, including Canada, China, Brazil, Switzerland, Belgium, Austria, and Russia Dutch weights and measures (NMi) has performed testing and published a statement that the flow calibration factor established on water transfers without field calibration to gas phase appiications within a tolerance determined in their testing relative to the transferability of a water host 2 aan catibration In spring of 2001, Measurement Canada granted type approval to Micro Motion Coriolis meters for use 1n fiscal transfer of natural gas. Shown below are two recent calibration curves on 3° custody transfer meters. These are, being used in “Industry Gate” applications in Australia and the USA Laboratory is Pigsar-Dorsten, with natural gas at 725, Psi. Flow rates ranged from 21 to 438 MSCFH (0.8 to 10.5 MMSCFD) “Accuracies were better than +/ 0.2% over the 20:1 test range Installation effects testing performed by Southwest Research Institute (SwRI) and sponsored by the Gas Research Institute (GRI) in 2002 confirmed bent tube Coriolis meters to be immune, within the uncertainty of the SwRI flow iab, to upstream installation effects. The test results can be found in GRI Topical Report GRI-01/0222 Some of the installation effects test data is shown below Im 2004 tne Colorado Engineenng Expenment Station Inc. (CEES!) performed testing on the transferability of water calibration data to the measurement of gases under the sponsorship of the Gas Research Institute (GRI). Their findings shown below lead CEES! to conclude that “The single fluid Calibration tests show that a water calibration of @ Coriolis mass flow meter can be used for natural gas applications without loss of accuracy [ 27 lero Motion Corote MatVibration and fluid puls: During product development, extensive analysis and testing have resulted in meter designs that are inherently stable under 4 wide range of mechanical vibration and fluid pulsation conditions. Although Coriolis meters are for the most part immune to ‘mechanical vibration and fluid pulsations, they are ‘sensitive to vibrations or pulsations at the resonant frequency of the flow tubes. The resonant frequency of the flow tubes is meter design and fluid density dependent. Testing of a Coriolis meter subjected to ‘mechanical vibrations is show in the graph below. Note that the area of sensitivity is only at the resonant frequency of the meter's flow tubes. Testing of a Coriolis meter subjected to fluid pulsations is shown in the following graph. Note that the area of sensitivity is only at the resonant frequency of the meter’s flow tubes. Jn applications where mechanical vibration or fluid pulsations are present it is recommended that the manufacturer be consulted to determine the fesonant frequency of the flow tubes at operating ‘conditions, ‘Sizing end Selection, Selection of & Coriolis meter for gas application is quite straight forward, but different than traditional technologies. The flow range of a Coriolis meter le ‘Gatermmined on the tow end by how much error is in its-weight measurement and on the high end by t ‘maximum allowable pressure drop across the meter up to a maximum velocity limit called out by the manufacturer where measurement becomes unstable, but does not damage the meter. This is Quite different from traditional flow technologies where the specified minimum and maximum flow is highly dependent on natural gas pressure and the maximum flow velocity where measurement is lost and/or flow damage occurs to the meter. The two ‘major considerations when sizing a Coriolis meter are: Pressure Drop @ Maximum Flow + Accuracy @ Minimum Flow A Coriolis meter’s minimum flow is dictated by the meters zero specification and the minimum acceptable accuracy for a particular application. The following equation is the most utilized method for, determining the minimum flow rate of a Coriolis = meter. _ ZeroStabilty MinFlow Accuracy The maximum flow through a Coriolis meter is dictated by allowable pressure drop across the meter, fluid density, and a set of reference test conditions often found in the manufacturers specifications. The equation for calculating the maximum allowable flow rate relative to allowable Pressure drop is as toliows Coriolis meters can be installed upstream of 2 Pressure regulator, resulting in a smaller and less ‘expensive primary (sensor) and increased turndown) sample of the retatienship of line pressure v2 ,' to turndown is shown in the chart below, where the change in tumdown with pressure for multiple meters is graphed. * Conolis flow meters for gas measurement are currently available in line diameters from 2.5mm (110°) to 150mm (6°) inches. Velocity in the Coriolis Meter Some Coriolis meters have performance limitations ‘at high gas velocities due to noise imposed on the meter signal. Such signal noise can affect meter accuracy and repeatability. Ihe gas velocity at Which signai noise becomes a problem Is design {vendor} specific Seldom is signal noise a concer wher the gas velocity in the meter 1s below approximately 200 fYsec. Some manufacturer's can achieve much higher gas velocities with the use of advanced signal processing techniques. To define the maximum recommended velocity a Mach limit is usually provided by the meter manufacturer. From the standpoint of a high velocity gas eroding the metal of the flow tubes, high gas velocities are not an issue. The reason for this is that Coriolis meters are made of nickel alloy metals. For gas to erode metals, the metal must oxidize from moisture in the gas and the high velocity gas then erodes the ‘axide fayer. This is why erosion on carbon steel pipe 18 of concem for many piping engineers. Carbon ‘tee! is susceptible to oxidation from the moisture in the gas and therefore is susceptible to erosion from ‘high velocity gas. A Coriolis meter’s immunity to high Yelocity gas erosion is similar to that of and orifice Plate oF sonic. nozzle, in that they are made of stamniess steel oF other nickel alloys. ‘M abrasive contaminants are present in the gas flow ‘Stream, erosion of the wetted meter components ‘fay be & concern when the meter is exposed to Biescist ee plea aceaen eee RS as oe EW high gas velocities. thie specific and when present filtration rence eucation ‘ation is recommenced, Zoro Stability The zero stability value defines 1 1 which the meter zero may dit during operation is constant over the operating range. it may be given a8 a value in flow rate units, or a percentage of a stated nominal mass flow rate. The zero stability value is the limiting factor when establishing meter turndown ratio. The stated zero stability value is achievable when the Coriolis flow meter is installed, and re-zeroed at operating conditions. Because process temperature will affect the meter zero stability, the estimated value of the zero stability is usually limited to meters at thermal equilibrium, The affect of changes in this value is typically stated by the manufacturer. In most gas applications changes in process temperature are negligible, but to minimize the effect it is recommended that a Coriolis meter be zeroed at normal process temperature conditions. ‘Temperature and Pressure Compensation Both pressure and temperature affect the meter vibration chet acteristics, hence the magnitude of the sensed Corlls force. in comparison to zero stability, these effects are small. but should be coniprnsated for to achieve opumum meter performance, Most meter designs compensate for temperature effect automatically by monitoring the temperature of the flow tube(s). The pressure effect can be continuously monitored and corrected for using an external pressure transmitter, or by entering a fixed adjustment for the known average pressure. Some Coriolis meter designs periodically check meter sensitivity by applying a waveform reference force to the tube(s), during field operation, and compare the system response to that achieved under reference flowing conditions. This system will compensate for both pressure and temperature effects, Errors and compensation methods for pressure and temperature effects should be stated in the manutacturer’s meter performance specifications and included, if necessary, when establishing meter performanceinstallation (Mounting) Proper mounting of the sensor is required Consideration should be given to the support of the ‘sensor and the alignment of the inlet and outlet piping flanges with the sensor. A spool piece should be used in place of the meter to align pipe-work prior to welding the Coriolis sensor mating flanges if piping is constructed in the field Piping should follow typical industry piping codes. Meter performance, specifically zero stability, can be affected by axial, bending, and torsion stresses. ‘When these stresses exist they can be amplified by their accuracies are not-interreiated For fi applications, the on-line density from the meter is used to flowing of actual volumie\ This 1s useful for fiscal transfers of liquid petroleum and is often corrected to base conditions, such as barreis of o at 60 deg F using API volume correction methods For gas applications, the meter output can be configured for standard or normal volumetric flow units, such as MMscfd or NM'hr Since the measurement accuracy of fluid density by a Coriolis meter is relative to a liquid densitometer's accuracy, this measurement does not meet the accuracies requited for gas measurement. Therefore the sone measurement with gas, rather the relative gravity or base density of the gas is entered into a flow computer as determined from either sampling methods, o on-line gas analysis. It should be note: that the gas physical property information (AGAS Gross Method 1, Gross Method 2, or Detail Method) and procedurai methods required by a Coriolis meter are identical to that which is required by volumetric meters; i.e. Turbine, Orifice, Rotary, and Ultrasonic. Coriolis technology uses the following calculations to output a highly accurate standard or normal volumetric output. = Mass em P. NCM Masso. Gn Pauw NCM eos = Pox M rw Zex Rx Ts In the accounting of flow volumes with a Conolis meter flow computers should log flow weighted Specific gravity or base density The purpose for doing 80 allows for simple gas compositional Fecalculation of logged volumes using the following ‘equations© & % . “8 NCM. j : c= rola = NCM. * onoldy Gy very Relate Density Recaleviation Method N aw x» Povo1ay CM ycxem = NCM coy Prinew Base Density Recalculation Method Seip cet Rms tse acm nme es ana # Gow Con Lees nen may ya pty ttt er Operation and Maintenance Considerations ‘Other than the vibrating sensor flow tubes), Coriolis meters have no moving parts, requiring minimal maintenance. ‘There are three common types of field verification checks, which include meter zero verification, sensor Gagnasc cherks and t diagnostic checks. Performing these verttication procedures will ‘confirm accurate performance of the Coriolis meter ‘and when an out of tolerance condition exists where D> meter Zero Stability ‘Should be checked periodically and reset if it does not meet the manufacturer's specifications. Drift in Zero Reading Product buildup, erosion or corrosion will affect the vance. Product buildup (coating) may bias the meter zero. It should be noted that a zero ‘snift will affect a Coriolis meter’s accuracy more at high flows. This is dictated by the “pinFlow" equation called out in the previous “Sizing ‘ond Selechon’ section of this document, If the causing @ zero dnt, cleaning and re- ‘meter should bring performance back to its orignat performance specification. If coating of the sensor continues, the zero will continue tordrift Although rare, erosion or corrosion will permanenth affect meter calibration and will compromise sensor integrity. When used within the specified fad and ambient condition limits, fatigue of the sensing tubes of a Coriolis meter due to vibration during the stated meter lifetime is not of concem, and does not need to be considered when inspecting a meter. However, operating the meter in more extreme corrosive or erosive applications will shorten the meter's expected lifetime. Secondary Element (Transmitter) ‘A diagnostic LED(s) and display may be provided to indicate operating status of the primary and secondary elements See the manufacturer's documentation for detailed description of secondary element diagnostic and trouble shooting procedures. Density Checks ‘As of this writing, operating density measured by the meter should not be used to convert mass flow rate fo volume flow rate when measuring gases. However, itis useful as a diagnostic tool to monitor changes in meter performance, corrosion, erosion, or change in operating conditions Chacking and Acjus Improper zeroing will result in measurement error. In order to adjust the zero of the meter there must tbe no flow through the flow sensor and the sensor must be filled with gas at process conditions. The meter zero must be established at process conditions of temperature, pressure and density Even though the stream is not flowing, the flow meter may indicate 2 small amount of flow, either positive of negative. Causes for the zero error are Usually related to the differences between the calibration conditions and the actual installation which include the following + Differences between the calibration media density and the gas density «Differences in temperature ‘+ Differing mounting conditions The meter should read a mass flow rate that is less than the manufacturer's zero stability specification under the no-flow condition‘he zeroing of the meter must be performed at ‘nominal operating condition with no flow through the meter. Once it has been confirmed that there is no flow through the meter, the zeroing procedure specified by the meter manufacturer should be followed. Application Examples Coriolis meters have been used in a wide variety of applications, from the “wellhead to the burner tip” Coriolis meters are primarily a smaller line size ‘meter, ideally suited to these “sweet spots” Line sizes 200mm (8°) and smaller 300 ANSI through $00 ANSI High turmdown requirements: Dirty, wet. or sour gas where maintenance can be an issue with other technologies There is no room for long straight-runs ‘Changing gas composition and density Sudden changes in gas flow velocity (fuel ‘gas applications) * Pulsating gas flows (fuel gas and Compression gas in the use of reciprocating compressors) ‘+ Applications were abnormally high flow rates can occur Coriolis meters can be sized for very low-pressure ‘rap (16C" #420), but cen also be installed upstream Of the pressure regulator with high pressure drop to! increased useable tumdown without concern of damage or malfunction due to regulator noise. For instance, in one application for custody transfer of nitrogen, @ 50-psid drop (1390" H,0) was allowed ‘across the Coriolis meter and the pressure regulator adjusted accordingly. This allowed the use of a 1° Coriolis meter instead of a 3" meter downstream of the regulator and a 40.1 useable turndown (Better than 1% accuracy at minimum flow and an average 0.45% base volume accuracy over 95% of the upper flow range) (gas: Saudi Aramco uses a number of Conolts meters on both the liquid and gas Side of Separators. This application 18 of particular Role because the gas stream is wet, with entrained hydrocarbon condensates Measurement of this eed accurately quantifying the value of - the fas/entrained liquid hydrocarbon stream uel Control, A major US vendor of gas turbine designs a high-efficiency, low emissions offen; This design utilizes a tno of Conols meters t measure the natural gas burned in each of thre combustion zones (fuel “rails"). The combination o high turndown, high accuracy, immunity to vidvatios ina very high vibration environment. along with ease of installation due to no straight oie run Fequirement, makes Coriolis technology a perfect ft ster. One specific example of gas measurement capability is at a natura utility in Western Australia Two 3° meters are in parallel with a third used as a “hot spare, The justification for using the Coniolis meters was based on installed and calibration/maintenance cost improvements over the more traditional. turbine metering systems Since Conolis meters require no Straight runs or flow conditioning the installed costs were reduced by tive times. even with the parallel meters required to handle the highest flows Additionally. periodic maintenance costs were much Feduced due to the intrinsic reliability of Coriolis meters (le no moving parts) Siniariy, reliability improvements had a very po: ettect on calibration and proving costsmemat checks by the customer have shown yagreement to better than 01% on all gas transfers over a6 year period Western Australia: Previous installation using turbine 3 meters for 50:1 turndown “After” installation since 1996, with two operating and one “hot spare” meter for 80:1 turndown. Custody transfer between a utility and cogeneration Plant at 0.3 ~ 24 MMSCFD at 500 psia A storage field in Hungary wulzes 27 Coriolis meters for the injection and withdrawal measurement of natural gas. The storage of natural gas, multiple small wells are required. The valhdrawal gas t9 also fully saturated. contains H2S and during high flows the wells produce sand. In this ficult application gnty Coriolis meters can provide Dedenchonal ‘measurement, long-term accuracy, and tt fed for reserv achieve tumdowns tequir oir The graph below shows performance testing on a Coriolis meter from an identical metering application in Redfield, lowa; where the meter tested was subjected to saturated gas laden with H2S, sand, and iron sulfide over @ 9 year period. The post 9 year data shows the meter is maintaining an accuracy of 0.5% or better and still performing within the manufacturers specifications (081108 Comprosas Ar Tet, 280 pele, 70°F ‘Sn 13008, ntaeg $993 Natura Ges Gaver Borage (creconal use) Proving, The data shown below was taken on natural gas, but the meter was calibrated (ie. the meter factor was established) on water at the factory. Based on an extensive database of water in ges calbration data, there is no change inby itself otters very affordable method cf inte energy flow rates. Since proving any gas meter in-situ is difficult, the stability of Coriolis meters makes them ideal for use on gas. By utilizing the transferability of water calitration to gas and the meter stability over time an extremely accurate and stable metering system can be established. The following methodology was proposed by the Australian utility in the previous ‘example to establish traceability for high-value gas transfers, Establish the meter factor on water Validate the meter factor on gas (i.e. natural gas at Pigsar) + Periodically remove the meter from service and verify the meter factor on water Altnough this methodology requires that the meter be removed from service, it defines very accurately n-titu performance of the mete: Since steps 1 & 2 establish the meter traceability between water and gas, verifying water performance in step 3 automatically validates the meter in-situ (gas) performance and eliminates the high cost of gas validations. After some experience, it is likely that the period to repeat step 3 would be lengthened from every year to every two or three years AA vanation of this proving methodology is to use a Coriolis meter as a master meter. By establishing the traceability between water and gas measurement on the master meter, it can be used to prove other meters (of any type). Eneroy Metering: Coriolis meters can be an ‘excellent reality check on energy consumption, Energy per SCF can vary as much as 10 times that of energy per @ unit weight for natural gas mixtures. ¥ composition varies and an average relative density ana’or eating value is utilized for energy measurement Coriolis can achieve total energy ‘accuracies unparelieled by volumetric meters uutlazng the same average values. A Coriolis meter Nerval Ose [deat Hosting [cet Hestra] an, a | Comonert_ vate Sorte |Vae ata] [4 7 6, [cor o 00 Nenana | 2601] 10123. Ema | 303 fa Propane —| 25.659. Teutane | 21232 | 32584 Neeutane | 21,300 | 3.2008 Eertane | —avo4s [40102 oecto | WPeniane—[—21,085 | 40182 | o.0000 | 0 co20 HHexane—| 20,983 | area] 00000 [00000 | [“Neateptane —| 20,839] “555 30000] 9.0000] NOcane | 70.759 | “62634 {00000 | 00000 Netonane—| 20,701] 10127 | 0.0000 | 00x00 Vanaven Hysocarons | 71s | 7477 Average Bef Heating vane Bib Heating vale st Average on But basi Yatton ® fom average iu Average on Suse Basi Varaton fam average Ota ‘Combustion control to boilers: In this application, Pulp mill in Quebec sought a more reliable way to meet EPA emissions requirements. Combustion control was easier, based on the mass (standard volume) ratio between the natural gas and combustion air, over wider turndowns with no flow conditioning , - mo Saree TES, ‘ Ethylene gas transfer: Ethylene is commonly viewed as a dificult to measure gas, due to its highly non- ideal nature. In this application, Coriolis meters are Used for intra-plant transfers attaining accuracies Uunattainable by volumetric meters, helping to meet both unit mass-balance goals, as well as reactor feed rate requirements. Ethylene is fed continuously to a polymerization reactor, where various grades of polyethylene (LDPE, etc) are made.Summary Although a relatively new technol logy for natural gas applications outside of compressed natural gas (CNG), Conolis meters have gained worldwide With @ worldwide installed base of around 400,000 units, Coriolis technology is seeing expanded use for both liquid petroleum and natural gas. A number of countries and groups have published standards or fare in the process of studying the technology. Most notably is AGA and API who have jointly published AGA Report No. 11 / API MPMS Chapter 14.9, Measurement of Natural Gas by Coriolis Meter. Technology limitations of earlier designs have been largely overcome, with high accuracy measurement how possible at lowpressure drop, tynically 150" we Conurs “Sweet spots” are mainly in lines of 200mm (8°) and smaller, where high turndown is needed, flow conditioning with other technologies to meet new AGA requirements is costly, and/or the gas is of dirty, sour, of of changing composition. Also, good potential exists for “simple” energy metering, using the Coriolis meter output directly, scaled for energy units. Coriolis technology merits serious consideration as a ‘bona fide contender to complement Ultrasonic in low cost of ownership metering for natural gas ‘applications. These two technologies overlap in the “100mm (47) to 200mm (8°) line size range. A" Coriolis and 12" ultrasonic in a fuel gas, metering installation. Third-party data from CEESI, Pigsar, SwRI, and others show little if any effect of flow profile, and transferability of a factory water calibration to natural gas measurement applications. Common Coriolis gas applications range from wellhead separator, medium to high pressure distribution metering, and fuel gas to power turbines, reciprocating engines, and boilers for combustion control. As users of gas meters investigate Coriolis they are finding it to be a fiscally responsible choice for gas measurement in today’s competitive business environment.ag weer in Liquid Meas, — Class # 2020 : James H. Smitn Technical Product Spe FMC Technologies 1602 Wagner Ave Erie.PA’ 16514 Background With the development of jet engines and liquid propellant rockets in the 1960's, the need arose for a reliable ‘quick response meter that couid be used on exotic fuels and oxdizers at extreme temperatures. The axial turbine meter met this nea Soon after, it was applied to many other industrial flow measurement applications ‘Turbine meters began to be applied extensively in the petroleum industry in the mid-1960's. Since the publication ‘of API Chapter 5.3 (formerly Standard 2534) “Measurement of Liquid Hydrocarbons by Turbine Meters” in March, ‘1970. the turbine meter has gained broad acceptance for the custody transfer of petroleum liquids such as liquefied petroleum gases (LPG's), light distillates and light crude oils. During this time. this technology was ‘applied primarily at large petroleum storage and transfer terminals, ‘Types of Meters Fiuic flow meters can. in general, be classified as either positive displacement or inference type meters. Positive aspiacement (PO) meters measure voiumetnc flow directly by continuously separating (isolating) a flow stream nto chscrete volumerre segments anc counting them. Inference meters infer 2 volumetnc flo by measuring some aynamec property of the tlow stream. Turbine meters fell im the latter category, .neasunng the fluia velocity parameter The most common examples of inference meters are onfice plates, Conolis mass meters, vortex ‘shedding meters. and ultrasonic meters. Construction ‘Turbine meters have three primary subassemblies that make up the meters simple design. There is the housing, intemal assembly and puise pick up (See Figure 1). The meter housing sub-assembly is normally constructed of 2 fanged ppe spool in sizes ranging from 1/4” to 24” with virtually any type of end connection and pressure rating The materials used to construct these meters are compatible with most petroleum based fluids (consult the manufacturer for other fluids) This avoids any corrosion or erosion that may otherwise take place amongst the wetted parts. Stainless steel is the matenal typically used for the meter tube, stators and rotor This is due to {ts high corrosion resistance properties. Helical meters use titanium rotors for use in crude service while Anumunum i$ applied in clean refined products such as in ioading racks. Titanium material is used primarily for three reasons. first, it 1s a light weight material which allows the rotor to respond quickly with changes in flowrate ang without carrying large inertia forces. Secondly the material hardness protects it from erosion caused by fluids containing sand and particulates and third because of its corrosion resistance when used with harsh crude oils. Turbine meter beanngs are typically of the ball bearing or journal bearing design The ball bearings can be made Ot stainiess steel or ceramic composite. The choice of bearing 's driven by the type of service the meter will see The Staintess siee! Cali bearing is optimum for use in low flow. clean refined products such as gasoline, diesel kerosene and LPG in severe applications such as crude oils, tungsten carbide journal bearings are used which are very hard resestant to wear and reduce friction giving the meter a long service life 146< e Figure 1 — Conventional Turbine Meter Assembly Princii ‘Turbine meters determine flow rate by measuring the velocity of a bladed rotor suspended in the flow stream ‘The volumetric flow rate is the product of the average stream velocity and the flow area at the rotor as related by the nasic eevation ‘Volume “low Rate = Area x Velocity The accuracy of a turbine meter is based on two assumptions: (1) the flow area remains constant and; (2) the rotor velocity accurately represents the stream velocity. The housing provides the constant inside diameter and cross section needed for the flowrate calculation. The effective rotor flow area, and thus the meter’s “K" factor (pulses / unit volume), can change for any one or a combination of the following reasons: + Erosion, Corrosion, Deposits — Even a seemingly small buildup or erosion of the bladed rotor can have & significant effect on meter performance. For example, a one mil (0.001") buildup on ali surfaces of a 4” rotor will decrease the flow area through the rotor, and increase the “K" factor, by about 0.5% + Boundary Layer Thickness - Boundary layer thickness is relatively constant and insignificant when operating on products with low viscosity But as the viscosity increases, the boundary layer increases which reduces the effective flow area. + Cavitation — (the local vaporization of product) substantially reduces flow area through the rotor, dramatically increasing rotor velocity and the meter's *k” factor. + Obstructions - Temporary obstructions, such as trash or “grass” immediately upstream of the bladed rotor can ‘substantially decrease effective flow area through the rotor as well as cause a pronounced shift in fluid velocity profile, The second basic assumption is that rotor velocity accurately represents the stream velocity. The relationship between these factors can be affected by * Rotor Blade Angle, Rotor Stability and Bearing Friction - Changes in these areas are normally caused by ‘damage from large foreign objects or fine grit in the fluid 147Jvelocity Profite and Swit! — Var, = ae ies that can be minimized bY proper ons in Velocity profile or Auid swiri are 'Nstallation and use of flow conditioners ‘Meter Flow Rate - % of Maximum 100% Figure 2 ~ Typical Performance Curve The performance characteristics ‘shown in Figure 2 ave essential to understand and Properly apply a turtine meter in an application. Some of the terms describing the curve are: " Turndown Ratio - the range over which the meter can accurately measure. Calculated by dividing the maximum flowrate capability of the meter by the minimum flowrate. Typically stated as a 10:1 or 5:1 turndown. " Accuracy ~ the difference from the actual measurement and the meter's indicated measurement. Typically eter accuracy 1s stated in tne folowing terms: * Repeatability — the meter's ability to reproduce the same measuiement for a set of constant conditions of flow rate. temperature, viscosity, density and pressure Repeatability of a ctistody transier wpe mest ieadeig sn Bisaily he willy 57% Im vung er -= 3.28% at age: other in 5 consecutive prove runs. * Linearity — the meter’s ability to maintain a meter factor triough-out the stavsd turmdown. Depending on meter size and application this is typically +/-0.18% of +/-0.25% for custody transfer applications, Turbine meters are generally applied in continuous duty operations and have the advantage of a tong service life with minimal meintenance. “With the proliferation of the microprocessor and the introduction of the electronic Preset device, turbines are also successfully applied in intermittent operations such as in truck loading tack applications. Fluid viscosity must be considered carefully in order to select the proper A general rule of thumb to consider for viscosity limits, keepin conventional and X's the meter size for helical. If we look at viscosity and flowrate as dex icted in Figure 3, we Scere Pee are {0 best apply the two turbine meter types. The helical provides tie most wane e weep eewcations that extend beyond the conventional meter capability, This chart shows thac rene meter technology has exparded its application crossing over into some of the traditional PD and Covi meter territory type of turbine (helical or conventional). 19 2 10:1 turndown, is 1X’s the meter size forVieosity Range (cST) 500 1,000 2,500 5,000 10,000 15,000 20,000 25,000 30,000 40,000 Flow Rate (bph) Figure 3 — Viscosity vs Flowrate for Various Measurement Technologies The application of a turbine meter can also be determined by a Reynolds number analysis. This provides more insight into the usable range of turbine meter technology. Reynolds number is the ratio of the inertia forces to viscous forces that quantifies the relative importance of these two types of forces for given ftow conditions. Thus, it is used to identify different flow regimes, such as laminar or turbulent flow. This is important because the turbine meter will perform differently in these regimes and is affected by flowrate and viscosity Pey-clds number is one of the most important dimensionless aumpers in fluid dynamics. It can be mathematically rer. esente Re= Where, D = inter..2i diameter of th pipe in feet ©.» velocity of flow in ieet per second. v = kinematic viscosity, square feet per second. Reynolds number can easily be calculated using the following derivation: Re = 2214 x Flow Rate (BPH) osity (inches) (cSt) The laminar flow range is typically defined as having @ Reynolds number of less than 2,100 while the turbulent flow range ‘s typically defined as having a Reynolds number of greater than 4,000. The range between these two values is called the “transition range" in which turbine meters should not be operated. This is Decause the flow profile is unstable affecting the meter performance. Turbine meters inherently produce a more linear response when operated in the turbulent fiow range It 's generally recommended that conventional turbine meters be applied in a Reynolds number of greater than 50,009. Figure 4 shows the response of a typical turbine meter ‘over a Reynolds number range. Meter Size x 149