Professional Documents
Culture Documents
Capacity Design of Gas Distribution Systems
Capacity Design of Gas Distribution Systems
Introduction
The designer of the gas distribution systems has many pipe layout options. Developing
alternative plans for extending or reinforcing gas distribution systems is the primary creative
activity in the capacity design process. Computer or manual simulation of these various plans at
design conditions in a process called network analysis can then be used to refine and optimize
them.
In choosing a plan, the designer’s objective is to develop an economic system that performs all
of its functions satisfactorily. That is, the system should provide a safe, reliable supply of gas to
all customers at sufficient pressure for proper operation of their appliances at all times. The task
of the capacity designer consists of balancing the possible performance penalties of a closely
sized system against the certain, but sometimes small, economic penalty of designing an
oversized system.
When use of the digital computer in the 1960s took most of the drudgery out of load gathering
and network analysis, the design engineer was freed to concentrate on creative approaches to
optimizing designs. Only then did designers realize they were not quite sure what comprised an
ideal distribution system or what were all the design variables. Answers to these basic questions
were developed largely through a series of annual computer applications
workshops jointly sponsored by the Computer and Distribution Design and Development
Committees of the AGA Operating Section from 1966-1972. Under the leadership of Samuel I.
Hyman of the Brooklyn Union Gas Company, these workshops provided an effective forum for
exchanging information among experienced gas system designers. Many of the ideas presented
in the first section of this chapter are from those workshops.
Operating Limitations
The primary operating limitation of a pipe system is the maximum allowable operating pressure
(MAOP) imposed by its weakest links. These usually consist of the pressure ratings of
components such as older pipe, new sections of plastic pipe, and valves or components, such as
threads, flanges, and bell and spigot joints in cast-iron systems. The MAOP is also set by the
internal relief capacity of its service regulators or by regulatory-agency code provisions such as
class location or specific orders.
Delivery-Pressure Commitments
The minimum permissible delivery pressure to residential and small-volume commercial
customers is a matter of company policy or regulatory-agency order. It is usually significantly
higher than the minimum pressure at which the customers' appliances will operate satisfactorily,
allowing for pressure drops in customers’ fuel lines beyond the meter.
Typical values of the minimum permissible delivery pressure are 4-7 in. wc (0.99-1 .7 kPa) in
low-pressure systems, ½-5 psig (3.4-34 kPa) in medium-pressure [1-15 psig (34-103 kPa)
systems, and 5-20 psig (34-138 kPa) in high-pressure [15-60 psig (103-414 kPa)] systems. Large
commercial and industrial customers often negotiate higher delivery pressures. Delivery pressure
commitments between 1 and 10 psig (6.9-69 kPa) are common and may range up to 150 psig
(1034 kPa) or more for a few customers. Industrial customers press for higher delivery pressures
to permit them to use less expensive smaller pipe in their fuel runs or to avoid installation of
compressors to obtain the high fuel pressure required by some industrial application,s such as
large gas turbines in cogeneration service.
In an existing distribution system, the settings of the source pressure regulators may be less
than the system’s maximum allowable operating pressure yet somewhat above the pressure
required to meet all pressure-delivery commitments for a design load. Of all these commitments,
one will require the highest source pressure to satisfy it. This location is the control point in the
system. If demands exceed peak load, it will be the first location at which the pressure will be
inadequate. The control point is often located at the service of a large-volume customer with a
relatively high delivery-pressure commitment or at a lower pressure regulator inlet. These are
often good locations for a permanent-recording pressure gage.
Source Pressures Available
The pressures available at the sources of a pipe system are a function of the operation of the
pipe system supplying the sources, whether it be a higher pressure distribution feeder system or
a transmission pipeline. These supply pipe systems also experience peak demands that create
flow or input pressure limitations on the sources of the system being studied. Pressure variations
in transmission lines also result from packing the line (building up the pressure to store gas in the
line in advance of peak demands) and unpacking the line (using the gas stored in the line as the
pressure declines).
Pipe Right-of-Way
Most distribution pipe systems are installed in the public right-of-way along streets.
Nevertheless, right-of-way must occasionally run through private property. In general, right-of-way
availability determines the alternative pipe routes. In rural areas, some companies prefer to install
their lines on private property just off the public right-of-way along highways. This avoids later
unreimbursed relocations resulting from road widening and other highway improvements. In
urban areas, the distribution designer attempts to avoid locating major supply mains where there
are plans for future large sewers, rapid-transit lines, or major water-main improvements.
Physical BARRIERS
Physical barriers—rivers, ravines, expressways—strongly influence the routing of pipe
systems. They are difficult and expensive to cross, and the crossings are expensive to maintain.
Insofar as possible, such crossings are avoided. Where crossings must be made, directional
drilling is the least disruptive to rivers and wetlands (marshes and swamps) and horizontal boring
can be used to cross under expressways and railroads.
Safety Considerations
Safety guidelines cover many areas. One important safety tactic is to pick locations for
distribution mains and regulators that minimize exposure to third-party damage. Most other safety
concerns relate more to the mechanical design of the pipe system rather than to capacity design.
Safety is not only a cornerstone of responsible design, but a legal necessity. The federal safety
regulations developed by the United States Department of Transportation are presented in Title
49 Code of Federal Regulations, Part 192, The Transportation of Natural and Other Gas by Pipe/
Minimum Federal Safety Standards. Shorthand notations such as 49 CFR 192.197 (which
identifies a specific paragraph of the code) are used when referring to this publication. Local
codes may sometimes specify more stringent guidelines.
The safety regulations that affect capacity design relate primarily to system operating
pressures. Here are some of the more significant ones:
Two-stage regulation
For pressures greater than 60 psig (414 kPa) per 49 CFR 192.197(c): two regulators in series
with a relief or automatic shut-off device between them or a service regulator and an upstream
monitoring regulator (two regulators in a series where first, the monitor is set at a higher
downstream pressure than the second, and thus, is normally wide open, but takes control if the
downstream regulator fails); a relief device, which may be a service regulator with an internal
relief valve; or an automatic shutoff device must be incorporated in all service lines
For all service lines connected to distribution systems operated at pressures greater than 125
psig (862 kPa): two-stage regulation, either series or monitoring,
Pressures greater than the MAOP of the system as per CFR 192.195: over-pressure
protection must be provided at all regulator stations in distribution system supplied through
pressure regulators from pipelines.
For a higher operating pressure in accordance with 49 CFR 192, Subpart K on
uprating: procedures are set forth that outline the steps that must be taken to qualify existing
distribution pipe systems.
A maximum operating pressure limit of 100 psig (669 kPa) for plastic and copper pipe
used in distribution systems in accordance with 49 CFR 192.123 and 192.125.
Over-pressure protection requirements—either two stage regulation or monitor regulators—
increase the pressure loss through regulator stations and service regulator sets. This reduces the
allowable pressure drop in mains and services at full loads and drives the design towards larger
diameters, more "looping," or larger regulators.
Reliability of Supply
The term reliability applied to a distribution system refers to continuity of service. Ideally, there
should be no disruption of service to any customer at any time. In practice, some scheduled
service interruptions in connection with construction and repair work are unavoidable; third-party
damage is another source of gas escapes that may require the shutdown of portions of a gas
system. The layout of the system, and the number and placement its valves, determines the
number of customers affected by shutdowns. Such emergency shutdowns also may reduce
system capacity during peak-demand periods.
Service interruptions also require that each customer be visited twice by skilled service
people: once to assure shut off of supply to the structure and again for proper relighting of
standing pilots and verifying the operation of appliances after gas service is restored. These
interruptions can be very expensive and time consuming, especially for large apartment buildings
where access often requires a locksmith and a law enforcement agent.
To increase reliability, an effort is made to have more than one source supplying each
distribution system. If a single source is used, two regulators can be installed in parallel rather
than a single regulator.
Standardization of Pipe Sizes
To minimize the expense of storing and handling a large variety of pipe and fitting sizes, most
gas system operators standardize on a relatively small number of nominal pipe sizes.
Consequently, the designer must normally choose from the available range of sizes except for
major projects.
Design Policy
Over the years, many companies have evolved policies that define their design philosophy.
Sometimes, they are codified in a distribution design manual; or if not, they can be inferred from
the results of network analyses. Because the design of large areas of existing distribution
systems probably conforms to the then existing company design philosophy, it is particularly
important to be aware of the constraints embedded in this philosophy when these same systems
are redesigned.
The capacity design parameters for a gas pipe system are the variables that the designer can
control, subject to the constraints previously identified. There are six design variables that relate
to system capacity:
The operating pressure level of the system
The minimum allowable pressure
Use of constant or remotely controlled pressure at the sources
The spacing between sources
The gradient pattern (the allocation of the allowable pressure drop between feeder mains,
distributor mains, service lines, and house lines)
The gas velocity.
An important aspect of a well-balanced system is the use of load shifting between source
regulators in low-pressure systems to improve the pressure distribution in the medium- and high-
pressure feeder systems that supply the low-pressure systems. A load can be shifted away from
a pressure regulator by decreasing its outlet set pressure; it’s shifted toward it by increasing its
outlet set pressure. The achievement of a well- balanced system requires considerable study of
its performance under various regulator outlet pressure settings.
Fig. 5-1. Ideal Piping Layout Developed by Gary
Source: Gary, J. .F., "The Effect of Heating Loads on a Distribution system." Pacific Coast
Gas Association Proceedings 45 (1954): 79-80.
Gradient Pattern
The source pressure selected and the delivery-pressure commitments to customers determine
the maximum pressure drop or difference of squares of pressure (rP 2) that can be tolerated in a
distribution system. This drop can be allocated uniformly per foot of distance traveled (uniform
gradient), or the designer can vary the pressure drop or r(P 2) per foot of pipe (gradient) for trunk,
feeder, distribution mains, and service pipe. The pattern used could well have a significant effect
on system cost, but the magnitude of this effect has not been studied systematically.
Gas Velocity
Designers have observed that limiting the velocity in all mains to a stated maximum value
produces results similar to those obtained by using a uniform gradient (Richwine, T. E., and D. W.
Schroeder, "Large-Scale Gas Network Design by Automated Means." AGA Operating Section
Proceedings 1986: 314-19). The appropriate limit varies with the source pressure, the maximum
source-to-load paths, and the minimum allowable tail-end pressure. Values of 40-60 feet/second
(12-18 m/second) are appropriate for a system operated at a source pressure of 15 psig (103
kPa).
Velocity limits were originated to avoid movement of dust and other deposits in older distribution
systems. A study by Fuidge and Smith in the 1960s indicated that, at atmospheric pressure, dust
begins to move at velocities of about 15 feet/second (4.6 m/second) and lifts at velocities of 30-40
feet/second (9.1-12 m/second) or greater (Fuidge, G. H. and A. V. Smith, "Gas-Borne Dust."
Institute of Gas Engineers Journal, May, 1966: 333-356).
The following are design variables that are related to safety and reliability and that can be
controlled by the system designer:
Use of a looped or radial system
Placement of valves
Layout of services (single or branched)
Design of regulator stations.
Only the first three of these relate to design of the pipe system.
Looped or Radial System
In a fully looped or gridded distribution system, as shown in Figure 5-2, mains are installed on
all or nearly all streets and connected together at most intersections. The other extreme is a
radial or tree system that has no closed loops, as shown in Figure 5-3. Most low-pressure
systems approximate fully looped systems, and blockages were fairly common in such systems at
the time most were installed. The most common cause was collection of water at dips in the
mains where they had settled away from the grade to which they were laid. This water is usually
ground water entering through leaks when the water table is high or through nearby water main
leaks. Accumulations of tar and naphthalene from manufactured gas were other blockage
culprits.
Fully looped systems minimized customer outages from blocked mains by providing many
alternative paths by which gas could reach each customer. Radial distribution systems are less
costly than looped systems; however, as others have it, the radial layout is analogous to a tree. If
the trunk is severed, the whole organism dies. Consideration of looped versus radial systems
thus involves a trade-off between economy and reliability of service.
Fig. 5-2. Fully Gridded Distribution Systems
Source: Richwine, T. E., and D. W. Schroeder, "Large-Scale Gas Network Design by
Automated Means." AGA Operating Section Proceedings 1986: 314-19.
Fig. 5-3. Radial Distribution System
Source: Richwine, T. E., and D. W. Schroeder, "Large-Scale Gas Network Design by
Automated Means." AGA Operating Section Proceedings 1986: 314-19.
A practical limitation of radial systems is the number of customers that the work force can get
back in service in a reasonable period. This raises the question of what is a reasonable down
time that does not seriously inconvenience customers. Most operators of distribution systems
would probably find 24 hours acceptable except for heating customers in severely cold weather.
Estimates of maximum permissible outage size range from 50 customers for small, isolated
distribution systems to 30,000 customers for one of the country's largest gas systems with a
highly developed service area and located in a warm climate.
The current trend in the design of new distribution systems is to use a mix of looped and radial
designs. Considerable, though not complete, looping is used for the headers and near the
sources; radial layout is used for the tail ends and extremities of each regulator feed area. Large
looped systems fed by a large number of regulator stations are seldom built.
System Layout and Valve Placement
Valves are installed in distribution systems to permit portions of the system to be shut down in
case of an emergency, such as escape of gas from a broken main or to facilitate construction. For
high-pressure distribution systems and transmission systems, valve placement may be mandated
by state and local code.
Considerable information on different approaches to pipe layout and valve placement is given in
a 1967 for the AGA, Variations Plat Piping Proposals. (Hunt, B. E., 'Variations - Plat Piping
Proposals." AGA Operating Section Proceedings 1967: 151-65.) It presents 13 different pipe
layouts for the same 95-customer sub-division, two of which are shown in Figures 5-2 and 5-3.
Though a source pressure of up to 125 psig (862 kPa) was available, only two designs used a
source pressure over 60 psig (414 kPa); the average was 40 psig (276 kPa). The minimum
pressure required in six of these designs was 3-5 psig (21-34 kPa). Five specified minimum
pressures of 10-15 psig (69-103 kPa), and one required 20 psig (27 kPa).
The main sizes installed ranged from 3/4 inch to 4 inches (19mm-102mm), but the 4 inch was
used in only one design and was installed to provide capacity for future customers surrounding
the subdivision. Main sizes of 1-1/4 in. (32 mm) and 2 in. (51 mm) predominated. The service
sizes were mostly 1/2 % in. or 1/4 in.
Only six of these designs included any closed loops, the rest were radial. Only seven of the
plans used any valves at all, and, of these, only four used more than one valve. Valves were used
sparingly, at least in part because of the cost of the maintenance required by federal safety
regulations per CFR 192.747.
The most novel layout is shown in Figure 5-4. The double main design avoids long side
services and lends itself to installation of the mains inside the sidewalk when the public right-of-
way extends that far. Both these features minimize exposure to third party damage. This design is
based on a 60-psig source pressure and a 5-psig minimum pressure. It consists mostly of ¼-inch
mains and uses ½-inch service lines.
Service Layout
Individual services are usually installed for each customer. An alternate approach is to branch
each service to single family homes to serve two or more customers from a single run of pipe
from the main to the building line. In areas with a high customer saturation, this can significantly
reduce the amount of service pipe required. Some reliability is sacrificed because service to more
than one customer is interrupted whenever the service line is shut down. Use of dual service runs
also limits the meter-set locations for each customer. In addition, it can create confusion for field
workers assigned to service-related work such as meter change out, emergency shutdown of
services, and repair work. These limitations have inhibited the wide use of dual services, although
some companies have used them with no problems.
Regulator Station Design
The design of a regulator station affects the capacity, safety, and reliability of the distribution
system that the station serves. The size and type of regulator used (conventional valve,
slotted/boot, or streamline flow), the station pipe size, and arrangement determine peak-load
pressure drop across the station. This limits how much of the upstream supply system pressure is
available for use in the distribution system. In some situations, this could limit the operating
pressure level of the system for peak load.
Regulator stations in the "critical pressure path" from the city gate station to the customer
should be designed for low-pressure drop under design conditions of flow. The cost of reducing
the pressure drop through a station is usually much less than the cost of reinforcing a system
through paralleling or replacing mains or making tie-ins to accommodate the station
drop. Expenditure of as low as $15,000 have reduced the drops across regulator stations from 25
psi (172 kPa) to 10 psi (69 kPa) and saved hundreds of thousands of dollars in downstream
reinforcements. When regulators are upsized to accommodate growth, the pressure drop in the
outlet piping connected to the street main should therefore be reviewed. A pipe size and
configuration that was adequate for the initial installation can severely reduce the pressure
available to the network.
Fig. 5-4. Double Main Systems
Richwine, T. E., and D. W. Schroeder, "Large-Scale Gas Network Design by Automated Means."
AGA Operating Section Proceedings 1986: 314-19.
Company design policy imposes additional constraints on the use of remote control of source
pressures, the gradient pattern, the placements of valves, and regulator station design. Policy
constraints should be regarded as expendable if there are good reasons for doing so.
Location of Mains
Gas mains are sometimes installed on both sides of wide streets to avoid long service lines
under the streets, as was shown in Figure 5-4. Sometimes, they are installed beyond the rear
property lines, particularly when a dedicated alley is present. Without an alley, access problems
result from shrubbery, fences, flower beds, gardens, and assorted customer structures. Favored
locations are in the grass strip between the sidewalk and the curb, between the sidewalk and the
property line when there is room,and under the roadway pavement if not avoidable.
Principles of Design Logic
The following principles of design logic significantly reduce the alternatives to be considered:
Pipe sizes should telescope from larger to smaller as the distance from the sources
increases along supply and feeder mains. This offsets or compensates for the flow rate that
decreases with distance from the sources.
Each load should be supplied via the shortest feasible path to a source.
The first of these principles rules out all alternatives in which increases in pipe size occur as
one moves away from the sources along trunk and feeder mains. Sometimes, though, the pipe
size along supply mains between sources is not telescoped in case adjacent sources need to pick
up the load if a source must be taken out of service.
The second is a common sense principle, or calculated index, for efficient, less-
costly systems. A quantitative measure of system efficiency arising from this principle is the Mcf-
miles (m3-km) or Mcf feet (m3-m) index. This index is the sum of the products of flow rate times
length for each pipe section in the system. The lower the value of this index for a layout to supply
a given system, the more effectively the second principle has been applied. If the index is divided
by the total system gas demand, the average length of the flow path per Mcf (km 3) of gas from
source to load is obtained. Boyer has noted that the Mcf-mile (km 3-m) index most effectively
defines what you should not do, rather than what you should do (Boyer, H. M., "Increasing the
Capacity of Existing Distribution Systems." AGA Operating Section Proceedings 1967: 151-
63). That is, a layout with a relatively high value of this index is an inefficient layout, whatever are
its other qualities.
Boyer advocated that a specified amount of reserve capacity, about 5 to 10 percent, be
designed into distribution systems. This is an alternative to choosing design loads with an
acceptably small probability of being exceeded. It is consistent with the observation in Chapter 3
that a 10 percent increase in design load drops the risk factor (the probability that the design load
will be exceeded one or more times in a year) by a factor of about 10.
Reinforcement/Redesign of Existing
Systems
The distribution design engineer faces reinforcement/redesign problems from the day that an
initial gas distribution system is installed. Now the redesign problem is characterized by many
constraints and a limited number of choices. In addition, the cost of installing pipelines in fully
developed urban areas several times as great and so requires more careful analysis than
installations in subdivisions being developed.
The following subsections describe the causes—or perceptions of events—that force
rethinking of the design of a distribution system.
Changes in Distribution System Load
Changes in the load imposed on a distribution system include growth of heating saturation;
changes in the heating value of the gas served; extension of the system away from the source
points; new subdivisions; redevelopment of land, especially changes in street systems; addition
or loss of significant industrial and commercial load; abandonment of residential and commercial
structures; and new types of gas utilization such as the fuel cell or cogeneration.
The impact of these changes on system performance can be checked in two ways: by pressure
surveys or by calculating system pressures from load information.
In a pressure survey, the gas pressure in the system is measured at several key locations in
the distribution system during peak demand. The pressure is measured by recording pressure
gages of the type illustrated in Figure 5-5 or by newer electronic recorders. The mechanism in
these gages is usually a pen arm attached to a bellows or to a Bourdon tube through a series of
levers. The pen records the pressure on a graduated paper chart, rotated by means of a
spring or battery powered clock mechanism. These clocks have a 24-hour, 7-day, or 30-day
movement, and the chart is divided into time units.
In most distribution systems, a number of pressure gages are installed permanently to record
the gas pressure continuously. These records often provide the first indication of irregularities in
system operation. Telemetering equipment may be used to transmit pressure readings from key
locations to a central location where they can be monitored continuously. In a pressure survey,
the permanently installed gages are supplemented with portable gages that record the pressures
at intervals along important trunk and feeder mains in areas where the lowest pressure is
expected. These gages may be connected to the mains through small-diameter piping brought to
the surface of the ground especially for the purpose. They are usually connected to the services
of low-consumption customers (ahead of the meter in a low-pressure system or ahead of the
service regulator in an intermediate-or high-pressure system). The low-pressure points of the
system and the pressure drops along important feeder mains can be determined from the records
provided by these gages.
Pressure surveys can be used to confirm estimated loads, to check the validity of the system
data base, to determine the adequacy of a distribution system for its present customers, or to
indicate the general areas that soon will need reinforcement. They cannot, however, provide
direct information on the adequacy of a system to carry anticipated loads, such as the peak load
for next winter. This can be predicted only by performance calculations for the system. These
calculations estimate the flow rates and pressures that will occur with anticipated loads. The
procedures and equipment used for these calculations are outlined later in this chapter.
Fig. 5-5. Portable Recording Pressure Gage.
Deterioration of the Existing Pipe Systems
In some soils, cast-iron mains corrode at a rate such that wholesale replacement or the
equivalent becomes necessary to minimize breakage at weakened locations. Most wrapped,
cathodically protected steel pipelines and plastic pipelines made of appropriate plastics do not
deteriorate to the point where replacement on a wholesale basis is necessary. Bare steel mains,
however, corrode at a rate such that additional cathodic protection or repairs are
uneconomical. Concerns for the viability of oxyacetylene welded pipe systems also may require
major pipe-replacement.
Public Works
Changes in street grades, street alignment, and the installation of new water or sewer systems,
express highways, and rapid transit systems can force the physical removal of important existing
gas mains.
Methods of Reinforcement
Distribution engineers in various parts of the country have developed different methods of
increasing system capacity while endeavoring to minimize cost. The following are some of the
methods commonly used:
Installing additional gas sources at key locations in the system. Substantial reinforcement
of systems can be accomplished most economically by extending higher pressure feeder mains
to feed gas into the system through pressure regulators, as illustrated in Figure 5-6.
Increasing gas pressure at the sources during peak load periods. This may be
accomplished by means of pressure-boosting regulators, which increase the source pressure with
increased demand or by remote control from a central location. In a low-pressure system, this
increase is limited to 2-3 in. wc (0.5-0.75 kPa) to assure proper operation of the appliances of
customers located near the source, unless service regulators are installed in the service lines of
these customers.
Converting all or part of a low-pressure system to a medium-pressure system. The
portion of the system to be converted must be disconnected or valved off from the remainder of
the system and a service regulator installed in each service in the area. It may also be necessary
to install a shut-off valve in each service. Unfortunately, in many older low-pressure systems the
use of increased pressure causes excessive leakage at the joints, and an extensive joint-
clamping or sealing program is needed. For this reason, the pressure is increased in several
stages and a leak survey is conducted after each increase. A study should be conducted first to
determine whether the system can withstand increased pressure without excessive renovation as
required by CFR 192, Sub-part K.
An alternative way to increase system pressures for a low-pressure system is through renewal—
by inserting a new main, often plastic, inside the old main and increasing the operating pressure
level.
Fig. 5-6. Reinforcement of a Low- Pressure System
Using a higher operating pressure in an existing medium- or high-pressure distribution
system. Figure 5-7 shows the effect that increasing the source pressure has on system
capacity. The three curves are for three initial source pressure values. The dashed line is the
estimated increase in capacity, 135%, obtained by increasing the source pressure of a
distribution system from 20 to 50 psig. It is apparent that additional pounds of source pressure
are equivalent to a lot of pipe. Thus, this method of reinforcement is normally used to the
maximum extent that system constraints permit. The primary limitations of the operating pressure
maintained in such systems are the available supply pressure, the MAOP* of existing system
facilities, operating codes**, safety considerations, and the leakage rate. Again, these
reinforcements should not be undertaken unless proper and safe operation of the revised system
can be assured.
Using tie-ins at key locations. An example of a tie-in is shown in Figure 5-8.
Replacing undersized mains. In distribution systems with appreciable space-
heating loads, mains of less than 2 in. (51 mm) diameter in medium-pressure systems, and of
less than 46 in. (102-152 mm) diameter in lower-pressure systems often have insufficient
capacity at peak loads.
Introducing additional mains parallel to existing mains. This method is relatively
expensive, but is used extensively where a less costly alternative is not available.
Closing loops on a radial system or a looped system not fully gridded. These ties often
are short and can provide an inexpensive way to increase network capacity.
The reinforcement method actually used for a system will depend on the layout and condition of
the existing piping, since the cost of reinforcement is closely related to this information. No hard-
and-fast rules can be given for the use of each method, but two or more reinforcement schemes
should be checked for adequacy and their costs compared to determine the most economical
method.
Fig. 5-7. System Design Capacity vs. Inlet Pressure with a Constant Percent Overload
Capacity.
Source: Place, J. C., "Effects of Multiple Pipeline Suppliers on the Design and Operation
of Distribution Systems," AGA Operating Section Proceedings 1990: Paper 90-DT-201, p
34244.
Fig. 5-8. Use of Tie-In to Increase the Capacity of a Distribution System.
Redesign Considerations
Low-Pressure Systems
In addition to the insertion problem, a re-evaluation needs to be made of the district regulator
stations that feed the low-pressure system. In some instances, these regulator stations are fed
from systems operating at such modest pressures that the reduced size of the replacement high-
pressure system does not produce the savings necessary to offset the cost of the reinforcements
needed in the supporting feeder system.
High-Pressure Systems
In general, high-pressure systems already have the investment in service regulators and the
necessary integrity to allow for some upgrading of operating pressures. The normal technique,
however, in the case of high-pressure systems requiring redesign is to increase the capacity by
installing additional supply points from the transmission system at locations distant from the
existing supply points. This permits operating at higher gradients (pressure drop per mile). Also,
loops can be closed to provide additional flow paths, and pipe sections with excessive gradients
can be paralleled. If a system’s operating pressure is increased, care must be taken that there is
internal relief capacity for existing service regulators and their vent piping. This may require fitting
these regulators with smaller orifices.
General Considerations
Historical distribution design practice has been to build redundancy in the main configuration of
a system. This reduces the sensitivity of the system to forced outages in any segment of the line.
Several decades of experience with steel or plastic construction has indicated that forced
outages of any given line may occur only once in many decades. With such high reliability, a
redesigned system can well tolerate much less redundancy in source-to-load paths and thus
fewer closed loops.
Large Loads
Earlier in this chapter, rational limits on the size of a radial section of a gas system were defined
in terms of the number of customers that the available work force can get back in service in a
reasonable time. Special treatment needs to be given to large loads such as industrial customers
and large apartment complexes. For industrial customers, the consequence of a forced outage is
generally economic. For large residential complexes, the consequences may be more serious,
even life-threatening, in extremely cold weather.
Where new large loads must be supplied, instead of feeding the new quantities of gas through
existing mains, it may be more economical to install new higher-pressure pipelines. Such
pipelines, in effect, parallel existing lines operating at much lower pressures. One technique often
used is to "unload" the existing system by transferring the services to the new higher-pressure
line and changing the service regulator for operation at higher pressure.
Routing the new pipeline past intersections where the existing distribution system has large
mains provides locations where future feeds from the new system to the older system can be
installed. Such a scheme postpones or possibly eliminates construction of additional distribution
system reinforcing mains. On occasion, a section of distribution main between a source and a
large customer can be stripped from the base system, converted to higher pressure, and used
efficiently as a direct feed to the customer. If necessary, the lost capacity in the distribution
system can be replaced by adding regulator feeds from the converted line.
Network AnalYsis
Calculation of the pressure drop along a service or single main requires only the application of a
flow equation as illustrated in Chapter IV. But most gas distribution systems consist of an
interconnected network of pipe as in Figure 5-2. In such a network, there are many alternate
paths for gas to flow from a source to a point of use. Consequently, the flow rates and direction in
each pipe section as well as the pressure drop are unknown. The solution of a network flow
problem involves the simultaneous evaluation of flow rates and pressures.
Several design problems that require network analysis were discussed earlier in this
chapter. The complex networks of pipe that comprise gas distribution systems present a
formidable flow analysis problem. Gas may be fed into a system from several transmission
pipeline city-gate stations, from one or more peak load gas production plants, and from natural
gas storage facilities as illustrated in Figure 1-22 in Chapter I. Many distribution systems consist
of several superimposed networks of piping operated at different pressure levels.
The run of pipe from one node (point of intersection of one or more pipes) to the next is called a
pipe section. A continuous series of pipe sections that starts and ends at the same node is
referred to as a loop. Networks of 200-300 loops and 500-600 pipe sections are very
common. Many networks contain over 1000 pipe sections, and some larger cities have
interconnected piping networks containing tens of thousands of pipe sections.
In the past, electrical analog models of a gas distribution system, such as the Mcllroy, were
used to determine the flow rates and pressures in a gas pipe network for a specified set of
conditions. Today, such analyses of pipe networks are usually done by the use of special purpose
computer software based on a mathematical model of the pipe system.
The method of load concentration often used in a network system is to assign the gas load for a
half block in each direction from an intersection or node to the node location as illustrated in
Figure 3-6. Large volume commercial and industrial loads as well as any load significantly larger
than the average load distributed along a block are shown separately at their point of connection
to the system.
The steady state flow equations discussed in Chapter IV are used to represent flow behavior in
the pipe sections comprising the network. After a flow equation has been selected, the procedure
outlined in Chapter IV can be used to calculate a resistance factor for each pipe section using
items 2, 3, and 4 in the above data list.
A schematic diagram such as that in Figure 5-9 can be used to summarize most of the data
required to define a network analysis problem. It shows how the pipe in the network is connected.
The intersection nodes have been numbered, from 1 to 11 The slanting arrows coming out of the
node indicate gas loads. The number at the tip of each arrow is the amount of load in Mcf/h. An
arrow pointing into a node indicates a source of gas. Sources can also be identified as negative
loads (gas flowing into rather than out of the system) by flagging them with a negative sign. The
numbers under or on the right of each pipe section are their flow equation resistance factors.
Fig. 5-9. Typical Schematic Diagram of a Gas Distribution System.
A network diagram may consist simply of a scale map of the piping system; however, it is
frequently desirable to prepare a special diagram for the larger and more complicated problems.
This diagram consists of the essential parts of the piping system, but excludes the stubs and
laterals. Its length scale may also be somewhat distorted to clarify the structure of the network
and to avoid congestion of the data.
When a computer program is used for network analysis, a schematic diagram is usually not
prepared. Instead, the nodes are systematically numbered on a system map: main intersections
mostly at street intersections; the point at which major load centers, mostly large volume
commercial and industrial customers are connected to the distribution mains; source locations
and locations where there is a change in pipe diameter. These node numbers are then used to
identify load centers and pipe sections, which are identified by the node numbers at each end.
The physical description of the system is compiled in two tables of data:
1. A pipe section list: the end node numbers and the length and diameter of each pipe
section.
2. A load list: the load center node numbers and such load center data as design loads;
elevation, if relevant; and minimum required pressure. Source nodes are identified and
the source pressures or the maximum gas input rate, whichever applies, are specified.
These data are then input into a computer. If the physical description data are stored in a
relational database, they can be called up directly into the computer.
The gas properties of gravity and viscosity as well as the flowing gas temperature are entered
directly into the computer in the format required by the network analysis program. The flow
equation is specified and the computer program calculates the pipe section resistance factors for
each pipe section from the gas properties, flowing temperature, and the pipe dimensions.
Solution Procedures
All methods of solving network flow problems are based on Kirchhoff’s two laws for steady-state
flow in a network, Figure 5-10:
1. FIRST KlRCHHOFF LAW:
Q1 + Q2 = Q3 + Q4 + L
or
P6) = (P7-P1)
or
The second law essentially states that only one pressure can exist at each point in the
network. As in single pipe problems, the pressure drop across a pipe section is obtained from the
flow rate by use of a flow equation.
The conditions imposed by these laws fix the flow rates for each section of pipe in a network.
That is, there is only one value for the flow rate in each pipe section for which Kirchhoff’s laws
will be satisfied throughout the network for a specified set of input conditions.
Trial and Error Solution
For many years, the only method used to solve network flow problems was a trial-and-error
procedure. Although no two people will follow exactly the same path in the trial-and-error solution
of a network problem, all will use variations of the following steps:
1. Assume a flow rate in all pipe sections that satisfy Kirchhoff's first law for all nodes
(Figure 5-10). In large networks, the procedure is to outline the area fed by each source
on the network diagram and work from the perimeter of each feed-area back toward the
source.
2. Calculate the pressure losses in each pipe section in the network by use of a flow
equation.
3. Sum the pressure losses around closed loops and along
continuous pipe sections between sources. Check the values of these summations
against Kirchhoff's second law.
4. Modify flow rates assumed in step 1 of the procedure in response to the results obtained
in step 3. Repeat steps 2, 3, and 4.
Modification of the flow rates is continued by the analyst until the calculated flow
rates and pressure losses satisfy Kirchhoff's second law within acceptable tolerances. This
procedure is very tedious and errors are difficult to avoid.
A skilled network analyst rarely attempts a complete solution of large problems by trial-and-
error, but works the problem only in sufficient detail to establish the adequacy of the piping layout
for the design load. First, loads are balanced loads between sources along trunk mains until
Kirchhoff’s second law is approximately satisfied. Then the areas in the network with the greatest
pressure losses and the areas of lowest pressure are identified to satisfy as closely as possible
Kirchhoff’s second law in those areas.
Computer Solution
If Kirchhoff’s laws are expressed in terms of node pressures (flow rates can be related to node
pressure by a gas flow equation), a set of first law equations are obtained that represent a
mathematical description of the network. The choice of node pressures as the basic flow variable
for network analysis results in automatic satisfaction of Kirchhoff’s second law equations, which
specify that there can be only one pressure at each node. A number of computer programs have
been developed by gas companies and commercial software firms to solve the set of Kirchhoff’s
law equations to determine the node pressures and pipe section flow rates for specified loading.
The results of network analysis by computer are usually expressed in tabular form and provide
this key information:
1. A summary of the input data.
2. The pressures and load or input at each node.
3. The flow rates and direction in each pipe section as well as its length and diameter.
From this information, the system designer can judge whether a new system need to be
designed or the existing system requires reinforcement. It also provides clues to appropriate
system modification by indicating where the lowest pressure areas are and which pipe sections
are bottle necks.
Because many distribution systems consist of hundreds and thousands of pipe sections, the
printout of the analysis results can be a stack of paper several inches thick. This has led to the
use of exception reports to summarize the results of most significance to the designer, such as
the ten nodes with the lowest pressure, the ten nodes with the highest pressure, and the ten pipe
sections with the highest gradients (rP2/ft), that is the sections working the hardest. Such
summaries provide the designer with enough information to guide modification of the preliminary
design into a workable one.
Interactive Design
High speed computers and versatile computer programs have made a new approach to design
feasible—interactive design. It now takes only a few minutes to solve even quite large network
problems. The most time-consuming step is printing out the complete results. Therefore, the
designer prints out only the exception report for his initial design. It gives him enough information
to tell him if it needs modification and guides him to the selection of sensible changes. Then he
directs the computer to resolve the system and print out the exception report for the solution. This
process continues until the designer is satisfied that the modified system has adequate, but not
excessive capacity. Only then does he print-out of the full results as a complete record of how the
system is expected to perform when the estimated future design load is applied to it.
Computer graphics has also substantially assisted the interactive design process. To provide
the input required to use it in network analysis, geographic coordinates are established for all of
the network nodes. If this information is in a geographic data base, it is called up into storage in
the computer. Otherwise, it is developed from a map of the system by use of digitizer
equipment. The high level of accuracy desired for mapping is not required for this application.
From this node coordinate data, a good schematic diagram of the network system is called up
on a display screen. It can color code ranges of values of design variables to graphically depict
the performance of the network. If the color-coded variable is pressure, red is usually selected for
the lowest range of pressures. The graphic reveals the areas of poorest pressure in the
system very clearly.
A black and white example of pressure coding based on a variety of line styles is shown in
Figure 5-11 for a low-pressure system. The lightest lines correspond to a pressure range of 3.5-
6.3 in. wc. Similarly, if the gradient (rP2/L) is the color-coded variable, the resulting graphic clearly
shows the bottle necks in the system as a guide to modification. Other variables that can be
color-coded are pipe diameters, flow rates, and gas heating values, when gases of different
heating valve feed into the system from different sources. Color graphics is a considerable aid to
the designer, and it is a very effective aid in communicating the results of network analysis to
administrative decision makers.
There are other network analysis applications in addition to those discussed earlier in this
chapter. Network simulations based on estimated loads at various ambient temperatures are
used by the operating departments of gas distribution systems to determine the minimum
temperature at which a facility, such as a gas main or regulator station, can be taken out of
service for maintenance work without creating low-pressure problems in the system.
System simulations under normal operating conditions, during extremely cold weather or under
emergency conditions such as a main damaged during third party construction work, are very
useful in training gas dispatchers, whose function will be described in Chapter XVI. System
simulations are also useful to the marketing department to identify areas of the system with
sufficient capacity to accept new loads without reinforcement or to establish the investment
required to permit adding large new commercial or industrial gas loads in a system.
Gas-supply planning is another activity that uses the results of system simulations. They help
determine the amount of supplementary gas, such as stored gas or propane-air, that will be
required on peak days. They also establish the way the system must be operated to satisfy
contractual obligations for both purchased gas and transportation service.
Fig. 5-11. Computer Graphic of Pressure Coded by Line Style
Network Analysis
When the Kirchhoff’s law equations are expressed in terms of pipe section flow rates, the
general form of the first law equations is:
åaQij = Lj - Sj (Eq. 5-1)
Where:
Lj = Rate of gas usage at node j, Mcf/h.
Qij = Gas flow rate in the pipe section between nodes I and j, Mcf/h.
Sj = Rate of gas input (if any) to node j, Mcf/h.
a = +1 when flow is toward node.
a = -1 when flow is away from node.
The summation is over all of the pipe sections connected to node j. For a network with a total of
n junctions, the total number of first law equations is n-I if the rates of gas input at all of the
sources are known, or n if two or more of these input rates are not known.
In the application of Kirchhoff’s second law, it is convenient to deal with the square of the
absolute pressures because this term occurs in all of the flow equations used for gas distribution
system flow calculations. Figure 5-12 will be used to illustrate the form of the second law
equations. It repeats Figure 5-9 with assumed flow rates added, the arrows indicate the direction
of flow. Application of Kirchhoff’s second law to loop B of Figure 5-12 in terms of the squares of
the absolute pressures gives:
(P62-P22) + (P72-P62)+(P32-P72) + (P22 -P32)=O (Eq. 5-2)
Where Pi is the absolute pressure at node i.
This equation is obviously true because each Pi2 is both added and subtracted so that their
algebraic sum is zero.
Fig. 5-12. Typical Schematic Diagram of a Gas Distribution System With Assumed Flow
Rates.
Substitution of the equivalent flow rate terms for each pair of terms in Equation 5-2 by use of a
flow equation gives:
-K2,6 (Q2,6)n + K6,7 (Q6,7)n - K3,7 (Q3,7)n - K2,3 (Q2,3)n = 0 (Eq. 5-3)
Where:
n = The value of the flow rate exponent when a flow equation is placed in the form:
r(P2) = Kij (Qij)n
The signs in Equation 5-3 result from the assumed flow directions and that the flow equation
gives the difference between the squares of the upstream and the downstream absolute
pressures. A simple sign convention proposed by Hardy Cross can establish the sign of each
term in Equation 5-3 without recourse to Equation 5-2. Flow rates in the clockwise direction about
a loop are considered positive, and flow rates in the counter-clockwise direction are considered
negative; each term in Equation 5-3 is given the same sign as that assigned to its flow-rate
term. The general form of the second law equations then becomes:
åekijêQijên = 0 (Eq. 5-4)
Where:
Kij = Resistance factor of the section of gas piping between junctions i and J.
e = +1 when the direction of the flow rate, Qij, is clockwise about the loop.
e = -1 when the direction of the flow rate, Qij, is counter-clockwise about the loop.
The vertical lines around the flow rate term, Qij signify that the value of the flow rate should be
treated as a positive number, even if its flow direction sign is negative. The e direction indicator
takes care of the flow direction.
The summation in each of these equations is over all of the pipe sections that form the k th
elementary closed loop or source connection. These source connections a continuous series of
pipe sections between gas sources maintained at a constant pressure, such as the outlet of a
pressure regulator. They provide a means of balancing loads between sources.
For a network made up of N elementary closed loops, in which the pressures at M nodes are
maintained at a known constant value with (M-A) maintained at the same pressure and A
maintained at different pressures, the total number of second law equations is (N+M-A-I). There
is one equation for each of the elementary closed loops plus one for each source connection
between the M-A source junctions maintained at the same pressure.
The general form of the A equations for the connections between sources maintained at
different pressures is:
åekijêQijên = C (Eq. 5-5)
Where:
C = Difference between the pressures (low-pressure system) or the squares of the pressures
(medium or high pressure system) maintained at the ends of the source connection.
The first systematic procedure for the solution of network flow problems was presented in 1936
by Hardy Cross, a professor at the University of Illinois. This method solves the second law
equations by an iterative, successive approximation procedure. The Hardy Cross method can be
used for the manual solution of small gas networks when a computer and network analysis
software are not available.
The logic and mechanics of this method are the subject of an AGA Technical Note provided
through the Distribution Engineering Committee of the AGA. Operating Section (American Gas
Association Technical Note, "The Hardy Cross Method for Network Analysis," Arlington, Va.: The
Association, 1994.) Reference should be made to this Technical Note for information on this
method.
Newton-Raphson Method
The best known general method of solving systems of non-linear, simultaneous equations, such
as Equations 5-1 or 5-4, is the Newton-Raphson method. Like the Hardy Cross method, it is also
an iterative method that requires initial assumptions for all values of the variables. In each
iteration, the system of non-linear equations is converted to a linear system of the same size and
then solved by an appropriate method, usually a special matrix technique. Each iteration is longer
and more complex than for the Hardy Cross method, but fewer iterations are required. The use of
this method for gas distribution network analysis was first suggested in 1955. Because the
mechanics of solution by the Newton-Raphson method are considerably more complex than for
the Hardy Cross method, its use was delayed until the late 1960s.
At that time, the method’s potential was recognized for integrated analysis of supply,
transmission, and distribution systems together and developed practical techniques for
implementing it. The operation of distribution system facilities other than pipe, such as regulators
and compressors, could now be represented as a function of pressure. Therefore, by choosing
node pressures as the variables to work with, the complete operation of a distribution system all
the way back to gas storage fields and city gate stations could be described by a system of
equations with pressure as the variable. This complete set of equations can then be solved
simultaneously to evaluate the performance of the entire gas supply system for a specified load
all at once. Pipe sections, compressors, and gas wells all become simply node
connecting elements whose performance can be described as a function of the pressures at the
two ends of the element. The general form of the equations for system elements other than pipe
sections are summarized below:
COMPRESSORS
REGULATORS
STORAGE WELLS
The work of the computer in network analysis begins with its use in developing design loads of
existing customers from consumption data and grouping these loads by load centers. The results
can be printed out in a load-node report and also stored on tape or disk for use as part of the
input required for network analysis. The analysis of networks of any size is usually done by digital
computer using a general purpose network analysis program. Programs have been developed to
do network analysis by both of the methods discussed.
The first computer program for network analysis was developed by D. V. Kniebes, of IGT, in
1955. It was based on the Hardy Cross method of analysis. Many gas companies soon
developed their own more sophisticated Hardy Cross programs. In the 1960s, the system-
planning engineers for a number of large distribution companies developed much more
sophisticated Hardy Cross programs. They aimed their improvement efforts in two primary
directions:
1. To minimize the clerical work required to prepare problems for solution.
2. To maximize the information content of the output.
This effort was extremely successful. The clerical work of data input was reduced, and very
close to the minimum possible. It consists primarily of a node-load list, source data, and a listing
of the end numbers, length, and diameter of all of the pipe sections in the system to be studied.
The node-load data for existing customers can be read directly into the computer from the tape or
disk load-node report generated by a computerized load estimation and gathering process. The
source data are simply a listing of the node number and input pressure (if one is specified) for
each source supplying the system.
Additional information on computer programs for network analysis based on the Hardy Cross
method can be obtained from the AGA Technical Note on the Hardy Cross Method. Several such
programs are still in use with both main frame and PC computers. These programs solve only for
pipe section flow rates and node pressures. Variations in other network flow parameters, such as
pipe diameter or regulator station characteristic constants, are handled on a case-by-case basis
by including the values of interest in the input data for individual solutions of each case. The
results of analysis are printed out in tabular form similar to those for the example in the next
section of this supplement.
Nawton-Raphson Programs.
Currently, the most extensively used computer programs for network analysis are based on the
Newton-Raphson method of analysis to solve the Kirchhoff's first law equations expressed in
terms of node pressures. Additional information on the fundamental basis for this type of network
analysis program and its applications is presented in the American Gas Association GEOP Series
(Stoner, Hunt, and Hyman, "Chapter 5 - Network Simulation," Distribution Book D-1: System
Design. Gas Engineering and Operating Practices Series, Vol. Ill, 203-248. Arlington, Va.:
American Gas Association. 1990).
The schematic diagram for a small, high pressure feeder system with two closed loops is shown
in Figure 5-13. As in Figure 5-12, the standing arrows coming out of the nodes indicate gas
loads. The number at the tip of each is the amount of the load in MCF/h. The squares with an R in
them represent regulator stations feeding a lower pressure system. The source is indicated by the
arrow pointing into node 22 at the upper right. The input pressure at this node has been set at 75
psig and the average atmospheric pressure is 14.4 psia. The numbers under or to the right of
each pipe section are its diameter and length.
These are the minimum allowable pressure requirements for this example system:
At the service to any customer: 3 psig
At the input to most service regulators: 3 psig
At the input to the regulator supplying gas at medium pressure at 25 psig
node 19:
At the input to all the other regulators serving the low pressure system: 3 psig
Pressure commitments at large industrial customers are:
Cargill Grain dryer at node 17: 36.8psig
Shoe factory at node 12: 11.1psig
Several improvements have been made to the system to handle the future loads expected to be
imposed on it. The objective of the network analysis was to determine if an input pressure of 75
psig at the source (node 22) was sufficient to satisfy all of the pressure requirements at peak load
conditions.
The analysis was conducted using Stoner Associates' GASSS Network Analysis Program and
the fundamental Moody (FM) flow relationship. With this program, the results are printed out in
three tables. One is a node report, Table 5-1, which gives the pressure and load at each node. A
quick review of the table indicates that the pressure at node 17, 32.87 psig, is below the pressure
commitment made to the Cargill Company. To meet this commitment, the source pressure can be
increased, if the maximum allowable operating pressure rating for the system is higher than 75
psig. Otherwise, additional reinforcements to the system must be made.
A second tabular report provided by this program is an exception report, Table 5-2, which
summarizes the following key results:
FIG. 5-13. Schematic Diagram of Feeder Systems
Table 5-1. Full Node Results from Pontiac Problem Analysis
Full Node Results for Pontiac Network Flow Problem – FM /Increased Corn Dryer Load on
System (Node 17 Load = 30 MCF/h)
NODE INFORMATION SOURCE
Node Pressure, psig Flow (MCF/h) Node Pressure, psig Flow (MCF/h)
10 44.09S 0.000 22 75.00 399.465 S
11 44.88S 0.000 23 45.48 S 56.800
SHOE FACTORY 24 42.96 S -1.500
12 45.90 S -19.500 25 42.94 S -1.635
13 41.27 S -21.100 26 42.82 S -1.635
14 38.04 S -1.000
15 36.57 S -1.200 3 48.16 S -3.800
16 34.66 S -2.300 31 53.09 S 0.000
CARGILL 4 47.03 S -5.200
17 32.87 S -30.000 5 46.02 S 0.000
18 64.07 S -6.000 6 43.45 S -19.600
19 66.16 S -2.500 7 42.97 S 0.200
STATION 157 8 42.80 S -3.00
2 49.10 S -144.800 STATION 142
20 71.06 S -2.000 9 42.57 S -75.000
21 72.19 S 0.800
Table 5-2. Exception Report for Pontiac Analysis
Exception Report for: Pontiac Network flow Problem - FM,lncreased
Corn Dryer Load on System (Node 17 Load = 30 MCF/h
Unknown Node flows (MCF/H)
22 399.465
SUM OF SYSTEM LOADS = -399.465 MCF/h
SUM OF SYSTEM SUPPLIES = 399.465
5 NODE(S) WITH LARGEST RESIDUALS (MCF/h)
21 -0.026
23 0.024
16 0.019
4 -0.014
10 0.013
5 NODE(S) WITH HIGHEST PRESSURES
22 75.00 psig
21 72.10
20 71.08
19 66.16
18 64.07
5 NODE(S) WITH LOWEST PRESSURES
17 32.87 psig
16 34.66
15 36.57
14 38.04
13 41.27
SYSTEM PERFORMANCE INDEXES
MCF-MILES INDEX = 634.08
INCH-MILES/MCF INDEX = 0.73437E -01
5 PIPE(S) WITH HIGHEST PSI /MI 2
31 2 6450. 4.124
31 2 6450. 4.124
9 10 -4627. 3.068
11 13 4195. 6.625
21 22 -3834.
5 PIPE(S) WITH HIGHEST VELOCITIES
31 2 102.90 fps
31 2 109.90
18 31 96.42
18 19 -84.21
19 20 -82.54
VOLUME OF LINE PACK (10 68 CF) = 0.0
1. The 5 nodes at which the incoming and outgoing flow rates are the farthest out of
balance (they should be equal).
2. The 5 nodes with the highest pressures.
3. The 5 nodes with the lowest pressures.
4. The five mains with the highest gradients (AP2ImiIe).
5. The five mains with the highest velocities.
These are the results likely to be of most interest and concern. Most of the locations at which
the pressure is inadequate are likely to show up in the lowest pressure tabulation. In this case,
the node at the Cargill grain dryer, node 17, is the first one listed.
Additional reinforcement options include paralleling or replacing mains working the hardest. The
gradient or rate of change in pressure per unit length of pipe is a good index of these mains. Of
the mains in the list, paralleling pipe section 21-22 or replacing it with a larger size would most
effectively increase the pressure at node 17 as well as everywhere else in the system. Pipe
sections 31-2 and 9-10 are not on the direct flow path to node 17. Pipe section 11-13 is on the
direct flow path, but upsizing or paralleling it would benefit only the branch of the system
downstream from it.
For larger systems, it is appropriate to increase the number of items in each sub-category to 10-
20 or perhaps even 50, depending on the size of the system. This variable can be specified by
the designer.
The third tabulation available from the GASSS program is the full node connecting element
report, Table 5-3. Node connecting elements (NC Es) are usually sections of pipe identified by
the node numbers at each end, but they can also be pressure regulators or compressors. This
report includes all of the system flow rates and pressures generated by the analysis. The from-
and-to designations of the first two node number columns indicate flow rate direction. If the flow
rate in the MCF/h column has a positive sign, the flow direction is away from the from node. If this
flow rate has a negative sign, the flow direction is toward the from node.
For pipe section 10-11, the flow rate is 10.194 MCF/h and the direction toward node 10. The F-
P and T-P columns give the pressures at the from and to nodes respectively. Thus, for pipe
section 10-11, the pressure at node 10 is 44 psig, and the pressure at node 11 is 45 psig. The S
label indicates that the connecting element between these two nodes is a pipe section. The series
of FM entries in the type column indicate that the fundamental Moody flow relationship was used
in all flow calculations.
The param. column lists the diameters of each pipe section in inches and the last column,
LEN= is the length of each pipe section in miles. The column of values of V= is the velocity of the
gas in each pipe section at flowing conditions and the RN- values are the Reynolds number
corresponding to the flow rate in each pipe section. It provides an archive report that isn't usually
printed until the study is completed.
For even a moderately large gas distribution system, the tabulation of the full node and NCE
results is a thick stack of paper. Here the exception report becomes particularly useful in pulling
together the results of significance to the system designer.
Table 5-3. Full Results For Pontiac Analysis
Full HCE Results for Pontiac Network
Flow Problem - FM Increased Corn
Dryer Load on System (Node 17, Load =30 Mcf/h)
FROM TO F-P T-F TYPE MCF/H PARAM REMARKS
10 11 44.S 45.S FM -10.914 3.068 V=-14.6 RN=98701. F=0.0192 LEN=0.472
11 12 45.S 46.S FM -16.517 4.124 V=48.5 RN=447505. F=0.0153 LEN=0.091
11 13 45.S 41.S FM 55.604 3.068 V=78.1 RN=502846. F=0.0157 LEN=0.099
12 31 46.S 53.S FM -86.015 4.124 V=-11.7 RN=578685. F=0.0149 LEN=0.421
13 14 41.S 38.S FM 34.505 3.068 V=51.4 RN=312045. F=0.0164 LEN=0.207
14 15 38.S 37.S FM 33.506 3.068 V=51.4 RN=303005. F=0.0165 LEN=0.095
15 16 37.S 35.S FM 32.306 3.068 V=51.5 RN=292154. F=0.0165 LEN=0.128
16 17 35.S 33.S FM 29.987 3.068 V=49.6 RN=271185. F=0.0167 LEN=0.133
18 19 64.S 66.S FM -394.177 6.625 V=84.2 RN=1650789. F=0.0130 LEN=0.089
18 31 64.S 53.S FM 388.175 6.625 V=96.4 RN= 1625651. F=0.0131 LEN 0.442
19 20 66.S 71.S FM -396.676 6.625 V=-82.5 RN=1661252. F=0.0130 LEN=0.216
2 3 49.S 48.S FM 157.376 6.249 V=47.4 RN=698735. F=0.0141 LEN=0.138
20 21 71.S 72.S FM -398.686 6.625 V=-78.2 RN=1669673. F=0.0130 LEN=0.046
21 22 72.S 75.S FM -399.460 6.625 V=-77.4 RN=1672914. F=0.0130 LEN=0.133
24 25 43.S 43.S FM 3.169 4.124 V=2.4 RN=21323. F=0.0259 LEN=0.310
25 26 43.S 43.S FM 1.531 2.067 V=4.6 RN=20549. F=0.0265 LEN=0.383
3 4 48.S 47.S FM 153.577 6.249 V=47.1 RN=681868. F=0.0141 LEN=0.l71
31 2 53.S 49.S FM 151.083 4.124 V=102.9 RN=1016441. F=0.0143 LEN=0.081
31 2 53.S 49.S FM 151.083 4.124 V=102.9 RN=1016441. F=0.0143 LEN=0.081
4 5 47.S 46.S FM 100.777 6.249 V=31.4 RN=447444. F=0.0148 LEN=0.334
4 10 47.S 44.S FM 47.613 4.124 V=35.2 RN=320323. F=0.0159 LEN=0.496
5 6 46.S 43.S FM 43.965 4.124 V=32.9 RN=295781. F=0.0161 LEN=0.497
5 23 46.S 45.S FM 56.824 6.249 V=17.9 RN=252296. F=0.0159 LEN=0.512
6 7 43.S 43.S FM 24.365 4.124 V=18.4 RN=163922. F=0.0174 LEN=0.267
7 8 43.S 43.S FM 19.500 4.124 V=14.7 RN=131189. F=0.0180 LEN=0.142
7 24 43.S 43.S FM 4.670 4.124 V=3.5 RN=31421. F=0.0237 LEN=0.155
8 9 43.S 43.S FM 16.499 4.124 V=12.5 RN=111000. F=0.0185 LEN=0.271
9 10 43.S 43.S FM -58.514 3.068 V=-80.3 RN=529162. F=0.0156 LEN=0.038
The restructuring of the gas industry mandated by the Federal Energy Regulatory Commission
(FERC) beginning in 1985 required gas distribution and especially transmission companies
to offer transportation service for gas users who purchase their gas directly from producers. If
only transportation gas is being supplied to a distribution system through a pipeline city gate
station, the design parameter for the station may be a daily take rather than a specified outlet or
set pressure. In that case, the peak demand period rate of use associated with the specified daily
take should be treated as a negative load in the analysis of the system.
Provision for combinations of flow and pressure control at take stations feeding the same
distribution system must also be incorporated in the network analysis process. (Place, J. C.,
"Effects of Multiple Pipeline Suppliers on the Design and Operation of Distribution Systems," AGA
Operating Section Proceedings 1990: Paper 90-DT-201, p 34244).
Data Management
One of the maxims of computer use is "garbage in, garbage out." That is particularly true for
network analysis. If the physical description data for a system does not accurately reflect the
facilities in service, network analyses based on these data will not accurately reflect its
performance. This is a particular problem for older systems for which records are incomplete. For
these systems, advantage must be taken of every opportunity to check the facilities in service
especially when they are excavated for maintenance or new construction. An important piece of
information is whether crossing mains are tied into each other.
The input data used for network analysis should also be thoroughly checked against the
existing physical description data. For older systems, it is a good idea to go back to the original
sources of such data like field notebooks. The status of the physical facilities as valves and
regulator stations also must accurately reflected in the input data for network analysis. Poor
pressure can be caused by a valve inadvertently left closed. The regulator set pressures used in
network analysis should correspond to those in current use.
It is common practice to build many checks for data consistency into network analysis
programs. This includes such things as checking for the satisfaction of Kirchhoff’s first law
at each node in Hardy Cross programs, or whether there are any mains or nodes not connected
to the system.
A good overall check on data accuracy is an analysis for the loads experienced on a cold day
with recording pressure gages set throughout the system, especially at the anticipated location of
lowest pressure and along principal feeder mains. Then, the calculated pressures are checked
against measured pressures. For 60 psig systems, the deviations should be no more than a few
psi, for 15 psig systems, less than 1 psi, and for the lowest pressure point in a low pressure
system, the pressures should match with 0.5 in. wc. Larger deviations will often indicate a
discrepancy between the physical description-data and the actual system. Comparison of the
calculated and measured pressures may also indicate the general area in which the discrepancy
exists.