ALBERTA EXPANSION-CONCLUSION TRANSIENT SIMULATIONS HELP DECIDE EXPANSION DESIGN

Stanley R. Kitt
Foothills Pipe Lines Ltd.
Calgary

Foothills Pipe Lines Ltd., Calgary, evaluated two approaches to solving pipeline-capacity delivery problems around the Empress, Alta., extraction plants (Part 1, OGJ, June 24, p. 38).

This capacity constraint resulted from the requirement for gas stripping by low-pressure Empress extraction plants which severely limited the operating pressure level of the high-pressure pipeline there.

The engineering evaluation employed off-line hydraulic transient analyses of alternative low-pressure and highpressure scenarios.

The problems arose as Foothills began planning an expansion to export more Canadian gas into the U.S. The location is about 160 miles north of Foothills' Monchy, Sask., delivery point to the Northern Border Pipeline in the U.S.

The segment of line around the Empress plants forms part of the 396-mile Eastern Leg prebuild section of the future Alaska Natural Gas Transportation System (Angts) pipeline project. (See Angts system map in Part 1, Fig. 1.)

The existing Foothills pipeline, although designed for a maximum operating pressure of 1,260 psig, was operationally integrated with the lower pressure NOVA Corp. of Alberta pipeline system in Alberta. Both pipelines delivered gas volumes to existing extraction plants at Empress at a fixed inlet pressure of approximately 610 psig.

The Foothills system, thus constrained in pressure at the Alberta-Saskatchewan border, was nevertheless required to transport gas the remaining distance to Monchy for delivery to Northern Border at maximum 1,260 psig. In the expansion case, this pressure was to be further increased to 1,435 psig.

Extensive engineering analysis, including off-line pipeline transient simulations to evaluate potential operational problems, indicated that the best approach for Foothills' system expansion entailed raising the operating pressure in the Foothills pipeline at Empress to nominal 1,000 psig and constructing the decompression-recompression facilities to maintain gas-stripping capabilities.

In late June 1989 Foothills received approval from the Canadian National Energy Board (NEB) to construct the facilities. Construction was completed and the facilities placed into service in November 1990.

The new facilities expand the gas from the nominal 1,000 psig pipeline pressure to the nominal 600 psig extraction-plant operating pressure then recompress the extraction-plant residue gas to original pipeline pressure.

With the gas directed through turboexpanders coupled to brake compressors, a substantial portion of the expansion energy is captured and used in gas recompression. Supplemental compression required to return gas to original pipeline pressure is provided by electricity-driven compressors,

TRANSIENT CONCERNS

Transient analyses of both the low-pressure, integrated and the high-pressure, segregated configurations were undertaken with an off-line, hydraulic transient simulator to examine the flow and pressure transients that would be experienced in these pipelines under particular compressor-unit outage conditions.

Such analysis 'enables the operator to find the greatest transient effects, evaluate the operational stability of a proposed system, and determine specifically whether any equipment additions or modifications (spare compressor unit, for example) might be needed to reduce transient concerns.

Steady-state hydraulic analysis offers an indication of the areas vulnerable to transients which may lead to system operational deficiency or instability.

Some serious design and operational concerns apparent from steady-state pressure profiles of the two system alternatives (particularly of the low-pressure, integrated option) included the following:

System stability. Would transients generated by unit outages successively knock off-line upstream or downstream stations so that operation of the entire pipeline could be jeopardized?
Spare units. Could system instability be overcome by the placement of spare compressor units at certain locations?
McNeill-Empress impact. What magnitudes of flow-pressure impact would the Empress plants, the TransCanada pipeline, and the NOVA pipeline experience as a result of station outages on the Foothills pipeline, and vice versa?
Transient absorption capability. Could transients be absorbed either wholly or in part by operator intervention?
Monchy deliverability. How would station upsets on the Foothills system affect the deliverability at Monchy?

STATION 369 SPARE?

The normal operating scenario for the low-pressure, integrated option based on steady-state hydraulics has Stations 369, 391, 392, and 394 (all single unit) operating, while Station 393 is on standby. Unresolved from steady-state analysis was the need for a second standing spare located at Station 369. (See Part 1, Fig. 2 for station locations.)

Transient analyses were therefore undertaken to evaluate both alternatives; that is, with and without the Station 369 spare. The case studies examined an outage of Station 369 (no-spare), an outage of Station 369 (with-spare), an outage of Station 392, and an outage of Station 394.

The first two case studies address the transient effects upon the pipeline and the interconnecting facilities under an outage of the first compressor downstream of the extraction plants. Because of the station's proximity to Empress, this outage would be expected to yield the most severe transients.

These cases also examined the survivability of the downstream pipeline and the contribution of the spare at Station 369 towards system stability.

The latter two cases focus more upon effects at Monchy of upsets of stations nearer Monchy and also upon the migration of transients upstream towards Empress.

In all cases, the station outage is assumed to have occurred at 0.5 hr (as plotted along the time or horizontal axis), and the station remained off-line for the duration of the simulation. At 1 hr, or 30 min after the outage of a particular station, the appropriate spare is activated.

In the case of Station 369 (no-spare) outage, the spare at Station 393 is activated. In the case of Station 369 (with-spare) outage, the second unit at Station 369 is activated instead.

For the latter two cases with Station 392 and Station 394 outages, the spare at Station 393 is activated.

Selected results of the transient simulations for the two cases of Station 369 outage, namely, no spare and with spare, have been superimposed for purposes of comparison and are presented in Fig. 1.

The general observations that emerge from an analysis of the study include the following:

The magnitude of the flow transient at McNeill (Fig. la) immediately after the outage of Station 369 is approximately equal; i.e., flow instantaneously drops from nominal 1,400 MMcfd essentially to 0 and remains there for approximately 12-15 min.
Flow will have recommenced, however, before a spare unit at Station 369 could be brought on line, recovering to approximately 5060% of original flow within 30 min after shutdown.
Flow recovery at McNeill after the initial surge remains slow in the case of no-spare at Station 369 and substantially below the original level for a prolonged period. On the other hand, flow recovery for the case of with-spare at Station 369 is full and instantaneous once the spare is brought on line, which is 30 min after shutdown.
The flow impact upon the Empress plants-NOVA pipeline (results not shown) is essentially the same for the two cases and quite significant as the flow drops from roughly 6,400 MMcfd to less than 5,200 MMcfd (a drop of 20%) instantaneously.
However, the duration of the transient is relatively short, so that the extraction plants should be able to cope by coming off-line and recycling. The upstream pipeline, however, will be unable to escape the transient.
The flow and pressure transients at downstream Stations 391 and 392 are severe, with Station 391 being the more adversely affected (Figs. 1 b and 1 c).
For the case of no-spare at 369, the model instability was so severe that operator intervention was required to produce a successful simulation. There was strong suspicion that Station 391 would have difficulty staying on-line during actual operations in the period immediately following the outage and before the spare unit could be activated, as compressor surge or stonewalling would probably be encountered.
If such were the case, downstream stations would be affected more adversely than shown, and system operation could become untenable. With a spare at Station 369, system deterioration would be arrested and stability restored, provided that Station 391 could be kept online until the spare was activated and its effects felt downstream.
The flow delivery at Monchy over time deteriorates to 90% of original level for the case of no-spare at Station 369 (on the assumption that the Saskatchewan stations could be kept online), while it remains unaffected for the case of with-spare at 369.

The conclusion that emerged from the comparative analysis of the Station 369 outage transient study was that a spare compressor unit at Station 369 was deemed essential to preserve system operational stability and integrity.

TRANSIENT COMPARISON

The normal operation of the high-pressure, segregated system has one unit operating at Station 367 in Alberta (with one unit on standby) and three units operating in Saskatchewan (Station 392 on standby).

Transient simulations were performed for the following station upsets: outage of Station 391, outage of Station 393, and outage of Station 394.

In each case, the station outage was assumed to occur at time 0.5 hr, and the spare at Station 392 activated at 1 hr. In the case of Station 391 outage, the control of upstream Station 367 was switched from discharge pressure to suction-pressure control to take advantage of the available absorption capability in that section of pipeline.

The relative operational performance of the highpressure, segregated option as compared to the alternative low-pressure, integrated (with Station 369 spare) option can be illustrated by examining the transients resulting from an outage of the first compressor station downstream of the Empress plants in each case; that is, under Station 391 outage in the high-pressure case vs. Station 369 outage (with spare) in the low-pressure case.

Selected superimposed graphs illustrating the transient effects for these two cases at two locations along the pipeline are presented in Fig. 2.

An analysis of the comparative results yields several general observations:

The transient impact at McNeill for the two options is approximately equal (Fig. 2a), conditional upon system stability for the low-pressure case being preserved by immediate activation of the Station 369 spare.
The magnitude of flow loss is greater for the low-pressure option, but recovery is somewhat faster as a result of the spare unit at Station 369 being activated.
Flow transients in the pipeline downstream of the station experiencing the outage tend to be more volatile and unstable for the low-pressure option than for the high-pressure option, but in both cases Monchy deliverability is maintained.
Pressure transients remain more of a concern for the low-pressure case. For example, the essentially instantaneous suction-pressure decline of approximately 180 psig at Station 391 raises concerns about that station's ability actually to remain on line and, therefore, about potential system instability.

Such concerns would become exacerbated should the Station 369 spare not be activated 30 min after the unit outage as assumed but in fact be delayed, perhaps because of operator uncertainty.

The conclusion that can be drawn from the comparison of transients generated by station outages is that although the severities of transients are approximately equal in the low-pressure and high-pressure options, the risk of system instability remains higher for the lowpressure case.

An additional advantage of the high-pressure, segregated option lies in the fact that the transients can generally be contained within the Foothills pipeline, the only interconnecting facility in the Empress area which would be affected being the Empress 11 plant.

The transients in this case can probably be absorbed in the upstream Foothills (Alta.) pipeline, but at worst they would not directly affect the NOVA pipeline until after migration to the Schrader Creek station at Caroline.

On the other hand, the transients in the low-pressure, integrated option would affect all of the extraction plants in the Empress area, as well as the NOVA and TCPL pipelines.

EXPANSION SYSTEM DESIGN

The extensive engineering and operations analyses performed for the alternative configurations for accommodating increased exports at Monchy resulted in the highpressure, segregated option being selected as the optimum manner for system expansion.

The factors involved in arriving at this conclusion included the following:

The segregation of the Foothills pipeline from NOVA enables the Foothills system to be operated at or near its design MAOP of 1,260 psig, with its attendant improved hydraulic efficiency.
The high arrival pressure at Empress greatly relieves the downstream pipeline capacity.
Concerns about pipeline operational instability due to flow and pressure transients are alleviated.
Because of the segregated nature of operation, pipeline transients would be essentially contained within the Foothills pipeline, whereas in the case of the low-pressure alternative, the interconnecting pipelines and plants at Empress would all be affected.
Transients generated on the NOVA-TCPL pipeline would likewise not affect the Foothills pipeline operation.
Foothills pipeline deliverability would be improved because the decompression-recompression facilities would be neutral to pipeline operation.
Capital costs for the decompression-recompression facilities are significantly lower than for the alternative compressor station located at Empress (Table 1).

The selection of the highpressure, segregated mode of expansion constitutes a significant departure from the established regime of Foothills pipeline operation which has been prevalent since the original Eastern Leg prebuild was placed into service.

Gas stripping, coupled with the operating restriction of the existing extraction plants at Empress, posed a serious obstacle to Foothills' being able to raise its pipeline operating pressure in the Empress area. The low-pressure, integrated manner of expansion on the other hand posed serious operating concerns.

Consequently, the concept of decompression-recompression facilities straddling the Foothills' pipeline offered a unique solution to Foothills' expansion needs.

Foothills applied to the NEB for permission to construct the decompression-recompression facilities; approval was granted in late June 1989.

Construction of these facilities was completed and the facilities placed into service in November 1990.

OPERATIONAL LINK

The purpose of the decompression-recompression facilities is to provide the link that would enable the Foothills pipeline to be operated at high pressure in the Empress area (950-1,000 psig), while allowing the existing, low-pressure extraction plants at Empress to continue stripping those volumes transported by Foothills in the traditional manner.

The general decompression-recompression process involves first the decompression or expansion of the gas from pipeline operating pressure to plant operating pressure by routing the gas through turboexpanders, by which a substantial portion of the expansion energy is recovered and used in recompressing extracted residue gas partly back to original pipeline pressure.

Supplemental compression facilities are then used to compress the residue gas the remaining way to original pipeline pressure. The decompression-recompression process also involves the conditioning of both the inlet and return gas streams to temperatures acceptable to the extraction plant and pipeline, respectively.

The facilities are located adjacent to the Empress 11 extraction plant. The decision to locate them at this site was made based on the proximity to the Foothills mainline; the availability of plant processing capacity; the availability of utilities such as fire and cooling water, instrument air, nitrogen, glycol, emergency power supply, and flare stacks, which could be shared by the facilities; and the suitability for common operatorship with the extraction plant.

The Empress 11 extraction plant was in all these respects the most suitable for location of the decompression-recompression facilities. It is the plant with the greatest available spare processing capacity, adequate to accommodate both the initial as well as the ultimate Foothills' volumes.

It is also the plant generally nearest the Foothills pipeline, and in addition, since it itself is adjacent to the Empress I plant and has common operatorship with the Empress I plant, offers the greatest economic benefits that would accrue from sharing common utilities and facilities as well as common operatorship.

DESIGN CONSIDERATIONS

The decompression-recompression facilities are the interface between the higher operating pressure Foothills main pipeline and the lower operating pressure Empress 11 extraction plant.

Design of the facilities, therefore, was required to conform to the requirements and constraints of both the pipeline and the extraction plant. Following are some of the considerations which guided the design of the facilities:

The facilities initially must be capable of accommodating 1,425 MMcfd of pipeline inlet gas and 1,350 MMcfd of processed, outlet gas.
The facilities must be capable of accommodating a relatively broad range of inlet and outlet volumes in accordance with projected seasonal flows in the Foothills pipeline.
The facilities must be capable of easy expansion for a future, ultimate pipeline volume of approximately 2,000 MMcfd.
The facilities must be designed so that pipeline return pressure after recompression will be generally equal to the pipeline inlet pressure before expansion under normal pipeline operation, i.e., the facilities must be essentially "neutral" to the pipeline operation.
The facilities must be capable of continuous operation under minor pipeline transient conditions. In the event of severe transient conditions in the pipeline, the facilities must be capable of being isolated temporarily from the pipeline.
The facilities must be designed to provide reliability equivalent to that of the extraction plant itself.
The facilities must condition the inlet gas to the extraction plant so that the inlet-gas temperature is in the range of 40 to 50 F. at all times.
The facilities must cool the gas stream returning to the pipeline to temperatures acceptable to the pipeline, for example, to 85 F. on an average July day.
The facilities must deliver gas to, and receive gas from, the extraction plant at a pressure of 615 psig.

THREE PARALLEL TRAINS

Based on the initial design capacity requirement of 1,425 MMcfd at the inlet and 1,350 at the outlet of the facilities and an eventual ultimate inlet capacity of approximately 1,800-2,000 MMcfd, coupled with the requirement that equipment for the eventual expansion be similar to that installed for initial operation, it was decided that the facilities be configured to comprise two parallel equipment trains initially, and three such trains ultimately.

Each individual equipment train comprises two gas-togas exchanger shells, one turboexpander-brake compressor, one recompressor, and two water-cooled heatexchanger shells.

Additionally, equipment common to all three trains included piping, inlet separator, cooling-water tower, electrical switchgear, and other miscellaneous equipment items, all of which are sized for the ultimate capacity.

The simplified process flow diagram provided in Fig.3 illustrates the process conditions for the design operation of the facilities.

Gas from the pipeline is received at the facilities at 1,000 psig and 75 F. After separation of entrained liquids and any solid particles in the inlet separator, the scrubbed gas is heated in shell-and-tube heat exchangers prior to being routed to the turboexpanders for expansion (decompression).

The heating medium in this exchange is the hot recompressed gas en route back to the pipeline. The heating of the inlet gas is controlled by regulation of the supply of the hot compressed gas to the exchangers. The heated inlet gas is expanded in the turboexpanders to pressure and temperature conditions acceptable to the extraction plant, namely 615 psig and 50 F., respectively.

The low-pressure residue gas, after processing in the extraction plant, is returned to the facilities for compression and cooling prior to return to the pipeline. The first stage of this recompression is achieved with the brake compressors which are directly coupled to the turboexpanders.

Approximately 50% of the pressure energy lost during the expansion of the inlet gas is recovered in the brake compressors. Supplemental compression provided by electric-drive recompressors is then used to pressure the residue gas back to the original pipeline pressure.

The hot, fully compressed gas is cooled first in the gas-to-gas exchangers where its energy is used to heat the inlet gas. Additional cooling of the gas stream to temperatures acceptable to the pipeline is achieved in the water-cooled, residue-gas coolers prior to its being returned to the pipeline.

ROTATING EQUIPMENT

The turboexpanders are radial inlet-axial outlet units equipped with actuator-controlled guide vanes for control of flow through the equipment. The guide-vane mechanism is designed to withstand full expander inlet pressures and can be varied from approximately 0 to 130% of the design mass flow rate.

The turboexpanders are designed to process 712.5 MMcfd of gas each. Under design conditions (suction pressure of 1,000 psig and discharge pressure of 615 psig), the expanders will generate approximately 11,100 hp at an isentropic efficiency of 85% and a speed of approximately 11,000 rpm. The machines are designed to produce a maximum of 14,650 hp at a speed of 13,000 rpm.

The directly coupled centrifugal compressors will, at design point, operate at an isentropic efficiency of 81%. The centrifugal recompressors for the facilities are axial-in and radial-out machines and equipped with actuator-operated inlet guide vanes to control inlet flow.

The units are fitted with dry gas seals and will be driven by constant-speed electric induction motors rated at 11,000 hp. The fixed-speed motors were chosen over variable-speed electric motors or gas-turbine drivers primarily on the basis of economics. The compressor guide vanes will provide the required operational flexibility to compensate for the fixed-speed motors.

All equipment items common to the current two trains and the future third train have been sized for the eventual maximum capacity of the facilities.