Sidney P. Santos
A Must in Gas Pipeline Design
Sidney Pereira dos Santos
Engineering Service of Petrobras - SEGEN
Rio de Janeiro - Brazil
Traditionally, a new project for a gas pipeline system starts in a technical economical evalua-tion phase which considers a steady state flow analysis with a lot of guesswork and assumptions, such as market delivery build up, minimum and maximum flow and estimated number of compres-sor stations. For a reasonable evaluation of the system a load factor is normally adopted for the gas pipeline, which means that the pipeline will be calculated with a constant flow which will be higher than the nominal flow considered to be supplied to the foreseen market. The main difficulty is to de-fine a duty load that would represent the system under analysis since different flow profiles, operat-ing pressures, diameter and length of the pipeline will affect the behavior of the complete system.
As soon as the technical economic evaluation shows an attractive return rate for the project and the companies involved decide on a go ahead position, the design phase takes place. It is in this phase that a transient analysis for the system has proved to be a must as we have experienced in the design of the Bolivia-Brazil Gas Pipeline. It is a 32" diameter and 1127 miles (1813 km) long pipeline from Rio Grande (Santa Cruz de La Sierra - Bolivia) to Campinas (São Paulo - Brazil) with 14 compression stations. It is designed for 1.043 Bcf/d (30 MM m3/d) of natural gas.
Why do a transient analysis in the design phase?
The first reason is related to the large investment that is required for such a system which in-cludes pipeline and compression stations. The system designed must be able to operate in the pre-dicted different scenarios, otherwise the transportation company would face penalties for not deliv-ering the contractual volumes of gas or make additional capital investment on the system, dramati-cally affecting the cash flow of the project.
A second reason is that in this phase normally we face the inclusion of new deliveries with dif-ferent profiles such as thermal generation that may interrupt the gas consumption to zero for a cer-tain time on a weekly basis, as will normally happen in Brazil. These scenarios must be taken into account since they directly affect the schedule of placing compression stations and units into opera-tion depending on the gas build up delivery.
Additionally we may use the transient analysis to make a pre-definition of the turbo compres-sors that best fit our system and also define the best arrangement for the units in the station, whether in a series (few units and bigger machines) or parallel arrangement (more units and smaller machines). In line with this we may do a failure analysis for a single unit or for a complete station and foresee how the system would cope with that, determine its remaining capacity, and also define a much better maintenance procedure, or even detect a necessity of a standby unit in the worst cases.
The transient analysis will also be very useful during the negotiations of ship or pay contracts that starts even before the design phase.
This paper will focus on a single line gas pipeline without storage facilities and with a flow de-mand that varies with respect to time in an hourly basis so as to show a behavior that could not be considered as a steady state flow. The software used was TGNET 5.3E from SSI.
The Bolivia-Brazil project consists of a 32" gas pipeline that goes from Rio Grande, Santa Cruz de La Sierra, in Bolivia, to the city of Campinas in the State of São Paulo, Brazil. From Campinas the pipeline continues going through the southern states of Brazil up to the city of Porto Alegre in the state of Rio Grande do Sul. The pipeline also connects with existing old and recently built gas pipelines and with the offshore gas production facilities, as can be seen from figure 1. The section of the gas pipeline with 32" diameter and 1127 miles (1813 km) was designed to handle up to 1.043 Bcf/d (30 MM m3/d, at 20C and 1 atm) of dry gas at Campinas.
The system has 14 compressor sta¬tions in the 32" sec¬tion with 4 turbocom¬pressor units (no stand by) of 7000 hp ISO per unit and 2 other compressor station in the south¬ern section with 2 or 3 units in parallel with approximately 1200 hp per unit.
The purpose of this paper is to pres¬ent the experience we gained at Petrobras as we designed the system from con¬ceptual design up to the basic design, us¬ing different tools that were available (based on load factor) in order to define the first outlook of the system. Later, with the transient analysis tool, a more accurate definition was obtained in the basic design which proved to be very helpful and a must in future designs.
2. System Configuration
The system configuration adopted for the project considers a MAOP (Maximum Allowable Op-erating Pressure) of 1420 psig (99.84 kgf/cm2g) for the section from Rio Grande to Campinas and extending up to Curitiba, and 1067 psig (75 kgf/cm2g) from Curitiba to Porto Alegre.
The sections of 24" and 32" are internally coated with epoxy paint with roughness of 350 micro inches (0.009 mm) and for other non coated pipelines the roughness considered is 700 micro inches (0.018 mm).
The pipeline profile varies from sea level to 3610 feet (1100 m).
2.1 Schematic Diagram
2.2 Elevation Profile
3. Gas Supply from Rio Grande, Bolivia
The gas supply from Bolivia for this project varies from 318.6 to 866 Bcf/d (9.16 to 24.91 MMm3/d at 20C and 1atm), including fuel as can be seen from the table below. Normal Build Up and TCO (Transportation Capacity Option) are two different scenarios considered for the project.
4. Design Criteria for the Bolivia-Brazil Pipeline
The criteria adopted to design the Bolivia-Brazil pipeline considered a delivery of 1.043 Bcf/d (30 MMm3/d) of natural gas at Campinas, State of São Paulo, in Brazil. With this steady state flow we defined the compressor stations required and also the spacing for the compressor stations. Af-ter that we considered the scenarios for each year of operation and also defined the number of sta-tions required for each year. The transient tool was of capital importance on these definitions.
5. Model Initial Simplifications
To initiate the system calculations, the input for steady state analysis based on load factor, was simplified without consideration of elevation profile. The flow was assumed to be isothermal and generic equipment were utilized.
6. Steady State Analysis
The steady state analysis with simplified model helped to perform a preliminary evaluation of the system requirement, to anticipate the quantity of compressor stations and locate them along the gas pipeline, and also produce information to begin selecting the turbo compressors. The target is to define the compressors and gas turbine drivers with the same size so as to simplify operation, maintenance, spare parts, station design, personnel training, and improve interchangeability be-tween the units.
For the Bolivia-Brazil project the criteria adopted was to define the configuration of the system for the design condition which is 1.043 Bcf/d (30 MM m3/d, at 20C and 1 atm). With station quantity and locations defined, we made the simulations backward to the first year of operation, taking the stations out of service where they were not required for operation.
7. Transient Analysis
In addition to the steady state analysis, the transient approach is fundamental to fine tune the system with regard to equipment sizing, as well as to define with accuracy and reliability, the gas transportation system. It is in this step that we consider everything that is available such as delivery profiles, pressure limits, performance maps for compressors and gas turbines, heat transfer be-tween pipe and soil, elevation profile, etc.
With every detail available introduced in the complete model with the exception of compressor and gas turbine performance maps, we can perform a preliminary transient study to define the per-formance maps for the design condition which will apply for every compressor station.
The main advantage of this analysis is to forecast, operate and detect any operation problem that may come up.
A complete transient analysis is then carried out from the design condition backward to the first year of operation. The performance of the compressor stations allocated from the steady state analysis is checked for each year of operation for the scenarios considered, but now considering the delivery profiles. In some cases we realized that the station allocation from steady state had to be changed to cope with transient scenarios.
This analysis was extremely important for the project because we could see well in advance the need of capital investment to guarantee the ship or pay and take or pay contracts without any risk for the future, as well as better define the transportation cost involved in those scenarios.
7.1 Delivery Profiles
7.2 Schedule of Compressor Stations Installation
7.3 Equipment Detailed Characterization
The turbo compressors previously selected from the transient analysis are shown in the fig-ures below. In the bid document we gave the option for the manufacturers to propose 3, 4 or 5 turbo compressor per station allowing different sizes for the compressor and driver, and the best choice proved to be the one defined from the transient analysis, which is 4 units per station in parallel ar-rangement. For performance maps see item 9.1.
7.4 Transthermal Evaluation
The transient analysis was performed with transthermal mode keeping the discharge gas temperature of the compressor station at 125 F (51.6 C) maximum.
8. Transient versus Steady State With Load Factor
One thing that becomes clear when doing transient analysis is that we can not afford to use only steady state tool in the design phase if an economic and reliable project is to be executed. For the purpose of comparison we have done some simulations in order to see what could happen in case this project would be designed only with steady state analysis based on load factor.
The load factor normally adopted by the engineering design offices for gas pipeline design var-ies from 0.8 to 0.95, which corresponds to an increase of the nominal flow from 1.25 to 1.05 respec-tively. The main problem with this approach is that we can not guarantee an economical and reliable design. For comparison purpose see item 8.1 which shows examples for the beginning of the gas pipeline (from Rio Grande to Station #1) and for the southern section (from Station #15 to Porto Ale-gre). The graphics show what could happen to the Bolivia-Brazil project in case we had kept the ap-proach using only load factor design.
8.1 Section From Rio Grande to Station #1 (Izozog)
This gas pipeline section from Rio Grande to Campinas was designed to deliver 30 MMm3/d (at Campinas) to Brazil, with a load factor of 100%. The purpose of this design condition was to de-fine the quantity of compressor stations and their location along the pipeline. The end pressure for this section is 71.06 kgf/cm2g at 31.24 MMm3/d.
8.1.1 Load Factor Design versus Transient Design
In order to make a comparison on what could happen to a pipeline design based on load factor when operating under transient conditions we considered the following assumptions for the section from Rio Grande to Station #1:
(a)L.F. @ 100% 869 MMcf/d (24.99 MMm3/d)
(b)L.F. @ 90% (90% of item (a) above) 966 MMcf/d (27.77 MMm3/d)
(c)L.F. @ 80% (Bolivia-Brasil design flow) 1086 MMcf/d (31.24 MMm3/d)
Item (a) above assumes a load factor of 100% which means no flow variation at the end point, or a pure steady state mode of operation. Item (b) is item (a) divided by 0.9 and item (c) is item (a) divided by 0.8.
The idea is to run a transient analysis for item (b) and see the results for both delivery profiles, compare the pressure variations with the ones we got from the Bolivia-Brazil design and see the benefit of the transient tool in the design phase.
8.1.2 Main Section - Delivery Profile Type 1
8.1.3 Main Section - Delivery Profile Type 2
As we can see from the above graphics, the delivery profile makes a marked influence on the end pressure of the section, and this considered alone would lead, even if we use of a transient simulator, to increase the diameter of the pipe or to shorten the stations spacing. However, taking into account all the supplies and deliveries of the system and fine tuning the system, we can im-prove the model and also optimize capital investment with regard to the pipeline and installed power.
Although the steady state result for items (b) and (c) shows pressures above 995.6 psig (70 kg/cm2g), the transient run for item (b) alone gives pressure around 853 (60) and 427 psig (30 kgf/cm2g) with the two (2) delivery profile considered, which would require future and unforeseen capital investment for contract obligations.
By considering the entire gas pipeline system operating in transient mode, taking advantage of line pack, the influence of other supplies spread along the system, and also the dynamic behavior of the compressor stations, the results may help us optimize the sizing of the pipelines. The graphic "Actual Readings From Complete Model" shown below proves this point.
8.2 Section From Station #15 (Araucária) and #16 (Biguaçu) to Porto Alegre
This is another example on the savings and reliability we acquire when using transient analy-sis in the design phase of a project.
The graphic below shows the Bolivia-Brazil design item 8.2.1 with the diameters selected for the sections and the end pressure profile that is kept above the minimum delivery pressure of 35 kgf/cm2g.
The following graphic item 8.2.2 shows what could happen if we had designed based on a load factor of 90% with smaller diameters for the pipe sections. The end pressure is too low, al-though in the steady state it is higher than the minimum required pressure.
8.2.1 Transient Design - South Section - Profile Type 2
8.2.2 Load Factor Design (90%) - South Section - Profile Type 2
9. Compressor Station Definition, Parallel versus Series Units Arrangement
It was discussed in the beginning of the project that the use of larger and fewer units in a se-ries arrangement would have advantages, with regard to power installed cost and operation cost (fuel and maintenance).
The failure analysis, in addition to the studies done to the first years of operation (which re-quire smaller installed power), helped decide on better arrangement for the compressor station. The following graphics show that the parallel arrangement was a better selection since, in case of a failure of one unit in a station, the decrease in flow was not as critical as in the series ar-rangement.
This kind of analysis has a close relationship with the ship or pay contract. The parallel ar-rangement in this project, proved to be more reliable and more flexible and also economically more attractive. Again, all of this analysis was possible by using a transient simulator.
9.1 Turbo Compressors
9.1.1 Parallel Arrangement
9.1.2 Series Arrangement
9.2 Parallel Arrangement Transient Output Graphics
9.2.1 Parallel Arrangement Failure Analysis
9.3 Series Arrangement Transient Output Graphics
9.3.1 Series Arrangement Failure Analysis
9.3.2 Comparable Table
10. Final Conclusions
This paper, which is based on a thorough and careful design of the Bolivia-Brazil gas pipeline, has the main purpose of stressing the advantages of using transient simulation. Not only as a tool for training operation personnel and act as a helpful tool in on-line systems, but also to emphasize its use in the design phase of a gas pipeline.
If a project has to be economic, flexible and reliable, a transient simulator must be one of the most impor-tant tool to be used.
The author wants to thank the following:
Service Engineering of Petrobras - for the permission to present this paper.
Sebouh Ohanian of Solar Turbines Incorporated - for the promptitude in supplying equipment
in-formation and professional skills.
12. About Petrobras S.A., SEGEN and the Author
Petrobras is a state owned integrated oil company. Petrobras has acquired a worldwide reputation from its achievement mainly in deep water exploration at 5607 feet (1709 meters) and production at 4593 feet (1400 meters), oil and gas research, pipeline design and construction management of re-lated projects. It produces 888,500 bbl/d of oil and condensate and 932 MMcf/d (26.8 MMm3/d) of natural gas.
SEGEN is the Service Engineering arm of Petrobras entitled to develop most of the oil and gas pro-jects necessary to keep Petrobras on the main stream with the renowned oil and gas companies, with updated technologies In addition to providing the engineering support to cope with Petrobras requirement, SEGEN is also capable to offer its expertise to the worldwide market.
Sidney Pereira dos Santos, the author, has a Mechanical Engineering degree and has 13 years of experience in shipbuilding design and construction, and also 10 years of experience in the oil and gas pipeline design at Petrobras . He has been deeply involved in most of the gas pipeline projects such as the Bolivia-Brazil project. He has been conducting technical-economic studies and ba-sic/conceptual design for the upcoming projects.
Phone: +55 21 8167-0134