Development of a fuel minimization program for natural gas transmission pipelines

Patrick Aaron Seamands, Louisiana Tech University

Abstract

An improved computer based method of scheduling multiple compressor operations for complex natural gas transmission pipeline networks which include operating loops, loops containing a compressor, has been developed. The technique minimizes compressor fuel consumption for various demand scenarios. The procedure, combining dynamic programming with a pattern search simplex algorithm, was developed by Stoner Associates, Inc. for Arkla Energy Resources (AER). The author envisioned the need for this improvement, received corporate approval, and acted as bidding, procurement, and project officer. The model was validated using real process historical data and recommended acceptance. Fuel savings of 1.5 to 25.6 percent were predicted.

AER was using a pipeline simulation model to review pipeline operations, both for historical evaluation and for operational planning purposes. The planning studies were being performed using the time consuming case study method. For fuel minimization studies, various operating strategies were simulated and the compressor fuel consumption for the cases were compared. The software was tested on increasingly more complex gas transmission pipeline configurations. Two of the five cases which were used were based on AER's real pipeline system and on actual historical operating data for a peak day and an average winter day. The peak day data was from the single operating day that had the highest volume of deliveries. The winter day was based on the daily average of data for the coldest month of the winter. The solution methodology that was used combines dynamic programming (DP) and a pattern search simplex algorithm. The operating loops are cut and optimized with the simplex algorithm and the primary path is optimized using DP. A flow simulation model is still used to reproduce the operating flows and pressures for each iteration.

Although the program cannot guarantee a global optimum since a pattern search is used, improved solutions could not be found by the case study method. The software which was developed provides an optimization program to guide the engineer in selection of compressor operation to minimize fuel consumption quickly by eliminating the need to do case studies. The software was general enough that it has been marketed and licensed to several natural gas transmission companies.