This paper describes a dynamic programming model which minimizes the cost of transporting natural gas through a pipeline network. The cost is minimized by selecting the combination of compressor engines along the pipeline which meet the system pressure and throughput requirements, yet minimize the cost of fuel gas and engine/pipeline maintenance. Given expected demands and pressure requirements on the current pipeline configuration, the model determines the discharge pressure, horsepower required and a recommended engine selection at each compressor station.
Michigan Wisconsin Pipe Line Company transports natural gas from gathering areas in Texas, Oklahoma, Louisiana and the Gulf of Mexico along mainlines to be distributed in Michigan and Wisconsin. The pipeline contains 11.000 miles of pipe located in 15 states with mainline pipe size ranging from 14 inches to 42 inches in diameter. The Company has 74 compressor stations with over 1 million horsepower available to pump gas. Gas is controlled in pipeline operations by the Gas Dispatch Department located in Detroit, Michigan. The Gas Dispatch Department has the responsibility of issuing discharge pressure orders (and compressor engine schedules if requested) to each compressor station along the pipeline system. A Gas Dispatch monitoring system in Detroit displays various information on a CRT screen retrieved from the field operation. A typical screen (Figure A) shows station information on suction and discharge pressure, suction and discharge temperature, total horsepower being used and total horsepower available. Based on this information and from information received directly from the station superintendents, dispatchers determine what discharge pressure orders should be given to each compressor station. For example, any or all of the following indicators require action from Gas Dispatch : excessive pressure, excessive temperature, insufficient pressure differential, engine taken down for maintenance and engine breakdowns. When a problem occurs, the dispatcher decides how to remedy the situation and informs the station(s) of the appropriate course of action. For example, excessive discharge pressure at Station A might prompt Gas Dispatch to request Station B to boost discharge pressure resulting in a lower suction pressure at Station A. This, in turn, results in a lower discharge pressure at Station A. Gas Dispatch must issue pressure orders in such a way as to meet all contractual (throughput and pressure) obligations without violating key physical constraints at any compressor station or pipeline segment. If these key physical constraints are violated, severe penalty costs to the company are incurred. For example, if a maximum discharge gas temperature is exceeded, the protective coating on the pipe may be damaged or if the maximum pressure differential between suction and discharge pressures is violated engine breakdown may occur. Therefore, the Gas Dispatch Department's primary concern is meeting demand without violating these physical constraints while the actual cost of gas delivery is secondary. For these reasons, it has been determined that a model which can determine the discharge pressures for each compressor station to meet all Company obligations without violating any physical constraints, yet minimize cost of delivery is required.
