Linear programming (LP) can be used in an engineering environment to optimize the design and operation of natural gas systems. A basic overview of the LP technique is given, along with the advantages and difficulties of using LP in an engineering analysis. Several design and operational problems are discussed including: fuel minimization, scheduling production, and system capacities. The LP technique is applied to each of these situations. An appendix provides an overview of the implementation of LP in the ISTAR model used by Michigan Wisconsin Pipeline Company.


Linear Programming is a relatively old methodology for optimizing complex systems. While it has been used for a number of years to analyze the financial aspects of pipeline operations, it has not seen widespread use in the engineering and design analysis of natural gas pipelines. However recent changes in the gas industry, such as rising costs and declining markets, have increased the need for effective design and planning optimization. Linear programming can help meet this need. This paper provides a short explanation of the Linear Programming (LP) technique and some of it basic advantages. The major discussion centers on a number of current operating and analytical problems common to much of the gas industry and how linear programming can be used to solve them. These problems range from short term operational planning to long term facility designs. Typically, the problems are complex and difficult to optimize using traditional analytical programs. The conclusion of the paper outlines the new areas that Michigan Wisconsin Pipeline Company is looking to apply LP analysis in the future. An appendix following the paper describes some of the methodologies used in ISTAR, the Michigan Wisconsin LP model.

Brief Description of LP

Linear Programming (LP) has been used for several decades as a means of optimizing complex systems with many variables. Its earliest applications included control, inventory control and modeling system costs. As early as 1965 LP was being used to help schedule gas by the Pacific Lighting system of California in conjunction with the AGA NX-37 project. This work emphasized the costs of gas operations rather than comprehensive engineering simulation. While cost optimizations have been around for a number of years, the modeling of the engineering attributes of a pipeline system; i.e pressure losses, compression, and flows, has seen rather limited development. One of the primary reasons for this is a requirement that Linear Programming equations be linear and continuous. Since almost nothing in pipeline analysis is both linear and continuous, this has restricted the development of LP models for true engineering pipeline simulations. The recent rise in the cost and complexity of natural gas system operations has led Michigan Wisconsin to examine the engineering uses of Linear Programming more closely. Before describing how LP is being used by Michigan Wisconsin, let me provide a brief description of the linear programming technique. The variables within the system (pressures and flows) have specified upper and lower limits.

This content is only available via PDF.
You can access this article if you purchase or spend a download.