Dynamic Programming Based Algorithm for Compressor Station Optimization

The process of determining how to operate a compressor station, providing that its volumetric flow rate, suction pressure and suction temperature have already been fixed, to meet a required discharge pressure with the minimum energy cost is called compressor station optimization. In this paper, the process of formulating a compressor station’s operation scheme is transformed into a multistage decision-making process and within each stage, the volumetric flow rate of a single compressor is decided. The multistage process starts from setting the flow rate of the first compressor and ends when that of the last one is set. With this basic idea in mind, a dynamic programming based algorithm is proposed to solve the optimization problem, which has advantages of tackling the non-linearity and discontinuity of the problem effectively and resulting in a global optimum. The algorithm is tested on an in-service compressor station, and a successful 5.1% energy cost reduction was achieved. Discretization step size analysis on the specific problem reveals that a 2% step size is suitable for stand-alone usage while the 7.5% step size is enough for embedded situation. Comparison with a simple genetic algorithm shows that the proposed method performs better under moderate and coarse step size, which is satisfactory from an engineering aspect.Copyright © 2014 by ASME