One-segment linearization modeling of electricity-gas system optimization

摘要

Linear models of energy systems are preferable in industries because linear optimization problems bring benefits in convergence, efficiency, and convenience for pricing. The nonlinear model of the natural-gas system is a major bottleneck for the fully-linear modeling of electricity-gas systems. In this paper, a one-segment linear gas flow model without the need of integer variables is presented. To facilitate the linear formulation of the gas flow model, a deep learning method is applied to predict the interval of the gas pressure. A special setting of the linearization interval is proposed, which avoids the influence of the range of the linearization interval on the optimal energy flow (OEF) solution. Based on the one-segment linear gas flow model, a fully-linear OEF model of the electricity-gas system is formulated. The efficiency of the proposed method is much improved compared with the traditional mixed integer linear programming approach.