机组耗量特性的混合整数模型建立与分析

Modeling and Analysis of Unit I/O Characteristics Based on Mixed-integer Programming

对机组耗量特性的建模是机组组合问题的一个重要组成部分。文中建立并分析了机组耗量特性的5种模型,并在IEEERTS96系统上,分别针对具有凸连续、非凸连续和非连续耗量特性曲线的机组,通过优化其在某日24时段的出力,比较各模型的求解效率,并根据各模型所引入变量数目对测试结果进行分析,从而为根据机组特性选择合适的模型以提高求解效率提供参考。结果表明:影响算法求解性能的主要因素是模型中整数变量的数目。对于凸连续耗量特性机组,尽量采用各区间出力叠加的建模方法,以避免引入整数变量;对于非凸和非连续耗量特性机组,引入辅助整数变量不可避免,此时则应采用连续变量较少的模型。

Modeling the I/O characteristic curves of the units is an important part of the unit commitment problem. Five models for I/O curves of units are presented and analyzed. Based on the IEEE RTS 96 system,the units with convex,non-convex and discontinuous I/O curves are scheduled respectively,and the results and computing times are compared based on different models in different cases. Furthermore,the number of integer variables,continuous variables and constraints based on different models are listed to analyze the differences in computing efficiency between the models,which is helpful in choosing the appropriate model according to the characteristic of unit for higher efficiency. The results show that the number of the integer variables affects the performance of the algorithm remarkably. For units with convex I/O curves,it is better to model the power by adding up the outputs in all the segments to avoid integer variables. For units with non-convex or discontinuous I/O curves,the assistant integer variables are inevitable and it is better to use a model with less continuous variables.