改进的混合整数算法求解机组组合问题  

Unit Commitment based on Improved Mixed Integer Programming

作　　者：夏新华[1] 高宗和[2] 徐帆[2] 邵烨楠 郑涛[2] 

机构地区：[1]国网电力科学研究院,南京211106 [2]国电南瑞科技股份有限公司,南京211106

出　　处：《电力学报》2016年第3期202-210,共9页Journal of Electric Power

摘　　要：为解决复杂的机组组合问题,在对比拉格朗日松弛法和混合整数算法的基础上,提出基于改进的混合整数算法求解机组组合问题。模型对混合整数非线性规划进行线性化处理,以机组的运行成本最小化为目标函数,考虑了包括爬坡约束在内的机组性能约束、机组初始状态约束等相关约束。该模型分别用于10机至100机测试系统进行计算分析,结果表明:相对于其他算法,该方法收敛速度更快,计算结果更优;而且,机组规模越大,越能体现该方法的优越性。To solve the complex unit commitment problem,the paper proposes an improved mixed integer programming（MIP）algorithm on the basis of comparing the advantages of LR and MIP algorithms.The problem is formulated to minimize the total operating cost of units.The model took the unit ramp constraint and initial state constraint and other relevant constraints into consideration by linearizing the mixed-integer nonlinear programming（MINLP）.Test systems with the number of machines in the range of 10 to 100are calculated and analysed.The result shows that the convergence rate of the method is faster and the calculation result is better.Besides,the superiority of the method is more apparent with the increase of the number of units.

关 键 词：机组组合 拉格朗日松弛法 混合整数算法 经济调度 

分 类 号：TM73[电气工程—电力系统及自动化]