【作者】 李春晓；

【导师】 舒隽；

【作者基本信息】 华北电力大学（北京） ， 电力系统及其自动化， 2011， 硕士

【摘要】 传统电力系统备用优化一般在给定备用总需求的前提下优化各机组的容量分配。这种备用优化方式不能体现不同预想场景对不同响应速度备用的要求,同时难于考虑复杂预想场景,本文针对这一不足,结合安全约束机组组合,建立了复杂预想场景下的备用优化模型。通过引入预想场景分类和预想场景集,建立了预想场景与不同响应速度备用的对应调节关系：瞬时响应备用(AGC)调节瞬时变化场景(瞬时变化负荷、机组故障、线路故障等),10min旋转备用主要调节缓慢变化场景(缓慢变化的负荷)。在此基础上,以最小化电能量成本与备用成本之和作为目标函数,以正常状态和预想场景下的机组性能,网络安全等为约束条件,综合求解机组组合、功率分配、备用分布以及预想场景下备用容量的优化调度,建立了备用、预想场景均不分类；备用分类、预想场景不分类；备用、预想场景均分类的三种预想场景下的备用优化模型。由此，不同响应速度的备用需求量和分布将不再由备用需求约束条件所决定,而是由预想场景的类型和要求所决定。针对该模型的复杂性,通过模型线性化,将其转化为混合整数线性规划问题。最后采用当前先进的通用代数建模系统GAMS(General Algebraic Modeling System)进行建模,并调用商用软件CPLEX的混合整数线性规划求解器进行求解。IEEE30节点测试系统上的算例结果表明：本文所提出模型的计算结果较传统方法更为合理和有效；同时本文方法不仅能考虑单重场景,还可以考虑多重场景以及场景的集合,为电力系统备用优化的研究提供了一种新思路。更多还原

【Abstract】 Traditional reserve optimization usually specify the requirements as fixed system constraints, which of two shortcomings:one is can not reflect the different scenarios’requirements to reserve with different corresponce rate, the other is hard to consider complicated contingency scenarios. According to the difficulty of traditional optimization methods, this paper presented an optimal reserve dispatch model considering contingency scenarios based on security constrained unit commitment (SCUC). By introducing scenarios classification and scenarios set, this paper establish a correspondence between contingency scenarios and reserve with different rate, that is:instantaneous reserve(AGC) adjust to instantaneous scenarios while 10min spinning reserve to slowly changing scene. Integrated with unit commitment, power dispatch, reserve distribution and the reserve capacity optimization of the contingency scenarios, the optimization model is to minimize the sum cost of the electricity and reserve while subjected to the security constraints conditions of base case and contingency scenarios. Therefore, the reserve demand and dispatch of different response rate are decided by the requirement of different types of scenarios, rather than the reserve demand constraints conditions. This complicated mixed-integer linear programming problem is modeled by GAMS and solved by CPLEX. The Numerical examples on IEEE30 bus test system show that: the results of the proposed model are more reasonable and effective than the traditional method; the proposed model can not only consider the single scenario, but also the multiple scenarios and the scenarios set, which provides a new idea for the set the reserve optimization.更多还原