【作者】 张成；

【导师】 张保勇； 谢云云；

【作者基本信息】 南京理工大学 ， 电力电子与电力传动， 2017， 硕士

【摘要】 电网大停电事故后的机组恢复问题是一个非线性组合优化问题。合理的机组恢复顺序能加快整个系统的恢复速度,减少停电损失。随着电网互联规模的扩大,系统安全运行面临考验,易发生大面积停电事故,提高大规模机组恢复控制策略的计算速度成了研究人员关注的重要问题。目前,机组恢复顺序的优化大多以运行经验和专家系统为基础。启发式算法,智能算法等方法也相继应用到机组恢复顺序问题之中。专家系统知识库的建立与维护工作量大,且具有局限性,难以全面获取系统信息。启发式算法和智能算法虽然能在一定程度上加快计算速度,但是对于大规模机组的计算时间仍然较长。因此,需要进一步研究提高机组恢复顺序问题计算速度的算法。本文提出了基于拉格朗日松弛法的停电系统机组恢复顺序优化的方法,主要完成了以下工作:1.针对当前停电系统机组恢复顺序数学模型复杂,形式多样的问题,从原理上分析了机组恢复的过程,并在此基础上简化了目标函数和约束条件,从而建立了停电系统机组恢复顺序的数学模型。数学模型以机组的状态变量作为目标函数和约束条件的决策变量,不再具体区分黑启动机组与非黑启动机组,形式上得到统一,方便算法的求解。2.针对传统算法计算速度上的不足,提出了将拉格朗日松弛法应用于求解大规模停电系统机组恢复顺序的优化方法,分析了拉格朗日松弛法的数学原理,在求解对偶问题的过程中,结合可分整数规划问题从原理上对计算速度的提高给出了合理的解释。3.针对拉格朗日松弛法收敛速度慢的问题,一方面给出了拉格朗日乘子初值的选取方法,另一方面在优化过程中采用次梯度法和自适应次梯度法相结合的方法迭代拉格朗日乘子,加快了算法的收敛速度,进一步减少了计算时间。4.通过对实际的新英格兰系统,广东电网部分分区系统,江苏电网部分分区系统进行计算分析,验证了算法的有效性。更多还原

【Abstract】 Generator restoration after large scale blackout is a nonlinear combinatorial optimization problem in the interconnected power grid.Reasonable generator restoration sequence play a very important role in speeding up the system restoration process and minimizing the outage lost.With the expansion of interconnected power grid,the safe operation of the system face test,and it is easy to occur large area blackout,it becomes an important problem for the researchers to improve the computing speed of restoration control strategy of the large scale units.At present,optimization of generator restoration sequence for blackout system is mostly based on the operation experience and expert system.Heuristic algorithm,intelligent algorithm and other methods subsequently are applied to generator restoration sequence optimization problem.The workload of establishment and maintenance of the knowledge base in expert system is very large,expert system is limited,it is difficult to get information of system in an all-round way.Although heuristic algorithm and intelligent algorithm can speed up the calculation in some degree,the computation time of large scale units is still long.Therefore,we need further study the algorithm which can improve the computing speed of generator restoration sequencing problem.This paper proposes a Lagrange Relaxation based optimization of generator restoration sequence for blackout system.The main work completed is as follows:1.For the current mathematical mode of generator restoration sequence for blackout system is complex and various,we analyzed generator restoration process in principle,on this basis we simplified objective function and constraint conditions,thus established the mathematical model of generator restoration sequence of blackout system.The unit state variables as the decision variables of objective function and constraint conditions,it isn’t to specifically distinguish black start units with non black start units,unified the form,and it is convenient for the algorithm to solve the problem.2.Considering the shortcomings of the traditional algorithm computing speed,proposed a generators restoration sequence optimization method for the large scale blackout system which is based on Lagrangian Relaxation,analyzes the mathematical principle of Lagrangian Relaxation algorithm,and in the process of solving the dual problem,combining with the decomposable integer programming problem to explain why this method can improve the computing speed.3.In view of the slow convergence speed problem of Lagrangian Relaxation algorithm,on the one hand the paper gives the selection method of the initial value of the Lagrange multiplier,on the other hand the subgradient algorithm and the adaptive subgradient algorithm are adopted to amend lagrangian multiplier in the process of optimization,speeding up the convergence speed of the algorithm,further reduce the computation time.4.Based on the actual New England system,partition system of the guangdong power grid,and partition system of the jiangsu power grid,we verified the effectiveness of the algorithm through simulation and analysis.更多还原