【作者】 易波；

【导师】 刘建华；

【作者基本信息】 长沙理工大学 ， 电力系统及其自动化， 2013， 硕士

【摘要】 配电系统是联接输电系统和用户的枢纽，在电力系统中占据重要地位，在保证配电系统安全稳定运行的情况下，才能保证人民的正常生活和国家经济的正常发展。配电网重构是在满足配电网正常运行的约束条件下，即满足馈线功率、电压以及配电网辐射状态运行等约束条件，通过变化配电网中的分段开关和联络开关开合状态，使整个网络实现降低线路损耗、消除过载以及改善电压质量等优化目标，最终使配电网在既经济又稳定的状态下运行。本文提出了基于模拟渔夫捕鱼算法和遗传算法为基础的二次协作优化方法对配电网重构网损最优问题进行了研究，该方法对于求解配电网重构网损最优问题达到了期望的结果，对配电网重构问题研究提供了一定的理论指导意义，而且，该方法也可以应用于智能电网的研究。首先，本文介绍了几种常见的配电网接线方式，建立了配电网网络结构的拓扑分析模型，并且采用树状数组来储存配电网的节点信息以及支路参数，这为配电网的辐射结构判断以及接下来的潮流计算都奠定了基础。在配电网重构潮流计算过程中，本文采用了基于支路电流方式的前推回代法计算算法过程中潮流变化，该方法能够快速地求解出目标函数的解。其次，在介绍模拟渔夫捕鱼算法和遗传算法的理论基础上，提出结合模拟渔夫算法和遗传算法优点的二次协作优化方法求解基于网损最优的配电网重构问题，模拟渔夫算法具备快速收敛局部最优解的特点，但是全局搜索能力差，遗传算法具备很强的全局寻优能力，但是遗传算法的初始种群是随机生成，导致其收敛速度较慢而且容易陷入局部最优解的问题，二次协作优化方法是利用模拟渔夫捕鱼算法来寻找遗传算法的的初始种群，因为渔夫捕鱼算法具备快速收敛局部最优解的能力，得出的局部最优解都是朝着全局最优解的方向发展的，极大地改善了遗传算法寻找全局最优解的效率和速度，由于这些局部最优解都是合理分布在全局最优解的周围，所以避免了遗传算法陷入早熟，最后利用遗传算法的比较强的全局寻优能力寻找全局最优解。为了测试二次协作优化方法的寻优性能，将模拟渔夫捕鱼算法、遗传算法和二次协作优化方法分别求解基本函数f1、f2和f3的全局最优解，在对求解基本函数f1、f2和f3的结果中，可以得知，二次协作优化方法的寻优能力是优于遗传算法和模拟渔夫捕鱼算法的。在算法运算过程可能出现不可行解，这些不可行解不仅会降低算法的运行效率，甚至造成算法最终不能收敛于全局最优解。本文首先生成一个连通性判断程序判断每一个染色体对应的网络结构图是否满足辐射状的约束条件（即网络结构中不存在“环网”和“孤岛”现象）。如果判断出该染色体对应的网络结构图不连通，说明它是不可行解，便采用基因疗法的“补”操作和“泻”操作将其改变成可行解，这样既不会丧失种群的多样性，也避免了算法陷入局部最优解的问题，又提高了算法的运算效率。最后，本文将基于二次协作优化方法应用于求解基于网损最优的配电网重构问题中，通过Matlab程序模拟求解算例IEEE69节点系统，结果表明所构建的配电网重构数学模型能有效地降低配电网中的网损；所提出的二次协作优化方法具有搜索效率高、性能好的特点。更多还原

【Abstract】 The power distribution system is the hub between transmission system and the user, itplays an important role in power system, under the condition of guarantee the safe and stableoperation of power distribution system, to ensure that people’s normal life and the normaldevelopment of national economy. Power distribution network reconfiguration is to meet theconstraint conditions of normal operation of distribution network, which satisfies the feederpower, voltage and power radiation condition of such as running, through changes in thedistribution network section switch and connective switch opening and closing state, makingthe whole network to reduce line loss, eliminate overload and improve voltage quality, etc.The optimization goal, eventually make the power distribution network in the condition ofboth economic and stable operation. Quadratic collaborative optimization method is proposedin this paper, based on the Simulation fisherman fishing algorithm and genetic algorithm fordistribution network reconstruction, the optimal network loss is discussed, the method forsolving the problem of distribution network reconfiguration the optimal network lossachieved the desired results, the distribution network reconfiguration problem study providescertain theoretical guiding significance, moreover, this method can be applied to the study ofthe smart grid.First of all, this article introduces several common distribution network connection mode,established a distribution network topology analysis model of network structure, and uses thetree to store an array of distribution network node and branch information parameters, for theradiation of the structure of distribution network as well as the next tide calculation laid afoundation. In this paper, reconstruction of distribution network power flow calculation in theprocess of power generation method based on branch current method calculation algorithmschange trend in the process, the method can solve the objective function of the solution morequickly.Secondly, the author introduces the Simulation fisherman fishing algorithm and based onthe theory of genetic algorithm, combined with the advantages of Simulation fishermanfishing algorithm and genetic algorithm is put forward second collaborative optimizationmethod to solve the problem of distribution network reconfiguration based on the optimalnetwork loss, and Simulation fisherman fishing algorithm with fast convergencecharacteristics of local optimal solution, but the global search ability is poor, geneticalgorithm has strong global optimization ability, but the genetic algorithm of the initial population is randomly generated, cause its convergence speed is slow and easy to fall intolocal optimal solution of the problem, secondary collaborative optimization method is to useSimulation fisherman fishing algorithm to find the initial population of genetic algorithm,because the fishing algorithm with fast convergence ability of local optimal solution, andconcluded that the local optimal solution is toward the direction of the global optimal solution,and greatly improve the efficiency of genetic algorithm to find the global optimal solution andspeed, as a result of these local optimal solution is reasonable distribution around the globaloptimal solution, so to avoid the genetic algorithm into puberty, finally using comparativestrong global optimization ability of genetic algorithm to find the global optimal solution. Inorder to test the second collaboration optimization performance optimization method,simulated the fisherman fishing algorithm, genetic algorithm and the second collaborationoptimization method respectively to solve the optimal solution of the basic functionf1、f2and f3,in the results of solving basic function, and learn that the secondary searchingcapability of the collaborative optimization method is superior to genetic algorithm andsimulation algorithm of fisherman fishingComputing process in algorithm may be infeasible solution, the infeasible solution willnot only reduce the efficiency of the algorithm, and even cause algorithm cannot converge toglobal optimal solution finally. At first, this paper generates a connectivity judge program tojudge each chromosome corresponding network structure whether meet the constraintconditions that the radial (that is, network structure does not has the "ring" and "island"phenomenon). If you determine that the chromosome corresponding network structure is notconnected, that it is infeasible solution, then using gene therapy to "repair" and "release" tochange it into operation feasible solution, such both neither can lose the diversity ofpopulation, also avoid the algorithm trapped in local optimal solution of the problem, andimprove the efficiency of the operation of the algorithm.Finally, this article will based on quadratic optimization method to solving the problemof distribution network reconfiguration based on the optimal network loss in IEEE69nodethrough the Matlab simulation numerical system, and the results show that the construction ofdistribution network reconfiguration mathematical model can effectively reduce the networkloss in distribution network; The proposed secondary collaborative optimization method havethe characteristics of high search efficiency and good performance.更多还原