配电网线损计算方法研究

发布时间：2018-04-04 23:00

本文选题：配电网　切入点：线损计算　出处：《湖南大学》2014年硕士论文

【摘要】：配电网线损计算是电力系统降损节能的重要技术手段，是线损管理科学化、规范化、制度化的实现基础。准确简便的线损计算有助于制定合理的降损措施，提高供电能力，增加电力企业经济效益。配电网结构复杂、分支线路多，往往缺乏准确、完整的线路和负荷资料，导致常规线损计算方法往往难以实施。针对这一问题，本文利用神经网络特有的非线性拟合特性，研究易于操作、可行性高，且满足工程计算精度的配电网线损计算方法。首先，，论文针对中压和低压配电网的特点，分析了中压和低压配电网现有线损计算方法，指出现有线损计算方法的适用条件与不足。其次，针对中压配电网配电线路结构复杂，运行数据不全，常规线损计算方法难以实施的问题，将径向基函数（Radial Basis Function，RBF）神经网络应用到中压配电网线损计算中，利用它的拟合特性，映射配电线路线损与特征参量之间复杂的非线性关系，记忆配电线路在结构参数和运行参数变化时线损的变化规律，建立了基于RBF神经网络的中压配电网线损计算模型。然后，对自适应二次变异差分进化（Adaptive Second Mutation DifferentialEvolution，ASMDE）算法进行了改进，采用了重构交叉概率因子思想和近似最优保存策略。利用改进的ASMDE算法对RBF神经网络的结构参数进行整体优化，克服了常规网络训练算法隐含层与输出层结构参数分开确定，输出层易陷入局部极小等缺点。实例仿真验证了所提中压配电网线损计算模型与算法的可行性和先进性。最后，研究低压配电网的线损计算方法。低压配电网供电方式复杂多样，分支线路众多，沿线用电负荷没有严格的规律，自动化程度不高，线路参数和负荷资料严重缺乏。基于上述问题，将BP神经网络用于低压配电网的线损计算与分析中，并对基于BP神经网络的线损计算模型的输入参数进行了详细地分析，找出了引起配电台区线损变化的主要参量，将其作为BP神经网络模型的输入参数。利用Matlab神经网络工具箱完成了网络的训练，通过实例对所建低压台区线损计算模型进行了仿真，结果验证了所建模型的准确性和实用性。
[Abstract]:Line loss calculation of distribution network is an important technical means to reduce loss and save energy in power system. It is the scientific, standardized and institutionalized realization foundation of line loss management.Accurate and simple line loss calculation is helpful to make reasonable loss reduction measures, improve power supply capacity and increase economic benefits of power enterprises.The structure of distribution network is complex and there are many branch lines, which often lack accurate and complete data of line and load, so it is difficult to carry out the conventional line loss calculation method.In order to solve this problem, this paper makes use of the special nonlinear fitting characteristic of neural network to study the distribution network line loss calculation method which is easy to operate, high feasibility and meets the engineering calculation accuracy.Firstly, according to the characteristics of medium voltage and low voltage distribution networks, this paper analyzes the existing line loss calculation methods of medium voltage and low voltage distribution networks, and points out the applicable conditions and shortcomings of the existing line loss calculation methods.Secondly, aiming at the problems that the distribution line structure is complex, the operation data is not complete and the conventional line loss calculation method is difficult to carry out, the radial basis function (RBF) Basis function neural network is applied to the line loss calculation of the medium voltage distribution network.By using its fitting characteristic, the complex nonlinear relationship between line loss and characteristic parameters of distribution line is mapped, and the variation law of line loss of distribution line when the structure parameter and operation parameter change are memorized.The line loss calculation model of medium voltage distribution network based on RBF neural network is established.Then, adaptive Second Mutation differential evolution (ASMDE) algorithm of adaptive quadratic mutation differential evolution is improved, and the idea of reconstructing crossover probability factor and approximate optimal preservation strategy are adopted.The improved ASMDE algorithm is used to optimize the structural parameters of the RBF neural network, which overcomes the shortcomings of the conventional network training algorithm, such as the structural parameters of the hidden layer and the output layer are determined separately and the output layer is prone to fall into local minima.The simulation results show that the proposed model and algorithm are feasible and advanced.Finally, the line loss calculation method of low voltage distribution network is studied.The power supply mode of low voltage distribution network is complex and diverse, the branch lines are numerous, there are no strict rules of power load along the line, the degree of automation is not high, and the data of line parameters and load are seriously lacking.Based on the above problems, the BP neural network is applied to the line loss calculation and analysis of the low voltage distribution network, and the input parameters of the line loss calculation model based on the BP neural network are analyzed in detail.The main parameters which cause the line loss change in the distribution station area are found out and used as the input parameters of the BP neural network model.The Matlab neural network toolbox is used to train the network. The model of line loss calculation in low voltage station is simulated by an example, and the accuracy and practicability of the model are verified.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM744

【参考文献】