|
Line loss as an important technical and economic indicators of the power system, has long been the widespread attention of the electricity enterprises and related departments. Especially since the reform of the electricity market, the line loss rate has been directly linked to the tariff impact on the operational efficiency of enterprises. Therefore, accurate and easy calculation of line losses, laying out the effective loss reduction measures is extremely important for assessment enterprise line loss management effectiveness. The theoretical line loss calculation method first in-depth analysis, pointed out the limitations of the various line loss calculation method. And has a large number of components for our distribution network, distribution complex, generally low degree of automation, the raw data is not easy to collect, a RBF neural network based on genetic algorithm for distribution network line loss calculation method . The method by RBF neural network fitting and radial basis function space local response characteristics mapping nonlinear complex relationship between the parameters and line loss of distribution lines, and for traditional RBF network learning methods, between the hidden layer and output layer structure parameters of determining independent output layer weight training is easy to fall into the local minimum and other shortcomings, the application of genetic algorithms throughout the RBF network carried optimize the RBF network different center and its corresponding width and various regulating the right to re-unification coding, to strengthen the cooperation between the hidden layer and output layer of RBF network, and genetic algorithm global search features, making the entire network model to reach the global optimum. In addition, the improvement of the genetic mechanism of the genetic algorithm itself, the genetic manipulation more perfect. In order to verify the practicality and feasibility of the proposed method, respectively, as a sample to a region 68 distribution lines and Tianjin Binhai Power Supply Bureau 67 line characteristics of the distribution line parameters, the line loss instance simulation. The Test Simulation results show that the genetic algorithm to optimize the RBF network, the network model is simple, fast training, computational accuracy, and has a strong practical and promotion. Capacity-expansion capabilities, the use of neural networks fit the nonlinear relationship between the distribution line characteristic parameters line line loss can be more accurate memory available, and thus a more accurate calculation of line losses. Finally, Borland C Builder 5.0 as a software development platform, based on the idea of ??object-oriented programming, to develop a set of adapted to the distribution network line loss calculation visualization software. The software has a the distribution line drawing, components run data entry, database management, line loss calculation, report output function. Friendly interface, and provide a lot of shortcuts, with international standard software interface style, easy to learn to master. The latter part of software design, conducted a number of tests and use, and to ensure the integrity, security and reliability of software.
|