Study On Radial Temperature Model Of High-Voltage Overhead Transmission Line Based On Artificial Neural Network

The radial temperature model of high-voltage overhead transmission line can be often used on transmission line design the carrying capacities estimation.Considering the fact that the study of the temperature of the transmission line on power system operation has been interested for several years,the temperature model of transmission line can be applied on the study of correcting line equivalent parameters,estimating risk of line operation,the electro-thermal coordination theory and so on,which puts a magnificent influence on the traditional power system analysis and its dispatch theory,and there are vital theoretical significance and engineering practical values in it.IEEE standard model and thermal-circuit model have been often applied in project for calculating the line temperature,and both of models take the line as an isothermal object.However,there are several air gaps in transmission lines consisting of aluminum and steel strands,so the lines are conductors with not good heat transfer.Therefore,there is an axial or radial temperature distribution in the line on its operation.It is hard to get the line temperature by traditional temperature model at that time.Although the Finite Element Method and the Distribution Parameter Thermal-Circuit Model for calculating line temperature distribution have been presented for several years,there are some disadvantages in both models.The first model can calculate the line temperature distribution by disperse grids,which can be equipped with high accuracy,but it needs a huge computation,and hence it is seldom applied in project.The second model has been usually applied in describing the axial temperature of the line,instead of the radial temperature.Considering this background,my paper is designed to propose a model for calculating the radial temperature of transmission line based on artificial neural network.Compared to traditional models,the model can acquire relatively accurate temperature differences in the core and surface of lines.The guidance for operation and maintenance of high-voltage transmission line system can be provided by enhancing the accuracy of the line temperature estimation.In this paper,some common models for calculating the temperature of transmission line are discussed,including IEEE Standard Model,Lumped-Parameter Thermal-Circuit Model and Distributed-Parameter Thermal-Circuit Model,and then the features and the calculation method of relative parameters of the models above are analyzed,which put a theoretical basis on the following study.Then,an experimental plate for the radial temperature of overhead transmission line has been built.As the conductor type has been chosen as LGJ-400/35,the influence of wind velocity,wind angle,ambient temperature and current on the radial temperature difference can be analyzed by experimental data.On this background,the BP(Back Propagation)neural network and the RBF(Radial Basis Function)neural network have been introduced,in which some parameters of wind velocity,wind angle,input ambient temperature,line current and its surface static temperature are put as the variables,and the parameters of the core static temperature of conductor and its thermal time constant are put as the output variables.Based on the two types of neural network,the model of calculating conductor radial temperature has been proposed.In the end,the temperature results by the proposed model and real measurement have been shown that the artificial neural network can be capable of relatively describing the process of the radial temperature rise on the static meteorological conditions,and compared to the model based on the BP neural network,the model based on the RBF neural network can possess higher accuracy.