Substation Voltage & Reactive Power Control Based On Load Forecasting Using Fuzzy Clustering Analysis And RBF Neural Network

To keep voltage quality and to reduce network loss are the most important two objects of power system operating safely and economically. For substations, voltage & reactive power control can make reactive power distribute reasonably by adjusting transformer taps and opening or closing compensators of reactive power, which can make for high voltage quality and low power loss.Nowadays, voltage & reactive power control equipments use 9-area-chart principle mostly, they can make transformer taps and compensators change frequently, and this may reduces their using time. In order to reduce tap operating times, a method for voltage & reactive power control in sections based on load forecasting using fuzzy clustering analysis and RBF neural network is proposed. Forecast substation active, reactive load and voltage of next day firstly, divide one day into several sections according to taps operating times in a day, then control in every section.The performance of controlling strategy lies on the accuracy of load forecasting. Considering synthetically the factors that influencing load, such as type of day, temperature, relative humidity, weather status, divide historical data into several sorts using fuzzy clustering analysis, then choose the same type days to forecast using RBF neural network. This method overcomes the disadvantage of the method only choose swatch by workday or weekend, and it improves the swatch's quality of RBF neural network, so the accuracy of load forecasting has been improved.Considering the disadvantage of dividing method only according to active or reactive load, in this paper, the dividing method considers active, reactive load and voltage. Two-winding-transformer and three-winding-transformer are all discussed in this paper.By the calculation of several different substations in Tianjin, the validity of the method has been proved.