Compensation Capacitor Based Short Term Load Forecast Optimized Switching

With the development of electric power industry and the scale-up of power networks, the requirement for the security and reliability of the network is becoming higher and higher. Power load is one of the most important factors related with the network's security and reliability and its sudden increase or decrease may have a bad effect on the security of the power system. So it is necessary to do some effective work on load forecast and track the load alteration.In this paper, factors that affect the load alteration and many load forecasting methods are introduced, among which two short-term load forecasting methods are analyzed in detail. One is based on the BP neural networks, the other is on the linear extrapolation. Both of these two methods nearly have the same effect. However, the first method has some disadvantages, such as a high requirement for the historical load data, a long operation time and some difficulties in converging. The second method conquers the shortcomings of the first and thereby is used in this thesis. A further research is required for the ignorance of the weather and the sudden incident which may affect the precision of the load forecast in some degree.The distribution and flow of the reactive power decide the level and characteristic of the voltage. It also affects the quality of the voltage and the loss of the network. Keeping the balance of the reactive power can improve the quality of the voltage and reduce the loss of the network. So reactive power compensation is important. But at present the control strategy isn't perfect, sometimes can even lead to oscillation of the capacitors' switch and the efficiency of the reactive power compensation is very low. This paper applies the short-term load forecasting to improve the control strategy of the compensation capacitor. The new strategy can reduce the times of the switch operation and prolong devices' life-span .