Research On Bus Load Forecasting For Distribution Network

Bus load forecasting can be used to analyze and predict all nodes’load in the power grid, and the results can be applied to develop the scheduling scheme in the power system. Moreover, Bus load forecasting can analyze power demand in the power system and predict the distribution of all loads in the future which assists manipulators to understand the running state of the power system and eliminates all hidden dangers in advance. Meanwhile, based on bus load forecasting, the level of standardized and meticulous management can be enhanced, which improves the security, quality and economical operation of the power grid. In conclusion, the meaning of bus load forecasting is significant and far-reaching.The research on bus load forecasting in this paper was divided into three main parts to set forth. In the first part, the methods of correlation analysis in the Gray theory was applied to quantitify the relationship between any two load series. Least square method was employed to fit the change rate of load and the load with minimum change rate was consider as reference load. And according to all of these above, the method of training sample grouping was proposed here. In the second part, because noise always objectively existed in load data, a forecasting model based on wavelete transform was put forward in this paper. Of course, walvelet discompositon model that used to optimize forecasting accuracy was also presented here. In the third part, because state estimation in the smart grid needed the results of bus load forecasting, after rigorous analysis about historic bus load and temperature information, a fast bus load forecasting model was concluded out. The training sample grouping model and the wavelet decomposition model had beed proved that preditcion accuracy obviously improved by experiments. The fast bus load forecasting model at present had been applied in a real project, and its correctness and effectiveness had also been proved.