Research And Application Of Ultra-Short Term Bus Load Forecasting

Bus load forecasting can be applied to dynamic state estimation, power system security checking, reactive power optimization, energy-saving power generation dispatching, etc. It improves the fine and intelligence level of grid dispatching. Ultra-short term load forecasting, whose most prominent feature is real-time operation, can follow the change of load, help take timely measures. Combined with the characteristics of both, Ultra-short term bus load forecasting will play a very important role in the future construction and development of smart grid, especially in the security early warning and intelligent scheduling techniques, etc. According to thorough analysis of the basic principle of the bus load, the paper privides a bus load forecasting method suited for practical application based on load derivation. In order to increase the prediction accuracy, the method firstly choose similay days through the load mode clustering and patern recognition of the history load data. At the same time, considering the small base, vulnerable to the influence of the users and poor stability of the bus load, the history forecasting error distribution need to be calculated and applied to probabilistic interval prediction to get prediction intervals under some confidence level, which meets the characteristic of the bus load better. The actual regional grid measuring load data in Hainan province has proved the method validity and practicability. Finally, the method proposed in this paper is introduced to real-time reactive power optimization and shows the ultra-short term load forecasting can reduce the number of control equipment in one concrete example.