一个综合智能化电力短期负荷预测系统的研究

介绍了一个模块化的综合智能化电力短期负荷预测系统，第一个愉采用人工神经网络建模，第二个模块采用自适应最优模糊逻辑系统建模，第三个模块是实现前地预测的基础上，针对其预测方法的不足，辅以模糊专家系统的修正机制，在天气变化不大且没有特殊事件发生时，可直接用自自适应最模糊逻辑系统预测方法和人工神经网络方法预测星期二到星期六的负荷，不必用模糊专家系统进行修正，对于星期日和星期一的负荷，或当天气突变，有特殊

A RESEACH OF AN INTEGRATED INTELLIGENT SYSTEM FOR SHORT-TERM ELECTRIC LOAD FORECASTING

An integrated intelligent modularization system for short termpower loadforecasting is presented in this paper. The first module of the system is based on artificial neural network. The second module is based on fuzzy logic system. The third module is an accessory fuzzy expert system that can be used to update the result of load forecasting in allusion to the shortcoming of the above mentioned technique. When the weather is steady and there isn't any special event, it is appropriate to forecast the load from Tuesday to Saturday using the method of both fuzzy logic system and artificial neural network without updating by fuzzy expert system. As to the load of Sunday and Monday, and that of other particular days on which remarkable change of weather occurs or some special event happen, it is necessary to update the forecasting result by use of fuzzy expert system. As an example, this system has been used to calculate the daily load of a provincial power network, the result shows that satisfactory forecasting accuracy can be achieved by this system. A practical software package of this system is put into test run.