论文部分内容阅读
介绍了一个模块化的综合智能化电力短期负荷预测系统 ,第一个模块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 ,第三个模块是在实现前二者预测的基础上 ,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 ,可直接用自适应最优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、有特殊事件发生时 ,就必须利用模糊专家系统修正方法对预测结果进行修正。对某省网的日负荷数据进行了具体计算 ,结果表明此负荷预测系统能取得满意的预测效果。该系统的实用化软件包已投入试运行。
A modular integrated intelligent power short-term load forecasting system is introduced. The first module is modeled by artificial neural network, the second module is modeled by adaptive optimal fuzzy logic system, and the third module is implemented in the first two Based on the prediction, the deficiency of the forecasting method is supplemented by the correction mechanism of fuzzy expert system. When there is not much change in the weather and no special event occurs, the adaptive optimal fuzzy logic system prediction method and artificial neural network method can be directly used to predict the load from Tuesday to Saturday without the need of a fuzzy expert system. For the load on Sunday and Monday, or when the weather is abrupt, when a special event occurs, the prediction result must be corrected by the fuzzy expert system correction method. The daily load data of a provincial network were calculated. The results show that this load forecasting system can achieve satisfactory forecasting results. The system’s practical software package has been put into trial operation.