基于模糊粗糙集和神经网络的短期负荷预测方法

针对采用神经网络进行电力系统短期负荷预测时其网络输入变量的选择是影响预测效果的关键问题，该文提出使用模糊粗糙集理论解决这一问题：对采集到的信息进行特征提取、形成决策表；利用模糊粗糙集理论进行属性约简、去除冗余信息；用得到的属性作为BP网络的输入进行训练预测。该方法既全面考虑了影响负荷预测的历史时间序列、气象等各种因素，为合理地选择神经网络的输入变量提供了一种新的方法，又避免了由于输入变量过多而导致神经网络拓扑结构复杂、训练时间长等不足。计算实例表明，文中提出的方法是有效且可行的。

A METHOD FOR SHORT TERM LOAD FORECASTING INTEGRATING FUZZY-ROUGH SET WITH ARTIFICIAL NEURAL NETWORK

Short-term load forecasting (STLF) plays a key role in power system operation and planning. In recent years, artificial neural networks (ANN) are more commonly used for load forecasting and it is important to select proper factors as input variables of ANN. An integrated approach based on fuzzy-rough sets theory and ANN for load forecasting is presented in this paper. Firstly, the fuzzy-rough set theory is applied to find relevant factors to the load among varied factors, then ANN module is trained using historical daily load and weather data selected before to perform the final forecast. This method provides a new way for ANN input variable selection and the testing results on a real power system show that the proposed model is feasible and promising for load forecasting.