模糊处理变结构神经网络日负荷预测方法研究

对于受不确定因素影响的日电力负荷，首次提出了基于模糊分类规则的变结构神经网络负荷预测模型，考虑从两方面改进预测精度，一个方面是通过模糊分类规则，使过去的负荷数据分为不同气候特征，选用同类特征数据进行预测，另一方面是通过神经网络变结构优化，确定最优网络和最优拟合逼近，从而得到最优的预测结果，这种新方法同时考虑了天气因素的影响和神经网络的最优确定，因此，较大提高了日负荷预测的精度。

The Research of the Load Forecasting Method of the Variable Structure Neural Network Based on the Fuzzy Treatment

For the daily electric power load with uncertainty influence factors,we first put forward the load forecasting model of the variable structure neural network based on the fuzzy classification rules.We consider improving the forecasting accuracy from two aspects.On the one hand through fuzzy classification rules,the past load data are divided into different network property.The data of the same property are used to forecast next day load.On the other hand,through the neural network making the variable structure optimization,we get the optimal network,the of optimal approximation and the optimal forecasting results.The new method considers the influence of weather factors and the optimal determination of the neural network.