混合模糊神经网络在短期负荷预测中的应用

Application of Hybrid Fuzzy Neural Network in Short-Term Load Forecasting

为提高电力系统短期负荷预测精度,将模糊逻辑和神经网络的长处融合在一起,构建了混合模糊神经网络短期负荷预测模型,用于预测预报日的负荷。其中针对模糊神经元的权值更新问题,采用了一种新的权值更新算法——一步搜索寻优法,进一步减小了预测误差。实际算例证明了该模型的有效性。

In order to improve the precision of electric power system short-term load forecasting, a hybrid fuzzy neural network based short-term load forecasting model is put forward to predict the loads by mixing the merits of fuzzy logic and neural network. And a novel approach to update the parameters of fuzzy neurons is proposed. Some experiential rules are applied to the adjustment procedure of load, which can reduce the forecast error effectively. The result of typical calculation example shows that the presented method is effective.