中长期负荷预测的模糊竞争学习聚类神经网络算法  

The clustering neural network based on fuzzy competitive learning algorithm for middle and long term load forecasting

作　　者：岳璐[1] 张尧[1] 

机构地区：[1]华南理工大学电力学院,广东广州510640

出　　处：《继电器》2008年第8期55-58,104,共5页Relay

摘　　要：电力系统中长期负荷预测受大量不确定因素的影响,聚类方法能够将各种影响因素综合引入预测模型,提高了预测精度。本文将神经网络引入到模糊聚类分析中,建立了中长期负荷预测的新方法,并且对聚类神经网络的学习算法进行了改进,利用模糊竞争学习完成网络运算,弥补了网络输出结果二值性的不足,使得学习规则中权值矩阵的改变速度加快,因而算法的收敛速度有很大提高。运用文中所述模型及算法综合考虑了历史负荷情况和未来不确定因素等对未来负荷变化的影响。通过与传统方法进行中长期负荷预测比较,结果表明该方法可以提高负荷预测的精度。Middle and long term load forecasting of power system is affected by various uncertain factors. Using clustering method, numerous relative factors can be synthesized for the forecasting model so that the accuracy of the load forecasting would be improved significantly. The new method introduces the neural network into the fuzzy clustering and establishes the new model of mid-long term load forecasting. The method also makes improvement in the learning algorithm. It adopts the fuzzy competitive learning to solve the binary results of the network output and makes the change rate of the weight matrix speed up. So the convergence speed is improved effectively. The proposed model considers the influences of both history and future uncertain factors. Compared with the traditional methods, the results show that the new algorithm improves the accuracy of load forecasting considerably.

关 键 词：中长期负荷预测 聚类神经网络 模糊竞争学习 信息扩充法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TM715[自动化与计算机技术—控制科学与工程]