机构地区: 广东电网公司
出 处: 《广东电力》 2013年第1期19-22,共4页
摘 要: 提高风电功率预报的准确率对电网的安全运行调度有着重要的意义。针对标准BP学习算法泛化能力不强的问题,设计了一种基于贝叶斯正则化算法修正权值的学习算法,用于风电的功率预测。仿真结果对比表明新的算法具有比标准BP算法和径向基神经网络具有更好的泛化能力,同时取得了良好的预测效果。 It is important to improve accuracy of wind electricity power prediction for safe operation and dispatch of power grid. Aiming at problem of poor generalization ability of standard BP learning algorithm, This paper designs a learning algo- rithm based on Bayes regularization algorithm correction weights which is used for wind electricity power prediction. The simulation result shows that the new algorithm is better of generalization ability than standard BP algorithm and radial basis function neural network and can get good prediction effect.
关 键 词: 神经网络 贝叶斯神经网络 风电场功率预报 泛化能力
领 域: [电气工程]