【摘要】 快速准确地进行配电网可靠性评估具有重要意义,然而传统的配电网可靠性评估方法并不适用于评估大规模配电系统的综合可靠性指标,对大规模电网的可靠性进行评估时往往会造成建模困难、计算量剧增的问题。因此,提出基于Improved-Elman（IElman）反馈型动态神经网络的配电网可靠性评估方法,即在Elman神经网络的承接层中加入自反馈连接增益系数来衡量历史信息对未来状态的影响程度,并通过思维进化算法对Elman神经网络的相关参数进行优化。在采用神经网络评估前,利用灰色关联度分析对神经网络的输入变量进行预处理。所提出的方法与普通神经网络评估模型相比,平均相对误差由5.43×10^-4降到7.32×10^-5,表明该方法能够有效简化计算,提高神经网络对复杂问题的评估精度。更多还原

【Abstract】 It is of great significance to carry out reliability evaluation of distribution network quickly and accurately.The traditional reliability assessment method of distribution network is not suitable for comprehensive reliability index evaluation of large-scale distribution system,and it often causes problems in modeling and the amount of calculations increases dramatically for large-scale power grid reliability evaluation.In this paper,the reliability evaluation method of distribution network is proposed based on improved-Elman（IElman）feedback dynamic neural network.The self-feedback connection gain coefficient is added to the acceptance layer of Elman neural network to measure the influence of historical information on the future state.The parameters of the Elman neural network are optimized by the thought-evolutionary algorithm.Before using neural network assessment,gray correlation analysis is adopted to preprocess the input variables of neural networks.Compared with the general neural network evaluation model,the average relative error is reduced from 5.43×10^-4 to 7.32×10^-5,which indicates that the proposed method can effectively simplify the analysis and improve the evaluation accuracy of neural networks for complex problems.更多还原