计及源荷不确定性的配电网可靠性高效评估方法

Efficient evaluation method of distribution network reliability considering real-time operation condition changes

基于输电线路故障率与服役时间、运行条件之间的耦合关系,本文提出了一种计及实时运行条件改变的配电网可靠性高效评估方法。为改进传统可靠性评估方法在系统运行条件改变时需要重复进行可靠性评估过程的弊端,基于模型-数据混合驱动提出了面向源荷多维不确定性因素的可靠性评估方法。首先,基于输电线路不可用率与运行条件及老化效应间的耦合关系,建立输电线路不可用率模型,并提出配电网关键元件选取方法。然后,构建配电网可靠性指标关于关键元件可靠性参数的解析模型。最后,使用多层前馈神经网络得到解析模型与可再生能源出力及配网负荷水平的映射关系,当源荷变化时快速更新配电网运行可靠性指标。算例测试对本文所提算法的准确性和有效性进行了验证。

Based on the coupling relationship between transmission line failure rate,service time and operating conditions,this paper presents a methodology for efficiently assessing the reliability of distribution networks that takes into account real-time changing operating conditions.In order to alleviate the disadvantages of the traditional reliability evaluation method that requires repeated reliability evaluation process when the operating conditions of the system change,a reliability evaluation algorithm for source-load multi-dimensional uncertainty factors is proposed based on the idea of model-data mixing.Firstly,based on the coupling relationship between transmission line unavailability rate,operating conditions and aging effect,a transmission line unavailability rate model is established,and a methodology for selecting key components for distribution networks is also proposed.Subsequently,an analytical model of the reliability parameters of the main components of the distribution network reliability index is developed.Finally,the mapping relationship between the analytical model and the renewable energy output and distribution network load level is obtained through the BP neural network,so as to enable rapid updating of distribution system reliability indicators in the event of changes in source loads.Sample tests have confirmed the accuracy and effectiveness of the proposed methodology.