计及全路径关联概率的连锁故障风险评估模型

Joint Probability for Full Path Based Risk Assessment Model of Cascading Failure of Grid

连锁故障包含了源发故障、状态转移、控制失当等偶然性与必然性交织的复杂因素,传统基于相邻事件概率关联的方法难以准确反映这种复杂性所带来的风险强弱。首先,通过耦合各级故障与后续事件的关联特征,建立了计及全路径关联概率的连锁故障风险评估模型;再利用贝叶斯网络推理算法,从概率拓扑角度推算出可能发生的连锁故障事件全路径关联概率,根据该关联概率方法可得到初始激励下最有可能发生故障的下一级线路及其影响范围与强度;然后从线路状态、节点电压以及网络拓扑等方面提出了5个后果严重度指标,并基于模糊分析方法对连锁故障后果进行综合评估,以最大隶属度原则量化所分析连锁故障路径风险等级;最后,新英格兰39节点标准系统验证了所提方法的有效性。

Cascading failure includes the source fault, state transfer, improper control by complex factors and inevitabil-ity of interwoven, traditional methods which can not adjacent event probability based on probability statistics reflect thiscomplexity of the actual risks. Based on the correlation characteristics of the coupled faults and the subsequent events, arisk assessment model of cascading failure is established. Through using the Bayesian network inference algorithm, thecascading failure can be calculated and may occur in the full path from the angle of topology association probability. Be-sides, according to the association probability, the most likely line tends to be the next one and can get its influencescope and strength under the excitation of the initial failures. Five severity index are proposed through the state of line,the node voltage, network topology and others to assess the risk of cascading failure with fuzzy analysis method is alsoused to evaluate cascading failure consequences. At last, the maximum membership degree principle is used to quantifypath risk level. It proved the effectiveness of this method by the new England 39-bus standard system.