SF_6高压断路器机械故障概率的非精确条件估计

Imprecise Estimation for Conditional Mechanical Outage Probabilities of SF_6 High Voltage Circuit Breakers

在小样本条件下,结合SF_6高压断路器自身特点,引入非精确概率概念,提出一种将非精确狄里赫雷模型(IDM)与贝叶斯网络相结合的断路器机械故障概率估计方法,对断路器机械故障可能性范围进行估计。首先,对表征断路器力学性能的特性参数进行状态划分;其次,根据历史数据,建立处理非精确条件概率推断问题的贝叶斯网络,并利用非精确狄里赫雷模型得到断路器机械故障样本缺乏条件下贝叶斯网络节点的非精确条件概率测度;最后运用贝叶斯网络的推理算法,根据断路器本次开断的参数估计得到在下次开断时其出现机械故障的非精确概率。该方法体现了短时间尺度下,断路器机械故障概率随其运行状态变化而时变的特点,并为小样本情况下断路器机械机构的可靠性评估以及状态检修提供了依据。通过算例验证,证明了所提方法的有效性。

Under the condition of limited outage samples and combined with the characteristics of the SF6 high voltage circuit breaker itself,this paper introduces the concept of imprecise probability and proposes a probability estimation method for the mechanical outage of the circuit breaker,which integrates imprecise Dirichlet model(IDM)and Bayesian network,to estimate the range of mechanical outage probability for the circuit breaker.Firstly,this paper divides the states for the characteristic parameters of mechanical properties characterizing the circuit breaker.Secondly,according to historical data,Bayesian network for processing the probability extrapolation in imprecise condition is established,and IDM is utilized to obtain the imprecise probabilistic dependences of Bayesian network nodes under the lacking condition of mechanical outage samples.Finally,by utilizing the reasoning algorithm of Bayesian network,the imprecise probability of mechanical outage for the next break is obtained according to the parameters of last break of the circuit breaker.This method not only reflects the time-varying feature of mechanical outage probability of the circuit breaker,but also provides reference for the reliability estimation and condition-based maintenance of the circuit breaker in limited outage samples.Through calculating the samples,the effectiveness of this method is verified.