基于T-S动态故障树的桁架机器人系统可靠性分析

Reliability Analysis of Truss Robot System Based on T-S Dynamic Fault Tree

为了提高桁架机器人系统的动态可靠性,并且在发生故障时提升排查故障的效率,提出一种对桁架机器人系统动态可靠性分析的方法。首先,对桁架机器人系统建立T-S动态故障树;然后,把T-S动态故障树和T-S动态门规则分别转化为离散时间贝叶斯网络(DTBN)与相应网络节点的条件概率表,根据DTBN正反向推理分别计算得到各时间段以及任务时间内系统失效概率和根节点后验概率,并获得各根节点的T-S概率重要度、T-S关键重要度和灵敏度;最后,应用Monte Carlo仿真法进行验证,结果显示相对误差为0.29%,证明所提方法可行。给出当系统发生故障时应优先排查的零部件,并确定了系统薄弱环节,得出应优先提高可靠度的零部件,为桁架机器人系统的动态可靠性分析提供了理论依据。

In order to improve the dynamic reliability of truss robot system and improve the efficiency of troubleshooting when faults occur,a method for dynamic reliability analysis of truss robot system is proposed.Firstly,the T-S dynamic fault tree was established for the truss robot system,and then the T-S dynamic fault tree and T-S dynamic gate rules were transformed into discrete time Bayesian network(DTBN)and conditional probability tables of corresponding network nodes,respectively.According to the forward and backward reasoning of DTBN,the system failure probability and the posterior probability of root node in each time period and task time were calculated.And the T-S probability importance,T-S critical importance and sensitivity of each root node are obtained.Finally,the Monte Carlo simulation method is used to verify the proposed method.The result shows that the relative error is 0.29%,which proves that the proposed method is feasible.The parts which should be preferentially checked when the system fails are given,and the weak links of the system are determined,and the parts which should be preferentially improved in reliability are obtained,which provides a theoretical basis for the dynamic reliability analysis of truss robot system.