系统可用性是综合反映系统可靠性及维修性的一种测度,如何利用现场获取的各组成单元的故障数据和维修数据进行系统可用性综合及评定恰是可靠性及维修性工程中的难题。考虑到串联模型具有普遍性以及指数信息更具通用性的工程背景,依据信息等价转换原则,从现场记录中提取、挖掘或折算出以时间和次数为特征的指数故障和指数维修的基本信息,并探讨了指数故障/指数维修的串联可修系统可用度概率估计的数值解析方法。基于极大似然估计深入研究了经典方法的置信区间;基于Bayes定理深入研究了单元和系统可用度一阶矩、二阶矩,又利用不完全Beta函数、β分布与F分布分位数间的关系,最终求解出串联可修系统可用度的经典近似置信限和Bayes近似置信限,研究成果有效地解决了串联可修系统可用性综合及评定的难题。这两种方法可适用于所有串联模式的复杂机电系统,而且伴随系统的运行,各组成单元故障/维修数据信息也在发生变化,据此可以实时地进行可修串联系统可用性综合及评定。以我国某风电厂风力发电机组为工程应用实例,验证了该方法的可行性与适用性。
Complex electromechanical system is developing towards multi-function,automation and intelligentization.For the complex electromechanical system characterized by intensive disciplines,complicated configuration and high fault risk,both safe operation and health management have become an important issue to both academic and engineering fields.System availability is a kind of measure reflecting system reliability and maintainability.How to conduct system availability synthesis and estimation using the unit reliability data and maintenance data obtained in the field operation process is a difficult issue in system reliability and maintainability engineering.Considering both the generality of series reliability model and universality of exponential information,the exponential fault time and maintenance time information characterized by time and number of occurrence is extracted,excavated or converted from the field record data according to information equivalent conversion principle.Furthermore,the numerical analytic method for the availability probability estimation of the series repairable system is discussed using exponential fault time and maintenance time information.The confidence intervals of classic method are studied deeply based on the maximum likelihood estimation.The first and second moments of the unit and system availability are studied based on Bayes theorem.Finally,the classical and Bayes approximate confidence limits of the series repairable system availability are derived according to the relationship among incomplete beta function,βand F distribution quantiles,the study result effectively solves the difficult subject of series repairable system availability synthesis and estimation.The two proposed methods are suitable to all the complex electromechanical systems with series reliability logistic relationship.As the complex electromechanical system is operating,the fault and repair data information of the system units is also changing,based on which the real-time synthesis and estimation of the syst