Importance Analysis In Reliability Engineering

Important analysis is one of the key aspects of system reliability analysis, which is usedto analyse the criticality of the particular component influence on a system quantitatively. It isone of the front edge and hot spot of sensitivity, risk, criticality and importance, and also apowerful tool for identifying imperfection within a system and improving the systemreliability, which has a positive meaning for the system reliability, security and fault diagnosis.At present, the existing importance measure techniques are lack of independence andcomprehensive, only as a concept of the reliability engineering. The study restricted to unicitymay lead to a shallow research. This paper attempts to take the reliability analysis of mouldsubsystem and oil supply subsystem for numerical control press as the research object,importance analysis is the master line. Combining the binary decision diagram BDD,multi-valued decision diagram MDD, logical differential calculus and Markov stochasticprocess method to discuss the component importance of the binary system and multi-statesystem, these innovations are expected to offer a new idea and method for system reliabilityassessment, life prediction, fault diagnosis and prevention. The presented method has apositive effect as a way to enrich and improve the existing reliability theories and methods.The main contents are as following:(1) The Binary Decision Diagram is presented to express the structure function of binarysystem, take the boolean theorems as basis to deduce the BDD computational rules. Thenpropose the probability path search and reliability calculation method based on BDD. At last,take advantage of recursive principle to obtain the Fault Tree transform the BDD, and utilizethe BDD to analysis the component importance of binary system instead of FT. The methodproposed can solve the difficulties of complication、inaccuracy and combinatorial explosion,provides an efficient and accurate method for the importance calculation.(2) Building on the BDD model to expand the Multi-valued Decision Diagram, whichallows to represent the structure function of MSS, drawing the mathematical tool——DirectPartial Logic Derivative(DPLD) in logical differential calculus into MSS, aiming to transformthe RBD of MSS. Combines MDD with DPLD to calculate the importance of multi-statecomponent, the advantages of the proposed method occupies the merits of time-saving andvisual through a comparison with existing methods——ascending method and descendingmethod. (3) Considering the performance requirement of component may vary according to thedaily and seasonal load demands, performance level can be used to analyse the effect ofcomponent on the system. Discuss the relationship between performance randomness andMarkov discrete stochastic process. The original system is divided into two new sub-systemsby a given level of performance, on this basis we analyse the component importance withrespect to a given performance level. The results have shown that the component importanceis different for different performance level, which can offer a practical reference frame forsystem detection and maintenance.