The Reliability Modeling, Analysis And Comprehensive Evaluation Method Of Complex Systems

ABSTRACT:With the great improvement of integration and electronization, modern systems become more and more complex and require higher reliability, which make it difficult to carry on comprehensive and effective research on modern complex systems’ reliability. This thesis mainly focuses on the reliability modeling and analysis method of load-balancing parallel system and complex working condition system, and comprehensive reliability evaluation method of complex functional module system.To overcome the defect of capacity flow model, the failure rate variation model of the component of load-balancing parallel system, a corrected capacity flow model considering the component’s load capacity is proposed. On the basis of the corrected capacity flow model, the Markov models of different conditions of system are rebuilt, and then the analytical expressions of some reliability indices are derived.By analyzing the characteristic of the effects of the working conditions on electronic product’s failure rate and repair rate, failure rate cloud model and repair rate cloud model under complex working condition are proposed. The failure rate cloud and repair rate cloud are generated from adaptability models and working condition models through X condition cloud generator. Since adaptability model and working condition model reflect the values of failure rate and repair rate of different working condition points and the variations of different working conditions respectively, the failure rate cloud and repair rate cloud contain information of actual working conditions and corresponding values of failure rate and repair rate, which can effectively reflect the dynamic effects of actual working conditions on product’s failure rate and repair rate.The traditional Markov model and fuzzy Markov model are of poor ability in reliability modeling and analysis of complex working condition system. To solve this problem, a reliability modeling and analysis method of complex working condition system based on cloud model and traditional Markov model is proposed. By substituting failure rate cloud and repair rate cloud for failure rate and repair rate in the reliability expression and availability expression derived from traditional Markov model, the reliability cloud and availability cloud are generated. The reliability cloud and availability cloud reflect actual working conditions’ effects on value range and distribution of the system’s reliability indices, which is intuitive and of strong dynamic description capability of working condition and makes the proposed method obviously superior to the others.Because the reliability modeling and analysis method of complex working condition system based on CHSPN model ignores the independent effects of actual working conditions on system’s reliability, a reliability modeling and analysis method of complex working condition system based on CTHSPN is proposed to remedy this defect. By dealing with each working condition separately, introducing control clock and control function of transition, extending the range of marking function and weighting function of arc, the proposed method maintains independent effect information of each working condition and has strong dynamic reliability modeling and analysis ability, which is skilled in describing each working condition’s independent effect on system’s reliability.The comprehensive reliability evaluation method of complex functional module system based on radar model and cloud model requires the relative values of function indices measurable and quantified and does not determine the weights of elements of each level, which limit its application scope. To get over this shortcoming, a comprehensive reliability evaluation method of complex functional module system based on cloud theory and analytic hierarchy process is proposed. By analytic hierarchy process, the proposed method can effectively build functional hierarchy of the system and determine the weights of elements of each level; and by using weighted model as system’s comprehensive model, cloud model to describe the relative values of function indices and evaluation sets of system, it is easy to deal with uncertainty information in the system, which can make the proposed method gain more applications.