Faults Diagnosis And Maintenance Optimization For Series-parallel Systems With Multi-faults

Many systems in the industrial field will be shut down due to their aging or failures of components.That the advanced equipment in the modern manufacturing system integrates mechanical and electronic technologies increases the complexity of fault diagnosis and the difficulty of maintenance decision.How to improve the quality and efficiency of fault diagnosis and scientifically arrange the maintenance become more and more important.Solving complex multi-fault diagnosis problems by rationally applying intelligent methods and constructing an effective maintenance decisionmaking model play key roles in modern industrial systems because reducing the cost of damages caused by the downtime has become one of the important indicators to judge whether the system is running well.The series-parallel system is a representative one of complex systems,therefore it is reasonable to set the series-parallel system as the object of the study.Based on the requirements of a series-parallel system with multifaults and the characteristics of multi-faults,a multi-faults diagnosis method combined with Elman network and ECOC-SVM is proposed.The study,based on the diagnosis of the cause of the failure,considers the economic dependence and structural dependence among components,and builds a maintenance optimization model with system reliability as the goal and cost and time limitations as constraints.The main research contents of this paper include the following aspects:(1)For the series-parallel system with multi-faults,the Elman network is used to construct the configuration of diagnosis model.Based on this model,the ECCO combined with SVM using the maximum likelihood estimation is used to screen the fault types to obtain the final diagnosis result.The effectiveness and advancement of the diagnostic method are verified by examples' calculation and analysis.(2)For the series-parallel system with failures and shutdowns due to multiple faults,after the causes of the faults is known,the maintenance optimization model is constructed,with the system reliability as the optimization goal,the economic dependency is considered,and the genetic algorithm is used to effectively solve the constrained optimization model,then the optimal maintenance plan is obtained and the effectiveness of the model is compared and analyzed by examples.(3)Considering the influence of components' structure dependence on system disassembly cost and disassembly time,a maintenance optimization model is built considering the economic and structural dependencies.Based on the reverse search algorithm for deriving the disassembly path set of specific components,the process for judging the structure dependence among components with faults is designed.It is a good opportunity to make full use of the downtime opportunities caused by system failures and to maximize system reliability under the premise of meeting the minimum system requirements and resource constraints.Finally,an example is given to verify that the model can help determine the optimal maintenance activity combination.