Research On The Application Of Fuzzy Petri Nets In Reliability Analysis And Fault Diagnosis Of Empty Container Stacker

Empty container stacker is the indispensable equipment in large transport industry. It plays an important role in the whole industry. The work system of empty container stacker is complex and it works frequently. It’s of great significance to conduct reliability analysis and fault diagnosis of the empty container stacker for evaluating the performance of the empty container stacker.Firstly，the failure phenomenon and fault factors are summarized through analyzing the basic structure and the working principle of the empty container stacker. It introduces the basic principle of fault tree analysis and the construction method of the fault tree model. The fault tree is established with the top event of the empty container system failure on the basis of previous obtained failure modes. The failure probability and reliability of system are calculated through qualitative analysis and quantitative analysis.Then, in order to solve the uncertain fault information of the system, the grey correlation analysis is combined with FTA and the grey fault diagnosis model is established. The correlation degree of fault feature vector and detecting fault vector is calculated. The priority of fault factors in fault diagnosis is judged according to the order of correlation degree.Afterward, due to complexity and diversity of the failure factor in empty container stacker, the failure diagnosis method based on fuzzy Petri net is proposed. The rule knowledge about fault diagnosis is established based on fuzzy knowledge rule. The transforming method between fault tree and fuzzy Petri net is applied to establish the fault diagnosis model based on FPN. Failure reasons are diagnosed by fuzzy reasoning algorithm quickly.Finally, this paper adopts PSO to optimize the parameters of FPN model of the empty container stacker, such as thresholds, weights and confidence. The basic flow chart of parameter optimization is given. Simulation experiment of parameter optimization is completed on basis of VB software program. The simulation results show that the parameters optimized by the particle swarm optimization algorithm are more close to the ideal value.