Reliability Analysis Of Magneto Of The Piston Engine Based On Bayesian Network

Through taking Slick 4371 aviation magneto equipped with widely used LYCOMING IO-360-L2 A piston engine in flight training of general aviation,as research target,this thesis tries to carry out systematic reliability analysis.First of all,carrying out a research on the existing reliability analysis method,then it determines to adopt the method of Bayesian network through considering the actual engineering situation of aviation magneto and bidirectional reasoning,as well as the advantages of dynamic characteristics.Next,through the research on the structure and working principle of magneto,it carries out the induction and hierarchical partitioning for failure mode,and then it establishes the topology model aviation magneto with the help of network building method based on fault tree.In addition,it carries out a research on the application of Bayesian network inference in reliability analysis.On the one hand,the use of bidirectional reasoning can help to find high risk sources of the system first;on the other hand,the time parameters should be introduced in it,and then it tries to analyze system reliability and risk factors of each period through discrete time series with the combination of data update;meanwhile,it adopts contrast verification through cases and Markov method,so as to confirm the feasibility of this method.Then,it gets the prior information in the light of the maintenance data of this type of magneto in a certain maintenance unit for many years.Through the Bayesian network inference in discrete time,it calculates the failure probability of the system in different time periods,the failure probability of different components in different time periods when there is a system failure,and the reliability index of each item for its importance and susceptibility.Finally,as for the problems existing in the current maintenance means,this thesis gives a method — optimizing its maintenance and management.Based on the reliability allocation method of posterior probability,it provides ideas for the optimization of the structure design on magneto,and then it gives the forecast method of maintenance cost based on gray theory,which helps to control the maintenance budget of general aviation enterprises.It also gives optimization method of examine and repair flow for reliability analysis based on the real time maintenance data,which is suitable for different requirements of engineering.