Maintenance Decision Based On Statistics And Its Application To Aeroengine System

Aircraft engine as the heart of the aircraft in the aviation industry occupies an extremely important role in the dynamic equipment,its reliability and performance to ensure the overall performance of aircraft and the flight safety and aircraft engine is more system failure rate,Aeroengine fault maintenance cost is very high,so the suitable maintenance plan,The maintenance mode of aero engine was originally after-service maintenance,and later developed to regular maintenance.Now most large aviation countries adopt maintenance according to the situation,the airlines has a great practical significance Based on system of aviation engine maintenance decision for the following three aspects:First,to make decisions based on single state parameters,give decision makers time to prepare for repairs,Most of the initial maintenance is based on single state parameters,in which some parameters for repair,after arriving in threshold based on the single status maintenance aspect,the thought of using data to predict,first using ARIMA time series model of exhaust gas temperature EGT prediction,accuracy is above ninety percent,due to the time series model can extract linear stationary characteristics,unable to effectively extract nonlinear features,consider the optimization,through ARIMA network LSTM combined with both short-term and long-term memory,time series model is utilized to extract the difference as both short-term and long-term memory network input,reoccupy LSTM error regression analysis,finally The error is added to the ARIMA model for prediction.Secondly,in the aspect of multi-state parameter decision making,for the historical fault data of the airline,firstly,the weibull model was used to make the fitting reliability decision based on the weibull model.Then use the weibull proportional hazards algorithm,calculate the instantaneous failure rate of historical failure data the most value,based on the instantaneous failure rate to make a decision,then time axis coordinates,covariate comprehensive state as a vertical coordinate,got a logarithm curve,the logarithmic curve is the risk of aeroengine time-state maintenance threshold finally consider the weibull proportional hazards model based on markov process,the processing,the continuous state discrete into normal general degenerate four state,serious degradation failure model based on semi-markov transition probability for aeroengine residual safe life.Third,this paper introduces the basic knowledge of decision theory and basic process of decision-making,analysis of several common methods,to achieve the goal of maintenance optimization for evaluation and then finish the last step of the decision theory of subjective evaluation method of grey system theory is introduced,itsapplication in addition to outside the scheme evaluation and sorting,used in small fuzzy information data is very wide,the last one are calculated by using grey target theory common scheme selection problem in engineering,and using the scheme for comprehensive distance to target in order.