Research On Condition Monitoring Technology Of Aviation Piston Engine Based On Flight Data Analysis

The aviation piston engine is the most important part of the general aviation aircraft,and its operation can have an extremely important impact on the safe operation of the general aviation aircraft,so it is necessary to master the running state of the engine and carry out engine condition monitoring.This can ensure that the engine is in a healthy state,and at the same time,condition based maintenance can be carried out according to the running condition of the engine,which helps to reduce unnecessary regular maintenance and maintenance costs.By monitoring the relevant engine parameters,the abnormal working state of the engine,the changing trend of performance and the potential faults of the engine can be found,which can help the maintenance personnel to evaluate the engine state scientifically and formulate a reasonable maintenance strategy,and achieve the goal of improving the safety and reliability of the aircraft.Taking part of the flight data recorded by the G1000 system of Cessna 172 R aircraft as the research object,the following contents are studied:(1)over limit detection and engine performance queuing based on single parameter;(2)performance trend analysis of aviation piston engine;(3)anomaly detection of aviation piston engine.Firstly,the relationship between engine parameters and engine performance and faults is analyzed,and the over limit detection of engine parameters is carried out,and the probability values of main engine parameters exceeding the limit are obtained.The method of queuing engine fleet performance based on exhaust gas temperature is studied.Then,the parameters which can explain the change of exhaust gas temperature are selected by multiple regression model,and the exhaust gas temperature is predicted by this regression model.A BP neural network for exhaust gas temperature prediction based on the above screening parameters and LM algorithm is constructed,which is proved by experiments to be significantly better than the regression model prediction algorithm,and the performance change of the engine is explored based on the changing trend of exhaust gas temperature over a long period of time.Finally,by analyzing the relationship between engine parameters and engine anomalies,a typical fault data model is constructed,and the fault information generated according to the fault data model is inserted into the actual flight record data to form a fault data set.On this data set,verify the effectiveness of the scheme of detecting engine typical faults based on the exhaust gas temperature prediction neural network algorithm.The experimental results show that the neural network model is effective for exploring the performance changes of aviation piston engines,and the engine anomaly detection algorithm designed can detect anomalies effectively and has higher detection rate and lower error detection rate.