Aeroengine Maintenance Decision Support System Development Based On Performance Parameters Prediction

Aeroengine is regarded as the heart of the aircraft, and its condition health caninfluence the airline’s economic and social benefits directly. In order to solve theexcessive maintenance and less maintenance problem during the work of themaintenance, the maintenance way of the aeroengine has changed into prognosis offault and health management etc. And the technology of the aeroengine performanceparameters monitoring has provided the technical support to this change.Statistical method and density method are used to identify abnormal aeroengineperformance parameters. Considering the masking effect of the statistical method andthe difficulty of identifying the lower bounds factor of the density method, a newmethod by combining these two methods is proposed to identify the abnormalaeroengine performance parameters. At the same time, in order to keep the local featureof the aeroengine performance parameters, the combination of the exponentialsmoothing method and the empirical mode decomposition (EMD) is used to smooth theaeroengine performance parameters.After preprocessing the aeroengine performance parameters, predicting the changetrend of the aeroengine performance parameters is beneficial to putting the aeroenginesafety monitoring window forward. According to these reasons, a time-weightmultilevel hidden neurons process neural network is proposed. At the same time,combining the characteristics of the aeroengine performance parameters，with theempirical mode decomposition(EMD) method, the combined prediction model for theprediction of the aeroengine performance parameters is established, and its efficiency isverified with the practical application.Under guaranteeing the safety, in order to improve the airlines’s economic benefits,after predicting the trend of the aeroengine performance parameters, the aeroengine’shealth condition needs to be made a scientific assessment. Therefore, the intervalnumber is taken to present the aeroengine performance parameters, and an aeroenginefleet health assessment method which is based on the interval grey correlation and D-Sevidence theory is proposed. Meantime, considering the heavy work of making theaeroengine maintenance schedule, in order to mine the further relationship between theaeroengine performance parameters and the aeroengine module maintenance level, a mining model is proposed which is based on the least square support vector machine,and the corresponding optimization algorithm based on the particle swarm optimizationis developed to optimize the mining model’s parameters. And the two methods areverified with the practical applications.Based on the above research results, orienting to the needs of the airlines, anaircraft engine maintenance decision support software prototype system based on theprediction of the performance parameters is developed, and the developed softwaresystem can provide technical support and software platform to the aeroengine faultdiagnosis and health management.