Research On Aeroengine Remaining Life Prediction And Maintenance Decision Based On Stochastic Process

This paper evaluates the operating state of the engine based on the health monitoring data.A performance degradation model based on stochastic Lévy subordinate process is established to realize the related research on engine fault diagnosis,residual life prediction and maintenance decision.Firstly,the data preprocessing related technology is studied for the problems of data error and different numerical units.When the fault diagnosis of aeroengine is carried out by traditional methods,there are many problems such as high data dimension and inaccurate diagnosis.The aeroengine fault detection method by combining improved KPCA and DBN algorithm is proposed.Compared with the traditional BP shallow neural network fault detection algorithm,the proposed KPCA+DBN combination method is better than the traditional BP neural network.Secondly,the life prediction method based on stochastic process assumes that the degradation process is progressively continuous degraded,and does not consider the random sporadic jump degradation caused by external shock or sudden factors.In order to predict more accurately and effectively,this paper proposes a new prediction method based on Lévy stochastic process,which is a Lévy subordinate process combining gamma process and composite Poisson process.Based on the eigenfunction and the inverse Fourier transform,we derive the reliability function and the lifetime probability density function,which are characterized by a comprehensive consideration of the progressive degradation and the jump degradation.Compared with the traditional Wiener process life prediction method,the accuracy of the model established in this paper is higher than that of the traditional method.Finally,the relationship between maintenance and spare parts is rarely considered in the existing maintenance decision research.The timing and joint decision models are established and optimized by genetic algorithm.Both models consider the impact of preventive maintenance costs and post-fault maintenance costs on the maintenance decision model.The impact of fixed spare parts costs on decision models is rarely considered for existing research.This paper also studies the optimization results of two decision models under different fixed spare parts costs,and analyzes the results.By comparing the optimization results of the two models,it is found that the joint decision model has less expected unit maintenance costs.