Engine Performance Evaluation Based On Muti-parameter

Aero-engine is regarded as the "heart" of the aircraft, its safety and reliability is greatly improved with the development of condition-based maintenance(CBM), airlines increasingly concerned about how to minimize the engine fleet management and maintenance cost under some reliability requirements. In order to effectively control the engine operating costs, we need to keep abreast of engines' performance,it can help to make reasonable operational and maintenance plan. Based on the analysis of domestic and international engine fleet management situation, establish baseline, data preprocessing and engine ranking are studied in this paper.To calculate the gas path parameter deviation which is critical to engine condition monitoring, this paper firstly studied how to establish engine baseline,and presents a new method. Initial cruise report data of several new engines is used to set up baseline equations.Consider the impact of flight condition, it is more reasonable to establish multi-parameter baselines. The Mean Impact Value(MIV) method based on Neural network is used to select variables from several relevant parameters.Aero-engine gas path parameter samples is a non-stationary time series which contains stochastic noise and mutational points, it is necessary to detect outlier and smooth before it is used to evaluate engine performance. First, a method by combining statistical and density is proposed to identify the abnormal data points. Then second exponential smoothing method is used to processing the data series. Particle swarm optimization is used to optimize the smoothing coefficient. The application shows that the effect of this method is good.The engine ranking result can help to decide the optimum quantity of spare engine,predict engine removal, make engine work-scope assessing &planning, arrange reasonable routes for engines and manage the engine operating cost, it can also remind the maintenance department to focus on the poor engines of failure. Given the lack of accurate and effective subjective expert information and reliability data for engine ranking, a method by combining entropy and TOPSIS is proposed to evaluate engine performance in this paper.Ranking is based on objective information included in the performance parameters. Examples of verification show that the proposed method is effective.