Effect Evaluation Of Civil Aeroengine Water Washand System Implementation

Supplying aircraft with power required for flight, aero-engine has been regarded asthe heart of an aircraft, whose performance state directly relates to the flight safety andmaintenance cost. Over time, dirt will buildup within the engine gradually, results in aperformance degradation of the engine. Water wash can restore engine performancepartially by removing accumulated dirt and contamination inside the engine. Airlineshave no concrete methods to evaluate the effect of water wash and wash engines atfixed period, which go against the idea of condition based aero-engine maintenance.Aiming at present problems about engine water washing, the quantitative evaluationmethod of the effect of water wash and wash timing decision from the perspective of theeffect of water wash are studied in this paper.Aero-engine state parameters reflect the performance of the engine, which are theimportant data base for evaluating the effect of water wash. However, the raw valuesoften contain interference information and have too many errors to be useful. In order toreduce errors, boxplot is applied to identify outliers according to the characteristic of thestate parameters. Then exponential smoothing method and singular value decomposition(SVD) are studied to reduce noise and extract the trend component.Based on aero-engine state parameters processing, the quantitative evaluationmethod of the effect of water wash is studied in two ways: extend time-on-wing andlower fuel consumption. Aero-engine performance deterioration model before and afterwash are established using the trend component extracted from state parameters dataand then the method of calculating time-on-wing extended is given. The calculationmethod of the quantity of fuel saved of cruise phase is presented. A fuel flow deviationforecasting model based on process neural networks (PNNs) optimized using particleswarm optimization (PSO) is established. Then, saved fuel consumption of cruise phaseis described with an example.The effect of water wash is the evaluation index of water wash timing. Aforecasting model for the effect of wash base on support vector machine (SVM) withprinciple components analysis (PCA) is established and validated with actual data. The proportional hazards model between aero-engine state parameters and water washinterval for an expected effect of water wash is established based on history wash data,and then the engine wash control limit is established. Application case is presented toprove the effectiveness of methods. Both ways can provide reference for enginewashing time determineFinally, according to the actual demand of Air China about engine water washmanagement, a system for engine water wash management is designed and developed.The requirement of system is analyzed, and the system function and workflow aredesigned, and then system architectural framework is presented. At last some runninginstances of the system are presented.