Research On Aero-engine Intelligent Monitoring Based On The Oil Analysis

As the core power of modern aircraft,aero-engine’s safety and reliability is essential. Theaero-engine operation and the fault diagnosis technology are always the top concerns for researchers.According to the statistics, more than80%of the mechanical equipment fault is concerned with wear,thus, the information getting from the oil monitoring which based on the theory of tribology candirectly reflect system’s operating condition.Based on the traditional physico chemicalanalysis,ferro-graph analysis and spectral analysis technology，and with the combine of statisticsand Artificial Intelligence technology，this article focus on the study of aviation engine abrasionobservation system.The research work mainly includes the following aspects.(1)Handle the initial oil monitoring data coming from the aero-engine with the traditionalmethods of mathematical statistics. The negative factors such as personal equation, environmentalfactors, lack of precision instruments, hiding in the data can be eliminate by datapreprocessing.(2)Considering the complexity of spectral data, the data is divided into groups using thefuzzy clustering method, some of data is abandoned according to the condition of the group to achievethe purpose of simplifying the data.(3)Using the analytic hierarchy process to assign the weight of allthe evaluation indicators, each indicator’s importance can be distinguished by the weight.(4)Thedeteriorative degree model is improved to evaluate the wear condition of Aero-engine. According tothe weight assigned in step3,the compositive deteriorative degree model can be built to appraise therunning state of aero-engine, combining all the indicators.(5)Applying the negative-selectionalgorithm of artificial intelligence methods to the aero-engine fault diagnosis, the system which isalready well trained can determine the fault type and the occurrence of the location.(6)Establish a timeseries prediction model with the parameter of compositive deteriorative degree, this model can predictthe tendency of wear status of aero-engine on the basis of historical data.(7)Developed a intelligentmonitoring system of aero-engine based on oil data which can forecast and diagnose the fault of theaero-engine judging by the data from the oil sample.