Research On Wear Condition Of Aeroengine Based On Oilspectrum Data

Aero-engine aircraft as the "heart" provide power for the aircraft flight. The structure ofaero-engine aircraft is very complex and long-term high temperature, high load conditions ofwork, very prone to wear and tear faults affecting the safety operation of the engine. To thisend the parts of aero-engine for wear monitoring is particularly important. If you can find thehidden faults and wear their removal in the early embryonic stage, which will ensure thesafety of aircraft engines, stable and reliable work of great significance.Lubricants can not only slow the wear between components, cooling components, whilealso taking the heat generated by mechanical rotation and the friction of the metal grains.These metal particles uniformly suspended in oil, in the same time these abrasive wearcondition of aircraft engines contains important information. First, this paper does notconsider non-isochronous building up the oil situation unvaried network of support vectormachines to predict the trends of Fe element. Then consider the non-isochronous building upthe oil situation multivariate GRNN network to predict trends of Fe element. The conclusionis, considering the non-isochronous up the oil in case the established network to better fit thetrend of Fe, better judgment aircraft engine wear.For Considering non-isochronous case of oil up the oil data set SVR model to fit thechange of Fe element, and the penalty factor c and kernel function parameter g using crossvalidation (K-CV) method to optimize the choice. Finally, this article uses thewell-established cross-validation optimal concentration of Fe SVR model to predict trendsand obtain good prediction.Finally, we use gray correlation analysis method, processing the acquired aero-engineflight hours under different oil spectroscopy data obtained in different metallic elements.From the horizontal relationships between elements for analysis, although the400hours flighttime away from the elements and the element of Fe gray correlation between the values hasdecreased in varying degrees intense. The judgment result is consistent with the actual working conditions, the cost of this method is low and the affected by human factors issimultaneously less.