基于LS-SVM的航空润滑油金属含量分析

Aviation lubricating oil metal content analysis based on LS-SVM

在航空润滑油光谱分析故障诊断中,对润滑油浓度的变化趋势进行预测是油液光谱分析故障诊断的重要内容之一。针对这一分析方法,介绍了最小二乘支持向量机并对航空发动机滑油系统铁元素浓度的变化趋势进行了预测;并采用遗传算法对最小二乘支持向量机的参数进行优化,通过与灰色模型的预测结果作比较,其结果表明,优化后的最小二乘支持向量机的预测精度高,具有很好的泛化能力和学习能力。

Because the aviation lubricating oil has an important effect to the aviation equipment,it is very necessary for us to monitor the aviation lubricating oil.In the aviation lubricant spectrum analysis fault diagnostic, estimate the variety trend of the lubricant density is one of the important contents that oil liquid spectrum analysis fault diagnosis .Aiming at the analysis method, this text introduces that LS-SVM forecast the change trend of the iron element density in the aviation engine slippery oil systemoin the paper,the hyper-parameters of LS-SVM could be optimized by genetic algorithm.The results show that the model has excellent learning ability and can provide more accurate data prediction compared with gray model.The error of the model is very little and the method is effective.