【摘要】 针对柴油发动机润滑油样光谱分析Fe元素数据中存在的野点、漏点、上下波动、"断层"现象等不利于柴油发动机磨损状态判别的问题,分析其产生的可能原因,对数据提出不同的修正处理方法。将原始光谱分析数据序列分成原油样数据、换油后数据、换滤后数据三段,对原油样数据采用灰度模型、线性回归、Gaussian模型、Hermite插值等方法进行修正处理,对换油和换滤后数据采用灰度模型进行预测修正和光滑处理,实现光谱分析数据的预处理。和原始序列相比,修正后数据在保持原有变化趋势的基础上实现平滑递增。更多还原

【Abstract】 For the problem of odd point,omitting point,fluctuating and fault phenomenon,which wasn’t beneficial to the wear identification of diesel engine,in the Fe element spectral analysis data of lubricating oil,the reason was analyzed and different amending methods for data were put forward.The original spectral analysis data were divided into the original data of oil,the data after oil change and the data after filter change.The pretreatment of Fe element spectral analysis data was realized by amending the original data of oil with the grey model,linear regression,Gaussian model and Hermite interpolation,by the forecast amending and smoothing treatment of the data after oil change and after filter change with the grey model.Compared with the original spectral analysis data,the amended data keep the original trend and the Fe element data achieve a smooth increment.更多还原