柴油机油液光谱分析中换油干扰的处理

摘要

介绍了油液监测方式中的光谱分析方法,分析了原子发射光谱分析过程中润滑油更换对光谱分析数据产生的影响.以润滑油样中光谱分析的Fe元素为例,就换油产生数据分析干扰,讨论了直接测量、线性回归、Hermite、灰度模型等干扰消除方法.结果表明:直接测量法最简单,但修正结果最不理想;线性回归法受样本数据量的影响;Hermite法对样本数据有稳定变化的要求;灰度法对样本数据的变化有较强的适应性,能够较理想地实现换油干扰的消除.%The spectrum analysis method of oil monitoring was introduced and the influence of lubricant oil changing on data analysis during the analyzing process of atom emission spectrum was analyzed. Taking Fe element as an example, the methods of direct measurement, linear regression, Hermite and grey model for eliminating the disturbance were discussed. The results show that the direct measurement method is the simplest but the most unsatisfactory. The linear regression method is much influenced by the sample quantity. The Hermite method has the requirement of steady variance for the sample data. The grey model method can be adaptable to different data and so can eliminate the disturbance caused by changing oil better.