非等间隔阶跃灰色模型在滑油光谱分析中的应用

Unequal Interval Jump Grey Modeling and Its Application to the Spectral Analysis of Lubricating Oil

针对非等间隔、具有阶跃趋势的原始数据序列建立了非等间隔阶跃灰色模型 ,模型参数的辨识采用了非线性最小二乘估计的Levenberg Marquardt算法。将所建模型用于 16V2 80ZJA型机车柴油机滑油的光谱分析数据的建模 ,得到了较高的拟合精度 ,尤其在阶跃点附近 ,大大改善了拟合精度。利用模型参数对换油后的测试数据进行修正 ,有利于提高光谱分析的准确度和可靠性。

Unequal interval jump grey model was built for raw data series with unequal interval and jump trend in this paper. Levenberg-Marquardt arithmetic that belongs to non-linear least-square estimation was used to recognize the parameters. The model built was used to fit spectrometric analysis values of diesel engine and the fitting precision is good. It is helpful to improve the accuracy and reliability of spectrometric analysis by revising test data after oil change with model parameter.