柴油十六烷值的预测——逐步多重线性回归—近红外分光光度法

将逐步多重线性回归（ＳＭＬＲ）技术与近红外光谱相结合，建立了预测柴油十六烷值的新方法。通过考察最佳实验条件建立了最佳的预测模型，测定了２９个柴油样品，其测量精密度优于０．４个十六烷值单位，相对误差≤２．８５％。实验和预测结果表明，该方法试样用量少，不破坏样品，测量时间少于３ｍｉｎ，在实际生产中测定柴油十六烷值方便易行、准确可靠。

Prediction of Diesel Catane Number--Stepwise MultipleLinear Regression-Near Infrared Spectroscopy

Stepwise multiple linear regression-near infrared spectroscopy is used to create a prediction model of diesel catane number. The prediction model are built by investigating the best experimental condtions, and 29 samples of diesel are tested. The experiment and prediction results show that the precision of this method is less than 0.4 catane unit, the relative error of prediction result is less than 2.85%, the amount of sample needed is less, sample is not destroyed, the time of determination is less than 3 minutes. It provides a simple, accurate and reliable new method of predicating diesel catane number for practical production.