近红外光谱的ELM校正模型测定柴油凝点

为了简便、快速地测定柴油凝点,建立了一个基于ELM算法的近红外光谱校正模型。首先选择KS法按4:1划分样本集,并以"一阶导数+矢量归一化"方法进行光谱预处理;以校正集数据训练ELM模型并进行参数优化后,代入测试集光谱数据完成预测。通过136个柴油样品数据建模验证,结果表明用近红外光谱的ELM校正模型测定柴油凝点是完全可行的,且其准确度、稳健性和速度均优于PLSR模型和LS-SVM模型。

Determination of Diesel Freezing Point by ELM Calibration Model of Near-infrared Spectrum

In order to conveniently and quickly determine diesel oil solidifying point,a NIRS calibration model based on ELM algorithm was established.Firstly,the KS method was used to divide the sample set by 4:1,and the spectrum was preprocessed by "first-order derivative + vector normalization" method.The ELM model was trained with the calibration set data and the parameters were optimized.The prediction was completed by substituting test set spectral data.The modeling validation of 136 diesel sample data was carried out.The results show that it is feasible to use near infrared spectroscopy ELM calibration model for determination of freezing point of diesel fuel.And its accuracy,robustness and speed are better than PLSR model and LS-SVM model.