蝙蝠算法优化近红外光谱校正模型测定柴油黏度  

Determination of Diesel Viscosity Based on NIRS Correction Model Optimized by Bat Algorithm

作　　者：胡振[1] 陈素彬[1] 张晓琪[1] 唐天国[1] 杨华[1] HU Zhen;CHEN Su-bin;ZHANG Xiao-qi;TANG Tian-guo;YANG Hua(Nanchong Vocational and Technical College, Sichuan Nanchong 637131, China)

机构地区：[1]南充职业技术学院,四川南充637131

出　　处：《当代化工》2019年第3期647-651,共5页Contemporary Chemical Industry

基　　金：四川省教育厅2018年度科研项目;项目号:18ZB0316;南充职业技术学院2017年度科研课题;项目号:ZRB1704

摘　　要：为了提高近红外光谱定量分析的精度和效率,提出以改迚二迚制蝙蝠算法同步优化参数和特征波长的LS-SVM模型。首先将动态速度权重和Cauchy随机扰动融入蝙蝠算法,幵用V-shaped函数将其离散化,得到WCBBA算法;然后以RMSECV为适应度函数,用WCBBA算法对LS-SVM模型的参数和特征波长迚行同步优化;最后根据优化结果建立近红外光谱定量校正模型,幵用于180个柴油样本的运动黏度测定。结果表明,该模型的各项性能指标优异,能够用于实际检测工作。In order to improve the accuracy and efficiency of the quantitative analysis of the near infrared spectrum, a LS-SVM model was proposed to improve the binary bat algorithm’s synchronization optimization parameters and characteristic wavelengths. First, the dynamic velocity weight and the Cauchy random disturbance were integrated into the bat algorithm, and the V-shaped function was used to discretize it, the WCBBA algorithm was obtained. Then the RMSECV was used as the fitness function and the WCBBA algorithm was used to synchronize the parameters and characteristic wavelengths of the LS-SVM model. Finally, a quantitative correction model of near infrared spectroscopy was established based on the optimization results, and 180 diesel samples were used to measure the kinematic viscosity. The results show that the model has excellent performance and can be used for practical testing.

关 键 词：蝙蝠算法 近红外光谱 定量模型 参数优化 特征波长 

分 类 号：TQ016[化学工程]