摘要
以近红外光谱法为基础测定方法,结合内模控制,论述了采用自适应神经网络建立校正模型测定石油化工产品组成的可行性.基于dSPACE硬件平台,实验以直馏柴油、加氢精制柴油和催化裂化柴油为校正模型的训练样本,对自适应神经网络校正模型进行了检验,实验结果表明:该方法响应快、误差小、鲁棒性强,在近红外长波区内,校正样品和验证样品的均方误差小于10-6.
A novel near infrared spectroscopy method for measuring the composition of chemical products is proposed.It is based on near infrared spectroscopy method and adaptive neural network internal model control.The feasibility of the method is discussed.The adaptive neural network calibration model is tested by using the experimental data of straight-run diesel,hydrofining diesel and catalytic cracking diesel as training samples.It is shown that the method has high response speed,little error and good robustness,and the mean squared error(MSE)of calibration samples and predicted samples is all the order of 10-6 in the spectral range of 800~2 300 nm.
出处
《西安石油大学学报(自然科学版)》
CAS
北大核心
2009年第1期56-60,共5页
Journal of Xi’an Shiyou University(Natural Science Edition)
基金
上海市教委自然科学基金(编号:050Z10)资助
关键词
近红外光谱
自适应神经网络
内模控制
定量分析
near infrared spectroscopy
adaptive neural network
internal model control
quantitative analysis