一种成品油模型校正集自动维护方法设计
|
摘要
近红外光谱定量分析技术用于成品油性质检测具有快速、高效的特点,但该方法对模型的质量要求很高,常需要定期维护。提出一种模型校正集自动维护的方法,首先采用主成分分析方法确定校正集样本与新样本的分布,再通过相似样本数量判断新样本是否处于稀疏区,并剔除异常样本,最后自动将处于稀疏区的新样本添加到模型校正集中。实验研究表明,该方法可有效避免建模一段时间后,因新样本偏离原校正集样本分布区域,导致模型预测精度降低的问题。
Near infrared spectroscopy based quantitative analysis is a rapid and efficient method for measuring the properties of refined oils. However, high quality of model is strongly needed, and regular maintenance of calibration set is required due to the change of working conditions for a period of time.A novel method for automatic maintenance is proposed in this paper.Firstly,principal component analysis is introduced to describe the distribution of calibration set samples and new samples.Secondly,the quantity of similar samples is employed to determine whether the new samples are in sparse area.Finally,the abnormal samples are eliminated and the remaining samples are automatically added to the calibration set. The experimental results indicate that this method can effectively improve the model prediction accuracy,which is of great significance in controlling production in refinery plants.
Near infrared spectroscopy based quantitative analysis is a rapid and efficient method for measuring the properties of refined oils. However, high quality of model is strongly needed, and regular maintenance of calibration set is required due to the change of working conditions for a period of time.A novel method for automatic maintenance is proposed in this paper.Firstly,principal component analysis is introduced to describe the distribution of calibration set samples and new samples.Secondly,the quantity of similar samples is employed to determine whether the new samples are in sparse area.Finally,the abnormal samples are eliminated and the remaining samples are automatically added to the calibration set. The experimental results indicate that this method can effectively improve the model prediction accuracy,which is of great significance in controlling production in refinery plants.
引文
[1]陆婉珍,褚小立,等.近红外光谱(NIR)和过程分析技术(PAT)[J].现代科学仪器,2007,(4):13-17.
[2]刘多强,关绍春,孙建章,等.中红外光谱和近红外光谱在油品分析中的技术比较[J].石油化工应用,2010,29(6):1-3.
[3]陈奕云,赵瑞瑛.结合光谱变换和Kennard-Stone算法的水稻土全氮光谱估算模型校正集构建策略研究[J].光谱学与光谱分析,2017,(7):2133-2139.
[4]阿尔达克·克里木.基于高光谱的ASTER影像土壤盐分模型校正及验证[J].农业工程学报,2016,(12):144-150.
[5] Jose A.,et al.Rapid Determination of Olive Oil Oxidative Stability and Its Major Quality Parameters Using Vis/NIR Transmittance Spectroscopy[J].J. Agric. Food Chem.2013,61(34):8056-8062.
[6]贾生尧.近红外光谱技术结合递归偏最小二乘算法对土壤速效磷与速效钾含量测定研究[J].光谱学与光谱分析,2015,(9):2516-2520.
[7]柏文良.烟叶四种成分近红外模型的建立、维护与转移[D].广州:中山大学,2010.
[2]刘多强,关绍春,孙建章,等.中红外光谱和近红外光谱在油品分析中的技术比较[J].石油化工应用,2010,29(6):1-3.
[3]陈奕云,赵瑞瑛.结合光谱变换和Kennard-Stone算法的水稻土全氮光谱估算模型校正集构建策略研究[J].光谱学与光谱分析,2017,(7):2133-2139.
[4]阿尔达克·克里木.基于高光谱的ASTER影像土壤盐分模型校正及验证[J].农业工程学报,2016,(12):144-150.
[5] Jose A.,et al.Rapid Determination of Olive Oil Oxidative Stability and Its Major Quality Parameters Using Vis/NIR Transmittance Spectroscopy[J].J. Agric. Food Chem.2013,61(34):8056-8062.
[6]贾生尧.近红外光谱技术结合递归偏最小二乘算法对土壤速效磷与速效钾含量测定研究[J].光谱学与光谱分析,2015,(9):2516-2520.
[7]柏文良.烟叶四种成分近红外模型的建立、维护与转移[D].广州:中山大学,2010.