协同区间PLS——紫外光谱法同时测定肉桂酸和苯乙酮

结合紫外光谱和化学计量学算法,建立了同时测定混合物中肉桂酸和苯乙酮的方法。常规全谱偏最小二乘(PLS)由于存在与组分无关或呈非线性关系的波长点,使模型预测性能降低。本文采用协同区间偏最小二乘(synergy interval PLS,siPLS)算法选择特征波段,并优化波段组合。与全谱PLS模型比较,结果显示,基于siPLS的方法可以明显简化模型,提高模型的预测精度。研究为肉桂酸和苯乙酮的同时分析提供参考,也为siPLS算法在光谱计量学中应用提供借鉴。     

基于KTA-LSSVM的青霉素发酵过程预测建模

为解决青霉素发酵过程预测建模中存在的输入变量选择问题,提出了基于核目标度量（kernel target alignment,KTA）和最小二乘支持向量机（least squares support vector machines,LSSVM）的青霉素发酵过程预测模型。首先,在分析影响青霉素产物浓度相关因素的基础上选取输入变量,采用KTA对输入变量进行尺度缩放,然后,利用Pensim仿真平台数据,采用混沌粒子群算法对LSSVM的参数寻优,建立青霉素发酵过程的KTA-LSSVM预测模型。青霉素浓度预测的KTA-LSSVM模型均方根误差为0.0179,LSSVM模型的均方根误差为0.0276,实验结果表明,本文提出的模型预测精度高,推广性能好。     

LSSVM过程建模中超参数选取的梯度优化算法

基于结构风险最小的最小二乘支持向量机（least squares support vector machine, LSSVM）为标准支持向量机（SVM）的约简;训练简易;性能良好。其模型精度受超参数影响;常规的网络搜索法很难搜得最佳超参数。在快速留一法的基础上;以全样本留一预测误差平方和最小化为目标;导出基于梯度的最优化算法;用以优选为LSSVM超参数;进而构建G-LSSVM模型。以柠檬酸发酵过程为算例对G-LSSVM进行检验;结果表明G-LSSVM的超参数选取耗时少;模型稳定性良好;且拟合和预报性能都优于标准SVM和神经网络。有望适用于机理不明、高度非线性、小样本的化工过程建模。     

基于非线性偏最小二乘回归的软测量建模及应用

针对复杂工业过程存在的多变量、相关性和非线性问题,提出一种新的基于非线性偏最小二乘(partial least squares,PLS)回归的软测量建模方法。该方法利用PLS作为模型的外部框架来提取输入输出主成分变量,同时消除变量间的相关性,然后用最小二乘支持向量机(least squares support vector machine,LSSVM)作为内部函数来描述主成分变量之间的非线性关系,并引入基于误差最小化的权值更新策略,来改进模型的预测精度。以pH中和过程的Benchmark模型来验证该方法的性能,并与其他建模方法比较,结果表明该方法预测精度较高,而且具有较强的泛化能力。将该方法应用于某电站燃煤锅炉的NOx排放软测量建模之中,取得了较好的预测效果。     

混沌最小二乘支持向量机及其在发酵过程建模中的应用

针对青霉素发酵过程的参数检测存在不确定因素,提出一种基于混沌最小二乘支持向量机的青霉素浓度预测方案。采用混沌优化算法对最小二乘支持向量机参数进行寻优,建立了一种混沌最小二乘支持向量机模型。首先,利用该模型对两种常规非线性函数曲线进行了仿真回归,结果表明,算法具有良好的建模精度;其次,基于Pensim仿真平台,运用文中方法预测青霉素发酵过程的产物量,实验仿真表明混沌优化算法具有良好的全局优化性能,在参数选择中可以有效避免陷入局部最小值,基于混沌优化的最小二乘支持向量机具有较高的建模精度。     

FiPLS在烟煤水分近红外检测中的应用

偏最小二乘是一个在近红外光谱解析中常用的计量学算法,结合变量筛选方法既可以提高模型的预测能力,也可以大大降低建模的难度。本文将前向区间偏最小二乘用于烟煤水分近红外光谱解析。提取出的区间数为2,变量个数从1557减少到54个。所提取的波长区间主要位于O-H一级泛频吸收带。预测平均绝对百分误差从0.0865降低到0.0818。研究结果表明,前向区间偏最小二乘可以显著减少变量数并提高预测准确度。     

基于LSSVM的溶剂油分离建模方法研究

针对传统支持向量机和单一模型建模的缺点,利用某炼油厂溶剂油分离过程中二侧线流量作为建模对象,对最小二乘支持向量机集成学习方法进行了研究。首先利用自适应系数加权模糊(AWFCM)聚类算法对训练样本进行聚类;然后对每一类数据使用最小二乘支持向量机建立子模型,并使用PLS合成函数得到最小二乘支持向量机集成模型;最后通过仿真实验来验证最小二乘支持向量机集成模型预测的精确性。结果表明,该算法在预测精度上有了较大的提高,对过程控制系统中分离效果的预测具有重要指导意义。     

基于最小二乘支持向量机回归的单桩竖向极限承载力预测

基于单桩载荷试验数据,采用最小二乘支持向量机(LSSVM)回归的方法,建立了单桩竖向极限承载力的预测模型.利用文献中桩的载荷试验数据来训练LSSVM模型,并确定了模型参数.研究结果表明,同常用的BP网络相比,LSSVM预测模型具有学习速度快、预测性能较好、选择参数少等优点,是一种有效的预测单桩极限承载力的方法.     

基于粗差判别的参数优化自适应加权最小二乘支持向量机在PX氧化过程参数估计中的应用

针对软测量建模过程中数据可能存在粗大误差以及粗差数据对模型的性能产生的影响,提出了一种基于粗差判别的自适应加权最小二乘支持向量机回归方法(WLS-SVM)。 该方法首先根据3δ法则检测出样本中的显著误差并加以剔除,然后根据样本误差的大小自适应地调整权值,使得非显著误差对模型性能的影响大大降低。另外,由于最小二乘支持向量机的正则化参数和核宽度参数对模型的拟合精度和泛化能力有较大的影响,一般依靠经验和试算的方法进行估计,耗时且不准确,本文将模型的参数作为进化算法的优化问题,应用自适应免疫算法(AIGA)对参数进行优化选择。仿真实验表明,该方法对非线性系统的建模具有很好的效果。同时,将该方法应用于工业PX氧化建模过程中动力学参数的估计中,结果表明,基于粗差判别的参数优化自适应最小二乘支持向量机预测精度高,取得了较好的效果。     

基于近红外光谱的汽油辛烷值在线分析仪

介绍自主开发研制的汽油辛烷值近红外(NIR)光谱在线分析仪。该分析仪包括NIR光谱在线测量、光谱预处理与实时建模等部分。对于原始的NIR光谱数据,采用多项式卷积算法进行光谱平滑、基线校正和标准归一化;通过模式分类与偏最小二乘进行实时建模。该分析仪已成功应用于某炼油厂重整反应器液相产物的辛烷值在线分析。     

变性燃料乙醇的快速近红外光谱(DA7200)分析

本文论述了采用近红外光谱DA7200的液体分析装置检测变性燃料乙醇主要成分含量的方法。讨论采用偏最小二乘法(PLS)建立校正模型过程中样品预处理及利用常规吸收峰优选波长的方法。经验证：水分、乙醇、甲醇、变性剂的预测值与浓度参考值具有良好相关性(相关系数大于0．96),测量重复性变异系数(CV)优于2％。结果表明,近红外光谱DA7200的液体分析装置可以满足变性燃料乙醇主要成分含量的实际测量要求,并可作为在线分析的参考模式。     

Unification of neural and statistical modeling methods that combine inputs by linear projection

Empirical modeling methods that combine inputs by linear projection include linear methods such as, ordinary least-squares regression, partial least-squares regression, principal components regression, and nonlinear methods such as, backpropagation networks with a single hidden layer, projection pursuit regression, nonlinear partial least-squares regression, and nonlinear principal components regression. In this paper, these popular modeling techniques are unified to yield a single method called nonlinear continuum regression (NLCR). This unification is based on the insight provided by a common framework for empirical modeling methods, and is achieved by using activation functions that adapt to the measured data, a common optimization criterion for finding the projection directions, and a hierarchical training methodology that allows efficient modeling. The adaptive-shape activation functions are determined by univariate smoothing in the space of the projected input versus output. The NLCR optimization criterion contains an adjustable parameter that controls the degree of overfitting or bias of the model, and spans the continuum of methods from projection pursuit regression or backpropagation networks to nonlinear principal components regression. Consequently, NLCR results in models that are usually more general and compact than those obtained by existing methods based on linear projection, while eliminating the need for arbitrary selection of an empirical modeling method based on linear projection for a given task. The improved modeling ability of NLCR and its performance on different types of training data are illustrated by examples based on simulated and industrial data.     

Implementation of soft computing approaches for prediction of physicochemical properties of ionic liquid mixtures

The main objective of this study was to develop soft computing approaches for prediction of physicochemical properties of IL mixtures including: density, heat capacity, thermal conductivity, and surface tension. The proposed models in this study are based on support vector machine (SVM), least square support vector machines (LSSVM), and group method of data handling type polynomial neural network (GMDH-PNN) systems. To find the LSSVM and SVM adjustable parameters, genetic algorithm (GA) as a meta-heuristic algorithm was utilized. The results showed that LSSVM is more robust and reliable for prediction of physicochemical properties of IL mixtures. The proposed GA-LSSVM model provides average absolute relative deviations of 0.38%, 0.18%, 0.77% and 1.18% for density, heat capacity, thermal conductivity, and surface tension, respectively, which demonstrates high accuracy of the model for prediction of physicochemical properties of IL mixtures.     

基于多尺度核JYMKPLS迁移模型的间歇过程产品质量的在线预测方法

针对过程数据不足,且具有强非线性和多尺度特性的新间歇过程,结合迁移学习方法与多尺度核学习方法的优势,提出了一种基于多尺度核JYMKPLS（Joint-Y multi-scale kernel partial least squares）迁移模型的间歇过程产品质量在线预测方法。该方法首先通过迁移学习利用相似源域的旧过程数据提高新间歇过程建模效率和质量预测的精度。然后,针对间歇过程数据的非线性和多尺度特性问题,引入了多尺度核函数以更好地拟合数据变化的趋势,从而提高模型的预测精度。此外,提出模型在线更新和数据剔除,通过在线持续改善迁移模型对新间歇过程的匹配程度,以消除相似过程间的差异性给迁移学习带来的不利影响,从而不断地提升预测精度。最后,通过仿真验证了所提方法的有效性,结果表明,与传统的数据驱动建模方法相比,本文所提方法能够有效提高建模效率和预测精度。     

Dynamic global feature extraction and importance-correlation selection for the prediction of concentrate copper grade and recovery rate

Large-scale industrial data have brought great challenges to data calculation and analysis. Feature extraction and selection have become one of the research emphases in data mining. To mine the dynamic characteristics of large-scale industrial data, a dynamic global feature extraction (DGFE) method integrating principal component analysis (PCA) and kernel principal component analysis (KPCA) is proposed such that the achieved feature set is not only dynamic but also contains linear and non-linear features. To ensure that the obtained feature set is optimal with the minimum redundancy, a new importance-correlation-based feature selection (ICFS) method is proposed. To verify the validity and feasibility of the proposed methods, the partial least square (PLS) and least square support vector machine (LSSVM) prediction models for the concentrate copper grade and the recovery rate are established. The effectiveness of the proposed methods is verified through data experiments on a copper flotation industrial process.     

基于分阶段的LSSVM发酵过程建模

发酵过程建模是研究微生物发酵的重要课题,基于模型可实现被测参量的软测量、系统的优化控制。鉴于引入混合核函数的最小二乘支持向量机在过程建模中具有优良表现,采用基于混合核函数的最小二乘支持向量机建模。但由于发酵过程周期较长,最小二乘支持向量机的全局模型预测精度难以保证,算法复杂度很高,因此提出一种分阶段建模方法。首先,选择表征阶段特性的辅助变量,利用模糊C均值聚类算法对样本数据聚类,将发酵过程分成不同的阶段,然后为各个阶段分别建立最优混合核最小二乘支持向量机局部模型,最后将局部模型合成构成过程的完整模型。将此方法应用于青霉素发酵过程和重组大肠杆菌发酵过程中,验证了该方法的有效性。     

Environmentally Friendly Determination of Quality Parameters of Biodiesel/Diesel Blends Using Fourier Transform Infrared Spectra

Multivariate calibration models based on data from mid‐infrared spectroscopy of biodiesel/diesel blends were obtained. The blends were prepared from diesel oil and esters of soybean oil, waste cooking oil, and hydrogenated vegetable oil in proportions ranging from 0 to 100 % biodiesel. The results showed that the multivariate regression models with interval partial least squares (iPLS), backward interval partial least squares (biPLS), and synergy interval partial least squares (siPLS) were able to determine the fractions of the infrared spectrum that contain the relevant information for estimating the values of physicochemical properties, flash point, specific gravity, and cetane number, which are used in quality control of the blends. In the best models, the values of determination coefficients were greater than 0.9500, proving their efficiency as an alternative to traditional analytical methods.     

Comparative study of estimation methods of NO<Subscript><Emphasis Type="Italic">x</Emphasis></Subscript> emission with selection of input parameters for a coal-fired boiler

This study focuses on estimation of NO_x emission and selection of input parameters for a coal-fired boiler in a 500 MW power generation plant. Careful selection of input parameters is required not only to improve accuracy of the estimation, but also to reduce the model dimensionality. The initial operating input parameters are determined based on operation heuristics and accumulated operation knowledge; the essential input parameters are selected by sensitivity analysis where the performance of the estimation model is assessed as one or some input parameters are successively eliminated from the computation while all other input parameters are retained. From the sequential input selection process, less than ten input parameters survived out of 36 initial input parameters. Auto-regressive moving average (ARMA) model, artificial neural networks (ANN), partial least-squares (PLS) model, and least-squares support vector machine (LSSVM) algorithm were proposed to express the relationship between the operating input parameters and the content of NO_x emission. Historical real-time data obtained from a 500 MW power plant coal-fired boiler were used to test the proposed models. It was found that principal components analysis (PCA) enhances the estimation performance of each model. Among the four proposed estimation models, the LSSVM model coupled with PCA scheme showed the minimum root-mean square error (RMSE) and the best R-square value.