间歇过程的统计建模与在线监测

现代过程工业正逐渐倚重于生产小批量、多品种、高附加值产品的间歇过程.基于多元统计模型的过程监测是保障生产安全和产品质量的重要工具.从间歇过程独特的数据特性出发,将现有的多元统计建模方法进行合理的分类,简要回顾了各类方法的起源、发展及延伸的历程.除了阐述每种方法的基本原理,还详细讨论了各种方法的适用背景,相互关联及优缺点等内容,并对这一领域中依然存在的问题以及研究前景给出中肯的评述.     

定向判别分析新算法及其在沉积化学中的应用

介绍了处理多元有序数据的定向判别分析新方法原理、建模流程、应用流程及其在沉积化学中的应用实例。这种判别分析将分类建模与判别归类分开,求解与专业知识结合。新方法用多组或逐步判别分析对多元有序数据建模,应用时根据应用领域的知识对样本归属作初步定向,然后选择模型的相关局部进行判别归类,从而实现有序判别。这种方法用于解决由于时间序列多元数据周期性造成的样本分类颠倒问题。在塔里木盆地沉积岩时间序列化学数据的应用实例中,解决了石油井下地层预测和归类问题。     

定向判别分析新算法及应用

本文介绍了多元有序数据定向判别分析新方法的原理、建模流程、应用流程和应用实例。这种判别分析将分类建模与判别归类分开。新方法用多组或逐步判别分析对多元有序数据建模，应用时根据应用领域的知识对样本归属作初步定向，然后选择模型的相关局部进行判别归类。这种方法解决了由于时间序列多元数据周期性造成的样本分类颠倒问
问题。     

高炉炼铁过程多元铁水质量非线性子空间建模及应用

高炉炼铁是一个物理化学反应复杂、多相多场耦合的大滞后、非线性动态系统,其关键工艺指标——铁水质量参数的检测、建模和控制一直是冶金工程和自动控制领域的难题.本文提出一种面向控制的数据驱动高炉炼铁多元铁水质量非线性子空间建模方法.首先,为了提高建模效率和降低计算复杂度,采用数据驱动典型相关性分析与相关性分析相结合的方法提取与铁水质量相关性最强的关键可控变量作为建模的输入变量;同时,为了更好地反映高炉非线性动态特性,将相关输入输出变量的时序和时滞关系在建模过程进行考虑;最后,采用基于最小二乘支持向量机（Least square support vector machine,LS-SVM）的非线性Hammerstein系统子空间辨识方法建立数据驱动的多元铁水质量非线性状态空间模型.同时,将核函数表示的模型非线性特性用多项式函数拟合,在仅损失很小模型精度的前提下大大降低模型的计算复杂度.基于实际数据的工业试验验证了所提建模方法的准确性、有效性和先进性.     

复杂高炉炼铁过程的数据驱动建模及预测算法

高炉炼铁过程的控制意味着控制高炉铁水温度及成份在指定的范围. 本文以高炉炉内热状态的重要指示剂---高炉铁水硅含量为研究对象, 针对机理建模难以准确预测、控制高炉铁水硅含量的发展变化, 利用数据驱动建模的思想, 建立了基于多元时间序列的高炉铁水硅含量数据驱动预测模型. 实例分析表明, 建立的数据驱动预测模型能够很好地预测高炉铁水硅含量, 连续预测167炉高炉铁水硅含量, 命中率高达83.23%, 预测均方根误差为0.07260. 这些指标均优于基于单一硅时间序列所建立的数据驱动模型, 对实际生产具有很好的指导作用.     

CAPP中基于随机方位层次表达法的箱体零件建模

针对箱体类零件单件小批量生产的计算机辅助工艺规程编制，在分析现有零件建模方法的基础上，提出了基于随机方位层次表达的零件建模方法。     

增强分子拓扑信息的多任务图神经网络算法

以分子毒性为代表的分子属性预测在以药物设计为主的多个领域的发展中发挥着重要作用，但直接利用分子结构信息快速且准确地预测分子毒性一直是一个挑战。目前，卷积网络和图网络等深度学习方法的出现在这个问题的解决上得到了一定的进展。而以图网络为主的深度学习方法在分子毒性预测中存在两个关键问题，影响预测性能：第一，数据驱动使得模型在面对小批量数据时依然没有可靠的性能。第二，建模分子结构只考虑了天然共价键，只能提供粗粒度的信息。为解决上述问题，给出了一种对分子结构的新型建模方式MT-ToxGNN。该方法将多任务的思想融入图神经网络中，使得不同任务在训练时可以互相学习不同数据的可靠分布，从而避免在小批量数据上的过拟合问题。将分子编码成拓扑图结构时同时考虑分子内共价键以及非共价作用，就是在使用分子共价键构建传统图的边集之后，再使用非共价作用构建新型图的边集，从而弥补传统图网络对分子结构信息表示的不足。使用特别设计的图网络分别处理分子的共价与非共价信息，充分学习不同的分子结构。在与大量先进方法的性能比较中，MT-ToxGNN在多个分子毒性数据集上皮尔森系数指标达到了最佳。     

灰色神经网络仿真及其在MCMQC中的应用

为了解决大规模定制生产质量控制中质量数据缺乏的灰色不确定性问题,将灰色理论引入到了神经网络建模中.以灰色理论的核心思想(信息递阶、时变和多维特性)为基础,构造了基于灰色体系的建模方法,并对其进行了简化,给出了经过简化的一般性建模方法--横向、纵向和混合建模法,并将其与神经网络建模相结合.以一个典型的灰色BP神经网络为例,以横向建模方式对其进行了实现,并将其应用到大规模定制生产过程质量控制(MCMQC)中.仿真实验结果表明,灰色神经网络解决大规模定制生产质量控制中的灰色不确定问题是有效可行的.     

基于遗传搜索算法热轧合同编制优化仿真研究

为解决钢铁企业多品种、小批量的热轧合同编制优化问题,针对规模大、约束复杂难以建模及求解等难点,以半旬为基本时间单位,在考虑各钢种炼钢能力、轧制能力等约束条件的基础上,建立以合同的提前期、拖期惩罚最小,各工序产能利用均衡,相邻排产合同的工艺约束惩罚费用最小以及各半旬的炼钢余材最少为优化目标的0-1非线性整数规划模型.由于所建模型具有多旅行商问题结构的特征及模型中约束条件复杂、数据规模较大,采用分段整数编码和启发式修复策略的遗传搜索算法进行求解.通过对实际生产数据进行仿真,验证了所提模型和算法的有效性,为科学合理地编制热轧合同计划提供了有效的解决方法.     

数据与未建模动态驱动的非线性PID切换控制

本文充分利用系统的数据信息和知识,把数据驱动控制、PID控制与一步超前最优控制策略相结合,提出了数据与未建模动态驱动的非线性PID切换控制方法.该方法首先利用被控对象往往运行在工作点附近的特点及系统丰富可测的数据信息,把被控对象表示成低阶控制器设计模型与高阶非线性项(未建模动态)和的形式.与以往方法的本质区别在于,所提的方法直接将未建模动态分解为前一拍数据与未知增量的和,并充分利用未建模动态可测数据信息补偿系统未知的非线性动态特性,设计非线性PID控制器,对未建模动态的未知增量采用自适应神经模糊推理系统(ANFIS)进行估计,从而设计带有未建模动态增量估计的非线性PID控制器.将控制器的跟踪误差引入切换指标,两个控制器通过切换机制协调控制系统,既保证系统的稳定,同时提高系统的性能.为解决PID控制器参数难以选择的问题,采用一步超前最优控制策略进行参数设计,从理论上给出了PID控制器参数选择的一般原则和方法,推导了保证闭环系统输入输出稳定性的条件;最后,通过数值仿真实验以及在水箱液位控制系统的物理对比实验,实验结果验证了所提算法的有效性和实用性.     

A survey on multistage/multiphase statistical modeling methods for batch processes

In industrial manufacturing, most batch processes are inherently multistage/multiphase in nature. To ensure both quality consistency of the manufactured products and safe operation of this kind of batch process, different multivariate statistical process control (MSPC) methods have been proposed in recent years. This paper gives an overview of multistage/multiphase statistical process control methods used for process analysis, monitoring, quality prediction and online quality improvement. Different types of phase divisions and modeling strategies are introduced and the method properties are discussed. For comparisons, a selection guide to these methods for different application purposes is provided. Finally, some promising research directions are suggested based on existing works.     

Implications of group technology in automated standard data systems

Group technology may be defined as collecting components into groups to exploit their similarities in the production process. While initially used as an organizational method in small-batch manufacturing, group technology has been recognized to have much broader applications in computer-aided design and manufacturing. A new area for application is automated standard data systems. Integration of automated process planning and standard development can result in more effective planning and can allow further breadth of application of group technology into more labor-intensive manufacturing operations.     

基于迭代主成分分析的过程监测方法的研究与实现

利用迭代主成分分析（PCA）算法，提出一种在线过程监测方法，根据实际生产过程经验，提供了由多元统计控制图判断过程是否正常的准则，实现了实时在线的PCA建模和过程监测，仿真例子验证了这种过程监测方法的有效性和可行性。     

Application of integrated recurrent neural network with multivariate adaptive regression splines on SPC-EPC process

The integration of statistical process control and engineering process control has been reported as an effective way to monitor and control the autocorrelated process. However, because engineering process control compensates for the effects of underlying disturbances, the disturbance patterns become very hard to recognize, especially when various abnormal control chart patterns are mixed and co-existed in the engineering process. In this study, a new control chart pattern recognition model which integrates multivariate adaptive regression splines and recurrent neural network is proposed to not only address the problem of feature selection (i.e., lagged process measurements) but also improve the pattern recognition accuracy. The performance of the proposed method is evaluated by comparing the recognition results of multivariate adaptive regression splines and recurrent neural network with the results of four competing approaches (multivariate adaptive regression splines-extreme learning machine, multivariate adaptive regression splines-random forest, single recurrent neural network, and single random forest) on the simulated individual process data. The experimental study shows that the proposed multivariate adaptive regression splines and recurrent neural network approach can not only solve the problem of variable selection but also outperform other competing models. Moreover, according to the lagged process measurements selected by the proposed approach, lagged observations that exerted significant impact on the construction of the control chart pattern recognition model can be identified successfully. This study has significant implications for research and practice in production management and provides a valuable reference for manufacturing process managers to better understand and develop strategies for control chart pattern recognition.     

Data mining model-based control charts for multivariate and autocorrelated processes

Process monitoring and diagnosis have been widely recognized as important and critical tools in system monitoring for detection of abnormal behavior and quality improvement. Although traditional statistical process control (SPC) tools are effective in simple manufacturing processes that generate a small volume of independent data, these tools are not capable of handling the large streams of multivariate and autocorrelated data found in modern systems. As the limitations of SPC methodology become increasingly obvious in the face of ever more complex processes, data mining algorithms, because of their proven capabilities to effectively analyze and manage large amounts of data, have the potential to resolve the challenging problems that are stretching SPC to its limits. In the present study we attempted to integrate state-of-the-art data mining algorithms with SPC techniques to achieve efficient monitoring in multivariate and autocorrelated processes. The data mining algorithms include artificial neural networks, support vector regression, and multivariate adaptive regression splines. The residuals of data mining models were utilized to construct multivariate cumulative sum control charts to monitor the process mean. Simulation results from various scenarios indicated that data mining model-based control charts performs better than traditional time-series model-based control charts.     

Developing empirical models from observational data using artificial neural networks

Computer-integrated manufacturing requires models of manufacturing processes to be implemented on the computer. Process models are required for designing adaptive control systems and selecting optimal parameters during process planning. Mechanistic models developed from the principles of machining science are useful for implementing on a computer. However, in spite of the progress being made in mechanistic process modeling, accurate models are not yet available for many manufacturing processes. Empirical models derived from experimental data still play a major role in manufacturing process modeling. Generally, statistical regression techniques are used for developing such models. However, these techniques suffer from several disadvantages. The structure (the significant terms) of the regression model needs to be decided a priori. These techniques cannot be used for incrementally improving models as new data becomes available. This limitation is particularly crucial in light of the advances in sensor technology that allow economical on-line collection of manufacturing data. In this paper, we explore the use of artificial neural networks (ANN) for developing empirical models from experimental data for a machining process. These models are compared with polynomial regression models to assess the applicability of ANN as a model-building tool for computer-integrated manufacturing.Operated for the United States Department of Energy under contract No. DE-AC04-76-DP00613.     

Construction of quality control charts by using probability and fuzzy approaches and an application in a textile company

A method that uses statistical techniques to monitor and control product quality is called statistical process control (SPC), where control charts are test tools frequently used for monitoring the manufacturing process. In this study, statistical quality control and the fuzzy set theory are aimed to combine. As known, fuzzy sets and fuzzy logic are powerful mathematical tools for modeling uncertain systems in industry, nature and humanity; and facilitators for common-sense reasoning in decision making in the absence of complete and precise information. In this basis for a textile firm for monitoring the yarn quality, control charts proposed by Wang and Raz are constructed according to fuzzy theory by considering the quality in terms of grades of conformance as opposed to absolute conformance and nonconformance. And then with the same data for textile company, the control chart based on probability theory is constructed. The results of control charts based on two different approaches are compared. It’s seen that fuzzy theory performs better than probability theory in monitoring the product quality.     

Quality assurance for micro manufacturing processes and primary-shaped micro components

The development of robust and flexible manufacturing processes represents the major challenge to be overcome by industrially applicable micro production. For robustness and flexibility to be ensured also for micro dimensions, the manufacturing processes need to be controlled and continuously improved by means of a quality assurance system that is fast to respond, matches the requirements of production and is tailored to the micro-specific demands. Up to now, micro manufacturing processes have been characterized by a high degree of variability resulting from an increased number of significant influencing factors. Besides, the tolerance specifications of components with micro structures are in the micro and sub-micrometer range, requiring measuring methods to meet the highest demands in terms of accuracy. For micro manufacturing processes to be controlled on the basis of measured data, the increased significance of measuring uncertainty and measurement variation in relation to the required tolerances is to be taken into consideration. Measured data will always reflect the effect of both manufacturing and measurement variation, which may lead to wrong decisions being taken on the approval or rejection of components if the most rigid tolerance specifications need to be met. Therefore, the focus of this research document, generated as part of the Collaborative Research Center (SFB) 499, has been put on the continuous monitoring, control and separation of manufacturing and measurement variation by means of statistical methods and tools. The two ideas presented and discussed are a novel design for quality control charts and a control loop for the combination of statistical process control and statistical modeling, developed as part of the SFB 499. This article concludes by giving first results regarding the use of statistical tools in the field of micro manufacturing technology.     

Digital twin modeling method based on biomimicry for machining aerospace components

High-performance aerospace component manufacturing requires stringent in-process geometrical and performance-based quality control. Real-time observation, understanding and control of machining processes are integral to optimizing the machining strategies of aerospace component manufacturing. Digital Twin can be used to model, monitor and control the machining process by fusing multi-dimensional in-context machining process data, such as changes in geometry, material properties and machining parameters. However, there is a lack of systematic and efficient Digital Twin modeling method that can adaptively develop high-fidelity multi-scale and multi-dimensional Digital Twins of machining processes. Aiming at addressing this challenge, we proposed a Digital Twin modeling method based on biomimicry principles that can adaptively construct a multi-physics digital twin of the machining process. With this approach, we developed multiple Digital Twin sub-models, e.g., geometry model, behavior model and process model. These Digital Twin sub-models can interact with each other and compose an integrated true representation of the physical machining process. To demonstrate the effectiveness of the proposed biomimicry-based Digital Twin modeling method, we tested the method in monitoring and controlling the machining process of an air rudder.     

Nonlinear diffusion process modeling using response surface methodology and variable transformation

The nonlinear process modeling is investigated using statistical design method and response surface methodology. Three input factors are examined with respect to the response factor. In order to minimize the joint confidence region of fabrication process with varying conditions, D-optimal experimental design technique is performed and diffusion rate is characterized by response model. Then, the statistical results are used to verify the fitness of the nonlinear process model. Based on the results, this modeling methodology can be optimized process condition for semiconductor manufacturing.