动态隐变量法及其在动态过程监控中的应用

针对动态系统模型难以获取的问题,以提高统计监控性能为目标,对现有的动态隐变量法进行了深入研究.首先指出动态隐变量可包含更多的动态信息,但仍具有自相关,为此提出了采用修正控制图的方法对动态隐变量空间进行检测,而对于自相关不显著的残差空间,指出可按照传统方法或是非参数方法建立控制图;接着将过程知识和经验受控平均运行长度的检验考虑在内,给出了一种时滞变量和时滞长度的确定方法;最后,提出了一种根据残差累积和以及递归特征消除算法(recursive feature elimination,RFE)进行故障变量辨识的方法.通过对双效蒸发过程的应用监控,表明了上述方法的有效性.     

基于SKICA的非线性过程缓变故障检测方法研究

针对非线性工业过程缓变型故障的检测问题,提出一种基于累积和核独立元分析(SKICA)的故障检测方法.通过核函数技术将观测数据从非线性空间映射到线性空间,然后对线性空间的数据应用独立元分析算法,提取观测数据中的非线性独立元.为了更好的检测过程中微小变化和缓变故障,进一步应用累积和控制图(CUSUM)的思想.建立累积和非线性独立元并以此构造统计量监控过程变化.在连续搅拌反应器(CSTR)上的仿真结果表明,SKICA方法能够比ICA方法更快的检测出非线性过程中的缓变型故障.     

基于检定数据自相关性的低压智能电能表局部异常点检测方法

低压电能表在运行过程中，若不能及时找出局部异常点故障，会直接影响智能电网的电力传输质量。为保证低压电能表的安全运行，提出基于检定数据自相关性的低压智能电能表局部异常点检测方法。该方法首先对电能表内部结构展开具体分析，获取电能表运行影响因素，并基于分析结构，结合检定自相关函数实施电能表数据的去噪处理；通过对低压智能电能表状态特征值的提取，利用BP-AdaBoost复合神经网络建立电能表的局部异常点检测模型，将获取的特征值输入模型中，根据模型输出实现低压智能电能表局部异常点的精准检测。实验结果表明，利用该方法开展低压智能电能表局部异常点检测时，可靠性高，检测效果好。     

基于冗余关系分析的传感器自诊断设计方法研究

工业过程中传感器数量众多且可靠性要求高,而传统定期检测评估其健康状况的方式不但费时费力且不能满足传感器 智能化的发展需求。 针对这一问题,提出了一种基于测量数据统计相关性的传感器自诊断设计方法。 利用传感器测量数据建 立其统计关系模型,借助自编码器提取传感器数据特征并将其编码为二进制形式。 在同时考虑传感器测量数据统计独立和统 计相关两种情况下,在有参考值时,通过引入故障检测概率和误检概率建立了独立统计模型实现传感器的故障自诊断;在无参 考值情况下,借助高斯 Copula 函数建立多元统计依赖模型评估参数之间的相关性,并利用贝叶斯理论在不依赖参考值的情况 下自学习获取传感器的健康状况。 本研究以镍闪速炉系统为例,两种模式下测量系统中健康传感器的故障检测后验概率达到 了 0. 92,即故障统计模型的参数与建模期望相符。 实验结果表明,所提方法在两种模式下均能准确识别出测量系统中的故障 传感器,验证了所提方法的有效性与可行性。     

方差监测的自适应累积和车身质量控制图研究

针对车身制造质量检测工作量大、数据处理方式简单等特点,提出一种基于方差的改进累积和控制图（CUSUM）方法,用于监测车身制造质量的方差波动。其基本思想是对控制图参数k动态更新和迭代,并与方差波动量相联系,以便实时监测车身焊接尺寸过程方差的微小波动。通过对控制图的平均运行链长进行分析和实际案例研究,并与常规累积和控制图和指数加权移动平均控制图作对比,表明该方法对过程方差微小变异更为有效和敏感。     

椭圆控制理论在轴承故障检测中的应用

本文将椭圆控制理论引入到轴承故障检测中,提出新的轴承故障检测方法。首先对轴承正常状态下的数据进行分段,以各段数据提取到的AR模型系数和残差为参考状态向量,然后将待测数据的AR系数和残差依次添加到参考状态向量中并计算其前两阶主成分,以得到的前两阶主成分构建控制椭圆,最后根据待测数据AR系数和残差的前两阶主成分在椭圆控制图中的分布来判断轴承是否发生故障。实验结果表明:该方法可有效地识别轴承是否出现故障。     

监控两阶段相关过程的触发CUSUM-CUSCORE控制图

为解决过程自相关情况下两阶段过程发生趋势性漂移的监控问题,设计了触发CUSUM-CUSCORE控制图。充分利用故障表征中隐藏的动态信息,提出采用触发CUSUM-CUSCORE控制图对两阶段相关过程进行监控,以探测过程输出质量特性发生的趋势性漂移。通过仿真,以平均运行链长为准则,比较了几种控制图的监控效果。仿真结果表明,基于触发CUSUM-CUSCORE控制图对过程输出质量特性具有较好的监控性能。     

过程控制中测量系统误差的性能影响分析

从控制图的建立和监控两个不同阶段的角度分别讨论了测量系统误差对过程质量控制监控性能的影响.以平均运行长度作为过程控制性能评价的指标,分析了两种不同类型的测量误差对监控性能的干扰模式.最后通过一个算例分析测量系统误差的危害性,表明一个好的测量系统在过程监控中的重要性.     

面向自相关过程的改进ARMAST控制图与Grubbs调和规则的集成

为解决多品种小批量生产方式存在过程输出自相关和机器设置的频繁调整通常导致过程均值偏离目标值两个问题。提出以改进的稳态自回归滑动平均控制图为监控技术、以基于稳态卡尔曼滤波器的Grubbs调和规则为调整技术的集成理论。其中,针对稳态自回归滑动平均控制图在高阶自回归滑动平均过程下难以求解稳态方差的问题,建立了通用的任意阶平稳自回归过程下稳态方差的求解方法,使其适用范围更广;针对稳态自回归滑动平均控制图的参数设计方法效率较低,提出了基于条件概率的参数设计方法,使其监控效率更好。当控制图预警后,采用Grubbs调和规则对生产过程进行调整。通过仿真实验和实例研究验证了改进的参数设计方法和集成方法的有效性。     

事件驱动的流水线建模与质检机器配置优化

数字孪生车间环境下获取生产数据更加便捷,使系统运行过程中收集机器故障、质检报废等扰动事件成为可能.为了科学评估并及时应对扰动事件对系统性能的影响,面向考虑质检报废的流水线,研究流水线性能评估与质检机器配置优化问题.考虑流水线中的质量检测操作,定义了质检报废事件,定量分析了机器故障与质检报废等扰动事件对系统产出的影响,建...     

Effects of imprecise measurement on the two dependent processes control for the autocorrelated observations

The observations from the process output are always assumed to be independent when using a control chart to monitor a process. However, for many processes the observations are autocorrelated and including the measurement error due to the measurement instrument. This autocorrelation and measurement error can have a significant effect on the performance of the process control. This paper considers the problem of monitoring the mean of a quality characteristic X on the first process, and the mean of a quality characteristic Y on the second process, in which the observations X can be modeled as an ARMA model and observations Y can be modeled as an transfer function of X since the state of the second process is dependent on the state of the first process. To distinguish and maintain the state of the two dependent processes with measurement errors, the Shewhart control chart of residuals and the cause-selecting control chart, based on residuals, are proposed. The performance of the proposed control charts is evaluated by the rate of true or false alarms. By numerical analysis, it shows that the performance of the proposed control charts is significantly influenced by the variation of measurement errors. The application of the proposed control charts is illustrated by a numerical example .     

Effects of imprecise measurement on the two dependent processes control for the autocorrelated observations

The observations from the process output are always assumed to be independent when using a control chart to monitor a process. However, for many processes the observations are autocorrelated and including the measurement error due to the measurement instrument. This autocorrelation and measurement error can have a significant effect on the performance of the process control. This paper considers the problem of monitoring the mean of a quality characteristic X on the first process, and the mean of a quality characteristic Y on the second process, in which the observations X can be modeled as an ARMA model and observations Y can be modeled as an transfer function of X since the state of the second process is dependent on the state of the first process. To distinguish and maintain the state of the two dependent processes with measurement errors, the Shewhart control chart of residuals and the cause-selecting control chart, based on residuals, are proposed. The performance of the proposed control charts is evaluated by the rate of true or false alarms. By numerical analysis, it shows that the performance of the proposed control charts is significantly influenced by the variation of measurement errors. The application of the proposed control charts is illustrated by a numerical example .     

An approach to controlling two dependent process steps with autocorrelated observations

The observations from the process output are always assumed independent when using a control chart to monitor a process. However, for many processes the process observations are autocorrelated. This autocorrelation can have a significant effect on the performance of the control chart. This paper considers the problem of monitoring the mean of a quality characteristic X on the first process step and the mean of a quality characteristic Y on the second process step, in which the observations X can be modeled as an AR(1) model and observations Y can be modeled as a transfer function of X since the state of the second process step is dependent on the state of the first process step. To effectively distinguish and maintain the state of the two dependent process steps, the Shewhart control chart of residual and the cause-selecting control chart are proposed. The proposed control charts’ performance is measured by the rate of alarm on the proposed charts. From numerical analysis, it shows that the performance of the proposed control charts is much better than the misused Hotelling T² control chart and the individual Shewhart X and Y control charts.     

The skipping strategy to reduce the effect of the autocorrelation on the <Emphasis Type="Italic">T</Emphasis><Superscript>2</Superscript> chart’s performance

In this article, we consider the T ² control chart for bivariate samples of size n with observations that are not only cross-correlated but also autocorrelated. The cross-covariance matrix of the sample mean vectors were derived with the assumption that the observations are described by a first-order vector autoregressive model—VAR (1). To counteract the undesired effect of autocorrelation, we build up the samples taking one item from the production line and skipping one, two, or more before selecting the next one. The skipping strategy always improves the chart’s performance, except when only one variable is affected by the assignable cause, and the observations of this variable are not autocorrelated. If only one item is skipped, the average run length (ARL) reduces in more than 30 %, on average. If two items are skipped, this number increases to 40 %.     

Process Monitoring Strategies for Surface Mount Manufacturing Processes

Establishing reliable surface mount assemblies requires robust design and assembly practices, including stringent process control schemes for achieving high yield processes and high quality solder interconnects. Conventional Shewhart-based process control charts prevalent in today's complex surface mount manufacturing processes are found to be inadequate as a result of autocorrelation, high false alarm probability, and inability to detect process deterioration. Hence, new strategies are needed to circumvent the shortcomings of traditional process control techniques. In this article, the adequacy of Shewhart models in a surface mount manufacturing environment is examined and some alternative solutions and strategies for process monitoring are discussed. For modeling solder paste deposition process data, a time series analysis based on neural network models is highly desirable for both controllability and predictability. In particular, neural networks can be trained to model the autocorrelated time series, learn historical process behavior, and forecast future process performance with low prediction errors. This forecasting ability is especially useful for early detection of solder paste deterioration, so that timely remedial actions can be taken, minimizing the impact on subsequent yields of downstream processes. As for the automated component placement process where very low fraction nonconforming frequently occurs, control-charting schemes based on cumulative counts of conforming items produced prior to detection of nonconforming items is more sensitive in flagging process deterioration. For the reflow soldering and wave-soldering processes, the use of demerit control charts is appealing as it provides not only better control when various defects with a different degree of severity are encountered, but also leads to an improved ARL performance. Illustrative examples of actual process data are presented to demonstrate these approaches.     

A neural network-based control scheme for monitoring start-up processes and short runs

Traditional control charts are commonly used as a monitoring tool in long-run processes. However, such control charts, due to the need for phase I analysis, are not suitable for start-up processes or short runs. Q control charts have been developed to help monitor start-up processes and short runs. In this article, a back propagation network is proposed for detecting a mean shift in start-up processes and short runs. In-control run length distribution of the control scheme is estimated using simulation study to provide information about the possibility of a false alarm within a specified number of observations. Performance of the proposed control scheme is assessed using different performance measures. It is shown numerically that the proposed control scheme outperforms the CUSUM of Q charts in detecting small to moderate mean shifts.     

A CUSUM scheme with variable sample sizes for monitoring process shifts

The cumulative sum scheme (CUSUM) and the adaptive control chart are two approaches to improve chart performance in detecting process shifts. A weighted loss function CUSUM scheme (WLC) is able to monitor both the mean shift and the increasing variance shift by manipulating a single chart. This paper investigates the WLC scheme with a variable sample sizes (VSS) feature. A design procedure is firstly proposed for the VSS WLC scheme. Then, the performance of the chart is compared with that of four other competitive control charts. The results show that the VSS WLC scheme is more powerful than the other charts from an overall viewpoint. More importantly, the VSS WLC scheme is simpler to design and operate. A case study in the manufacturing industry is used to illustrate the chart application. The proposed VSS WLC scheme suits the scenario where the strategy of varying sample sizes is feasible and preferable to pursue a high capability of detecting process variations.     

Fault detection of cycle-based signals using wavelet transform in FAB processes

This paper presents a wavelet multiresolution analysis based process fault detection algorithm to improve the accuracy of fault detection. Using Haar wavelet, coefficients that well reflect the process condition are selected and Hotelling’s T2 control chart that uses the selected coefficients is constructed for assessing the process condition. To enhance the overall efficiency and accuracy of fault detection, the following two steps are suggested: First, a denoising method that is based on wavelet transform and soft-thresholding. Second, coefficient selection methods that use the difference in the variance. For performance evaluation, various types of abnormal process conditions are simulated and the proposed algorithm is compared with other methodologies. Also, We apply the proposed algorithm to the industrial data of the dry etching process, which is one of the FAB processes. Our method has a better fault-detection performance for various sections and various changes in mean than other methods.     

Double sampling \overline{X} control chart for a first-order autoregressive moving average process model

In this paper, we consider the double sampling (DS) $\overline{X} $ control chart for monitoring processes in which the observations can be represented as a first-order autoregressive moving average (ARMA(1, 1)) model. The properties of the DS $\overline{X} $ control chart with the sampling intervals driven by the rational subgroup concept are studied and compared with the Shewhart chart and the variable sample size (VSS) chart, both properly modified to account for the serial correlation. Numerical results show that the correlation within subgroups has a significant impact on the properties of the charts. For processes with low to moderate correlation levels, the DS $\overline{X} $ chart is substantially more efficient in detecting process mean shifts.