• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.34 no.4
• /
• pp.57-65
• /
• 2011

Monitoring auto correlated processes is prevalent in recent manufacturing environments. As a proactive control for manufacturing processes is emphasized especially in the semiconductor industry, it is natural to monitor real-time status of equipment through sensor rather than resultant output status of the processes. Equipment's sensor data show various forms of correlation features. Among them, considerable amount of sensor data, statistically autocorrelated, is well represented by Box-Jenkins autoregressive moving average (ARMA) model. In this paper, we present a design method of statistical process control (SPC) used for monitoring processes represented by the ARMA model. The proposed method shows benefits in the power of detecting process changes, and considers robustness to ARMA modeling errors simultaneously. We prove benefits through Monte carlo simulation-based investigations.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.22 no.51
• /
• pp.1-10
• /
• 1999

In recent industry society, it is revealed that, as an increase in the use of automated manufacturing and process inspection technology, the data from mass production system exhibits some degrees of autocorrelation. The operation characteristics of traditional control charts developed under the independence assumption are adversely affected by the presence of serial correlation. Therefore, when autocorrelated construction contacted with time-series models explain, the time-series models are the Box-Jenkins forecast models which have been proposed as the best forecasting tool which allows for partitioning of variation into result from the autocorrelation structure and variation due to unusual but assignable causes. In this paper, for the AR(1) process of Box-Jenkins forecast models, when the constant term ξ are zero and different from zero, I want to analyze the sensitivity of (equation omitted), CUSUM and EWMA control chart for forecast residuals.

• The Korean Journal of Applied Statistics
• /
• v.37 no.2
• /
• pp.191-207
• /
• 2024

Many studies have been conducted to quickly detect out-of-control situations in autocorrelated processes. The most traditionally used method is a residual control chart, which uses residuals calculated from a fitted time series model. However, many procedures for monitoring autocorrelated processes using statistical learning methods have recently been proposed. In this paper, we propose a monitoring procedure using the latent vector of LSTM Autoencoder, a deep learning-based unsupervised learning method. We compare the performance of this procedure with the LSTM Autoencoder procedure based on the reconstruction error, the RNN classification procedure, and the residual charting procedure through simulation studies. Simulation results show that the performance of the proposed procedure and the RNN classification procedure are similar, but the proposed procedure has the advantage of being useful in processes where sufficient out-of-control data cannot be obtained, because it does not require out-of-control data for training.

• Journal of Korean Society for Quality Management
• /
• v.36 no.4
• /
• pp.87-92
• /
• 2008

It is common to use CUSUM charts for detecting small level shifts in processes control, in which reference value(k) and decision interval(h) are the design parameters to be determined. To control process with autocorrelation, CUSUM charts could be applied to residuals obtained from fitting ARIMA models. However, constant level shifts in processes lead to varying mean shifts in residual processes and thus standard CUSUM charts may need to be modified. In this paper, we study the performance of CUSUM charts with various design parameters applied to autocorrelated processes, especially focussing on ARMA(1,1) models, and propose how they can be determined to get better performance in terms of the average run length.

• Communications for Statistical Applications and Methods
• /
• v.5 no.2
• /
• pp.373-381
• /
• 1998

In this paper a new control chart which accounts for correlation between process subgroups will be proposed. We consider the case where the process fluctuations are autocorrelated by a stationary AR(1) time series and where n($\geq1$) items are sampled from the process at each sampling time. A simulation study is presented and shows that for correlated subgroups, the proposed control chart makes a significant improvement over the traditionally employed X-bar chart which ignores subgroup correlations. Finally, we illustrate the proposed chart by comparing the standardized residuals and X-bar chart on a data set of motor shaft diameters.

• Journal of the Korean Data and Information Science Society
• /
• v.25 no.6
• /
• pp.1539-1547
• /
• 2014

In this paper we apply the autoregressive process to the nonlinear quantile regression in order to infer nonlinear quantile regression models for the autocorrelated data. We propose a kernel method for the autoregressive data which estimates the nonlinear quantile regression function by kernel machines. Artificial and real examples are provided to indicate the usefulness of the proposed method for the estimation of quantile regression function in the presence of autocorrelation between data.

• Proceedings of the Korean Society for Quality Management Conference
• /
• 2006.04a
• /
• pp.313-318
• /
• 2006

Many control charting methods for both i.i.d and autocorrelated data can be viewed as charting the output of a linear filter applied to the process data. We propose a generalization of this concept, in which the filter parameters are optimally selected to minimize the out-of-control ARL while constraining the in-control ARL to some desired value. A number of interesting characteristics of the optimal fitters are discussed.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.35 no.3
• /
• pp.52-61
• /
• 2012

The design method for cumulative sum (CUSUM) control charts, which can be robust to autoregressive moving average (ARMA) modeling errors, has not been frequently proposed so far. This is because the CUSUM statistic involves a maximum function, which is intractable in mathematical derivations, and thus any modification on the statistic can not be favorably made. We propose residual-based robust CUSUM control charts for monitoring autocorrelated processes. In order to incorporate the effects of ARMA modeling errors into the design method, we modify parameters (reference value and decision interval) of CUSUM control charts using the approximate expected variance of residuals generated in model uncertainty, rather than directly modify the form of the CUSUM statistic. The expected variance of residuals is derived using a second-order Taylor approximation and the general form is represented using the order of ARMA models with the sample size for ARMA modeling. Based on the Monte carlo simulation, we demonstrate that the proposed method can be effectively used for statistical process control (SPC) charts, which are robust to ARMA modeling errors.

• Proceedings of the Korean Society for Quality Management Conference
• /
• 2007.04a
• /
• pp.302-309
• /
• 2007

To conduct statistical process control needs the assumption that the process data are independent. However, most of chemical processes, like a semi-conduct processes do not satisfy the assumption because of autocorrelation. It causes abnormal out of control signal in the process control and misleading process capability. In this study, we introduce that Shore's method to solve the problem and to find the optimal subgroup size to estimate variance for AR(l) model. Especially, we focus on finding an actual subgroup size for small samples using simulation. It may be very useful for statistical process control to analyze process capability and to make a Shewhart chart properly.

• Journal of Korean Society for Quality Management
• /
• v.35 no.2
• /
• pp.106-112
• /
• 2007

In statistical process control, it is assumed that the process data are independent. However, most of chemical processes such as semi-conduct processes do not satisfy the assumption because of presence of autocorrelation between process data. It causes abnormal out of control signal in the process control and misleading estimation in process capability. In this study, we adopted Shore's method to solve the problem and propose an optimal subgroup size to estimate the variance correctly for AR(1) processes. Especially, we focus on finding an actual subgroup size for small samples based on simulation study.