• Journal of Applied Reliability
• /
• v.18 no.1
• /
• pp.20-32
• /
• 2018

Purpose: The purpose of this study is to set up an anomaly detection criteria for sensor data coming from a motorcycle. Methods: Five sensor values for accelerator pedal, engine rpm, transmission rpm, gear and speed are obtained every 0.02 second from a motorcycle. Exploratory data analysis is used to find any pattern in the data. Traditional process control methods such as X control chart and time series models are fitted to find any anomaly behavior in the data. Finally unsupervised learning algorithm such as k-means clustering is used to find any anomaly spot in the sensor data. Results: According to exploratory data analysis, the distribution of accelerator pedal sensor values is very much skewed to the left. The motorcycle seemed to have been driven in a city at speed less than 45 kilometers per hour. Traditional process control charts such as X control chart fail due to severe autocorrelation in each sensor data. However, ARIMA model found three abnormal points where they are beyond 2 sigma limits in the control chart. We applied a copula based Markov chain to perform statistical process control for correlated observations. Copula based Markov model found anomaly behavior in the similar places as ARIMA model. In an unsupervised learning algorithm, large sensor values get subdivided into two, three, and four disjoint regions. So extreme sensor values are the ones that need to be tracked down for any sign of anomaly behavior in the sensor values. Conclusion: Exploratory data analysis is useful to find any pattern in the sensor data. Process control chart using ARIMA and Joe's copula based Markov model also give warnings near similar places in the data. Unsupervised learning algorithm shows us that the extreme sensor values are the ones that need to be tracked down for any sign of anomaly behavior.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.20 no.42
• /
• pp.73-85
• /
• 1997

In the product liability age the demand on quality is extremely high and inspection and test are automated. The process capability indices $C_p, {\;}C_{pk}$ and p control chart widely used to provide unitless measure of process performance and process control. Traditional process capability indices $C_p, {\;}C_{pk}$ do not represent the process variation from target value. The convention p chart for control of fraction nonconforming becomes inadequate when the fraction nonconforming becomes very small such as PPM level production system. This paper proposes process performance measure considering quadratic loss function and cumulative counts control chart for control of PPM level production system.

• Journal of Korean Institute of Industrial Engineers
• /
• v.27 no.3
• /
• pp.289-296
• /
• 2001

This paper proposes multivariate control chart for autocorrelated data which are common in chemical and process industries and lead to increase in the number of false alarms when conventional control charts are applied. The effect of autocorrelated data is modeled as a vector autoregressive process, and canonical analysis is used to reduce the dimensionality of the data set and find the canonical variables that explain as much of the data variation as possible. Charting statistics are constructed based on the residual vectors from the canonical variables which are uncorrelated over time, and therefore the control charts for these statistics can attenuate the autocorrelation in the process data. The charting procedures are illustrated with a numerical example and Monte Carlo simulation is conducted to investigate the performances of the proposed control charts.

• Journal of Korean Society for Quality Management
• /
• v.25 no.1
• /
• pp.100-115
• /
• 1997

Shewhart control chart is a basic technique to monitor the state of a process. We observe samples of size four or five and plot some statistic(e.g., mean or range) of each sample on the chart. When setting up the chart, we need to obtain u, pp.r and lower control limits. It is common practice that those limits are calculated from the preliminary 20-40 samples presumed to be homogeneous. However, it may ha, pp.n in practice that the samples are contaminated by outlying observations caused by various reasons. The presence of outlying observations make the control limits wider and hence decrease the sensitivity of the charts. In this paper, we introduce robust control charts with tighter control limits when outlying observations are present in the preliminary samples. Examples will be given via simulation study.

• Journal of the Korean Data and Information Science Society
• /
• v.15 no.2
• /
• pp.285-295
• /
• 2004

Statistical process control(SPC) and engineering process control(EPC) are two complementary strategies for quality improvement. An integrated process control(IPC) can use EPC to reduce the effect of predictable quality variations and SPC to monitor the process for detection of special causes. In this paper we assume an IMA(1,1) model as a disturbance process and an occurrence of a level shift in the process, and we consider the economic performance for applying an EWMA chart to monitor MMSE-controlled processes. The numerical results suggest that the IPC scheme in an IMA(1,1) disturbance model does not give additional advantages in the economic aspect.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.21 no.45
• /
• pp.365-374
• /
• 1998

Control Chart is a powerful tool to detect that process is in control or out of control. CIM can have real effect when CIM involve automated quality control. A neural network approach is used for unnatural pattern detecting of control chart. The previous moving window method uses all unnatural pattern that is detected as moving time window. Therefore, It trains a large number of unnatural pattern and takes training time long. In this paper, the proposed method tests a small number of training unnatural pattern which modifies test data without repeating time. We shows that the proposed method has differences In training time and identification rate on the previous moving windows method. As results, we reduced training time and obtain the same identification rate.

• Convergence Security Journal
• /
• v.8 no.3
• /
• pp.33-39
• /
• 2008

Software failure time presented in the literature exhibit either constant, monotonic increasing or monotonic decreasing. For data analysis of software reliability model, data scale tools of trend analysis are developed. The methods of trend analysis are arithmetic mean test and Laplace trend test. Trend analysis only offer information of outline content. In this paper, we discuss exponentially weighted moving average chart, in measuring failure time. In control, exponentially weighted moving average chart's uses are efficiency case of analysis with knowing information, Using real software failure time, we are proposed to use exponentially weighted moving average chart and comparative analysis of software failure time.

• Journal of the Korean Data and Information Science Society
• /
• v.28 no.2
• /
• pp.339-348
• /
• 2017

One of the major issues of statistical process control for variables data is monitoring both the mean and the standard deviation. The traditional approach to monitor these parameters is to simultaneously use two seperate control charts. However there have been some works on developing a single chart using a single plotting statistic for joint monitoring, and it is claimed that they are simpler and may be more appealing than the traditonal one from a practical point of view. When using these control charts for variables data, estimating in-control parameters and checking the normality assumption are the very important step. Nonparametric Shewhart-Lepage chart, proposed by Mukherjee and Chakraborti (2012), is an attractive option, because this chart uses only a single control statistic, and does not require the in-control parameters and the underlying continuous distribution. In this paper, we introduce the Shewhart-Lepage chart, and propose the design procedure to find the optimal diagnosis limits when the location and the scale parameters change simultaneously. We also compare the efficiency of the proposed method with that of Mukherjee and Chakraborti (2012).

• Journal of Korea Society of Industrial Information Systems
• /
• v.6 no.3
• /
• pp.65-70
• /
• 2001

This Paper Propose the control chart Pattern to provide a more comprehensive scheme for detecting process shifts using individual observations in start-up process. In this paper, which uses the backpropagation algorithm two samples are fed into the trained neural network to provide outputs ranging from 0 to 1. The main advantage of using neural networks approach with a control chart is that the neural network has almost no delay in detecting small shift. This paper illustrates how neural networks can provide a useful method for optimizing parameter(connection weights) that affect process control. Simulation results show that the performance of the proposed control chart using the neural network (NNCC) is quite promising.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.12
• /
• pp.187-194
• /
• 2020

In today's process-oriented industries, such as semiconductor and petrochemical processes, autocorrelation exists between observed data. As a management method for the process where autocorrelation exists, a method of using the observations is to construct a batch so that the batch mean approaches to independence, or to apply the EWMA (Exponentially Weighted Moving Average) statistic of the observed value to the EWMA control chart. In this paper, we propose a method to determine the batch size of UBM (Unweighted Batch Mean), which is commonly used as a management method for observations, and a method to determine the optimal batch size based on ARL (Average Run Length) We propose a method to estimate the standard deviation of the process. We propose an improved control chart for processes in which autocorrelation exists.