• Journal of the Korean Operations Research and Management Science Society
• /
• v.36 no.1
• /
• pp.39-55
• /
• 2011

Process monitoring has been emphasized for the monitoring of complex system such as chemical processing industries to achieve the efficiency enhancement, quality management, safety improvement. Recently, ICA (Independent Component Analysis) based MSPC (Multivariate Statistical Process Control) was widely used in process monitoring approaches. Moreover, DICA (Dynamic ICA) has been introduced to consider the system dynamics. However, the existing approaches show the limitation that their performances are strongly dependent on the statistical distributions of control variables. To improve the limitation, we propose a novel approach for process monitoring by integrating DICA and LOF (Local Outlier Factor). In this paper, we aim to improve the fault detection rate with the proposed method. LOF detects local outliers by using density of surrounding space so that its performance is regardless of data distribution. Therefore, the proposed method not only can consider the system dynamics but can also assure robust performance regardless of the statistical distributions of control variables. Comparison experiments were conducted on the widely used benchmark dataset, Tennessee Eastman process (TE process), and showed the improved performance than existing approaches.

• Proceedings of the KWS Conference
• /
• 2002.10a
• /
• pp.589-594
• /
• 2002

In manufacturing of welded parts which require a proof on acceptance close observation of the process is mandatory. The manufacturer is obliged to document the welding process. An approach for monitoring and recording the process of laser beam welding combining theoretical knowledge, process monitoring and welding experience will be presented. The range of application of a device for co-axial process monitoring was extended up to 20 mm sheet thickness welded in one pass at beam powers of up to 20 kW. Crucial features of the process which refer to the formation of failures were observed. The results presented give rise to the general applicability of the approach for process monitoring and recording as well as failure management.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.14 no.6
• /
• pp.1-6
• /
• 2015

Tool condition monitoring plays one of the most important roles in the improvement of both machining quality and productivity. In this regard, various process signals and monitoring methods have been developed. However, most of the existing studies used cutting force or acoustic emission signals, which posed risks of interference with the machining system in dynamics, fixturing, and machining configuration. In this study, a feed motor current signal is used as a process signal representing process and tool states in tool breakage monitoring based on an adaptive autoregressive model and unsupervised neural network. From the experimental results using various cases of tool breakage, it is shown that the developed system can successfully detect tool breakage before two revolutions of the spindle after tool breakage.

• Proceedings of the KIEE Conference
• /
• 1998.07g
• /
• pp.2390-2392
• /
• 1998

In this paper, we discuss the development of monitoring system with data process equipment which transfers data from Remote Terminal Unit(RTU) to monitoring computer. The RTUs sense temperature, pressure and PLC(Programmable Logic Controller) nodes conditions of scrubber in semiconductor manufacture processing. The data Process equipment is connected every RTU and a monitoring computer through serial communication. This equipment receives informations from RTU, process data, and transfers them to monitoring computer. To avoid congestion in data communication, task scheduling algorithm used RT O/S(Real-Time Operating System) is embedded in ROM which is a part of data Process equipment.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.17 no.7
• /
• pp.374-379
• /
• 2016

On-line process monitoring schemes are designed to give early warnings of process faults. In the artificial intelligence and machine learning fields, reliable approaches have been utilized, such as kernel-based nonlinear techniques. This work presents a kernel-based empirical monitoring scheme with a small sample problem. The measurement data of normal operations are easy to collect, whereas special events or faults data are difficult to collect. In such situations, noise filtering techniques can be helpful in enhancing the process monitoring performance. This can be achieved by the preprocessing of raw process data and eliminating unwanted variations of data. In this work, the performance of several monitoring schemes was demonstrated using three-dimensional batch process data. The results showed that the monitoring performance was improved significantly in terms of the detection success rate.

• IE interfaces
• /
• v.18 no.3
• /
• pp.350-360
• /
• 2005

This paper considers a strategy for planning of rework in semiconductor monitoring burn-in test process. The equipment error in monitoring burn-in test process generates many defects. These defects are transformed into good products by rework process, i.e. retest. Rework has the advantage of saving production costs. But rework increases holding costs and incurs rework costs. In monitoring burn-in test process, rework depends on operator's experience with no pre-defined specification. In practice, a number of rework activities are performed with respect to the product importance and inventory quantity. Moreover, disregard for order jobs schedule have caused due date penalties. So a strategy for planning of rework by which order jobs schedule are not affected is suggested. Futhermore, production costs, rework costs and inventory costs for planning of rework are considered.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.15 no.3
• /
• pp.1-7
• /
• 2016

The cutting processes used for monitoring engineering includes analysis and feedback about strange conditions, tools collision and tools wear in real time, for improving the working ratio of equipment and productivity. In this study, we proposed monitoring using profibus to increase the reliability as the most important factor for cutting monitoring. The profibus can increase the reliability of cutting monitoring for cutting torque of a main spindle motor and a feed motors through PLC-based interface.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.102.1-102
• /
• 2001

In this work, extensions to partial least squares (PLS) for wastewater treatment (WWT) process monitoring are discussed. Conventional data gathered by monitoring WWT systems are usually time varying, high dimensional, correlated and nonlinear, PLS has been shown to be an efficient approach in modeling and monitoring high dimensional and correlated data. To represent dynamic and nonlinear features of the data several kinds of dynamic nonlinear PLS (DNLPLS) models have been proposed. However, the complexity and ambiguity of the models make them unsuitable for WWT monitoring, Recently, dynamic fuzzy PLS (DFPLS) was proposed ...

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.105.4-105
• /
• 2001

This paper describes a newly designed multipoint process monitoring system based on a NIR acousto-optic tunable filter. The NIR multipoint process monitoring system consists of a NIR AOTF device for wavelength selection, an InGaAs array sensor, and a specially designed iin-line type of optical fiber probe. Unlike a FTS(Fourier Transform Spectrometry) or grating based monitoring system, an AOTF has no moving parts, and it can be rapidly tuned to any wavelength in its operating range within microseconds. Thus, the AOTF is advantageous in terms of faster spectral imaging capability and rigidity required for industrial monitoring environment. In the current feasibility evaluation, an enhanced optical fiber probe with 3 monitoring points was used. However, ...

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.16 no.1
• /
• pp.156-161
• /
• 1992

In this research a monitoring system for turning using NC information and the current of feed motor as a monitoring signal was developed. The overall system consists of modules such as learning process, NC data transmission, generation of forecast information, signal acquisition, monitoring and post process. In the learning process, the reference data and the cutting force equation necessary for monitoring are obtained from the accumulated monitoring results. In the generation of forecast information, the information of forecasted cutting forces is acquired from the cutting force equation and NC program and appended to each NC block as a monitor code. Reliability of monitoring is improved by using the monitor code in the real-time monitoring. Monitoring module is divided into two parts : the off-line monitoring where errors of NC program are checked and the on-line monitoring where the level of motor current is monitored during cutting operations. If the actual current level exceeds the limit value provided by the monitor code in the level monitoring, it is recognized as abnormal. In the event of abnormal status, the post processor sends the emergency stop signal to NC controller to stop the operation. Actual experiments have shown that the developed monitoring system works well.