• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.19 no.11
• /
• pp.94-101
• /
• 2020

In the age of industry 4.0, artificial intelligence is being widely used to realize machinery condition monitoring. Due to their excellent performance and the ability to handle large volumes of data, machine learning techniques have been applied to realize the fault diagnosis of different equipment. In this study, we performed the failure mode effect analysis (FMEA) of an aluminum electrolytic capacitor by using deep learning and big data. Several tests were performed to identify the main failure mode of the aluminum electrolytic capacitor, and it was noted that the capacitance reduced significantly over time due to overheating. To reflect the capacitance degradation behavior over time, we employed the Vanilla long short-term memory (LSTM) neural network architecture. The LSTM neural network has been demonstrated to achieve excellent long-term predictions. The prediction results and metrics of the LSTM and Vanilla LSTM models were examined and compared. The Vanilla LSTM outperformed the conventional LSTM in terms of the computational resources and time required to predict the capacitance degradation.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.21 no.9
• /
• pp.1531-1540
• /
• 1997

Mordern industrial society pursues unmanned system and automation of manufacturing process. Abreast with this tendensy, production of goods which requires advanced accuracy is increasing as well. According to this, the work sensing time of dressing by monitoring and diagnosing the condition of grinding, which is th representative way in accurate manufacturing, is an important work to prevent serious damages which affect grinding process or products by wearing grinding wheel. Computer vision system was composed, so that grinding wheel surface was acquired by CCD camera and the change of cutting edge ratio was measured. Then we used automatic thresholding technique from histogram as a way of dividing grinding cutting edge from grinding surface. As a result, we are trying to approach unmanned system and automation by deciding more accurate time of dressing and by visualizing behavior of grinding wheel by making use of computer vision.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.19 no.9
• /
• pp.47-52
• /
• 2020

With the 4th industrial revolution, condition monitoring using machine learning techniques has become popular among researchers. An overload due to complex operations causes several irregularities in MOSFETs. This study investigated the acquired voltage to analyze the overcurrent effects on MOSFETs using a failure mode effect analysis (FMEA). The results indicated that the voltage pattern changes greatly when the current is beyond the threshold value. Several features were extracted from the collected voltage signals that indicate the health state of a switched-mode power supply (SMPS). Then, the data were reduced to a smaller sample space by using a principal component analysis (PCA). A robust machine learning algorithm, the support vector machine (SVM), was used to classify different health states of an SMPS, and the classification results are presented for different parameters. An SVM approach assisted by a PCA algorithm provides a strong fault diagnosis framework for an SMPS.

• Journal of the Korean Society for Precision Engineering
• /
• v.33 no.6
• /
• pp.453-458
• /
• 2016

CFRP (Carbon Fiber Reinforced Plastic) and CFRP-metal stacks have recently been widely used in the aerospace and automobile industries. When CFRP is machined by a brittle fracture mechanism, defect generation behaviors are different from those associated with metal cutting. The machining quality is strongly dependent on the properties of CFRP materials. Therefore, process control for CFRP machining is necessary to minimize the defects of differently manufactured CFRPs. In this study, defects in drilling of CFRP substrates with a variety of fiber directions and resin types are compared with respect to thrust force. An experimental study on material interface detection is carried out to investigate its benefits in process control.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.10 no.4
• /
• pp.1-7
• /
• 2011

The mode analysis of end-milling was introduced using recursive parametric modeling. Also, a numerical mode analysis of FRF in end-milling at different conditions was performed systematically. In this regard, a REIVM(recursive extended instrumental variable method) modeling algorithm was adopted and natural modes of real and imaginary part were discussed. This recursive approach can be used for the on-line system identification and monitoring of an end-milling for this purpose. For acquiring a cutting force, an experimental practice was performed. And these end-milling forces were used for the calculation of FRF(Frequency response function) and its mode analysis. Also, the FRF was analysed for the prediction of end-milling system. As a results, this algorithm was successful in each condition for the detection of natural modes of end-milling. After numerical analysis of the FRF, the tooth passing frequency was discriminated in their FRF, power spectrum and mode calculation.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.21 no.1
• /
• pp.41-48
• /
• 2022

Vehicle vibration was introduced in the time and frequency domains using fast Fourier transform (FFT) analysis. In particular, a vibration mode analysis and characteristics of the frequency response function (FRF) in a sport utility vehicle (SUV) passing over a bump barrier at different speeds was performed systematically. The response behavior of the theoretical acceleration was obtained using a numerical method applied to the forced vibration model. The amplitude and frequency of the external force on the vehicle cause various power spectra with individual intrinsic system frequencies. In this regard, several modes of power spectra were acquired from the spectra and are discussed in this paper. The proposed technique can be used for monitoring the acceleration in a vehicle passing over a bump barrier. To acquire acceleration signals, various experimental runs were performed using the SUV. These acceleration signals were then used to acquire the FRF and to conduct mode analysis. The vehicle characteristics according to the vehicle condition were analyzed using FRF. In addition, the vehicle structural system and bump passing frequencies were discriminated based on their power spectra and other FRF spectra.

• Proceedings of the Korean Society of Machine Tool Engineers Conference
• /
• 1997.10a
• /
• pp.59-64
• /
• 1997

This paper investigates the relationship between cutting conditions and Acoustic Emission(AE) signals; AEavg, AErms, AEmode, as the base working to monitor the tool wear with in-process. For this purpose, cutting tests were conducted on a CNC lathe with comprehensive cutting conditions.. It is known that AEavg and AErms are proportionaly increased as the increasing of cutting velocity and depth of cut respectively. It is also known that AEmode among three kinds of AE signals may be applied for in-process monitoring to make the self diagnosis system because of its stability to the variation of cutting condition.

• Journal of the Institute of Convergence Signal Processing
• /
• v.23 no.2
• /
• pp.84-90
• /
• 2022

Machine tool state monitoring is a process that automatically detects the states of machine. In the manufacturing process, the efficiency of machining and the quality of the product are affected by the condition of the tool. Wear and broken tools can cause more serious problems in process performance and lower product quality. Therefore, it is necessary to develop a system to prevent tool wear and damage during the process so that the tool can be replaced in a timely manner. This paper proposes a method for diagnosing five tool states using a deep learning-based hierarchical convolutional neural network to change tools at the right time. The one-dimensional acoustic signal generated when the machine cuts the workpiece is converted into a frequency-based power spectral density two-dimensional image and use as an input for a convolutional neural network. The learning model diagnoses five tool states through three hierarchical steps. The proposed method showed high accuracy compared to the conventional method. In addition, it will be able to be utilized in a smart factory fault diagnosis system that can monitor various machine tools through real-time connecting.

• Analytical Science and Technology
• /
• v.27 no.5
• /
• pp.243-247
• /
• 2014

Formaldehyde is defined as carcinogen causing leukaemia, lymphoma or nasopharyngeal carcinoma at high level of exposure. Furniture-manufacturing workers can be exposed to formaldehyde, which implies serious impact on health of the workers. The authors carried out ambient monitoring of formaldehyde in the field, and identified the source of formaldehyde generated during the working process by testing the condition in the laboratory settings. After sampling formaldehyde in the air with 2,4-DNPH (2,4-dinitrophenylhydrazine) coated silica gel, we extracted formaldehyde derivative with acetonitrile and analyzed the extract using HPLC with UV detector at 360 nm. Formaldehyde was separated by ACQUITY UPLC BEH $C_{18}$ column at a flow rate of 0.5 mL/min using 45% acetonitrile as mobile phase. The workers were exposed to higher level of formaldehyde than normal air. Formaldehyde up to 0.31 ppm was detected in the process of veneer attachment, which exceeded 0.3 ppm, the ceiling value of ACGIH standard. The laboratory test of measuring formaldehyde generated from the glue and veneer used in the attachment process resulted in more formaldehyde generation as the temperature increased, and more from the veneer. Heating the veneer to $100-150^{\circ}C$ following the real condition of the manufacturing site generated 1.14-2.70 ppm of formaldehyde from the sample, which was 2-5 times higher level than Korean limit of exposure (0.5 ppm). As the workers handling and processing the veneer which was produced by wet process had high possibility to be exposed to formaldehyde, urgent improvement and management of working environment of furniture manufacturer is demanded.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.10 no.6
• /
• pp.128-133
• /
• 2011

In this study, the steel material for shipbuilding(LR-A class) was used, and FCAW was taken advantage of 3G attitude and they are welded by different welding ways. As a result of analyzing wave with welding monitoring system, the stable values are obtained which are the first floor(electronic current 164~182 A, voltage 24 V), the second floor(electronic current 174~190 A, voltage 22~25 V), the third floor(electronic current 158~188 A, voltage 22~25 V), and fourth floor(electronic current 172~184 A, voltage 22~25 V), at this time, the stable wave standard deviation and changing coefficient could be obtained. When the welding testing through nondestructive inspection was analyzed know defect of welding, there was no defect of welding in A, D, E, but some porosities in B, and slag conclusion near the surface in C, because the length of arc was not accurate, and the electronic current and voltage was not stable. After observing the change of heat affect zone through micro testing, each organization of floor formed as Grain Refinement, so welding part was fine, the distance of heat affect zone is getting wider up to change the values of the electronic current and voltage. As a result of degree of hardness testing, the hardness orders were the heat affect zone(HAZ), Welding Zone(WZ), and Base Metal(BM). When the distribution of degree of hardness is observed. B is the highest degree of hardness The reason why heat effect zone is higher than welding zone and base metal, welding zone is boiled over melting point($1539^{\circ}C$) and it starts to melt after the result of analysis through metal microscope, so we can know that delicate tissue is created at the welding zone. Therefore, in order to get the optimal conditions of the welding, the proper current of the welding and voltage is needed. Furthermore the precise work of welding is required.