• 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.

• Journal of the Korean Data and Information Science Society
• /
• v.20 no.5
• /
• pp.933-939
• /
• 2009

The bivariate data in clinical research fields often has two types of failure times, which are mark variable for the first failure time and the final failure time. This paper showed how to generate bootstrap data to get Bayesian estimation for the joint distribution of bivariate survival times. The observed data was generated by Frank's family and the fake date is simulated with the Gamma prior of survival time. The bootstrap data was obtained by combining the mimic data with the observed data and the simulated fake data from the observed data.

• Journal of Biosystems Engineering
• /
• v.28 no.6
• /
• pp.537-544
• /
• 2003

The objectives of this study were to investigate the failure characteristics of a total of 90 parts of tractor driveline, and to predict their average annual demands required to perform the after-sales service. The failure characteristics such as failure mode, mean time between failures, characteristic life and reliability were analyzed using the data collected through the experienced mechanics at the part centers of the tractor manufacturers. The analysis was based on the assumption that the failure distribution follows the Weibull distribution. The average annual demands were also predicted for the replacement parts using the mean time between failures and the renewal theory based on the Weibull distribution. The results of the study revealed that the driveline parts failure was mostly from wearout and their average characteristic life is about 1.760 hours. The estimated mean time between failures was in a range of 670∼3,740 hours and reliability in a range of 40∼60％. The annual replacement demands were in a range of 4∼45 for a service of 100 tractors.

• Proceedings of the Korean Nuclear Society Conference
• /
• 1996.05b
• /
• pp.691-698
• /
• 1996

This paper propose a procedure to estimate the system lifetime distribution using simulation method in a parametric framework and also develop the criterion for terminating the simulation. We assume that a system is composed of many components whose lifetime and repair time distributions are general, and repair of each component is imperfect or not. General simulation algorithms can not be adopted for this case, due to the dependency of successive operating times and the discontinuity in base line intensity function of failure process. Then we propose algorithms for generating failure times subject to imperfect repair. We develop the event time tracking logic for identifying the system failure time, and also develop the criterion for terminating the simulation. Our procedure is composed of two phases. The first phase of the procedure is to generate the system failure times from the inputs. The second phase is to estimate the lifetime distribution of the system. The best model is selected by a fully automated procedure among well-known parametric families, and the required parameters are estimated. We give examples to show the accuracy of our procedure and the effect of repair effect of components to system MTTF(Mean Time To Failure).

• IE interfaces
• /
• v.16 no.1
• /
• pp.27-33
• /
• 2003

The maintenance cost in K Steelworks has been continuously increased in proportion to the production cost. However, there seems to be a possibility of reducing cost through the optimization of maintenance actions. The failure types of the equipment in steelworks ate various with different failure cost. Thus the failure rate and cost of each type of failures should be considered simultaneously when the optimum maintenance period is to be determined. It is considered that the equipment undergoes periodic replacement and a specified number of incomplete preventive maintenance actions are performed during a replacement period. Assuming that the time to failure follows a Weibull distribution, the parameters of the failure rate are estimated using the maximum likelihood estimation. The optimal replacement period is determined to minimize the average cost per unit time. As the result of analysis it is suggested that the existing maintenance period for a hot-rolling equipment can be extended significantly.

• International Journal of Control, Automation, and Systems
• /
• v.2 no.1
• /
• pp.1-14
• /
• 2004

This paper surveys some existing direct adaptive feedback control schemes for linear time-invariant systems with actuator failures characterized by the failure pattern that some inputs are stuck at some unknown fixed or varying values at unknown time instants, and applications of those schemes to aircraft flight control system models. Controller structures, plant-model matching conditions, and adaptive laws to update controller parameters are investigated for the following cases for continuous-time systems: state tracking using state feed-back, output tracking using state feedback, and output tracking using output feedback. In addition, a discrete-time output tracking design using output feedback is presented. Robustness of this design with respect to unmodeled dynamics and disturbances is addressed using a modified robust adaptive law.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.42 no.4
• /
• pp.194-202
• /
• 2019

Reliability analysis of the components frequently starts with the data that manufacturer provides. If enough failure data are collected from the field operations, the reliability should be recomputed and updated on the basis of the field failure data. However, when the failure time record for a component contains only a few observations, all statistical methodologies are limited. In this case, where the failure records for multiple number of identical components are available, a valid alternative is combining all the data from each component into one data set with enough sample size and utilizing the useful information in the censored data. The ROK Navy has been operating multiple Patrol Killer Guided missiles (PKGs) for several years. The Korea Multi-Function Control Console (KMFCC) is one of key components in PKG combat system. The maintenance record for the KMFCC contains less than ten failure observations and a censored datum. This paper proposes a Bayesian approach with a Dirichlet mixture model to estimate failure time density for KMFCC. Trends test for each component record indicated that null hypothesis, that failure occurrence is renewal process, is not rejected. Since the KMFCCs have been functioning under different operating environment, the failure time distribution may be a composition of a number of unknown distributions, i.e. a mixture distribution, rather than a single distribution. The Dirichlet mixture model was coded as probabilistic programming in Python using PyMC3. Then Markov Chain Monte Carlo (MCMC) sampling technique employed in PyMC3 probabilistically estimated the parameters' posterior distribution through the Dirichlet mixture model. The simulation results revealed that the mixture models provide superior fits to the combined data set over single models.

• The Transactions of the Korean Institute of Electrical Engineers P
• /
• v.59 no.2
• /
• pp.154-157
• /
• 2010

In this paper, the Time-varying Failure Rates(TFR) of power distribution system components are extracted from the recorded failure data of KEPCO(Korea Electric Power Corporation) and the reliability of power distribution system is evaluated using Mean Failure Rate(MFR) and TFR. The TFR is approximated to bathtub curve using the exponential and Weibull distribution function. In addition, Kaplan-Meier estimation is applied to TFR extraction because of incomplete failure data of KEPCO. Also the reliability of the real power distribution system of Korea is evaluated using the MFR and TFR extracted from real failure data, respectively and the results of each case are compared with each other. As a result, it is proved that the reliability evaluation using the TFR is more realistic than MFR. In addition, it is presented that the application method at power distribution system maintenance and repair using the result of TFR.

• Journal of Korean Society for Quality Management
• /
• v.44 no.2
• /
• pp.373-388
• /
• 2016

Purpose: This paper suggests a hierarchical time delay model to evaluate failure risks in FMEA(failure modes and effects analysis). In place of the conventional RPN(risk priority number), a more reasonable and objective risk metric is proposed under hierarchical failure cause structure considering time delay between a failure mode and its causes. Methods: The structure of failure modes and their corresponding causes are analyzed together with the time gaps between occurrences of causes and failures. Assuming the severity of a failure depends on the length of the delayed time for corrective action, a severity model is developed. Using the expected severity, a risk priority metric is defined. Results: For linear and quadratic types of severity, nice forms of expected severity are derived and a meaningful metric for risk evaluation is defined. Conclusion: The suggested REM(risk evaluation metric) provides a more reasonable and objective risk measure than the conventional RPN for FMEA.

• Transactions of The Korea Fluid Power Systems Society
• /
• v.3 no.3
• /
• pp.8-13
• /
• 2006

It is very important for the manufacturers to predict the life of hydraulic system components according to the results of life tests. Since it takes too much time to test the hydraulic system components until failure, the no-failure test method is applied for the life test of them. If the shape parameter of Weibull distribution, the number of samples, the confidence level, and the assurance life are given, the no-failure test times of hydraulic system components can be calculated by given equation. Here, the procedures to obtain the no-failure test times of the hydraulic system components such as hydraulic motors and pumps, hydraulic cylinders, hydraulic valves, hydraulic accumulators, hydraulic hoses, and hydraulic filters are described briefly.