• 한국공작기계학회논문집
• /
• 제14권1호
• /
• pp.15-23
• /
• 2005

For reliability assessment for machine tools, failure mode analyses by two viewpoints were studied in this paper. First, this study developed the reliability data analysis program, which searches f3r optimal failure distribution like failure rate or MTBF(Mean Time Between Failure) using failure data and reliability test data of mechanical parts in the web. Moreover, this data analysis program saves both failure data or reliability data and their failure rate or MTBF for database establishment. Second, this paper conducted failure mode analysis through such performance tests as circular movement test and vibration testing for machine tools when reliability data is not available. A developed web-based analysis program shows correlations between failure mode and performance test result and also accumulates all the data. These kinds of data analysis programs and stored data furnish valuable information for improving the reliability of mechanical system.

• Journal of information and communication convergence engineering
• /
• 제10권1호
• /
• pp.11-20
• /
• 2012

In a marine ubiquitous sensor network (USN) system using expensive sensors in the harsh ocean environment, it is very important to discover failures and devise recovery techniques to deal with such failures. Therefore, in order to perform failure modeling, this study analyzes the USN-based real-time water quality monitoring service of the Gaduri Aqua Farms at Songdo Island of Yeosu, South Korea and devises methods of discovery and recovery of failure by classifying the types of failure into system element failure, communication failure, and data failure. In particular, to solve problems from the perspective of data, this study defines data integrity and data consistency for use in identifying data failure. This study, by identifying the exact type of failure through analysis of the cause of failure, proposes criteria for performing relevant recovery. In addition, the experiments have been made to suggest the duration as to how long the data should be stored in the gateway when such a data failure occurs.

• ETRI Journal
• /
• 제36권6호
• /
• pp.1023-1031
• /
• 2014

Failure analysis is necessary to clarify the root cause of a failure, predict the next time a failure may occur, and improve the performance and reliability of a system. However, it is not an easy task to analyze and interpret failure data, especially for complex systems. Usually, these data are represented using many attributes, and sometimes they are inconsistent and ambiguous. In this paper, we present a scalable approach for the analysis and interpretation of failure data of high-performance computing systems. The approach employs rough sets theory (RST) for this task. The application of RST to a large publicly available set of failure data highlights the main attributes responsible for the root cause of a failure. In addition, it is used to analyze other failure characteristics, such as time between failures, repair times, workload running on a failed node, and failure category. Experimental results show the scalability of the presented approach and its ability to reveal dependencies among different failure characteristics.

• 한국안전학회지
• /
• 제35권1호
• /
• pp.79-86
• /
• 2020

Failure data that has systematically managed for a long time has high reliability to an estimated volume. But since much cost and effort are needed to secure reliability data, data from overseas country is used in quantitative risk analysis in many workplaces. Reliability of the data that can be collected in workplaces can be dropped because of insufficient sample or lack of observation time. Therefore, estimated data is difficult to use as it is and environment and characteristic of the workplace cannot be reflected by using data from overseas country. So this study used Bayesian method that can be used reflecting both reliability data from overseas country and workplace failure data that has less samples. As a setting toward difficult situation that securing sufficient failure data cannot be achieved, we composed workplace failure data equivalent to mass observation time 20%(t=17000), 40%(t=24000), 60%(t=31000), 80%(t=38000) and IEEE data by using Bayesian method.

• 한국수자원학회논문집
• /
• 제40권1호
• /
• pp.39-50
• /
• 2007

본 논문에서는 대수-선형 파손율 모형(log-linear ROCOF)과 와이블 파솔율 모형(Weibull ROCOF)을 이용하여 상수도 주철 배수관로의 파손율을 모형화하고, '수정된 시간 척도'를 이용하여 최적교체시기를 산정할 수 있는 방법이 개발되었다. 두 ROCOF의 모형화를 위하여 개별 관로의 파손시간을 기록한 '파손 시간자료(failure-time data)'와 일정 시간간격 사이에서 발생하는 파손횟수를 기록한 '파손 횟수자료(failure-number data)'를 이용하였고, 최대로그우도 추정값을 이용하여 두 ROCOF의 각 파손자료 유형에 대한 모형화 수행 능력을 검증하였다. 또한 두 ROCOF를 이용한 관로의 최적교체시기 방정식은 ROCOF의 매개변수 추정에 있어서 수렴성을 보장하기 위하여 '수정된 시간 척도'를 적용하여 유도하였다. 연구대상 주철 배수 관로들의 '파손 시간자료'와 '파손 횟수자료'에 두 파손율 모형을 적용시켜 본 결과 파손 시간자료를 이용할 경우 대수-선형 ROCOF가 와이블 ROCOF 보다 적합한 모형인 것으로 나타났다. 또한 두 모형 모두 '파손 시간자료'를 이용하는 것이 '파손 횟수자료'를 이용하는 것보다 모형화 수행 능력이 높아지는 것으로 나타나서, 분석에 사용된 관로의 파손율 모형화와 최적교체시기 산정을 위해서는 일정 시간간격 동안의 관로 파손횟수를 기록하는 것보다 관로의 파손시간을 기록하는 것이 더욱 우수한 모형화 결과를 낳는 것으로 나타났다.

• 산업경영시스템학회지
• /
• 제43권2호
• /
• pp.33-38
• /
• 2020

Anomaly detection of Machine Learning such as PCA anomaly detection and CNN image classification has been focused on cross-sectional data. In this paper, two approaches has been suggested to apply ML techniques for identifying the failure time of big time series data. PCA anomaly detection to identify time rows as normal or abnormal was suggested by converting subjects identification problem to time domain. CNN image classification was suggested to identify the failure time by re-structuring of time series data, which computed the correlation matrix of one minute data and converted to tiff image format. Also, LASSO, one of feature selection methods, was applied to select the most affecting variables which could identify the failure status. For the empirical study, time series data was collected in seconds from a power generator of 214 components for 25 minutes including 20 minutes before the failure time. The failure time was predicted and detected 9 minutes 17 seconds before the failure time by PCA anomaly detection, but was not detected by the combination of LASSO and PCA because the target variable was binary variable which was assigned on the base of the failure time. CNN image classification with the train data of 10 normal status image and 5 failure status images detected just one minute before.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2001년도 춘계학술대회 논문집
• /
• pp.88-91
• /
• 2001

The reliability data analysis for components of CNC machining center is studied in this paper. The failure data of mechanical part is analyzed by Exponetial, Weibull, and Log-normal distributions. And then, the optimum failure distribution model is selected by goodness of fit test. The reliability data analysis program is developed using ASP language. The failure rate, MTBF, life, and failure mode of mechanical parts are estimated and searched by this program. The failure data and analysis results are stored in the database.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제4권1호
• /
• pp.9-18
• /
• 2015

소프트웨어 신뢰도 성장 모델은 고장 데이터를 사용해서 소프트웨어 출시일 또는 추가 테스트 노력을 결정하는 데 사용된다. 소프트웨어 신뢰도 성장 모델을 사용할 때 특정 소프트웨어 신뢰도 성장 모델을 모든 소프트웨어에 사용할 수 없는 문제가 있다. 또한 신뢰도를 평가하기 위해 이미 많은 수의 소프트웨어 신뢰도 성장 모델이 제안되었다. 따라서 특정 조건에 맞는 최적의 소프트웨어 신뢰도 성장 모델을 선택하는 것은 중요한 이슈가 되었다. 기존 소프트웨어 신뢰도 성장 모델 선택 방법에서는 수집된 고장 데이터 전체를 사용하고 있다. 그런데 초기에 수집된 고장 데이터는 미래 고장 예측에 영향을 주지 않을 수도 있고 경우에 따라서는 미래 고장 예측 과정에서 왜곡된 결과를 초래할 수도 있다. 이를 해결하기 위해서 이 논문에서는 부분 고장 데이터를 이용하여 적합도 평가를 수행하는 방법에 기반을 둔 소프트웨어 신뢰도 성장 모델 선택 방법을 제안한다. 이 논문에서는 고장 데이터에서 과도하게 불안정한 데이터를 제외한 부분 데이터를 사용한다. 소프트웨어 신뢰도 성장 모델 선택에 사용될 부분 데이터는 전체 고장 데이터와 고장 데이터의 일부를 제외한 부분 고장 데이터의 미래 고장 예측 능력의 비교를 통해서 찾는다. 연구의 타당성을 보이기 위하여 실제 수집된 고장 데이터를 사용해서 전체 데이터를 적용한 경우보다 부분 데이터를 사용한 경우의 미래 고장 예측 능력이 더 정확함을 보인다.

• 산업경영시스템학회지
• /
• 제41권2호
• /
• pp.84-94
• /
• 2018

In recent years, the diminishing of operation and maintenance cost using advanced maintenance technology is attracting many companies' attention. Especially, the heavy machinery industry regards it as a crucial problem since a failure of heavy machinery requires high cost and long downtime. To improve the current maintenance process, the heavy machinery industry tries to develop a methodology to predict failure in advance and to find its causes using usage data. A better analysis of failure causes requires more data so that various kinds of sensor are attached to machines and abundant amount of product usage data is collected through the sensor network. However, the systemic analysis of the collected product usage data is still in its infant stage. Many previous works have focused on failure occurrence as statistical data for reliability analysis. There have been less works to apply product usage data into root cause analysis of product failure. The product usage data collected while failures occur should be considered failure cause analysis. To do this, this study proposes a methodology to apply product usage data into failure cause analysis. The proposed methodology in this study is composed of several steps to transform product usage into failure causes. Various statistical analysis combined with product usage data such as multinomial logistic regression, T-test, and so on are used for the root cause analysis. The proposed methodology is applied to field data coming from operated locomotive and the analysis result shows its effectiveness.

• 한국신뢰성학회지:신뢰성응용연구
• /
• 제17권4호
• /
• pp.305-315
• /
• 2017

Purpose: The purpose of this research is to propose procedure and methodology for developing failure rate databook which is suitable for Korean operation environment. Methods: To this end, we investigate failure databooks used in foreign countries and study the procedure and methodology for collecting failure data, organizing the data, estimating failure rate and summarizing results. Results: We develop the procedure of development of failure databook, the items for data collection, database schema of part details and part summary and contents of failure databook by considering the application environment in Korea. Conclusion: The results of our research could be utilized for the development of Korean failure rate databook and research of reliability prediction model and could ultimately contribute to improve the accuracy of reliability prediction.