학술논문
베이지안 네트워크를 이용한 출입문 장치의 확률론적 고장 분석에 관한 연구 / A Study on probabilistic failure analysis of a door system using Bayesian networks
Document Type
Dissertation/ Thesis
Author
Source
Subject
Language
Korean
Abstract
철도차량 내 여러 부품들은 장시간, 장거리 운행으로 품질 및 내구연한 저하가 가속화 되고 있으며, 이를 예방하기 위해 노후화된 부품에 대한 신속한 교체 혹은 신형철도차량의 도입이 순차적으로 이루어지고 있다. 이처럼 고장이 발생한 각 장치에 대한 이상적인 해결책은 모든 시스템의 모든 요소를 신품으로 교체 혹은 도입하는 것일 수 있지만, 현실은 재원 및 예비품이 제한되어 있다는 것이다. 재원 및 예비품 등 경제적 측면과 유지보수, 정비 등 효율성 측면을 고려했을 때 고장확률에 근거한 신뢰성 분석은 고장에 영향을 미치는 중요한 구성요소의 식별, 검사, 수리 및 교체와 관련된 판단으로써 중요한 역할을 한다. 따라서 본 논문에서는 철도차량시스템 중 고객안전과 밀접하게 연관된 장치인 2호선 신형철도차량 출입문장치에 대한 신뢰도 및 고장확률을 추정하였다. 우선 출입문 장치에 대한 구조분석수행 후 고장유형 및 치명도 분석(Failure Mode Effects and Criticality Analysis, FMECA)을 통해 13개의 고장모드 및 19개의 하위부품요소를 추렸다. 이 후 치명도 우선순위(Risk Priority Number, RPN) 결정을 통해 고장모드 대한 수용가능 수준의 정량적 평가를 마쳤다. 또한 고장모드의 논리적 연결 관계를 명확히 보여주기 위해 결함수분석(Fault Tree Analysis, FTA)을 하였고, 하위부품요소의 고장률을 여러 예측규격(NPRD, NSWC 등)과 필드데이터(Field Failure Data)를 활용하여 추정하였다. 최종적으로 결함수 분석과 하위요소부품의 고장률을 베이지안 네트워크(Bayesian Network)로 통합 맵핑(Mapping)하여, 상위 및 하위 부품 요소 간 종속관계에 대한 직관적인 도식화뿐만 아니라 출입문 장치라는 최종시스템의 신뢰도 및 고장확률을 추정하는 방법론을 제시하였다.
Machinery and electronic components in railway vehicles are being operated for a long time, accelerating their durability and deterioration in quality. In order to prevent such problems, replacement of aging parts or introduction of new railway vehicles are being carried out continuously. The ideal solution for each device that has such a failure may be to replace or introduce all elements of all systems, but the reality is that resources and reserves are limited. When considering economic aspects such as financial resources and reserves, and efficiency aspects such as maintenance, reliability analysis based on failure probabilities plays an important role as a judgement related to identification, inspection, repair and replacement of critical components that affect the failure. Therefore, in this paper, the reliability and probability of failure were estimated for a passenger door system of the new railway vehicle on line 2, this system closely related to customer safety among the railway vehicle systems. First of all, 13 failure modes and 19 subparts elements were selected through the Failure Mode Effects and Criticality Analysis(FMECA) after structural analysis of the door system. The quantitative assessment of acceptable levels for failure modes was then completed by determining the risk priority number(RPN). The Fault Tree Analysis(FTA) was then performed to clearly demonstrate the logical connection of failure modes, and the failure rate of the lower component elements was estimated using various prediction standards(NPRD, NSWC, etc.) and field failure data. Finally, the fault tree analysis and failure rate of the sub-element components were integrated mapping into the Bayesian Network. This paper provided a methodology to estimate the reliability and probability of failure of the system called the final door system as well as the visual and schematic of the dependencies between the upper and lower component elements.
Machinery and electronic components in railway vehicles are being operated for a long time, accelerating their durability and deterioration in quality. In order to prevent such problems, replacement of aging parts or introduction of new railway vehicles are being carried out continuously. The ideal solution for each device that has such a failure may be to replace or introduce all elements of all systems, but the reality is that resources and reserves are limited. When considering economic aspects such as financial resources and reserves, and efficiency aspects such as maintenance, reliability analysis based on failure probabilities plays an important role as a judgement related to identification, inspection, repair and replacement of critical components that affect the failure. Therefore, in this paper, the reliability and probability of failure were estimated for a passenger door system of the new railway vehicle on line 2, this system closely related to customer safety among the railway vehicle systems. First of all, 13 failure modes and 19 subparts elements were selected through the Failure Mode Effects and Criticality Analysis(FMECA) after structural analysis of the door system. The quantitative assessment of acceptable levels for failure modes was then completed by determining the risk priority number(RPN). The Fault Tree Analysis(FTA) was then performed to clearly demonstrate the logical connection of failure modes, and the failure rate of the lower component elements was estimated using various prediction standards(NPRD, NSWC, etc.) and field failure data. Finally, the fault tree analysis and failure rate of the sub-element components were integrated mapping into the Bayesian Network. This paper provided a methodology to estimate the reliability and probability of failure of the system called the final door system as well as the visual and schematic of the dependencies between the upper and lower component elements.