earticle

영어

Product safety management means the various activities to protect the safety of the consumer against the hazard arising from product accidents in the whole process of manufacture, import and distribution of goods including industrial and electrical products. In other words, it includes the implementation of the various activities including production and importation of safe products, product safety assessment, safety certification and confirmation by the safety assessment, and recall the accident and illegal products through market surveillance activities to market products. This study presents a new product risk assessment model based on bayesian network model as a tool for risk assessment needed for product safety that reflect the casual relationship affecting the product risk and shows an application example. For the risk assessment, the structure of the network is designed hierarchically to factor layer, accident layer, and injury layer and the probabilities and the conditional probabilities of each layers are derived. The factor layer is composed of the product defect factors and the use environment factors including factors such as subject of use, hours of use, and place of use. The accident and injury layer is composed of accidents caused by the factor layer and injuries caused by the accidents. Finally, the product risk is determined based on the probabilities and severities of injuries by reference to PI(probability impact) matrix. The severity of injury is obtained by calculating ISS using AIS presented in IISC. For the injury, damages to property as well as personal injuries are considered. An example applying the risk assessment model presented in this paper to the stroller is shown and the model is compared with EU RAPEX RAG and Japan’s R-Map. The two methods are representative method that is currently being presented related to the risk assessment. The EU RAPEX RAG, as a method based on the scenario, composes possible scenarios and assesses the scenarios and adopts the higher risk to the final risk. As a result, it is shown that the EU RAPEX RAG and the model presented in this paper gives the same results each other. Japan’s R-Map, on the other hand, gives one step higher result compared to the result of this paper. This can be seen because it follows the R-Map manual regulation when different types of injuries occurred together the strictest definition is adopted because R-Map does not assess separately each incident of injury. The model of this paper overcomes the subjectivity of probability of the scenario steps and infinity of the scenario creation that scenario-based assessment methods have. It also have the advantage that a more elaborate assessment is possible as the risk is assessed by deriving the probability and severity of injury by the types of injury. At present, the classification basis for safety management of the products in Korea is weak because there do not exist standardized risk assessment methods to classify the products into the categories for safety management. The model in this paper is a systematic and leading assessment model in Korea and can be utilized to review the current products managed by regulations and classify them into the new categories or derive reasonable and objective safety management levels for the new products. Also the model can be utilized to support the development of inherently safe products in the process of the development of the new products in the companies by letting themselves to assess the risk of the products. Further research should be the problems to ensure the reasonableness of the derivation process of the subjective product defect probabilities and the conditional probabilities if there does not exist data relating accidents. In addition, there may be problems reviewing the appropriateness of classification of the products by applying multi-variate cluster analysis if there exist data relating accidents and comparing it with the results of this study.

한국어

제품 안전관리는 공산품 및 전기제품 등의 제조 및 수입, 유통의 전 과정에서 제품사고로 인해 발생하는 위해로부터 소비자의 안전을 보호하는 제반 활동을 의미한다. 즉, 안전한 제품의 생산 및 수입, 제품안전성 평가, 안정성 평가결과에 의한 안전인증 및 확인, 시장 출시 제품에 대한 시장감시 활동을 통한 사고 및 불법제품의 리콜 등 다양한 활동의 수행을 포함하고 있다. 본 연구에서는 제품 안전관리에 필요한 제품 리스크 평가를 위하여 리스크에 영향을 미치는 인과관계를 반영한 새로운 평가기법으로서 베이지안네트워크 기반의 확률적 모델을 적용한 리스크 평가모델을 제시하고 적용사례를 보인다. 리스크 평가를 위하여 요인계층, 사고계층, 부상계층 등으로 인과네트워크 구조를 설계하고 각 계층확률 및 조건부확률을 도출한다. 요인계층은 제품결함요인과 사용주체, 사용시간, 사용장소 등을 포함하는 사용환경요인으로 구성한다. 사고계층과 부상계층은 요인계층에 의해 발생하는 사고와 사고에 의해 발생하는 부상으로 구성한다. 최종적으로 부상확률과 부상심각도를 바탕으로 리스크를 결정한다. 본 모델의 적용 예를 보이기 위하여 유모차 사례분석을 통해 평가모델 프로토타입을 제시하고 유럽(EU RAPEX RAG)과 일본(NITE R-Map)의 기법과 비교한다. 본 논문의 모델은 시나리오 기반의 평가기법이 지니고 있는 시나리오 생성과정의 무한성과 확률 도출과정의 주관성을 극복하고 있다. 또한 부상 유형별로 부상확률과 부상심각도를 도출하여 리스크를 평가함에 따라 더욱 정교한 평가가 가능한 장점을 가지고 있다.