텍스트 데이터 기반 고장 지식 추출 및 재사용

During the entire product lifecycle, a large amount of design knowledge is stored in documents. Engineers can make better decision based on reuse of design knowledge in documents during the product design stage. The challenge for reuse of design knowledge from documents is to quickly gather the right information from the unstructured text in the documents and synthesize it into structured knowledge. In this paper, we proposed approaches for failure knowledge extraction and reuse based on text data in failure analysis documents. First, failure information and design information are extracted and structured using failure context and failure knowledge representation model. Second, quantitative analysis between failure modes and components is conducted. Then, rules are generated using clusters between failure modes and components. Lastly, potential failure is predicted based on current design information and pre-generated rules.