An Ontology-based Comprehensive D-matrix Construction for Accurate FDD

There is need to gather information regarding servable symptoms and failure modes to modify the fault dependency matrix which can be helpful to build Accurate and efficient fault diagnosis. Dependency matrix is an organized diagnostic model to pinch the graded system-level fault diagnostic information. Organizing this information is typically based on the previously known knowledge and research. Its not enough to collect the information like symptoms, related diagnostic mechanism only once since technology is improving day by day. It is a challenging task to regularly update the D-matrix to have best result. To represent unstructured knowledge, ontology based data mining technology can be helpful which deals with grouping of unstructured data considering similarities and differences between them. Ontology is constructed which will describe the commonly observed correlation in fault diagnostic domains. By using different text mining algorithm, necessary artifacts like symptoms, failure mode and their relation with unstructured data can be discovered In this paper, various ways are mentioned to create D-matrix with the help of available data like engineering design, data source and also by using text mining techniques like document annotation, term extraction, and phrase merging. The association in faults and their causes are mapped into D-matrix by using engineering knowledge. Engineering knowledge like engine control, mode of failure, data analysis, effects after failure etc. scene development of diagnostic matrix, takes much of efforts and time as compared to text mining techniques.