Open AccessThe Inference System for Fault Diagnosis
Author(s) -
Yongeui Hong,
Taewoo Lee,
Kyoungdyuk Rho,
Hyun-Seung Cha,
Jonghyo Lee
Publication year - 2019
Publication title -
international journal of engineering sciences
Language(s) - English
Resource type - Journals
ISSN - 0976-6693
DOI - 10.36224/ijes.110402
Subject(s) - fault (geology) , computer science , context (archaeology) , inference , data mining , fault model , feature (linguistics) , artificial intelligence , machine learning , engineering , paleontology , linguistics , philosophy , electrical engineering , seismology , electronic circuit , biology , geology
A knowledge-based fault diagnosis system uses prior knowledge and context knowledge for prediction to improve fault diagnosis performance. The proposed representative fault diagnosis system consists of three levels. With this structure, fault diagnosis can be flexibly performed even in a complicated environment. The three-level consists of the fault diagnosis level, the learning level, and the information processing level. The Fault diagnosis level is able to express the correlation by using the data obtained from the controller and diagnose the fault by logically inferring it. The learning level links the logical language perceived by humans and the numerical data processed by the computer, and keeps it consistent with the situation. The information processing level acquires the feature value required by the higher level in the candidate region among the numerous data obtained from the controller and sends it to the higher level. The proposed algorithm can effectively diagnosis faults by using additional prior knowledge and situation data.