| 摘要 |
PURPOSE:To automatically and quickly process also a fluctuation event whose abnormality identification is difficult by identifying the abnormality of an objective apparatus and estimating its cause by fuzzy inference based upon its abnormality history data. CONSTITUTION:This abnormal apparatus diagnostic system is provided with a history data base 5 storing the abnormal states of the apparatus 4 experienced in the past, their causes and corresponding processing, an abnormality inference part 6 for inputting abnormality history data simultaneously with the input of measuring data from a measuring system 1 for always measuring the state of the apparatus 4 and executing the identification of the abnormality of the apparatus 4 and the estimation of the cause of the abnormality by fuzzy inference and a data updating part 7 for processing a series of abnormal experiences by a neural network and updating the abnormality history data of the apparatus 4. Thus, the apparatus state is diagnosed from the measuring system information of the apparatus 4 and the past abnormality history data of the apparatus 4 without depending only upon operator's decision and the identification of the cause and the determination of its corresponding operation are executed. Even in the case of an even whose abnormality identification is difficult, its abnormality can be quickly and automatically identified. |
