| 摘要 |
<p>PURPOSE:To allow a system itself to teach a recovery procedure while learning it and to teach the exact recovery procedure corresponding to a lot of abnormal states with a small scale data base by successively storing the respective part states of a facility at the time of occurrence of abnormality while classifying them according to the recovery procedure, and comparing those states at every abnormality occurrence. CONSTITUTION:An abnormality processing data recording part 11 records respective part state data and recovery procedure data as a pair of abnormality processing data at the time of each occurrence of abnormality. An abnormality processing data analystic part 12 divides the respective part state data corresponding to the same recovery procedure data among the plural pairs of abnormality processing data into the same groups, analyzes those data inside the respective groups, generates abnormality feature data including a common abnormal state item and representatively expressing the feature of an abnormal state belonging to that group and stores those data while pairing them with the correspondent recovery procedure data. When the abnormality occurs, a recovery procedure deciding part 13 compares those state data with the stored feature information and outputs recovery procedure data in the group with the highest degree of coincidence.</p> |
