| 摘要 |
<p>PURPOSE:To enable all observation values which are observed by a target to be diagnosed to be utilized effectively and at the same time a plurality of failures to be diagnosed simultaneously. CONSTITUTION:A parameter of normal state to be diagnosed and a parameter of an observation value to be diagnosed are given to a deviation detection part 3, thus enabling a qualitative deviation to be detected as a symptom. According to this detected symptom, a parallel qualitative propagation is executed by using a qualitative cause and effect model data base 5 regarding a qualitative behavior to be diagnosed which is constructed in a network, thus enabling a characteristic point to be collected. A set of the characteristic point thus collected is classified into a cluster by a clustering part 6 and at the same time the result is evaluated by an evaluation part 7 for obtaining a solution.</p> |
