| 摘要 |
<p>A system comprising a memory element (20) in which is stored, for each possible fault on the plant, a rule (21) specifying the procedure to be adopted for determining the presence of the fault. Each rule (21) is assigned a factor which takes into account the probability of the fault occurring and the cost of the relative check procedure, and the rules (21) are formed into groups in each of which they are arranged on a priority basis according to the aforementioned factor. For identifying a fault, the system acquires alarm recognition data for determining the type of alarm and identifying the part of the machine from which the alarm message or fault originates. On the basis of this information, the system addresses the relative highest priority rule (21) for determining the check procedure to be performed, and, in the event of a negative outcome of the check procedure, the next check procedure down in order of priority is addressed until the fault is identified or until no further hypotheses remain. <IMAGE></p> |
