| 摘要 |
<p>PURPOSE:To positively acquire and apply suitable diagnostic knowledge in a simple manner by linking nodes representing a phenomenon within a process with an arc representing the correlation of the nodes on the basis of the structure of devices for the process thereby to obtain an oriented graph, and constructing a network model for each arc which sets an influential factor indicating the degree of influences between the nodes. CONSTITUTION:Such an abnormality is considered that an indication LI1 of a level meter 61A is not coincident with an actual level LA1. Therefore, nodes 71, 72 which represent the actual level LA1 and indication LI1 of the level meter, and the arc 71A are set. A node 73 of a PID attribution is set for an adjuster 64 corresponding to its function. Moreover, when the actual level LA1 becomes low and a pump 66 sucks in vain, the actual flow rate FA2 is turned to zero and therefore, a node 81 of a transient attribution (MORE attribution) is set. An influential factor is set for each arc. The influential factor is distinguished between positive and negative in accordance with the influential relation between the phenomena of nodes, and moreover indicated by a continuous value of -1.0-1.0. Accordingly, a network model 70 is formed.</p> |
