=-TINE EXF'ERT SYSTEMS FOR DIAGNOSTICS AND CLOSJD-LOOP CONTROL

The next generation of real-time expert systems for use on industrial plants will incorporate closed-loop control capabilities. This may be achieved through the explicit representation of process knowledge, such as in a production rule system or through implicit techniques such as adaptive plant parameter-based algorithms. This paper identifies the relative merits of both approaches. In order to demonstrate the strengths and limitations of both approaches, real-time trials are conducted on an industrial hydraulic drive system.