Monitoring process control systems

Recently, there has been considerable interest in the applications of artificial intelligence to process control. To date, most work has been focused on how to enhance the process control loop itself or how to interact directly with the process and the control system. For near term applications, we claim that artificial intelligence should be used purely in a monitoring role. Rather than being involved in the active control of the system, artificial intelligence should be used to identify and assist with rapidly correcting any problems which cause the process control system to fail. We also claim that it is too early to be concerned with real time issues. Rather, the emphasis should be on on-line systems, with particular emphasis placed on aids in knowledge base development and trending capabilities to provide early prediction of failure. We have developed a system called Annie that embodies these principles into a working process monitoring system.