In this work, an Automated Operating Procedure System (ATOPS) is developed. ATOPS is an automation system to operate a nuclear power plant (NPP) by monitoring signals, diagnosing statuses, and generating control actions according to corresponding operating procedures, without any human operator’s help.
ATOPS has two types of functions: Anomaly detection functions and procedure execution functions. The anomaly detection functions include plant status monitoring and detecting anomaly status when alarms occur. The procedure execution function is, as it were, an automated operating procedure. If an abnormal status is detected, this function starts operating the plant automatically according to the corresponding operating procedures. This work is on the first step yet, so ATOPS is implemented about the procedure execution function.
The procedure execution function consists of two parts: a procedure database and an inference engine. First, text operating procedures are converted to a procedure database using Fuzzy Colored Petri Nets (FCPN). FCPN has more expressive power than other Petri Nets (PN). In order to make procedure database, we need not only powerful expressive methods but also various methods which can handle vague or ambiguous information. Current operating procedures still have vague and ambiguous expressions which are difficult to implement in a computer-based software. To cope with this problem, the FCPN in this work has four types of place according to the complexity of the tasks which operators should do. The first type is the normal place which does nothing or only has a jump action. The second type is the basic place which has simple actions to need no inference. The third type is the simple place which has actions including ambiguous information. The last type is the complex place which has actions including information needs supplementary data. The ATOPS needs some methods to implement the information of third and last type places, so each ...