Design of a fault tolerated intelligent control system for load following operation in a nuclear power plant

• Fault detection of a nuclear power plant has been investigated via parameter identification.
• A combination of extended Kalman filter and recursive least square is used for fault detection.
• An adaptive neuro-fuzzy control based on emotional learning is designed for the system.
• The proposed controller has fault tolerant property and is robust against failure.

Fault detection has always an important role in maintaining the system stability and assuring satisfactory and safe operation. In this paper a method based on system identification is used for fault detection on a nuclear reactor. The combination of Extended Kalman Filter and recursive Least Square are used to identify the system parameters. Another goal of this paper is design of a fault tolerant control system for a nuclear reactor during power change operation. The proposed controller is an adaptive neuro-fuzzy controller based on emotional learning. Performance of the controller in term of transient response and robustness against failure is very good and outstanding.