Development of fault detection and diagnosis for reactor cooling system by using artificial intelligent techniques

• Main Authors: Abdul Rahman, R. Z., Syafiee Anuar, M. A., Mohd Aziz, M. A. F., Che Soh, A., Mohd Noor, S. B., Abdul Karim, J.
• Format: Article
• Published: International Advance Journal of Engineering Research 2023
• Online Access: http://psasir.upm.edu.my/id/eprint/107309/; https://www.iajer.com/volume-06-issue-12/
• Institution: Universiti Putra Malaysia

Reactor Cooling System (RCS) equipped with a safety system that will trigger when the reading from the sensor exceeds the threshold of normal operation. Fault Detection and Diagnosis (FDD) system is one of the safety measures that have been in ensuring the safety of the reactor. Act in giving immediate response when the faults occur and have the capability to identify the faults location. This allows the operator to react swift and according if any disturbance were to happen. In realizing this, a model-based FDD system, a system modelling and fault diagnosis algorithm need to be studied. For this study, two artificial intelligence techniques have been applied which are Adaptive Neuro Fuzzy Inference System (ANFIS) for system modelling and Artificial Neural Network (ANN) to diagnose the fault on a reactor cooling system. The ability of neural networks to learn from experience or previous data has demonstrated a significant improvement in fault detection efficiency. Additionally, a history-based strategy that is based on historical data has been shown to improve the accuracy of fault identification. As a result, complete FDD systems that successfully detect and classify 8 fault classes with performance of 96 accuracy have been developed.