No abstract

The 1MWth Reaktor TRIGA PUSPATI (RTP) in Malaysia Nuclear Agency has been in operation more than 37 years. The existing core power control uses a conventional controller known as Feedback Control Algorithm (FCA). It is technically challenging to keep the core power output stable and operate within the acceptable error bands for the safety demand of the RTP. At present, the power tracking performance of the system could be considered unsatisfactory where constant gains of power change rate constraint and control rod speed constraint are used. Hence, a study of a new power change rate constraint design to achieve safe control rod speed range is conducted to improve the current performance. In this paper, a new power change rate constraint (PCRC) method using fuzzy logic is proposed to control the core power. The Takagi-Sugeno (T-S) type Fuzzy model is chosen due to its capability to work well with linear controller and making the computational control algorithm efficient. The model for core power control consists of mathematical models of the reactor core, FCA, and control rods selection algorithm. The mathematical models of the reactor core are based on point kinetics model, thermal-hydraulic models and reactivity models. The performance of power tracking and actuation signal for control rod drive input are compared between the conventional PCRC (cPCRC) and Fuzzy PCRC using MATLAB. In conclusion, the proposed Fuzzy PCRC has satisfactory performance in core power tracking for controlling the nuclear reactor with high reliability and safety.

The 1 MWth TRIGA PUSPATI Reactor known as RTP undergoes more than 37 years of operation in Malaysia. The current core power control utilized Feedback Control Algorithm (FCA) and a conventional Control Rod Selection Algorithm (CRSA). However, the current power tracking performance suffers and increase the workload on Control Rod Drive Mechanism (CRDM) if the range between minimum and maximum rod worth value for each control rod has a significant difference. Thus, it is requiring much time to keep the core power stable at the power demand value within the acceptable error bands for the safety requirement of the RTP. In conventional CRSA, regardless of the rod worth value, the lowest position of the control rod is selected for up-movement to regulate the reactor power with 2% chattering error. To improve this method, a new CRSA is introduced named Single Control Absorbing Rod (SCAR). In SCAR, only one rod with highest reactivity worth value will be selected for coast tuning during transient and the lowest reactivity worth value will be selected for fine-tuning rod movement during steady-state. The simulation model of the reactor core is represented based on point kinetics model, thermal-hydraulic models and reactivity model. The conventional CRSA model included with control rod position dynamic model and actual reactivity worth curve data from RTP. The FCA controller is designed based on Proportional-Integral (PI) controller using MATLAB Simulink simulation. The core power control system is represented by the integration of a reactor core model, CRSA model and FCA controller. To manifest the effectiveness of the proposed SCAR algorithm, the results are compared to the conventional CRSA in both simulation and experimentation. Overall, the results shows that the SCAR algorithm offers generally better results than the conventional CRSA with the reduction in rising time up to 44%, workload up to 35%, settling time up to 26% and chattering error up to 18% of the nominal value.

<span lang="EN-US">Reactor TRIGA PUSPATI (RTP) is the only research nuclear reactor in Malaysia. Maintenance of RTP is crucial which affects its safety and reliability. Currently, RTP maintenance strategies used corrective and preventative which involved many sensors and equipment conditions. The existing preventive maintenance method takes a longer time to complete the entire system’s maintenance inspection. This study has investigated new predictive maintenance techniques for developing RTP predictive maintenance for primary cooling systems using machine learning (ML) and augmented reality (AR). Fifty papers from recent referred publications in the nuclear areas were reviewed and compared. Detailed comparison of ML techniques, parameters involved in the coolant system and AR design techniques were done. Multiclass support vector machines (SVMs), artificial neural network (ANN), long short-term memory (LSTM), feed forward back propagation (</span><span lang="EN-US">FFBP), graph neural networks-feed forward back propagation (GNN-FFBP) and ANN were used for the machine learning techniques for the nuclear reactor. Temperature, water flow, and water pressure were crucial parameters used in monitoring a nuclear reactor. Image marker-based techniques were mainly used by smart glass view and handheld devices. A switch knob with handle switch, pipe valve and machine feature were used for object detection in AR markerless technique. This study is significant and found seven recent papers closely related to the development of predictive maintenance for a research nuclear reactor in Malaysia.</span>