Optimal efficiency on nuclear reactor secondary cooling process using machine learning model

Abstrak

This review delves into the quest for optimal efficiency in the secondary cooling process of nuclear reactor water plant coolant systems. Modeling secondary cooling nuclear processes is hardly performed. Thus, Neural networks with traditional statistical methodologies are integrated to innovate a hybrid model to revolutionize nuclear reactor cooling systems' performance, reliability, and safety. A total of 63 indexed papers were reviewed in the nuclear field that analyzed critical research gaps, including the need for uncertainty modeling and resilience against external hazards. Insights into sensor technologies, data analytics, and real-time monitoring underscore the importance of continuous optimization and predictive maintenance were reviewed. A descriptive analysis for a month of sampling data was presented for the parameters of temperature for TT003 and TT004 and pressure for PT002 and PT003 of the secondary process. The confidence level of 95.0% is identified for the temperature and pressure parameters. The lowest standard error was recognized at 0.00032 and 0.01691, respectively. The review culminates with a forward-looking perspective, recognizing the pivotal role of hybrid machine learning models in shaping the future of secondary cooling processes for nuclear reactor water coolant plants to improve the efficiency and sustainability of nuclear reactor systems.