فهرست نشریات علمی دسترسی آزاد ایران

This study investigates the dynamic effects of the steam valve coefficient on the performance of a Pressurized Water Reactor. The novelty of this research is centered on its specific application: leveraging a Whale Optimization Algorithm-optimized Artificial Neural Network framework to conduct an in-depth analysis of the steam valve coefficient —a key but often overlooked parameter. We demonstrate the framework's superior ability to capture the complex, nonlinear effects of steam valve coefficient variations on critical reactor outputs (fuel temperature and steam pressure). The system is modeled using nonlinear differential equations, and a feedforward Artificial Neural Network is trained to simulate reactor behavior under varying operating conditions. Three optimization algorithms—Levenberg-Marquardt, Gradient Descent with Momentum, and Whale Optimization Algorithm —were applied and compared. The Whale Optimization Algorithm achieved the lowest mean squared error (MSE = 0.0015) and the highest prediction accuracy (98.37%). This result demonstrates a significant improvement in process efficiency compared to traditional algorithms, such as Levenberg-Marquardt (94.21% accuracy) and Gradient Descent with Momentum (91.02% accuracy). This hybrid modeling–optimization framework enables the accurate and real-time prediction of dynamic responses.