Robust Prediction of Pressurized Water Reactor Dynamics Under Core Inlet Temperature Transients Using a Whale Optimization Algorithm-Based Artificial Neural Network
سال انتشار: 1405
نوع سند: مقاله ژورنالی
زبان: انگلیسی
مشاهده: 35
فایل این مقاله در 26 صفحه با فرمت PDF قابل دریافت می باشد
- صدور گواهی نمایه سازی
- من نویسنده این مقاله هستم
استخراج به نرم افزارهای پژوهشی:
شناسه ملی سند علمی:
JR_IJCCE-45-11_015
تاریخ نمایه سازی: 21 آبان 1405
چکیده مقاله:
This study presents a Whale Optimization Algorithm-optimized Artificial Neural Network framework for predicting the dynamic response of a Pressurized Water Reactor under core inlet temperature transients. While the Artificial Neural Network–Whale Optimization Algorithm methodology itself is well established in the optimization literature, its application to core inlet temperature-driven reactor dynamics under uncertainty conditions and its systematic comparison with Levenberg–Marquardt and Gradient Descent with Momentum training strategies constitute the novelty contribution of the present work. A validated nonlinear system dynamics model based on coupled differential equations was employed to generate transient datasets under both nominal and uncertainty-driven operating conditions. Five reactor state variables, including reactivity, reactor core thermal power, core inlet temperature, steam generator inlet temperature, and steam valve coefficient, were used as inputs to predict fuel temperature and steam pressure as safety-related outputs. The proposed framework was developed to capture the strongly coupled nonlinear interactions between neutronic feedback mechanisms and thermo-hydraulic behavior induced by core inlet temperature perturbations. Three training and optimization approaches, namely the Levenberg–Marquardt algorithm, Gradient Descent with Momentum, and the Whale Optimization Algorithm, were systematically evaluated. Among the investigated methods, the Whale Optimization Algorithm achieved the lowest mean squared error (۰.۰۰۱۴) and the highest prediction accuracy (۹۸.۶۱%), outperforming the Levenberg–Marquardt algorithm (۹۴.۸۱%) and Gradient Descent with Momentum (۹۰.۴۰%). Validation under deterministic and uncertainty-driven transient scenarios demonstrated excellent agreement with the reference model and confirmed the robustness of the proposed framework. The results establish the proposed approach as an accurate and computationally efficient surrogate model for robust prediction of Pressurized Water Reactor dynamics, with potential applications as a computationally efficient surrogate model for reactor transient analysis, with future applicability to real-time monitoring and digital twin environments following additional validation studies.
کلیدواژه ها:
نویسندگان
- -
Department of Nuclear Engineering, Science and Research Branch, Islamic Azad University, Tehran, I.R. IRAN
- -
Department of Nuclear Engineering, Science and Research Branch, Islamic Azad University, Tehran, I.R. IRAN
- -
Reactor and Nuclear Safety Research School, Nuclear Science and Technology Research Institute (NSTRI), Tehran, I.R. IRAN
مراجع و منابع این مقاله:
لیست زیر مراجع و منابع استفاده شده در این مقاله را نمایش می دهد. این مراجع به صورت کاملا ماشینی و بر اساس هوش مصنوعی استخراج شده اند و لذا ممکن است دارای اشکالاتی باشند که به مرور زمان دقت استخراج این محتوا افزایش می یابد. مراجعی که مقالات مربوط به آنها در سیویلیکا نمایه شده و پیدا شده اند، به خود مقاله لینک شده اند :