View in EDS

A self-organized fuzzy neural model for the pressurizer system in nuclear power plants.

Saved in:
Bibliographic Details
Title: A self-organized fuzzy neural model for the pressurizer system in nuclear power plants.
Authors: Ali, Elsayed H.1 (AUTHOR) sayedmahdy@yahoo.com, Sheta, Amal A.1 (AUTHOR), El-Araby, Sayed M.1 (AUTHOR), Mahmoud, Mohammed I.2 (AUTHOR), Mahmoud, Tarek A.2 (AUTHOR)
Source: Annals of Nuclear Energy. Feb2026:Part A, Vol. 227, pN.PAG-N.PAG. 1p.
Subjects: Nuclear power plants, Fuzzy neural networks, Adaptive control systems, Mathematical optimization, Statistical models, Nonlinear mechanics, Predictive control systems, Pressure control
Abstract: • Proposed a self-organized fuzzy neural network for modeling nonlinear pressurizer dynamics. • Enabled automatic rule generation and adaptive structure refinement for real-time accuracy. • Integrated SOFNN with generalized predictive control for multivariable regulation in PWRs. • Used hybrid KRLS–gradient optimization for rapid data-driven structural adaptation. • Outperformed PID and fuzzy-PID controllers under diverse operating conditions. This paper presents a data-driven Self-Organized Fuzzy Neural Network (SOFNN) model for modeling and controlling the pressurizer (PZR) system in pressurized water reactors (PWRs). The proposed model is designed to adaptively capture the nonlinear and time-varying dynamics of the PZR by automatically adjusting its structure and rule base for the developed model in response to data complexity. A systematic structure identification process for the model is employed, combining the Two-Stage Fuzzy Curves and Surfaces method for input selection, subtractive fuzzy clustering for rule generation, and particle swarm optimization for initializing membership function parameters. Network parameter adaptation is achieved using a hybrid learning algorithm that integrates the sliding-window Kernel Recursive Least Squares (KRLS) method with gradient-based optimization. The model's performance is evaluated through a comparative analysis with benchmark mathematical models and a Takagi-Sugano fuzzy neural (TSKFNN) network approach, demonstrating superior prediction accuracy and computational efficiency. Furthermore, SOFNN is used to develop a Generalized Predictive Control (GPC) strategy for the PZR system. Simulation results under various operational scenarios confirm the effectiveness of the proposed GPC controller, which outperforms both conventional PID and fuzzy PID controllers in terms of dynamic response and control precision. The results highlight the potential of the SOFNN-GPC framework as a robust and adaptive solution for real-time control of complex nonlinear systems. [ABSTRACT FROM AUTHOR]
Copyright of Annals of Nuclear Energy is the property of Pergamon Press - An Imprint of Elsevier Science and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
Description
Abstract:• Proposed a self-organized fuzzy neural network for modeling nonlinear pressurizer dynamics. • Enabled automatic rule generation and adaptive structure refinement for real-time accuracy. • Integrated SOFNN with generalized predictive control for multivariable regulation in PWRs. • Used hybrid KRLS–gradient optimization for rapid data-driven structural adaptation. • Outperformed PID and fuzzy-PID controllers under diverse operating conditions. This paper presents a data-driven Self-Organized Fuzzy Neural Network (SOFNN) model for modeling and controlling the pressurizer (PZR) system in pressurized water reactors (PWRs). The proposed model is designed to adaptively capture the nonlinear and time-varying dynamics of the PZR by automatically adjusting its structure and rule base for the developed model in response to data complexity. A systematic structure identification process for the model is employed, combining the Two-Stage Fuzzy Curves and Surfaces method for input selection, subtractive fuzzy clustering for rule generation, and particle swarm optimization for initializing membership function parameters. Network parameter adaptation is achieved using a hybrid learning algorithm that integrates the sliding-window Kernel Recursive Least Squares (KRLS) method with gradient-based optimization. The model's performance is evaluated through a comparative analysis with benchmark mathematical models and a Takagi-Sugano fuzzy neural (TSKFNN) network approach, demonstrating superior prediction accuracy and computational efficiency. Furthermore, SOFNN is used to develop a Generalized Predictive Control (GPC) strategy for the PZR system. Simulation results under various operational scenarios confirm the effectiveness of the proposed GPC controller, which outperforms both conventional PID and fuzzy PID controllers in terms of dynamic response and control precision. The results highlight the potential of the SOFNN-GPC framework as a robust and adaptive solution for real-time control of complex nonlinear systems. [ABSTRACT FROM AUTHOR]
ISSN:03064549
DOI:10.1016/j.anucene.2025.111947