{"title":"A method fusing self-attention-SPN and NSGA-III for multi-objective radiation shielding design optimization","authors":"Li He , Guangyao Sun , Yican Wu","doi":"10.1016/j.anucene.2025.111507","DOIUrl":null,"url":null,"abstract":"<div><div>High-performance compact nuclear facilities require radiation shielding designs that balance safety and weight considerations. Current intelligent shielding design methods typically combine Evolutionary Algorithms (EA) for optimization with neural networks for evaluation. However, the neural networks used, primarily BP or DNN models composed of Fully Connected (FC) layers, require large datasets and extensive computation resources. A novel method fusing Self-Attention-based Sequence Prediction Network (Self-Attention-SPN) and Non-dominated Sorting Genetic Algorithm III (NSGA-III) was proposed in this paper for multi-objective radiation shielding design optimization. By reformulating dose rate calculation as a sequence prediction problem, the SPN of lightweight network structure leverages the multi-physics feature projection and multi-head self-attention mechanism to effectively capture the inter-layer physical feature relationships, ensuring high prediction accuracy with small datasets. The method is validated using the Savannah reactor case, where SPN achieves Monte Carlo (MC)-level accuracy with significantly reduced computational cost. Comparative experiments show that training data with additional physical parameters can reduce SPN training loss, underscoring the importance of physical information. Furthermore, SPN outperforms BP in prediction accuracy, validating the effectiveness of the multi-head self-attention mechanism. Sensitivity analysis of NSGA-III coupled with SPN prediction perturbation confirms the robustness of the proposed method. The optimization solutions effectively converge to the Pareto front, demonstrating the method’s efficiency and reliability for multi-objective radiation shielding design.</div></div>","PeriodicalId":8006,"journal":{"name":"Annals of Nuclear Energy","volume":"219 ","pages":"Article 111507"},"PeriodicalIF":1.9000,"publicationDate":"2025-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Annals of Nuclear Energy","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S030645492500324X","RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"NUCLEAR SCIENCE & TECHNOLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
High-performance compact nuclear facilities require radiation shielding designs that balance safety and weight considerations. Current intelligent shielding design methods typically combine Evolutionary Algorithms (EA) for optimization with neural networks for evaluation. However, the neural networks used, primarily BP or DNN models composed of Fully Connected (FC) layers, require large datasets and extensive computation resources. A novel method fusing Self-Attention-based Sequence Prediction Network (Self-Attention-SPN) and Non-dominated Sorting Genetic Algorithm III (NSGA-III) was proposed in this paper for multi-objective radiation shielding design optimization. By reformulating dose rate calculation as a sequence prediction problem, the SPN of lightweight network structure leverages the multi-physics feature projection and multi-head self-attention mechanism to effectively capture the inter-layer physical feature relationships, ensuring high prediction accuracy with small datasets. The method is validated using the Savannah reactor case, where SPN achieves Monte Carlo (MC)-level accuracy with significantly reduced computational cost. Comparative experiments show that training data with additional physical parameters can reduce SPN training loss, underscoring the importance of physical information. Furthermore, SPN outperforms BP in prediction accuracy, validating the effectiveness of the multi-head self-attention mechanism. Sensitivity analysis of NSGA-III coupled with SPN prediction perturbation confirms the robustness of the proposed method. The optimization solutions effectively converge to the Pareto front, demonstrating the method’s efficiency and reliability for multi-objective radiation shielding design.
期刊介绍:
Annals of Nuclear Energy provides an international medium for the communication of original research, ideas and developments in all areas of the field of nuclear energy science and technology. Its scope embraces nuclear fuel reserves, fuel cycles and cost, materials, processing, system and component technology (fission only), design and optimization, direct conversion of nuclear energy sources, environmental control, reactor physics, heat transfer and fluid dynamics, structural analysis, fuel management, future developments, nuclear fuel and safety, nuclear aerosol, neutron physics, computer technology (both software and hardware), risk assessment, radioactive waste disposal and reactor thermal hydraulics. Papers submitted to Annals need to demonstrate a clear link to nuclear power generation/nuclear engineering. Papers which deal with pure nuclear physics, pure health physics, imaging, or attenuation and shielding properties of concretes and various geological materials are not within the scope of the journal. Also, papers that deal with policy or economics are not within the scope of the journal.