摘要：
Researchers have utilized neural networks as agent models to predict shielded data sets for reactor shielding designs. However, previous models faced challenges related to unbalanced multi-scale characteristics in input and output data sets, resulting in significant generalization errors. Additionally, these models did not consider source term parameters like energy spectrum probability density distribution parameters and angle sine and cosine distribution parameters. To address these challenges, this study proposes an improved neural network model with the Swish-NNTA learning scheme. This learning scheme incorporates the high-performance Swish activation function and the enhanced Nadam-Nadamax Transition Algorithm (NNTA) learning algorithm. Experimental results demonstrate that the mean absolute percentage error (MAPE) for the upper Savannah shielding model is 0.748%, while the MAPE for the H.B.ROBINSON-2 PWR shielding model is 1.428%. The proposed shielding calculating agent model achieves higher accuracy and efficiency, presenting a promising solution for enhancing reactor shielding design and analysis.
作者：
Biao, Zhang;Jinjia, Cao;Shuang, Lin;Yingming, Song
期刊：
Journal of Instrumentation,2023年18(1):P01017 ISSN：1748-0221
作者机构：
[Jinjia, Cao; Biao, Zhang; Yingming, Song] Univ South China, Sch Nucl Sci & Technol, Hengyang 421001, Peoples R China.;[Shuang, Lin] China Nucl Power Engn Co LTD, Instrument Control Design Inst Beijing Nucl Engn R, Beijing 100142, Peoples R China.
关键词：
Crystals;Gamma rays;Ionization;Monte Carlo methods;Neutron irradiation;Nuclear medicine;Scintillation counters;Detection efficiency;Detector modeling;Detector modeling and simulation I (interaction of radiation with matter, interaction of photon with matter, interaction of hadron with matter, etc);Detector simulations;Emission process;Interaction of radiation with matter;Liquid scintillator;Model and simulation;Scintillator, scintillation and light emission process (solid, gas and liquid scintillator);Neutrons
期刊：
Annals of Nuclear Energy,2023年194:110097 ISSN：0306-4549
通讯作者：
Song, YM;Li, ZF
作者机构：
[Zhang, Biao; Li, Xiaomeng; Gui, Long; Gao, Jinjun; Song, Yingming] Univ South China, Sch Nucl Sci & Technol, Hengyang 421001, Peoples R China.;[Song, Yingming] CNNC, Nucl Ind Coll, Beijing 102618, Peoples R China.;[Yuan, Weiwei] Univ South China, Radon Prov Key Lab, Hengyang 421001, Peoples R China.;[Han, Song; Li, Zhifeng] China Nucl Power Technol Res Inst, Shenzhen 518026, Peoples R China.;[Li, Zhifeng] Xi An Jiao Tong Univ, Sch Nucl Sci & Technol, Xian 710049, Peoples R China.
通讯机构：
[Li, ZF ] C;[Song, YM ] U;Univ South China, Sch Nucl Sci & Technol, Hengyang 421001, Peoples R China.;CNNC, Nucl Ind Coll, Beijing 102618, Peoples R China.;China Nucl Power Technol Res Inst, Shenzhen 518026, Peoples R China.
摘要：
In this paper, a many-objective optimization method for reactor shielding design coupled with NSGA -III and neural network is proposed. By establishing the many-objective optimization model, using Monte Carlo method to calculate samples to train neural network, and the prediction results of neural network are used as the parameters of fitness function for many-objective optimization. The coupling between neural network and NSGA -III is realized, and the Pareto optimal solution of many-objective optimization of reactor shielding design is obtained. The results show that the method of neural network coupling NSGA -III performs well in solving the many-objective optimization problem, and can be applied to the many-objective optimization engineering design of reactor radiation shielding.(c) 2022 Elsevier Ltd. All rights reserved.
作者机构：
[Luo, W.; Luo, W; Zhu, Z. C.; Lan, H. Y.; Song, Y. M.; Sun, X. Y.] Univ South China, Sch Nucl Sci & Technol, Hengyang 421001, Peoples R China.;[Gao, Q. Y.] Chinese Acad Sci, Inst Modern Phys, Lanzhou 730000, Peoples R China.;[Cai, X. Z.; Chen, J. G.; Chen, JG] Chinese Acad Sci, Shanghai Inst Appl Phys, Shanghai 201800, Peoples R China.
通讯机构：
[Luo, W ] U;[Chen, JG ] C;Univ South China, Sch Nucl Sci & Technol, Hengyang 421001, Peoples R China.;Chinese Acad Sci, Shanghai Inst Appl Phys, Shanghai 201800, Peoples R China.
摘要：
<jats:title>Abstract</jats:title><jats:p>Disposal of long-lived fission products (LLFPs) produced in reactors has been paid a lot attention for sustainable and clean nuclear energy. Although a few transmutation means have been proposed to address this issue, there are still scientific and/or engineering challenges to achieve efficient transmutation of LLFPs. In this study, we propose a novel concept of advanced nuclear energy system (ANES) for transmuting LLFPs efficiently without isotopic separation. The ANES comprises intense photoneutron source (PNS) and subcritical reactor, which consist of lead–bismuth (Pb-Bi) layer, beryllium (Be) layer, and fuel, LLFPs and shield assemblies. The PNS is produced by bombarding radioactive cesium and iodine target with a laser-Compton scattering (LCS) γ-ray beam. We investigate the effect of the ANES system layout on transmutation efficiency by Monte Carlo simulations. It is found that a proper combination of the Pb-Bi layer and the Be layer can increase the utilization efficiency of the PNS by a factor of ~ 10, which helps to decrease by almost the same factor the LCS γ-beam intensity required for driving the ANES. Supposing that the ANES operates over 20years at a normal thermal power of 500 MWt, five LLFPs including <jats:sup>99</jats:sup>Tc, <jats:sup>129</jats:sup>I, <jats:sup>107</jats:sup>Pd, <jats:sup>137</jats:sup>Cs and <jats:sup>79</jats:sup>Se could be transmuted by more than 30%. Their effective half-lives thus decrease drastically from ~ 10<jats:sup>6</jats:sup> to less than 10<jats:sup>2</jats:sup>years. It is suggested that this successful implementation of the ANES paves the avenue towards practical transmutation of LLFPs without isotopic separation.</jats:p>
期刊：
Annals of Nuclear Energy,2022年175:109248 ISSN：0306-4549
通讯作者：
Yingming Song
作者机构：
[Zhang, Gema; Song, Yingming] Univ South China, Sch Nucl Sci & Technol, Hengyang 421001, Peoples R China.;[Zhang, Zehuan] Tsinghua Univ, Inst Nucl Energy & New Energy Technol, Beijing 100084, Peoples R China.;[Zhang, Zehuan] Beijing Key Lab Nucl Detect & Measurement Technol, Beijing 100084, Peoples R China.;[Yuan, Weiwei] Univ South China, Radon Prov Key Lab, Hengyang 421001, Peoples R China.
通讯机构：
[Yingming Song] S;School of Nuclear Science and Technology, University of South China, Hengyang 421001, China
关键词：
Deep learning;Nuclear facility decommissioning;Raspberry Pi;Source term reconstruction;γ-ray dose detector
摘要：
The radiation damage weakens the efficiency of decommissioning nuclear facilities, and the source term is hard to be located and measured because it often distributes in or inside objectives. In practice, the hot spot and source intensity are measured by the gamma camera then the radiation field is estimated by particle transport algorithm as a reference in scheme planning. This technology is expensive in measuring and simulating. Hence, this paper presents a source term activity reconstruction (hereinafter referred to as STAR) method based on deep learning to solve this problem. Firstly, a UNET liked framework is constructed to establish the correlation between the radiation field and source activity. Secondly, a source activity reconstruction problem is used to validate the framework. Then we deploy it to the Raspberry Pi with a c-ray detector to verify it can be applied in practical works. The verification results show that Raspberry Pi can complete source term activity reconstruction in a few seconds without consuming too much computing resources. In framework validating, 5000 samples consist of randomly generated activity distribution and its grid dose value which is calculated by the Monte-Carlo program. The results show that the average reconstruction error is less than 15% and the trained framework is performing well in Raspberry Pi. This method reduces the requirements for instruments and the dose detection is parallelable. Therefore, it can be widely used in nuclear facility decommissioning to improve the efficiency of source reconstruction. (C) 2022 Elsevier Ltd. All rights reserved.
作者机构：
[Li, Chao; Ma, Shaohang; Song, Yingming; Zhang, Zehuan] Univ South China, Sch Nucl Sci & Technol, Hengyang 421001, Peoples R China.;[Song, Yingming] Univ South China, Natl Exemplary Base Int Sci & Tech Collaborat Nuc, Hengyang 421001, Peoples R China.;[Guo, Yaping] China Inst Atom Energy, Beijing 102413, Peoples R China.
通讯机构：
[Yingming Song] S;School of Nuclear Science and Technology, University of South China, Hengyang 421001, China<&wdkj&>National Exemplary Base for International Sci & Tech. Collaboration of Nuclear Energy and Nuclear Safety, University of South China, Hengyang 421001, China
作者机构：
[Chen, Yang; Huang, Hudie; Mao, Jie; Song, Yingming; Zhang, Zehuan] Univ South China, Sch Nucl Sci & Technol, Hengyang 421001, Peoples R China.;[Song, Yingming] Univ South China, Natl Exemplary Base Int Sci & Tech, Collaborat Nucl Energy & Nucl Safety, Hengyang 421001, Peoples R China.;[Yuan, Weiwei] Univ South China, Radon Prov Key Lab, Hengyang 421001, Peoples R China.
通讯机构：
[Yingming Song] S;School of Nuclear Science and Technology, University of South China, Hengyang 421001, China<&wdkj&>National Exemplary Base for International Sci & Tech. Collaboration of Nuclear Energy and Nuclear Safety, University of South China, Hengyang 421001, China
关键词：
Dimension reduction of Redundant objectives;Many-objective optimization;Multi-objective optimization;PCA;Shielding design
摘要：
Safety and lightweight optimization is significant in shielding design of compact nuclear reactors. The optimal solutions obtained by the existing intelligent optimization algorithm of shielding design is not ideal when there are redundant optimization objectives. In this paper, we propose a novel multi-objective shielding optimization method which couples NSGA-II, deep neural network (DNN) and principal component analysis (PCA) for optimization by eliminating redundant objectives. The efficacy of the method is demonstrated by solving up to a plate shielding model and Savannah marine nuclear power reactor shielding model. (C) 2021 Published by Elsevier Ltd.
作者机构：
[毛婕; 宋英明; 张泽寰; 杨力] School of Nuclear Science and Technology, University of South China, Hengyang;421001, China;[韩嵩; 赵均] China Nuclear Power Technology Research Institute, Shenzhen;518026, China;[毛婕; 宋英明; 张泽寰; 杨力] 421001, China
期刊：
Science and Technology of Nuclear Installations,2021年2021:1-15 ISSN：1687-6075
通讯作者：
Song, Yingming
作者机构：
[Wang, Bo; Li, Chao; Mao, Jie; Song, Yingming; Zhang, Zehuan] Univ South China, Sch Nucl Sci & Technol, Hengyang 421001, Hunan, Peoples R China.;[Song, Yingming] Univ South China, Natl Exemplary Base Int Sci & Technol Collaborat, Hengyang 421001, Hunan, Peoples R China.;[Yuan, Weiwei] Univ South China, Radon Prov Key Lab, Hengyang 421001, Hunan, Peoples R China.
通讯机构：
[Song, Yingming] U;Univ South China, Sch Nucl Sci & Technol, Hengyang 421001, Hunan, Peoples R China.;Univ South China, Natl Exemplary Base Int Sci & Technol Collaborat, Hengyang 421001, Hunan, Peoples R China.
摘要：
<jats:p>In the field of radiation protection, the point-kernel code method is a practical tool widely used in the calculation of 3-D radiation field, and the accuracy of the point-kernel integration method strongly depends on the accuracy of the build-up factor. It is well known that calculation of the build-up factor for single-layer shields is composed of single material, but it is very complicated to calculate the build-up factor for multilayer shields (MLBUF). Recently, a novel and high-precision method based on the deep neural network (DNN) for calculating MLBUF has been proposed. In this paper, the novel method is described completely by slab models. Through the study of photon transport in multilayer shields, the parameters that mainly affect the calculation of build-up factor are analyzed. These parameters are trained by DNN as the input vectors, and the build-up factor for multilayer shields is predicted based on the trained DNN. The results predicted by DNN confirm that the method can calculate the build-up factor for multilayer shields quickly and accurately. The method has been preliminarily applicated into a 3-D radiation field calculation software, and it has proved that the method for calculating MLBUF has a broad application prospects in 3-D radiation field calculation.</jats:p>
作者机构：
[张泽寰; 宋英明; 毛婕] School of Nuclear Science and Technology, University of South China, Hengyang;Hunan;421001, China;[卢川; 唐松乾; 肖锋; 吕焕文; 杨俊云] Science and Technology on Reactor System Design Technology Laboratory, Nuclear Power Institute of China, Chengdu;610213, China
作者机构：
[宋英明; 王岩; 付孟婷; 沈格宇] School of Nuclear Science and Technology, University of South China, Hengyang;Hunan;421001, China;[肖锋; 吕焕文] Science and Technology on Reactor System Design Technology Laboratory, Nuclear Power Institute of China, Chengdu;610213, China
