Physics-informed reinforcement learning optimization of nuclear assembly design
Optimization of nuclear fuel assemblies if performed effectively, will lead to fuel efficiency improvement, cost reduction, and safety assurance. However, assembly optimization involves solving high-dimensional and computationally expensive combinatorial problems. As such, fuel designers’ expert jud...
Main Authors: | , , , , , , , |
---|---|
Other Authors: | |
Format: | Article |
Published: |
Elsevier BV
2021
|
Online Access: | https://hdl.handle.net/1721.1/130571 |
_version_ | 1811075151473672192 |
---|---|
author | Radaideh, Majdi I. Wolverton, Isaac Joseph, Joshua Mason Tusar, James J. Otgonbaatar, Uuganbayar Roy, Nicholas Forget, Benoit Robert Yves Shirvan, Koroush |
author2 | Massachusetts Institute of Technology. Department of Nuclear Science and Engineering |
author_facet | Massachusetts Institute of Technology. Department of Nuclear Science and Engineering Radaideh, Majdi I. Wolverton, Isaac Joseph, Joshua Mason Tusar, James J. Otgonbaatar, Uuganbayar Roy, Nicholas Forget, Benoit Robert Yves Shirvan, Koroush |
author_sort | Radaideh, Majdi I. |
collection | MIT |
description | Optimization of nuclear fuel assemblies if performed effectively, will lead to fuel efficiency improvement, cost reduction, and safety assurance. However, assembly optimization involves solving high-dimensional and computationally expensive combinatorial problems. As such, fuel designers’ expert judgement has commonly prevailed over the use of stochastic optimization (SO) algorithms such as genetic algorithms and simulated annealing. To improve the state-of-art, we explore a class of artificial intelligence (AI) algorithms, namely, reinforcement learning (RL) in this work. We propose a physics-informed AI optimization methodology by establishing a connection through reward shaping between RL and the tactics fuel designers follow in practice by moving fuel rods in the assembly to meet specific constraints and objectives. The methodology utilizes RL algorithms, deep Q learning and proximal policy optimization, and compares their performance to SO algorithms. The methodology is applied on two boiling water reactor assemblies of low-dimensional ( ~2 x 10⁶ combinations) and high-dimensional ( ~10³¹ combinations) natures. The results demonstrate that RL is more effective than SO in solving high dimensional problems, i.e., 10 × 10 assembly, through embedding expert knowledge in form of game rules and effectively exploring the search space. For a given computational resources and timeframe relevant to fuel designers, RL algorithms outperformed SO through finding more feasible patterns, 4–5 times more than SO, and through increasing search speed, as indicated by the RL outstanding computational efficiency. The results of this work clearly demonstrate RL effectiveness as another decision support tool for nuclear fuel assembly optimization. |
first_indexed | 2024-09-23T10:01:39Z |
format | Article |
id | mit-1721.1/130571 |
institution | Massachusetts Institute of Technology |
last_indexed | 2024-09-23T10:01:39Z |
publishDate | 2021 |
publisher | Elsevier BV |
record_format | dspace |
spelling | mit-1721.1/1305712023-02-07T04:48:33Z Physics-informed reinforcement learning optimization of nuclear assembly design Radaideh, Majdi I. Wolverton, Isaac Joseph, Joshua Mason Tusar, James J. Otgonbaatar, Uuganbayar Roy, Nicholas Forget, Benoit Robert Yves Shirvan, Koroush Massachusetts Institute of Technology. Department of Nuclear Science and Engineering MIT Intelligence Initiative Optimization of nuclear fuel assemblies if performed effectively, will lead to fuel efficiency improvement, cost reduction, and safety assurance. However, assembly optimization involves solving high-dimensional and computationally expensive combinatorial problems. As such, fuel designers’ expert judgement has commonly prevailed over the use of stochastic optimization (SO) algorithms such as genetic algorithms and simulated annealing. To improve the state-of-art, we explore a class of artificial intelligence (AI) algorithms, namely, reinforcement learning (RL) in this work. We propose a physics-informed AI optimization methodology by establishing a connection through reward shaping between RL and the tactics fuel designers follow in practice by moving fuel rods in the assembly to meet specific constraints and objectives. The methodology utilizes RL algorithms, deep Q learning and proximal policy optimization, and compares their performance to SO algorithms. The methodology is applied on two boiling water reactor assemblies of low-dimensional ( ~2 x 10⁶ combinations) and high-dimensional ( ~10³¹ combinations) natures. The results demonstrate that RL is more effective than SO in solving high dimensional problems, i.e., 10 × 10 assembly, through embedding expert knowledge in form of game rules and effectively exploring the search space. For a given computational resources and timeframe relevant to fuel designers, RL algorithms outperformed SO through finding more feasible patterns, 4–5 times more than SO, and through increasing search speed, as indicated by the RL outstanding computational efficiency. The results of this work clearly demonstrate RL effectiveness as another decision support tool for nuclear fuel assembly optimization. 2021-05-11T21:34:31Z 2021-05-11T21:34:31Z 2021-02 2020-09 Article http://purl.org/eprint/type/JournalArticle 0029-5493 https://hdl.handle.net/1721.1/130571 Radaideh, Majdi I. et al. "Physics-informed reinforcement learning optimization of nuclear assembly design." Nuclear Engineering and Design 372 (February 2021): 110966. http://dx.doi.org/10.1016/j.nucengdes.2020.110966 Nuclear Engineering and Design Creative Commons Attribution-NonCommercial-NoDerivs License http://creativecommons.org/licenses/by-nc-nd/4.0/ application/pdf application/pdf Elsevier BV Prof. Roy |
spellingShingle | Radaideh, Majdi I. Wolverton, Isaac Joseph, Joshua Mason Tusar, James J. Otgonbaatar, Uuganbayar Roy, Nicholas Forget, Benoit Robert Yves Shirvan, Koroush Physics-informed reinforcement learning optimization of nuclear assembly design |
title | Physics-informed reinforcement learning optimization of nuclear assembly design |
title_full | Physics-informed reinforcement learning optimization of nuclear assembly design |
title_fullStr | Physics-informed reinforcement learning optimization of nuclear assembly design |
title_full_unstemmed | Physics-informed reinforcement learning optimization of nuclear assembly design |
title_short | Physics-informed reinforcement learning optimization of nuclear assembly design |
title_sort | physics informed reinforcement learning optimization of nuclear assembly design |
url | https://hdl.handle.net/1721.1/130571 |
work_keys_str_mv | AT radaidehmajdii physicsinformedreinforcementlearningoptimizationofnuclearassemblydesign AT wolvertonisaac physicsinformedreinforcementlearningoptimizationofnuclearassemblydesign AT josephjoshuamason physicsinformedreinforcementlearningoptimizationofnuclearassemblydesign AT tusarjamesj physicsinformedreinforcementlearningoptimizationofnuclearassemblydesign AT otgonbaataruuganbayar physicsinformedreinforcementlearningoptimizationofnuclearassemblydesign AT roynicholas physicsinformedreinforcementlearningoptimizationofnuclearassemblydesign AT forgetbenoitrobertyves physicsinformedreinforcementlearningoptimizationofnuclearassemblydesign AT shirvankoroush physicsinformedreinforcementlearningoptimizationofnuclearassemblydesign |