基于深度学习增强的降阶集成卡尔曼滤波器的参数偏微分方程贝叶斯数据同化

IF 7.2 2区 物理与天体物理 Q1 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS
Yanyan Wang , Liang Yan , Tao Zhou
{"title":"基于深度学习增强的降阶集成卡尔曼滤波器的参数偏微分方程贝叶斯数据同化","authors":"Yanyan Wang ,&nbsp;Liang Yan ,&nbsp;Tao Zhou","doi":"10.1016/j.cpc.2025.109544","DOIUrl":null,"url":null,"abstract":"<div><div>Bayesian data assimilation for systems governed by parametric partial differential equations (PDEs) is computationally demanding due to the need for multiple forward model evaluations. Reduced-order models (ROMs) have been widely used to reduce the computational burden. However, traditional ROM techniques rely on linear mode superposition, which frequently fails to capture nonlinear time-dependent dynamics efficiently and leads to biases in the assimilation results. To address these limitations, we introduce a new deep learning-enhanced reduced-order ensemble Kalman filter (DR-EnKF) method for Bayesian data assimilation. The proposed approach employs a two-tiered learning framework. First, the full-order model is reduced using operator inference, which finds the primary dynamics of the system through long-term simulations generated from coarse-grid data. Second, a model error network is trained with short-term simulation data from a fine grid to learn about the ROM-induced discrepancy. The learned network is then used online to correct the ROM-based EnKF, resulting in more accurate state updates during the assimilation process. The performance of the proposed method is evaluated on several benchmark problems, including the Burgers' equation, the FitzHugh-Nagumo model, and advection-diffusion-reaction systems. The results show considerable computational speedup without compromising accuracy, making this approach an effective tool for large-scale data assimilation tasks.</div></div>","PeriodicalId":285,"journal":{"name":"Computer Physics Communications","volume":"311 ","pages":"Article 109544"},"PeriodicalIF":7.2000,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Deep learning-enhanced reduced-order ensemble Kalman filter for efficient Bayesian data assimilation of parametric PDEs\",\"authors\":\"Yanyan Wang ,&nbsp;Liang Yan ,&nbsp;Tao Zhou\",\"doi\":\"10.1016/j.cpc.2025.109544\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><div>Bayesian data assimilation for systems governed by parametric partial differential equations (PDEs) is computationally demanding due to the need for multiple forward model evaluations. Reduced-order models (ROMs) have been widely used to reduce the computational burden. However, traditional ROM techniques rely on linear mode superposition, which frequently fails to capture nonlinear time-dependent dynamics efficiently and leads to biases in the assimilation results. To address these limitations, we introduce a new deep learning-enhanced reduced-order ensemble Kalman filter (DR-EnKF) method for Bayesian data assimilation. The proposed approach employs a two-tiered learning framework. First, the full-order model is reduced using operator inference, which finds the primary dynamics of the system through long-term simulations generated from coarse-grid data. Second, a model error network is trained with short-term simulation data from a fine grid to learn about the ROM-induced discrepancy. The learned network is then used online to correct the ROM-based EnKF, resulting in more accurate state updates during the assimilation process. The performance of the proposed method is evaluated on several benchmark problems, including the Burgers' equation, the FitzHugh-Nagumo model, and advection-diffusion-reaction systems. The results show considerable computational speedup without compromising accuracy, making this approach an effective tool for large-scale data assimilation tasks.</div></div>\",\"PeriodicalId\":285,\"journal\":{\"name\":\"Computer Physics Communications\",\"volume\":\"311 \",\"pages\":\"Article 109544\"},\"PeriodicalIF\":7.2000,\"publicationDate\":\"2025-02-17\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Computer Physics Communications\",\"FirstCategoryId\":\"101\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S0010465525000475\",\"RegionNum\":2,\"RegionCategory\":\"物理与天体物理\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Computer Physics Communications","FirstCategoryId":"101","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0010465525000475","RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS","Score":null,"Total":0}
引用次数: 0

摘要

参数偏微分方程(PDEs)控制系统的贝叶斯数据同化由于需要多个正演模型评估,计算量很大。降阶模型(ROMs)被广泛用于减少计算量。然而,传统的ROM技术依赖于线性模式叠加,这往往不能有效地捕获非线性时相关动力学,并导致同化结果的偏差。为了解决这些限制,我们引入了一种新的深度学习增强的降阶集成卡尔曼滤波(DR-EnKF)方法用于贝叶斯数据同化。所建议的方法采用两层学习框架。首先,利用算子推理对全阶模型进行约简,通过粗网格数据生成的长期仿真发现系统的主要动力学特性;其次,使用来自细网格的短期模拟数据训练模型误差网络,以了解rom引起的误差。然后在线使用学习到的网络来校正基于rom的EnKF,从而在同化过程中获得更准确的状态更新。在Burgers方程、FitzHugh-Nagumo模型和平流-扩散-反应系统等基准问题上对所提方法的性能进行了评估。结果表明,在不影响精度的情况下,该方法具有相当大的计算速度,使该方法成为大规模数据同化任务的有效工具。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Deep learning-enhanced reduced-order ensemble Kalman filter for efficient Bayesian data assimilation of parametric PDEs
Bayesian data assimilation for systems governed by parametric partial differential equations (PDEs) is computationally demanding due to the need for multiple forward model evaluations. Reduced-order models (ROMs) have been widely used to reduce the computational burden. However, traditional ROM techniques rely on linear mode superposition, which frequently fails to capture nonlinear time-dependent dynamics efficiently and leads to biases in the assimilation results. To address these limitations, we introduce a new deep learning-enhanced reduced-order ensemble Kalman filter (DR-EnKF) method for Bayesian data assimilation. The proposed approach employs a two-tiered learning framework. First, the full-order model is reduced using operator inference, which finds the primary dynamics of the system through long-term simulations generated from coarse-grid data. Second, a model error network is trained with short-term simulation data from a fine grid to learn about the ROM-induced discrepancy. The learned network is then used online to correct the ROM-based EnKF, resulting in more accurate state updates during the assimilation process. The performance of the proposed method is evaluated on several benchmark problems, including the Burgers' equation, the FitzHugh-Nagumo model, and advection-diffusion-reaction systems. The results show considerable computational speedup without compromising accuracy, making this approach an effective tool for large-scale data assimilation tasks.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
Computer Physics Communications
Computer Physics Communications 物理-计算机:跨学科应用
CiteScore
12.10
自引率
3.20%
发文量
287
审稿时长
5.3 months
期刊介绍: The focus of CPC is on contemporary computational methods and techniques and their implementation, the effectiveness of which will normally be evidenced by the author(s) within the context of a substantive problem in physics. Within this setting CPC publishes two types of paper. Computer Programs in Physics (CPiP) These papers describe significant computer programs to be archived in the CPC Program Library which is held in the Mendeley Data repository. The submitted software must be covered by an approved open source licence. Papers and associated computer programs that address a problem of contemporary interest in physics that cannot be solved by current software are particularly encouraged. Computational Physics Papers (CP) These are research papers in, but are not limited to, the following themes across computational physics and related disciplines. mathematical and numerical methods and algorithms; computational models including those associated with the design, control and analysis of experiments; and algebraic computation. Each will normally include software implementation and performance details. The software implementation should, ideally, be available via GitHub, Zenodo or an institutional repository.In addition, research papers on the impact of advanced computer architecture and special purpose computers on computing in the physical sciences and software topics related to, and of importance in, the physical sciences may be considered.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信