Identification of physical processes via combined data-driven and data-assimilation methods
From MaRDI portal
Publication:2222261
DOI10.1016/J.JCP.2019.05.008zbMATH Open1453.62797arXiv1810.11977OpenAlexW2898608201WikidataQ127940854 ScholiaQ127940854MaRDI QIDQ2222261FDOQ2222261
Authors: Haibin Chang, Dongxiao Zhang 
Publication date: 26 January 2021
Published in: Journal of Computational Physics (Search for Journal in Brave)
Abstract: With the advent of modern data collection and storage technologies, data-driven approaches have been developed for discovering the governing partial differential equations (PDE) of physical problems. However, in the extant works the model parameters in the equations are either assumed to be known or have a linear dependency. Therefore, most of the realistic physical processes cannot be identified with the current data-driven PDE discovery approaches. In this study, an innovative framework is developed that combines data-driven and data-assimilation methods for simultaneously identifying physical processes and inferring model parameters. Spatiotemporal measurement data are first divided into a training data set and a testing data set. Using the training data set, a data-driven method is developed to learn the governing equation of the considered physical problem by identifying the occurred (or dominated) processes and selecting the proper empirical model. Through introducing a prediction error of the learned governing equation for the testing data set, a data-assimilation method is devised to estimate the uncertain model parameters of the selected empirical model. For the contaminant transport problem investigated, the results demonstrate that the proposed method can adequately identify the considered physical processes via concurrently discovering the corresponding governing equations and inferring uncertain parameters of nonlinear models, even in the presence of measurement errors. This work helps to broaden the applicable area of the research of data driven discovery of governing equations of physical problems.
Full work available at URL: https://arxiv.org/abs/1810.11977
Recommendations
- Data-driven identification of 2D partial differential equations using extracted physical features
- Robust data-driven discovery of governing physical laws with error bars
- Discovering governing equations from data by sparse identification of nonlinear dynamical systems
- Data-driven discovery of PDEs in complex datasets
- Data-driven identification of parametric partial differential equations
Applications of statistics to physics (62P35) PDEs with randomness, stochastic partial differential equations (35R60) Numerical methods for inverse problems for initial value and initial-boundary value problems involving PDEs (65M32) Applications to the sciences (65Z05)
Cites Work
- Markov chain Monte Carlo. Stochastic simulation for Bayesian inference.
- Discovering governing equations from data by sparse identification of nonlinear dynamical systems
- Data Assimilation
- Levenberg-Marquardt forms of the iterative ensemble smoother for efficient history matching and uncertainty quantification
- Hidden physics models: machine learning of nonlinear partial differential equations
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Machine learning of linear differential equations using Gaussian processes
- Model selection for dynamical systems via sparse regression and information criteria
- Learning partial differential equations via data discovery and sparse optimization
Cited In (13)
- Methods to recover unknown processes in partial differential equations using data
- An unsupervised data completion method for physically-based data-driven models
- PDE-Net 2.0: learning PDEs from data with a numeric-symbolic hybrid deep network
- Machine learning subsurface flow equations from data
- Deep-learning based discovery of partial differential equations in integral form from sparse and noisy data
- A composite neural network that learns from multi-fidelity data: application to function approximation and inverse PDE problems
- Data-driven identification of 2D partial differential equations using extracted physical features
- Robust data-driven discovery of governing physical laws with error bars
- Sequential model identification with reversible jump ensemble data assimilation method
- DL-PDE: Deep-Learning Based Data-Driven Discovery of Partial Differential Equations from Discrete and Noisy Data
- Coercing machine learning to output physically accurate results
- DLGA-PDE: discovery of PDEs with incomplete candidate library via combination of deep learning and genetic algorithm
- Model selection of chaotic systems from data with hidden variables using sparse data assimilation
Uses Software
This page was built for publication: Identification of physical processes via combined data-driven and data-assimilation methods
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2222261)
