Physics-informed multi-LSTM networks for metamodeling of nonlinear structures
From MaRDI portal
long short-term memory (LSTM)metamodelingnonlinear structuresPhyLSTM\(^2\)PhyLSTM\(^3\)physics-informed deep learning
Numerical methods for ordinary differential equations (65L99) Learning and adaptive systems in artificial intelligence (68T05) Mathematical modeling or simulation for problems pertaining to mechanics of deformable solids (74-10) Mathematical modeling or simulation for problems pertaining to fluid mechanics (76-10)
Abstract: This paper introduces an innovative physics-informed deep learning framework for metamodeling of nonlinear structural systems with scarce data. The basic concept is to incorporate physics knowledge (e.g., laws of physics, scientific principles) into deep long short-term memory (LSTM) networks, which boosts the learning within a feasible solution space. The physics constraints are embedded in the loss function to enforce the model training which can accurately capture latent system nonlinearity even with very limited available training datasets. Specifically for dynamic structures, physical laws of equation of motion, state dependency and hysteretic constitutive relationship are considered to construct the physics loss. In particular, two physics-informed multi-LSTM network architectures are proposed for structural metamodeling. The satisfactory performance of the proposed framework is successfully demonstrated through two illustrative examples (e.g., nonlinear structures subjected to ground motion excitation). It turns out that the embedded physics can alleviate overfitting issues, reduce the need of big training datasets, and improve the robustness of the trained model for more reliable prediction. As a result, the physics-informed deep learning paradigm outperforms classical non-physics-guided data-driven neural networks.
Recommendations
- Deep autoencoders for physics-constrained data-driven nonlinear materials modeling
- An intelligent nonlinear meta element for elastoplastic continua: deep learning using a new time-distributed residual U-net architecture
- A metalearning approach for physics-informed neural networks (PINNs): application to parameterized PDEs
- Enhanced physics‐informed neural networks for hyperelasticity
- Physically recurrent neural networks for path-dependent heterogeneous materials: embedding constitutive models in a data-driven surrogate
- Multifidelity modeling for physics-informed neural networks (PINNs)
- Physics-informed neural networks for shell structures
- Learning nonlocal constitutive models with neural networks
- Deep learning-accelerated computational framework based on physics informed neural network for the solution of linear elasticity
Cites work
- A continuation multi level Monte Carlo (C-MLMC) method for uncertainty quantification in compressible inviscid aerodynamics
- A multiscale multi-permeability poroplasticity model linked by recursive homogenizations and deep learning
- Data-driven reduced order modeling for time-dependent problems
- Deep hidden physics models: deep learning of nonlinear partial differential equations
- scientific article; zbMATH DE number 3994822 (Why is no real title available?)
- Kriging metamodeling in simulation: a review
- Krylov subspace techniques for reduced-order modeling of large-scale dynamical systems
- Machine learning in cardiovascular flows modeling: predicting arterial blood pressure from non-invasive 4D flow MRI data using physics-informed neural networks
- Meta-modeling game for deriving theory-consistent, microstructure-based traction-separation laws via deep reinforcement learning
- Metamodeling: Radial basis functions, versus polynomials
- Metamodels for computer-based engineering design: Survey and recommendations
- Model order reduction for large-scale structures with local nonlinearities
- Neural networks for nonlinear dynamic system modelling and identification
- On the limited memory BFGS method for large scale optimization
- Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Prediction of the mechanical behavior of nonlinear heterogeneous systems by multi-level finite element modeling
- Reliability-based structural optimization using neural networks and Monte Carlo simulation
- Surrogate modeling for fluid flows based on physics-constrained deep learning without simulation data
Cited in
(49)- Deep generative modeling for mechanistic-based learning and design of metamaterial systems
- A physics-informed deep learning framework for inversion and surrogate modeling in solid mechanics
- Data-driven simulation for general-purpose multibody dynamics using deep neural networks
- PhyCRNet: physics-informed convolutional-recurrent network for solving spatiotemporal PDEs
- Physics-informed graph neural Galerkin networks: a unified framework for solving PDE-governed forward and inverse problems
- Physics-based self-learning recurrent neural network enhanced time integration scheme for computing viscoplastic structural finite element response
- PhyGeoNet: physics-informed geometry-adaptive convolutional neural networks for solving parameterized steady-state PDEs on irregular domain
- CAN-PINN: a fast physics-informed neural network based on coupled-automatic-numerical differentiation method
- Scientific machine learning through physics-informed neural networks: where we are and what's next
- Surrogate modeling of elasto-plastic problems via long short-term memory neural networks and proper orthogonal decomposition
- Deep autoencoders for physics-constrained data-driven nonlinear materials modeling
- FEA-Net: a physics-guided data-driven model for efficient mechanical response prediction
- A-PINN: auxiliary physics informed neural networks for forward and inverse problems of nonlinear integro-differential equations
- Physics-informed neural networks for data-driven simulation: advantages, limitations, and opportunities
- Transfer learning based physics-informed neural networks for solving inverse problems in engineering structures under different loading scenarios
- An unsupervised latent/output physics-informed convolutional-LSTM network for solving partial differential equations using peridynamic differential operator
- Physically motivated structuring and optimization of neural networks for multi-physics modelling of solid oxide fuel cells
- scientific article; zbMATH DE number 6401662 (Why is no real title available?)
- PhySR: physics-informed deep super-resolution for spatiotemporal data
- A structure-preserving neural differential operator with embedded Hamiltonian constraints for modeling structural dynamics
- Variable linear transformation improved physics-informed neural networks to solve thin-layer flow problems
- A novel physics-informed deep learning strategy with local time-updating discrete scheme for multi-dimensional forward and inverse consolidation problems
- Physics-based self-learning spiking neural network enhanced time-integration scheme for computing viscoplastic structural finite element response
- SeismicNET: physics-informed neural networks for seismic wave modeling in semi-infinite domain
- Transfer learning of recurrent neural network‐based plasticity models
- Embedding physical knowledge in deep neural networks for predicting the phonon dispersion curves of cellular metamaterials
- A non-intrusive approach for physics-constrained learning with application to fuel cell modeling
- Physics informed and data-based augmented learning in structural health diagnosis
- Bayesian synergistic metamodeling (BSM) for physical information infused data-driven metamodeling
- Deep learning-enhanced efficient seismic analysis of structures with multi-fidelity modeling strategies
- Data generation framework for inverse modeling of nonlinear systems in structural dynamics applications
- Neural-integrated meshfree (NIM) method: a differentiable programming-based hybrid solver for computational mechanics
- Probabilistic graph networks for learning physics simulations
- Physics-informed deep learning of rate-and-state fault friction
- Deep learning in computational mechanics: a review
- Physics-informed probabilistic slow feature analysis
- Prediction of spatiotemporal dynamics using deep learning: coupled neural networks of long short-terms memory, auto-encoder and physics-informed neural networks
- Predictive modeling of nonlinear system responses using the residual improvement deep learning algorithm (RIDLA)
- Multi-fidelity enhanced few-shot time series prediction model for structural dynamics analysis
- Combining physics-informed graph neural network and finite difference for solving forward and inverse spatiotemporal PDEs
- Multi-benchmark adaptive sampling physics-informed neural network for complex and coupled equations
- Physics-informed neural operators for efficient modeling of infiltration in porous media
- Initialization-enhanced physics-informed neural network with domain decomposition (IDPINN)
- SK-PINN: accelerated physics-informed deep learning by smoothing kernel gradients
- Microstructural evolution prediction in the phase field crystal model with VAE-LSTM and ensemble Kalman filtering
- Physics-infused deep neural network for solution of non-associative Drucker-Prager elastoplastic constitutive model
- Physics-encoded graph deep learning for physical diffusion systems
- A physics-informed machine learning framework for inverse multi-solution problems
- A physics-informed framework for three-dimensional forward and inverse elastostatic problems in functionally graded materials
This page was built for publication: Physics-informed multi-LSTM networks for metamodeling of nonlinear structures
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2236167)
