Abstract:In this paper, we present a Newton-like method based on model reduction techniques, which can be used in implicit numerical methods for approximating the solution to ordinary differential equations. In each iteration, the Newton-like method solves a reduced order linear system in order to compute the Newton step. This reduced system is derived using a projection matrix, obtained using proper orthogonal decomposition, which is updated in each time step of the numerical method. We demonstrate that the method can be used together with Euler's implicit method to simulate CO$_2$ injection into an oil reservoir, and we compare with using Newton's method. The Newton-like method achieves a speedup of between 39% and 84% for systems with between 4,800 and 52,800 state variables.

What problem does this paper attempt to address?

The paper attempts to address the issue of computational efficiency when solving large-scale nonlinear ordinary differential equations (ODEs) using implicit numerical methods. Specifically, the authors propose a Newton-type method based on model order reduction techniques, aiming to improve computational efficiency by solving a low-order linear system at each iteration. This method is particularly suitable for stiff systems, as these systems typically involve a large computational workload. ### Main Issues: 1. **Computational Efficiency**: Traditional Newton's method requires solving a large-scale linear system at each iteration, which is very time-consuming when dealing with systems with a large number of state variables. 2. **Model Order Reduction**: How to effectively reduce a high-order system to a low-order system, thereby significantly reducing computation time while maintaining accuracy. ### Solution: - **Newton-type Method**: The authors propose a Newton-type method based on model order reduction techniques. In each iteration, this method reduces the Jacobian matrix of the original system to a low-order system using a projection matrix and solves this low-order linear system to compute the Newton step. - **Projection Matrix**: The projection matrix is computed using Proper Orthogonal Decomposition (POD) and is updated at each time step to adapt to changes in the system. - **Adaptive Update**: If the convergence rate is too slow, the method will revert to solving the original full-order linear system. ### Application Example: - **CO2 Injection Reservoir Simulation**: The authors used this method to numerically simulate the process of CO2 injection into a reservoir. This process is described by 4 coupled partial differential equations and discretized using the finite volume method. Experimental results show that this method can achieve a speedup of 39% to 84% compared to the traditional Newton's method. ### Summary: The main contribution of this paper is the proposal of an efficient Newton-type method that significantly improves the computational efficiency of implicit numerical methods for solving large-scale nonlinear ODEs through model order reduction techniques. This method is particularly effective in handling stiff systems, capable of significantly reducing computation time while maintaining accuracy.

A Newton-like Method based on Model Reduction Techniques for Implicit Numerical Methods

Fast Model Order Reduction Via Nonlinear Optimization

Automatic Decoupling and Index-aware Model-Order Reduction for Nonlinear Differential-Algebraic Equations

Non-intrusive model reduction of large-scale, nonlinear dynamical systems using deep learning

Improving Newton's method performance by parametrization: the case of Richards equation

NKS Method for the Implicit Solution of a Coupled Allen-Cahn/Cahn-Hilliard System

A fast neural hybrid Newton solver adapted to implicit methods for nonlinear dynamics

Newton's Method Applied to Nonlinear Boundary Value Problems: A Numerical Approach

A Fully‐implicit Numerical Algorithm of Two‐fluid Two‐phase Flow Model Using Jacobian‐free Newton–Krylov Method

Active-Set Reduced-Space Methods with Nonlinear Elimination for Two-Phase Flow Problems in Porous Media.

Truncated Model Reduction Methods for Linear Time-Invariant Systems Via Eigenvalue Computation.

Model reduction of coupled systems based on non-intrusive approximations of the boundary response maps

Continuous Newton Method for Nonlinear Partial Differential Equation

Arbitrarily high order implicit ODE integration by correcting a neural network approximation with Newton's method

Data-driven model reduction via non-intrusive optimization of projection operators and reduced-order dynamics

Model reduction on inertial manifolds for N–S equations approached by multilevel finite element method

Asymptotically Newton-Type Methods without Inverses for Solving Equations

A Fully Implicit Method for Lattice Boltzmann Equations.

Neural empirical interpolation method for nonlinear model reduction

A Matrix-Free Newton Method

Model reduction of convection-dominated partial differential equations via optimization-based implicit feature tracking