We propose a novel type of nonlinear solver acceleration for systems of nonlinear partial differential equations(PDEs)that is based on online/adaptive learning.It is applied in the context of multiphase flow in porous...We propose a novel type of nonlinear solver acceleration for systems of nonlinear partial differential equations(PDEs)that is based on online/adaptive learning.It is applied in the context of multiphase flow in porous media.The proposed method rely on four pillars:(i)dimensionless numbers as input parameters for the machine learning model,(ii)simplified numerical model(two-dimensional)for the offline training,(iii)dynamic control of a nonlinear solver tuning parameter(numerical relaxation),(iv)and online learning for time real-improvement of the machine learning model.This strategy decreases the number of nonlinear iterations by dynamically modifying a single global parameter,the relaxation factor,and by adaptively learning the attributes of each numerical model on-the-run.Furthermore,this work performs a sensitivity study in the dimensionless parameters(machine learning features),assess the efficacy of various machine learning models,demonstrate a decrease in nonlinear iterations using our method in more intricate,realistic three-dimensional models,and fully couple a machine learning model into an open-source multiphase flow simulator achieving up to 85%reduction in computational time.展开更多
We study the long-time asymptotic behaviour of viscosity solutions u(x,t)of the Hamilton-Jacobi equation u_(t)(x,t)+H(x,u(x,t),Du(x,t))=0 in T^(n)×(0,∞)with a PDE approach,where H=H(x,u,p)is coercive in p,non-de...We study the long-time asymptotic behaviour of viscosity solutions u(x,t)of the Hamilton-Jacobi equation u_(t)(x,t)+H(x,u(x,t),Du(x,t))=0 in T^(n)×(0,∞)with a PDE approach,where H=H(x,u,p)is coercive in p,non-decreasing in u and strictly convex in(u,p),and establish the uniform convergence of u(x,t)to an asymptotic solution u_(∞)(x)as t→∞.Moreover,u_(∞) is a viscosity solution of Hamilton-Jacobi equation H(x,u(x),Du(x))=0.展开更多
We present a parallel numerical method of simulating the interaction of atoms with a strong laser field by solving the time-depending Schr?dinger equation(TDSE) in spherical coordinates. This method is realized by com...We present a parallel numerical method of simulating the interaction of atoms with a strong laser field by solving the time-depending Schr?dinger equation(TDSE) in spherical coordinates. This method is realized by combining constructing block diagonal matrices through using the real space product formula(RSPF) with splitting out diagonal sub-matrices for short iterative Lanczos(SIL) propagator. The numerical implementation of the solver guarantees efficient parallel computing for the simulation of real physical problems such as high harmonic generation(HHG) in these interaction systems.展开更多
In this paper, we shall be concerned with the numerical solution of parabolic equations in one space variable and the time variable t. We expand Taylor series to derive a higher-order approximation for U<sub>t&l...In this paper, we shall be concerned with the numerical solution of parabolic equations in one space variable and the time variable t. We expand Taylor series to derive a higher-order approximation for U<sub>t</sub>. We begin with the simplest model problem, for heat conduction in a uniform medium. For this model problem, an explicit difference method is very straightforward in use, and the analysis of its error is easily accomplished by the use of a maximum principle. As we shall show, however, the numerical solution becomes unstable unless the time step is severely restricted, so we shall go on to consider other, more elaborate, numerical methods which can avoid such a restriction. The additional complication in the numerical calculation is more than offset by the smaller number of time steps needed. We then extend the methods to problems with more general boundary conditions, then to more general linear parabolic equations. Finally, we shall discuss the more difficult problem of the solution of nonlinear equations.展开更多
This paper develops a residual-based adaptive refinement physics-informed neural networks(RAR-PINNs)method for solving the Gross–Pitaevskii(GP)equation and Hirota equation,two paradigmatic nonlinear partial different...This paper develops a residual-based adaptive refinement physics-informed neural networks(RAR-PINNs)method for solving the Gross–Pitaevskii(GP)equation and Hirota equation,two paradigmatic nonlinear partial differential equations(PDEs)governing quantum condensates and optical rogue waves,respectively.The key innovation lies in the adaptive sampling strategy that dynamically allocates computational resources to regions with large PDE residuals,addressing critical limitations of conventional PINNs in handling:(1)Strong nonlinearities(|u|^(2)u terms)in the GP equation;(2)High-order derivatives(u_(xxx))in the Hirota equation;(3)Multi-scale solution structures.Through rigorous numerical experiments,we demonstrate that RAR-PINNs achieve superior accuracy[relative L^(2)errors of O(10^(−3))]and computational efficiency(faster than standard PINNs)for both equations.The method successfully captures:(1)Bright solitons in the GP equation;(2)First-and second-order rogue waves in the Hirota equation.The RAR adaptive sampling method demonstrates particularly remarkable effectiveness in solving steep gradient problems.Compared with uniform sampling methods,the errors of simulation results are reduced by two orders of magnitude.This study establishes a general framework for data-driven solutions of high-order nonlinear PDEs with complex solution structures.展开更多
基金MUFFINS,MUltiphase Flow-induced Fluid-flexible structure InteractioN in Subsea applications(EP/P033180/1)the PREMIERE programme grant(EP/T000414/1)SMARTRES,Smart assessment,management and optimization of urban geothermal resources(NE/X005607/1).
文摘We propose a novel type of nonlinear solver acceleration for systems of nonlinear partial differential equations(PDEs)that is based on online/adaptive learning.It is applied in the context of multiphase flow in porous media.The proposed method rely on four pillars:(i)dimensionless numbers as input parameters for the machine learning model,(ii)simplified numerical model(two-dimensional)for the offline training,(iii)dynamic control of a nonlinear solver tuning parameter(numerical relaxation),(iv)and online learning for time real-improvement of the machine learning model.This strategy decreases the number of nonlinear iterations by dynamically modifying a single global parameter,the relaxation factor,and by adaptively learning the attributes of each numerical model on-the-run.Furthermore,this work performs a sensitivity study in the dimensionless parameters(machine learning features),assess the efficacy of various machine learning models,demonstrate a decrease in nonlinear iterations using our method in more intricate,realistic three-dimensional models,and fully couple a machine learning model into an open-source multiphase flow simulator achieving up to 85%reduction in computational time.
基金the National Natural Science Foundation of China(11971344)Jiangsu Graduate Science Innovation Project(KYCX20-2746)。
文摘We study the long-time asymptotic behaviour of viscosity solutions u(x,t)of the Hamilton-Jacobi equation u_(t)(x,t)+H(x,u(x,t),Du(x,t))=0 in T^(n)×(0,∞)with a PDE approach,where H=H(x,u,p)is coercive in p,non-decreasing in u and strictly convex in(u,p),and establish the uniform convergence of u(x,t)to an asymptotic solution u_(∞)(x)as t→∞.Moreover,u_(∞) is a viscosity solution of Hamilton-Jacobi equation H(x,u(x),Du(x))=0.
基金Project supported by the National Natural Science Foundation of China(Grant Nos.11534004,11627807,11774131,and 11774130)the Scientific and Technological Project of Jilin Provincial Education Department in the Thirteenth Five-Year Plan,China(Grant No.JJKH20170538KJ)
文摘We present a parallel numerical method of simulating the interaction of atoms with a strong laser field by solving the time-depending Schr?dinger equation(TDSE) in spherical coordinates. This method is realized by combining constructing block diagonal matrices through using the real space product formula(RSPF) with splitting out diagonal sub-matrices for short iterative Lanczos(SIL) propagator. The numerical implementation of the solver guarantees efficient parallel computing for the simulation of real physical problems such as high harmonic generation(HHG) in these interaction systems.
文摘In this paper, we shall be concerned with the numerical solution of parabolic equations in one space variable and the time variable t. We expand Taylor series to derive a higher-order approximation for U<sub>t</sub>. We begin with the simplest model problem, for heat conduction in a uniform medium. For this model problem, an explicit difference method is very straightforward in use, and the analysis of its error is easily accomplished by the use of a maximum principle. As we shall show, however, the numerical solution becomes unstable unless the time step is severely restricted, so we shall go on to consider other, more elaborate, numerical methods which can avoid such a restriction. The additional complication in the numerical calculation is more than offset by the smaller number of time steps needed. We then extend the methods to problems with more general boundary conditions, then to more general linear parabolic equations. Finally, we shall discuss the more difficult problem of the solution of nonlinear equations.
基金supported by the National Natural Science Foundation of China(Grant Nos.12575003 and 12235007)the K.C.Wong Magna Fund in Ningbo University。
文摘This paper develops a residual-based adaptive refinement physics-informed neural networks(RAR-PINNs)method for solving the Gross–Pitaevskii(GP)equation and Hirota equation,two paradigmatic nonlinear partial differential equations(PDEs)governing quantum condensates and optical rogue waves,respectively.The key innovation lies in the adaptive sampling strategy that dynamically allocates computational resources to regions with large PDE residuals,addressing critical limitations of conventional PINNs in handling:(1)Strong nonlinearities(|u|^(2)u terms)in the GP equation;(2)High-order derivatives(u_(xxx))in the Hirota equation;(3)Multi-scale solution structures.Through rigorous numerical experiments,we demonstrate that RAR-PINNs achieve superior accuracy[relative L^(2)errors of O(10^(−3))]and computational efficiency(faster than standard PINNs)for both equations.The method successfully captures:(1)Bright solitons in the GP equation;(2)First-and second-order rogue waves in the Hirota equation.The RAR adaptive sampling method demonstrates particularly remarkable effectiveness in solving steep gradient problems.Compared with uniform sampling methods,the errors of simulation results are reduced by two orders of magnitude.This study establishes a general framework for data-driven solutions of high-order nonlinear PDEs with complex solution structures.
