A data-driven reduced-order model for stiff chemical kinetics using dynamics-informed training  

作　　者：Vijayamanikandan Vijayarangan Harshavardhana A.Uranakara Shivam Barwey Riccardo Malpica Galassi Mohammad Rafi Malik Mauro Valorani Venkat Raman Hong G.Im 

机构地区：[1]CCRC,King Abdullah University of Science and Technology,Thuwal,Saudi Arabia [2]Argonne Leadership Computing Facility,Argonne National Laboratory,Lemont,IL 60439,USA [3]Mechanical and Aerospace Engineering Department,Sapienza University of Rome,Via Eudossiana,18,Rome 00184,Italy [4]Department of Aerospace Engineering,University of Michigan,Ann Arbor,MI 48109,USA

出　　处：《Energy and AI》2024年第1期181-192,共12页能源与人工智能（英文）

基　　金：support from the Argonne Leadership Computing Facility,which is a U.S.Department of Energy Office of Science User Facility operated under contract DE-AC02-06CH11357;support of ONR,United States Grant No.N00014-21-1-2475 with Dr.Eric Marineau as Program Manager.

摘　　要：A data-based reduced-order model(ROM)is developed to accelerate the time integration of stiff chemically reacting systems by effectively removing the stiffness arising from a wide spectrum of chemical time scales.Specifically,the objective of this work is to develop a ROM that acts as a non-stiff surrogate model for the time evolution of the thermochemical state vector(temperature and species mass fractions)during an otherwise highly stiff and nonlinear ignition process.The model follows an encode-forecast-decode strategy that combines a nonlinear autoencoder(AE)for dimensionality reduction(encode and decode steps)with a neural ordinary differential equation(NODE)for modeling the dynamical system in the AE-provided latent space(forecasting step).By means of detailed timescale analysis by leveraging the dynamical system Jacobians,this work shows how data-based projection operators provided by autoencoders can inherently construct the latent spaces by removing unnecessary fast timescales,even more effectively than physics-based counterparts based on an eigenvalue analysis.A key finding is that the most significant degree of stiffness reduction is achieved through an end-to-end training strategy,where both AE and neural ODE parameters are optimized simultaneously,allowing the discovered latent space to be dynamics-informed.In addition to end-to-end training,this work highlights the vital contribution of AE nonlinearity in the stiffness reduction task.For the prediction of homogeneous ignition phenomena for H2-air and C2H4-air mixtures,the proposed ROM achieves several ordersof-magnitude increase in the integration time step size when compared to(a)a baseline CVODE solver for the full-chemical system,(b)statistical technique–principal component analysis(PCA),and(c)computational singular perturbation(CSP),a vetted physics-based stiffness-reducing modeling framework.

关 键 词：Stiff system Chemical kinetics Reacting flows Autoencoders Neural ODE 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]