Learning transient evolution of multidimensional reacting flows by multiscale Fourier neural operators
H Zhang, Y Weng, Z Zhao, D Zhou - Proceedings of the Combustion …, 2024 - Elsevier
Numerical stiffness and high dimensional scalars in transient reacting flow lead to a strong
trade-off between computational cost and solution accuracy for reacting flow simulations …
trade-off between computational cost and solution accuracy for reacting flow simulations …
A pre-partitioned adaptive chemistry of hydrogen for supersonic combustion with pre-exponent adjustment
H Liu, M Zhu, Y Rao, B Zhang, J Le - International Journal of Hydrogen …, 2024 - Elsevier
The chemistry mechanism has great significance in the simulations of supersonic
combustion flow. A pre-partitioned adaptive chemistry with pre-exponent adjustment (PPAC …
combustion flow. A pre-partitioned adaptive chemistry with pre-exponent adjustment (PPAC …
Artificial intelligence as a catalyst for combustion science and engineering
Combustion and energy conversion play critical roles in all facets of environmental and
technological applications, including the utilization of sustainable energy sources for power …
technological applications, including the utilization of sustainable energy sources for power …
A combined PCA-CSP solver for dimensionality and stiffness reduction in reacting flow simulations
In reduced-order model (ROM) development for reacting flow simulations, data-driven
principal component analysis (PCA) for dimensionality reduction and the physics-based …
principal component analysis (PCA) for dimensionality reduction and the physics-based …
Analysis of droplet evaporation dynamics using computational singular perturbation and tangential stretching rate
Computational singular perturbation (CSP) has been successfully used in the analysis of
complex chemically reacting flows by systematically identifying the intrinsic timescales and …
complex chemically reacting flows by systematically identifying the intrinsic timescales and …
Understanding Latent Timescales in Neural Ordinary Differential Equation Models for Advection-Dominated Dynamical Systems
The neural ordinary differential equation (ODE) framework has shown promise in
developing accelerated surrogate models for complex systems described by partial …
developing accelerated surrogate models for complex systems described by partial …
tLaSDI: Thermodynamics-informed latent space dynamics identification
We propose a data-driven latent space dynamics identification method (tLaSDI) that embeds
the first and second principles of thermodynamics. The latent variables are learned through …
the first and second principles of thermodynamics. The latent variables are learned through …
InVAErt networks for amortized inference and identifiability analysis of lumped parameter hemodynamic models
Estimation of cardiovascular model parameters from electronic health records (EHR) poses
a significant challenge primarily due to lack of identifiability. Structural non-identifiability …
a significant challenge primarily due to lack of identifiability. Structural non-identifiability …
Chemical Timescale Effects on Detonation Convergence
Numerical simulations of detonation-containing flows have emerged as crucial tools for
designing next-generation power and propulsion devices. As these tools mature, it is …
designing next-generation power and propulsion devices. As these tools mature, it is …
[PDF][PDF] A Physics-Constrained Autoencoder-NeuralODE Framework for Learning Complex Hydrocarbon Fuel Chemistry: Methane Combustion Kinetics
Computational fluid dynamics (CFD) modeling of turbulent combustion remains
computationally demanding, which is attributed to the complex interaction of multiple …
computationally demanding, which is attributed to the complex interaction of multiple …