In order to study the effect of compressibility on Rayleigh-Taylor (RT) instability, we numerically simulated late-time evolution two-dimensional single-mode RT instability for isothermal background stratification with different Mach numbers and Atwood (At) using a high-order central compact finite difference scheme. It is found that initial density caused by plays stabilizing role, while expansion-compression flow destabilizing role. For case small number, between two sides interface small,...

Hybrid neural-physics modeling frameworks through differentiable programming have emerged as powerful tools in scientific machine learning, enabling the integration of known physics with data-driven learning to improve prediction accuracy and generalizability. However, most existing hybrid rely on explicit recurrent formulations, which suffer from numerical instability error accumulation during long-horizon forecasting. In this work, we introduce Im-PiNDiff, a novel implicit...

This study numerically analyzes the two-dimensional (2D) compressible multi-mode Rayleigh–Taylor instability at different Atwood numbers (At) and stratification parameters (Sr), corresponding to levels of flow compressibility. It is found that growth bubble thickness suppressed with increase in Sr due density small At, whereas it enhanced large because expansion compression motions. The ratio spike increases any At. effects compressibility on molecular mixing fraction, Taylor Reynolds...

In compressible Rayleigh--Taylor instability, flow compressibility plays an important role in the generation of large-scale kinetic energy, which mainly comes from conversion potential energy for small stratification parameter (Sr) and internal through pressure-dilatation work large Sr. The latter leads to bubble heights increasing rapidly bubbles that are bigger at overall statistics normalized subgrid-scale (SGS) flux is nearly independent Sr, but reverse SGS increases significantly with...

The hybrid neural differentiable models mark a significant advancement in the field of scientific machine learning. These models, integrating numerical representations known physics into deep networks, offer enhanced predictive capabilities and show great potential for data-driven modeling complex physical systems. However, critical yet unaddressed challenge lies quantification inherent uncertainties stemming from multiple sources. Addressing this gap, we introduce novel method,...

Abstract Non-diffusive phonon transport presents significant challenges in micro/nanoscale thermal characterization, compounded by the limitations of experimental-numerical techniques and presence measurement noise. Additionally, inverse modeling uncertainty quantification for submicron remain under-explored. In this study, we introduce a physics-informed Bayesian deep learning framework designed to address Boltzmann equation (BTE)-based forward problems leveraging limited noisy data. Our...

Large-eddy simulations (LES) and implicit LES (ILES) of three-dimensional compressible Rayleigh–Taylor turbulence with miscible fluids are performed compared direct numerical simulation (DNS) at the Atwood number At=0.5 stratification parameters Sr = 1.0 4.0. Three sub-grid-scale (SGS) models including constant-coefficient spatial gradient model (CSGM), dynamic Smagorinsky (DSM), mixed (DMM) considered. The CSGM achieves a high accuracy by using velocity gradients neighboring grids. priori...

The Fourier neural operator (FNO) framework is applied to the large eddy simulation (LES) of three-dimensional compressible Rayleigh-Taylor (RT) turbulence with miscible fluids at Atwood number $A_t=0.5$, stratification parameter $Sr=1.0$, and Reynolds numbers $Re=10000$ 30000. FNO model first used for predicting turbulence. different magnitudes physical fields are normalized using root mean square values an easier training models. In \emph{a posteriori} tests, outperforms velocity gradient...

The Fourier neural operator (FNO) framework is applied to the large eddy simulation (LES) of three-dimensional compressible Rayleigh–Taylor turbulence with miscible fluids at Atwood number At=0.5, stratification parameter Sr = 1.0, and Reynolds numbers Re 10 000 30 000. FNO model first used for predicting turbulence. different magnitudes physical fields are normalized using root mean square values an easier training models. In a posteriori tests, outperforms velocity gradient model, dynamic...

In China, increasing the application ratio of hot-rolled H-shapes has become a severe problem that government, academia, and engineering circles must vigorously address. Research on reasonable built-up columns is one primary methods. After reviewing various combination in existing research, paper proposes new flanged cruciform (FCHCs) made three H-shapes. Using comprehensive imperfections given by design standard, GB50017-2017, analyzes global buckling FCHCs subjected to axial compression...

Chemical vapor infiltration (CVI) is a widely adopted manufacturing technique used in producing carbon-carbon and carbon-silicon carbide composites. These materials are especially valued the aerospace automotive industries for their robust strength lightweight characteristics. The densification process during CVI critically influences final performance, quality, consistency of these composite materials. Experimentally optimizing processes challenging due to long experimental time large...