arXiv subjects

Xingjian Gao

Publications and source records attributed to Xingjian Gao.

4 recordsLinked to original sources

arXiv2026

A Three-Dimensional Two-Temperature Gas-Kinetic Scheme with Generalized Kinetic Boundary Condition for Hypersonic SBLI

Accurate prediction of aerothermal loads in hypersonic flows is critical yet challenging due to the coupling of Shock-Wave/Boundary-Layer Interactions (SBLI) and thermal non-equilibrium. This work presents the development of a three-dimensional two-temperature Gas-Kinetic Scheme (3D 2T-GKS) on unstructured meshes. The scheme resolves translational-rotational and vibrational energy modes within a unified kinetic framework. A key innovation is the integration of a Generalized Kinetic Boundary Condition (GKBC), which physically decouples the thermal accommodation of vibrational energy from the translational-rotational mode, thereby offering a more accurate model for gas-surface interactions. Additionally, a Discontinuity Feedback Factor (DFF) is employed to capture strong shock waves with reduced numerical dissipation compared to classical limiters. The method is rigorously validated against standard experimental benchmarks, including the sharp double-cone and hollow cylinder-flare configurations. Numerical results demonstrate that the proposed solver, augmented by the GKBC, accurately captures complex wave structures, separation topologies, and surface heat flux distributions. These findings confirm the robustness and fidelity of the 3D 2T-GKS for simulating complex hypersonic non-equilibrium flows.

physics.flu-dyn↗

arXiv2026

A two-temperature gas-kinetic scheme for hypersonic nonequilibrium flow computations

Accurate aerodynamic and aerothermodynamic predictions are crucial for numerous hypersonic applications. This paper proposes a gas-kinetic scheme (GKS) coupled with a two-temperature kinetic model, which distinguishes between the translational-rotational and vibrational modes of temperature. Compared with one-temperature model and the translational-rotational multi-temperature model, the proposed model provides a more physically accurate simulation of real gas effects when vibrational energy modes of air are excited. On the other hand, it is computationally simpler than multi-temperature model with independent translational, rotational and vibrational modes. The scheme is implemented on both structured and unstructured grids. To further improve the robustness for strong shock and rarefaction waves, the discontinuity feedback factor is employed instead of traditional limiters. Numerical verifications are conducted on one-dimensional shock structure, two-dimensional (2D) hypersonic flow over a cylinder, 2D hypersonic flow over a wedge and 2D Edney Type IV shock/shock interaction. Compared with experimental data, the reference results from direct simulation Monte Carlo (DSMC) method and Navier--Stokes (NS) solvers, the present method demonstrates accurate prediction of the thermally non-equilibrium shock wave structures and hypersonic flow fields.

physics.flu-dyn↗

arXiv2023

Closed-Loop Transcription via Convolutional Sparse Coding

Autoencoding has achieved great empirical success as a framework for learning generative models for natural images. Autoencoders often use generic deep networks as the encoder or decoder, which are difficult to interpret, and the learned representations lack clear structure. In this work, we make the explicit assumption that the image distribution is generated from a multi-stage sparse deconvolution. The corresponding inverse map, which we use as an encoder, is a multi-stage convolution sparse coding (CSC), with each stage obtained from unrolling an optimization algorithm for solving the corresponding (convexified) sparse coding program. To avoid computational difficulties in minimizing distributional distance between the real and generated images, we utilize the recent closed-loop transcription (CTRL) framework that optimizes the rate reduction of the learned sparse representations. Conceptually, our method has high-level connections to score-matching methods such as diffusion models. Empirically, our framework demonstrates competitive performance on large-scale datasets, such as ImageNet-1K, compared to existing autoencoding and generative methods under fair conditions. Even with simpler networks and fewer computational resources, our method demonstrates high visual quality in regenerated images. More surprisingly, the learned autoencoder performs well on unseen datasets. Our method enjoys several side benefits, including more structured and interpretable representations, more stable convergence, and scalability to large datasets. Our method is arguably the first to demonstrate that a concatenation of multiple convolution sparse coding/decoding layers leads to an interpretable and effective autoencoder for modeling the distribution of large-scale natural image datasets.

cs.CV↗

arXiv2022

Revisiting Sparse Convolutional Model for Visual Recognition

Despite strong empirical performance for image classification, deep neural networks are often regarded as ``black boxes'' and they are difficult to interpret. On the other hand, sparse convolutional models, which assume that a signal can be expressed by a linear combination of a few elements from a convolutional dictionary, are powerful tools for analyzing natural images with good theoretical interpretability and biological plausibility. However, such principled models have not demonstrated competitive performance when compared with empirically designed deep networks. This paper revisits the sparse convolutional modeling for image classification and bridges the gap between good empirical performance (of deep learning) and good interpretability (of sparse convolutional models). Our method uses differentiable optimization layers that are defined from convolutional sparse coding as drop-in replacements of standard convolutional layers in conventional deep neural networks. We show that such models have equally strong empirical performance on CIFAR-10, CIFAR-100, and ImageNet datasets when compared to conventional neural networks. By leveraging stable recovery property of sparse modeling, we further show that such models can be much more robust to input corruptions as well as adversarial perturbations in testing through a simple proper trade-off between sparse regularization and data reconstruction terms. Source code can be found at https://github.com/Delay-Xili/SDNet.

cs.CV↗