Linlin Tian

Dr. Linlin Tian

orcid

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
3 March 2022 in Renewable Energy
Verified
Journal Article
Closed access
LTLinlin Tian
Linlin Tian
YSYilei Song
Yilei Song
PXPengcheng Xiao
Pengcheng Xiao
NZNing Zhao
Ning Zhao
WSWen Zhong Shen
Wen Zhong Shen
CZChunling Zhu
Chunling Zhu

1 February 2023 in Physics of Fluids
Verified
Journal Article
Closed access
PDF file
HTHongwei Tang
Hongwei Tang
YWYan Wang
Yan Wang
TWTongguang Wang
Tongguang Wang
LTLinlin Tian
Linlin Tian
This work introduces a novel data-driven framework to formulate explicit algebraic Reynolds-averaged Navier–Stokes (RANS) turbulence closures. Recent years have witnessed a blossom in applying machine learning (ML) methods to revolutionize the paradigm of turbulence modeling. However, due to the black-box essence of most ML methods, it is currently hard to extract interpretable information and knowledge from data-driven models. To address this critical limitation, this work leverages deep learning with symbolic regression methods to discover hidden governing equations of Reynolds stress models. Specifically, the Reynolds stress tensor is decomposed into linear and non-linear parts. While the linear part is taken as the regular linear eddy viscosity model, a long short-term memory neural network is employed to generate symbolic terms on which tractable mathematical expressions for the non-linear counterpart are built. A novel reinforcement learning algorithm is employed to train the neural network to produce best-fitted symbolic expressions. Within the proposed framework, the Reynolds stress closure is explicitly expressed in algebraic forms, thus allowing for direct functional inference. On the other hand, the Galilean and rotational invariance are craftily respected by constructing the training feature space with independent invariants and tensor basis functions. The performance of the present methodology is validated through numerical simulations of three different canonical flows that deviate in geometrical configurations. The results demonstrate promising accuracy improvements over traditional RANS models, showing the generalization ability of the proposed method. Moreover, with the given explicit model equations, it can be easier to interpret the influence of input features on generated models.

1 May 2023 in Physics of Fluids
Verified
Journal Article
Open access
PDF file
HTHongwei Tang
Hongwei Tang
YWYan Wang
Yan Wang
TWTongguang Wang
Tongguang Wang
LTLinlin Tian
Linlin Tian
YQYaoru Qian
Yaoru Qian
The past few years have witnessed a renewed blossoming of data-driven turbulence models. Quantification of the concomitant modeling uncertainty, however, has mostly been omitted, and the generalization performance of the data-driven models is still facing great challenges when predicting complex flows with different flow physics not seen during training. A robust data-driven Reynolds-averaged turbulence model with uncertainty quantification and non-linear correction is proposed in this work with the Bayesian deep neural network. In this model, the Reynolds stress tensor is decomposed into linear and non-linear parts. The linear part is taken as the usual linear eddy viscosity model while the non-linear counterpart is learned by a Bayesian deep neural network. Independent tensor bases of invariants and tensors constituted by mean strain rate tensor and rotation rate tensor are embedded into the neural network to effectively consider key turbulence features in different flows. The proposed model is well validated through numerical simulations of four canonical flows that significantly deviate in geometrical configurations and/or Reynolds numbers from those in the training data. With the non-linear corrections of embedded invariants and tensors representing key features of turbulence, the proposed model not only improves the predictive capabilities of Reynolds-averaged turbulence models on the same mesh but also has better generalization performance when simulating complex turbulent flows with large scale separation. In addition, this model allows us to quantitatively demonstrate the confidence interval of the predicted flow quantities that are originated from the model itself.

26 September 2022 in Applied Thermal Engineering
Verified
Journal Article
Closed access
YXYunjie Xu
Yunjie Xu
LTLinlin Tian
Linlin Tian
Chunling ZhuChunling Zhu
Chunling Zhu
Ning ZhaoNing Zhao
Ning Zhao

1 December 2023 in Physics of Fluids
Verified
Journal Article
Open access
PDF file
YXYunjie Xu
Yunjie Xu
LTLinlin Tian
Linlin Tian
QBQingyong Bian
Qingyong Bian
WGWei Guo
Wei Guo
Chunling ZhuChunling Zhu
Chunling Zhu
Ning ZhaoNing Zhao
Ning Zhao
Droplet impact and freezing on cold surfaces is a widely encountered multi-physical phenomenon involving droplet deformation and the liquid–solid phase change. Due to its complexity in nature, it is challenging to simulate the three-dimensional (3D) droplet impact and freezing process. Furthermore, due to the limitation of experimental techniques, it is not easy to experimentally investigate the impact of liquid droplets on ultra-cold superhydrophobic surfaces, which is crucial in some applications. Thus, in the present work, a 3D lattice Boltzmann (LB) method is developed to simulate the droplet impact and freezing on an ultra-cold superhydrophobic surface, in which an enhanced cascaded LB method is used to solve the multiphase flow field, and a multi-relaxation-time scheme is applied to solve the liquid–solid phase change model. The previous experimental results are numerically reproduced, proving that the present model can satisfactorily describe the droplet impact and solidification. The surface temperatures have no significant influence on droplet spreading. However, during the droplet retraction, a rim of ice first appears near the three-phase contact line, and then, the droplet bottom will completely solidify into ice. The occurrence of solidification at the bottom of the droplet will lead the droplet to break at a lower impact velocity, which can only be observed at a high Weber number on the room-temperature superhydrophobic surface. In addition, the effects of surface temperatures and Weber numbers on the evolution of spreading factors and space-averaged heat flux are also quantitatively analyzed in detail.

6 May 2022 in Advances in Aerodynamics
Verified
Journal Article
Open access
PDF file
XQXinyu Qi
Xinyu Qi
YYYuchen Yang
Yuchen Yang
LTLinlin Tian
Linlin Tian
ZWZhenming Wang
Zhenming Wang
Ning ZhaoNing Zhao
Ning Zhao
The combination of Cartesian grid and the adaptive mesh refinement (AMR) technology is an effective way to handle complex geometry and solve complex flow problems. Some high-efficiency Cartesian-based AMR libraries have been developed to handle dynamic changes of the grid in parallel but still can not meet the unique requirements of simulating flow around objects. In this paper, we propose an efficient Cartesian grid generation method and an information transmission approach for the wall boundary to parallelize the implementation of ghost-cell method (GCM). Also, the multi-valued ghost-cell method to handle multi-value points is improved to adapt to the parallel framework. Combining the mentioned methodologies with the open-source library p4est, an automatic and efficient simulation of compressible flow is achieved. The overall performance of the methodology is tested through a wide range of inviscid/viscous flow cases. The results indicate that the capability and parallel scalability of the present numerical methodology for solving multiple types of flows, involving shock and vortices, multi-body flow and unsteady flows are agreeable as compared with related reference data.