Mengtao Yan

Mengtao Yan

Auto-generated profile, based on publicly available information

Mengtao Yan

Gaoling School of Artificial Intelligence, Renmin University of China, Beijing, China

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
1 November 2022 in Physics of Fluids
Verified
Journal Article
Open access
PDF file
MYMengtao Yan
Mengtao Yan
ZZZhiming Zhang
Zhiming Zhang
Shangce GaoShangce Gao
Shangce Gao
Shuyang CaoShuyang Cao
Shuyang Cao
A masked gated recurrent unit (GRU) model is proposed to establish the mapping relationship between surface pressures on a square cylinder and wake velocities, which can be used to predict statistical and instantaneous aerodynamic pressure fields on a square cylinder from its wakefield. A novel mask net is proposed to figure out one or two wake points where the velocities contribute dominantly to the surface pressure field. A three-dimensional unsteady large-eddy simulation of flow around a square cylinder is performed at Re = 22 000 to generate data for training and validating the proposed models. Results show that local mean pressure coefficients can be well predicted from velocities at even one wake point, but the accuracies of predicting fluctuating pressure coefficients and time-series of local pressure coefficients depend on both the model and the surface pressure location, with more satisfactory predictions achieved in the cross-flow direction. High correlation coefficients of pressure coefficient distributions around a square cylinder between predicted and real distributions are achieved except for the masked GRU model with one wake point. Meanwhile, in terms of the temporal correlation coefficient, all models exhibit good prediction of time-series of pressure coefficients on the side and back surfaces where they are strongly affected by vortex shedding and lower accuracy on the front surface where the pressure coefficients deviate somewhat randomly around the mean value. Large prediction error occurs at the corners of the square cylinder. This study has potential application to risk analysis of structures subject to flow-induced loads.

27 May 2024 in IEEE/CAA Journal of Automatica Sinica
Verified
Journal Article
Closed access
ZZZhiming Zhang
Zhiming Zhang
Shangce GaoShangce Gao
Shangce Gao
Mengchu ZhouMengchu Zhou
Mengchu Zhou
MYMengtao Yan
Mengtao Yan
Shuyang CaoShuyang Cao
Shuyang Cao
Accurately predicting fluid forces acting on the surface of a structure is crucial in engineering design. However, this task becomes particularly challenging in turbulent flow, due to the complex and irregular changes in the flow field. In this study, we propose a novel deep learning method, named mapping network-coordinated stacked gated recurrent units (MSU), for predicting pressure on a circular cylinder from velocity data. Specifically, our coordinated learning strategy is designed to extract the most critical velocity point for prediction, a process that has not been explored before. In our experiments, MSU extracts one point from a velocity field containing 121 points and utilizes this point to accurately predict 100 pressure points on the cylinder. This method significantly reduces the workload of data measurement in practical engineering applications. Our experimental results demonstrate that MSU predictions are highly similar to the real turbulent data in both spatio-temporal and individual aspects. Furthermore, the comparison results show that MSU predicts more precise results, even outperforming models that use all velocity field points. Compared with state-of-the-art methods, MSU has an average improvement of more than 45% in various indicators such as root mean square error (RMSE). Through comprehensive and authoritative physical verification, we established that MSU's prediction results closely align with pressure field data obtained in real turbulence fields. This confirmation underscores the considerable potential of MSU for practical applications in real engineering scenarios. The code is available at https://github.com/zhangzm0128/MSU.

3 August 2025
Verified
Conference Paper
Open access
MYMengtao Yan
Mengtao Yan
QWQi Wang
Qi Wang
HWHaining Wang
Haining Wang
RCRuizhi Chengze
Ruizhi Chengze
YZYi Zhang
Yi Zhang
HSHao Sun
Hao Sun

8 August 2026
Verified
Conference Paper
Open access
BZBocheng Zeng
Bocheng Zeng
RZRui Zhang
Rui Zhang
RMRunze Mao
Runze Mao
MYMengtao Yan
Mengtao Yan
XBXuan Bai
Xuan Bai
HSHao Sun
Hao Sun
Efficiently solving Poisson equations on complex, irregular domains remains a fundamental challenge in scientific computing, as classical iterative solvers often suffer from prohibitive runtime due to ill-conditioned systems. While neural operators offer a fast alternative, they typically rely on large-scale labeled datasets or struggle with unstable training dynamics when using physics-informed residual losses. We propose NPSolver, a neural Poisson solver trained without solution labels via iterative physics supervision. Instead of relying on fully converged numerical solutions or raw PDE residuals, NPSolver utilizes a small number of preconditioned conjugate gradient (PCG) steps to refine its own predictions, providing a more stable and well-scaled training signal. Theoretical analysis confirms that this iterative supervision serves as a well-conditioned error proxy and that a stop-gradient design is essential for optimization stability. To better capture boundary-driven features under mixed boundary conditions, we further introduce the Boundary-Aware Transolver (BA-Transolver ) architecture that explicitly separates interior and boundary tokenization. Extensive evaluations on 2D and 3D irregular geometries demonstrate that NPSolver outperforms both physics-informed and data-driven baselines. Furthermore, a downstream thermal control task highlights the model's capability for conducting efficient and reliable gradient-based boundary control. We will release our codes and data at https://github.com/intell-sci-comput/NPSolver.