Qinyuan Li

Qinyuan Li

Auto-generated profile, based on publicly available information

Qinyuan Li


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 December 2023 in Physics of Fluids
Verified
Journal Article
Open access
PDF file
QLQinyuan Li
Qinyuan Li
Jiajian ZhuJiajian Zhu
Jiajian Zhu
YTYifu Tian
Yifu Tian
MSMingbo Sun
Mingbo Sun
MWMinggang Wan
Minggang Wan
HWHongbo Wang
Hongbo Wang
The ignition and flame propagation in an axisymmetric supersonic combustor were investigated. The laser-induced plasma was employed to ignite the supersonic inflow with a speed of Mach 2.5 and a total temperature of 1486 K. A direct-connect axisymmetric model scramjet with a fully transparent glass combustor was built, which enabled the circumferential and axial flame propagation in the cavity-based axisymmetric supersonic combustor to be visualized by the high-speed photography from the endoscopic and external views, respectively. An initial flame kernel is produced by the laser-induced plasma and propagates to the cavity leading edge along the axial direction. The establishment of the cavity shear-layer flame facilitates circumferential flame propagation. The circumferential flame propagation is coupled with the axial propagation, eventually generating a loop-shaped flame with a central-hole. Acceleration of the flame propagation can be observed, especially when the global equivalence ratio is increased. A plausible explanation for the flame propagation in the axisymmetric supersonic combustor was found using URANS numerical simulation. The axisymmetric cavity generates a low-speed loop-shaped recirculation region and thickened cavity shear-layer with an appropriate local equivalence ratio, resulting in the simultaneous axial and circumferential flame propagation. The increased temperature in the cavity and the thickened cavity shear-layer during the flame propagation produce a more intense heat release and mass transfer, leading to faster flame propagation.

24 May 2022 in arXiv
Verified
Preprint
Open access
PDF file
LBLu Bai
Lu Bai
WJWeixing Ji
Weixing Ji
QLQinyuan Li
Qinyuan Li
XYXilai Yao
Xilai Yao
XWXin Wei
Xin Wei
WZWanyi Zhu
Wanyi Zhu
Deep learning is attracting interest across a variety of domains, including natural language processing, speech recognition, and computer vision. However, model training is time-consuming and requires huge computational resources. Existing works on the performance prediction of deep neural networks, which mostly focus on the training time prediction of a few models, rely on analytical models and result in high relative errors. %Optimizing task scheduling and reducing job failures in data centers are essential to improve resource utilization and reduce carbon emissions. This paper investigates the computational resource demands of 29 classical deep neural networks and builds accurate models for predicting computational costs. We first analyze the profiling results of typical networks and demonstrate that the computational resource demands of models with different inputs and hyperparameters are not obvious and intuitive. We then propose a lightweight prediction approach DNNAbacus with a novel network structural matrix for network representation. DNNAbacus can accurately predict both memory and time cost for PyTorch and TensorFlow models, which is also generalized to different hardware architectures and can have zero-shot capability for unseen networks. Our experimental results show that the mean relative error (MRE) is 0.9% with respect to time and 2.8% with respect to memory for 29 classic models, which is much lower than the state-of-the-art works.