Pu Gan

Pu Gan

Auto-generated profile, based on publicly available information

Pu Gan

Institute of Earth System Science Faculty of Geographical Science Beijing Normal University Beijing China

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

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

Representative Publications

View All Publications
1 October 2023 in Physics of Fluids
Verified
Journal Article
Open access
PDF file
XWXiaofei Wu
Xiaofei Wu
PGP. Gan
P. Gan
JLJ. Li
J. Li
Fangxin FangFangxin Fang
Fangxin Fang
XZX. Zou
X. Zou
JZJiang Zhu
Jiang Zhu
Adaptive meshes are pivotal in numerical modeling and simulation, offering a means to efficiently, precisely, and flexibly represent intricate physical phenomena, particularly when grappling with their intricacies and varying scales. However, the transition from two dimensions (2D) to three dimensions (3D) poses a substantial challenge, as the computational demands of dynamically adaptive mesh techniques increase exponentially. Addressing this challenge effectively, we turn to the cutting-edge realm of artificial intelligence and neural networks. In our study, we harness the innovative power of a long short-term memory (LSTM) neural network as an error estimator for adapting unstructured meshes in both 2D and 3D scenarios. This LSTM network predicts the evolution of the adaptive grid based on specified variables, presenting itself as an artificial intelligence-driven architecture to optimize the adaptive criterion for the target variable. This is achieved by establishing a direct correspondence between the Riemann metric and these variables. To demonstrate the practical applicability of our approach, we seamlessly integrate the LSTM error estimator into the 3D adaptive atmospheric model Fluidity-Atmosphere (Fluidity-Atmos), thereby enabling real-time mesh adaptation during numerical simulations. We assess the effectiveness of this method in terms of simulation precision and computational efficiency through a series of experiments in both 2D and 3D settings. Our results not only reveal that the mesh patterns generated by the LSTM error estimator within Fluidity-Atmos closely resemble those produced by traditional error estimators but also underscore its superior performance in enhancing simulation accuracy. Notably, as the number of nodes increases, the LSTM mesh generator substantially reduces CPU time requirements by up to 50% in 3D cases compared to the conventional mesh generator within Fluidity-Atmos, highlighting its remarkable computational efficiency.

11 March 2026 in Journal of Geophysical Research: Atmospheres
Verified
Journal Article
Closed access
PGPu Gan
Pu Gan
JLJ. Li
J. Li
XWXiaofei Wu
Xiaofei Wu
QWQizhong Wu
Qizhong Wu
XZXun Zou
Xun Zou
Fangxin FangFangxin Fang
Fangxin Fang

12 May 2023
Verified
Conference Paper
Closed access
LTLiangxing Tang
Liangxing Tang
TZTingdong Zhou
Tingdong Zhou
YCYong Chen
Yong Chen
PLPanpan Li
Panpan Li
PGPing Gan
Ping Gan
The single-phase ground fault of the distribution network is a relatively high failure leading factor and includes many fault causing types. If the different types of grounding faults cannot be properly handled, the reliability and availability of the distribution network will be hazard in the long run. How to distinguish different fault types according to the detection of waveform map of the power grid within a short duration has always been a research hotspot. In this study, a fault type detection method based on recurrent neural network is proposed to solve the problem that the current and voltage characteristics of high resistance grounding fault are not obvious and the detection of high resistance grounding fault is difficult. Firstly, the zero-sequence voltage data in 2 cycles after the single-phase ground fault is extracted, and the zero-sequence voltage is mapped into the two-dimensional time-frequency map. Secondly, the LSTM network is built to identify the fault type and the detection accuracy is therefore optimized. To verify the effectiveness of the algorithm, we built a 10 kV distribution network simulation model, by which the experimental results show that our method can effectively detect the arc high resistance grounding fault signal and identify such grounding fault types.

7 February 2026 in Journal of Computational Physics
Verified
Journal Article
Closed access
PGPu Gan
Pu Gan
JLJ. Li
J. Li
Fangxin FangFangxin Fang
Fangxin Fang
XWXiaofei Wu
Xiaofei Wu
JZJiang Zhu
Jiang Zhu
XZXun Zou
Xun Zou

9 January 2019
Verified
Book Chapter
Open access
PDF file
JWJianwen Wang
Jianwen Wang
Zhenhai LiZhenhai Li
Zhenhai Li
XXXu Xingang
Xu Xingang
HZHongchun Zhu
Hongchun Zhu
Haikuan FengHaikuan Feng
Haikuan Feng
XXXu Xiaobin
Xu Xiaobin
Nitrogen nutrition index (NNI) can monitor winter wheat nitrogen status precisely. Current studies by remote sensing data are to construct the above-ground biomass (AGB) and plant nitrogen concentration (PNC) with spectral indices, respectively, and then substitute them into established NNI equation. This leads to an accumulation of unavoidable error. Therefore, the objective in the study was to construct a direct NNI equation with remote sensing data to reduce this error. Field measurements data including AGB, PNC and canopy hyperspectral at different winter wheat growth stages during 2012/2013, 2013/2014, 2014/2015, 2015/2016 growing seasons in Beijing, China were collected. This study was endeavored to establish a vegetation index critical N dilution curve (Nvic) with two different spectral indices, RTVI (Red edge Triangular Vegetation Index) and NDVI/PPR (the ratio of the normalized difference vegetation index to the plant pigment ratio), which are sensitive to AGB and PNC, respectively. The vegetation index NNI (NNIvi) was calculated from the ratio between the NDVI/PPR and Nvic. Results showed that (1) Nvic can be described by an equation, Nvic = 1106.4(VIRTVI)−1.512, where RTVI ranged from 2.39 to 22.14; the determination coefficient (R2) was 0.57; (2) The NNI based on the above Nvic dilution curve was in good accordance with the classical NNI, with the root mean square error (RMSE), normalized RMSE (nRMSE) and normalized average error (NAE) of 0.194, 22%, and 11%, respectively. The critical nitrogen dilution model constructed in this study was available for winter wheat nitrogen status monitoring. Thus, this study offers a new method which was suitable and convenient for estimating the NNI of the winter wheat and it can reduce quadric error for constructing NNI through indices directly instead of inversing AGB and PNC.