Shanlin Chen

Shanlin Chen

Auto-generated profile, based on publicly available information

Shanlin Chen

Department of Systems Engineering, City University of Hong Kong (CityU), Kowloon, Hong Kong

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

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

Representative Publications

View All Publications
25 April 2024 in Renewable and Sustainable Energy Reviews
Verified
Journal Article
Closed access
YCYinghao Chu
Yinghao Chu
YWYiling Wang
Yiling Wang
DYDazhi Yang
Dazhi Yang
SCShanlin Chen
Shanlin Chen
MLMengying Li
Mengying Li

21 October 2024 in IEEE Transactions on Sustainable Energy
Verified
Journal Article
Closed access
TJTao Jing
Tao Jing
SCShanlin Chen
Shanlin Chen
David Navarro-AlarconDavid Navarro-Alarcon
David Navarro-Alarcon
YCYinghao Chu
Yinghao Chu
MLMengying Li
Mengying Li
Solar forecasting has emerged as a cost-effective technology to mitigate the negative impacts of intermittent solar power on the power grid. Despite the multitude of deep learning methodologies available for forecasting solar irradiance, there is a notable gap in research concerning the automated selection and holistic utilization of multi-modal features for ultra-short-term regional irradiance forecasting. Our study introduces SolarFusionNet, a novel deep learning architecture that effectively integrates automatic multi-modal feature selection and cross-modal data fusion. SolarFusionNet utilizes two distinct types of automatic variable feature selection units to extract relevant features from multichannel satellite images and multivariate meteorological data, respectively. Long-term dependencies are then captured using three types of recurrent layers, each tailored to the corresponding data modal. In particular, a novel Gaussian kernel-injected convolutional long short-term memory network is specifically designed to isolate the sparse features present in the cloud motion field derived from optical flow. Subsequently, a hierarchical multi-head cross-modal self-attention mechanism is proposed based on the physical-logical dependencies among the three modalities to investigate the coupling correlations among the modalities. The experimental results indicate that SolarFusionNet exhibits robust performance in predicting regional solar irradiance, achieving higher accuracy than other state-of-the-art models and a forecast skill ranging from 37.4% to 47.6% against the smart persistence model for the 4-hour-ahead forecast.

9 September 2026 in Energy and AI
Verified
Journal Article
Open access
TJTao Jing
Tao Jing
SCShanlin Chen
Shanlin Chen
YCYinghao Chu
Yinghao Chu
MLMengying Li
Mengying Li

6 May 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
SCShanlin Chen
Shanlin Chen
TJTao Jing
Tao Jing
JFJiannong Fang
Jiannong Fang
MLMengying Li
Mengying Li
FPFernando Porté-Agel
Fernando Porté-Agel
Solar irradiance forecasts from numerical weather prediction (NWP) models are indispensable data sources in photovoltaic (PV) power forecasting, especially over long time horizons. It is well accepted that post-processing NWP forecasts is beneficial for the irradiance-to-power conversion process. However, this is primarily focused on the global horizontal irradiance (GHI), as GHI separation has been extensively used to generate the direct normal irradiance (DNI) and diffuse horizontal irradiance (DHI). Considering that there are some accumulated uncertainties in the GHI-separated DNI and DHI, and these irradiance components are available as NWP forecasts, this work directly incorporates all three post-processed NWP irradiance forecasts for improved PV power forecasting with the physical model chain. Compared with GHI separation, PV forecasts converted from all irradiance components are found to have better performance. Post-processing NWP irradiance forecasts generally reduces errors in the irradiance-to-power conversion process, the use of all post-processed irradiance components shows more benefits than GHI separation for PV power forecasting. Furthermore, an in-depth sensitivity analysis is performed to quantify the impact of each irradiance component on PV power forecasting with the physical model chain. It is found that DNI generally has a higher influence on the PV power output, while DNI also poses greater challenges for the post-processing step. Unexpectedly, DHI exhibits a considerably stronger impact on PV power forecasting than GHI when all irradiance components are used in the physical model chain. Given the increasing demand for accurate PV power forecasts, this work highlights the practical value of post-processed NWP irradiance forecasts for all three components. Such post-processing is essential for the efficient integration of solar energy systems.

14 March 2026 in Energy and AI
Verified
Journal Article
Open access
SCShanlin Chen
Shanlin Chen
TJTao Jing
Tao Jing
DYDazhi Yang
Dazhi Yang
YCYuntian Chen
Yuntian Chen
MLMengying Li
Mengying Li
YCYinghao Chu
Yinghao Chu