Skip to content
Conference

A spatiotemporal attention network for wind-power grid-connection fluctuation forecasting

Sep 2026 · International Conference on Intelligent Transportation Systems and Automation Control · Vol 14368, pp. 143681Y - 143681Y-9 · 0 citations · 18 references
Engineering

TL;DR

A spatiotemporal attention network for wind-power fluctuation forecasting, denoted STAN-WF, provides an interpretable and deployable forecasting module for renewable-energy integration, dispatch automation, and transport-hub energy systems under volatile operating conditions.

Abstract

High-penetration wind generation introduces steep grid-connection power fluctuations that challenge dispatch reserve allocation, automatic generation control, and short-term balancing in energy facilities. This paper proposes a spatiotemporal attention network for wind-power fluctuation forecasting, denoted STAN-WF. The model constructs a dynamic wind-farm graph from historical supervisory-control data, numerical weather prediction variables, grid-operation states, and optional optical-sensing features from sky cameras or visual inspection systems. Spatial attention learns directed propagation relations among wind farms, temporal attention selects horizon-dependent lag patterns, and a fluctuation-aware decoder jointly predicts active power, power change, and ramp-event probability. Case-study results on an anonymized 12-farm wind-power dataset show that STAN-WF reduces RMSE by 11.9% compared with a strong spatiotemporal graph-convolution baseline and improves ramp-event F1-score from 0.842 to 0.891. The method provides an interpretable and deployable forecasting module for renewable-energy integration, dispatch automation, and transport-hub energy systems under volatile operating conditions.

View source

Similar papers

Open access Oct 2026

Weather-Transition-Aware Dynamic Mixture-of-Experts for Day-Ahead Probabilistic Wind Power Forecasting

Wind power forecasting under transitional weather conditions is crucial for the secure operation and dispatch of power systems. However, existing methods have limited ability to perceive and adapt to the rapid, nonstationary changes in the weather–power mapping caused by transitional weather, resulting in substantially...

Yi-Mei Wang, Jia-Xi Huang, Yi-Fei Dou et al. · 0 citations
Open access Oct 2026

Temporal Convolutional Network with a Physics-Informed Decoder for 15-Minute-Ahead Solar Power Forecasting

As renewable energy grows, photovoltaic power integration into power grids has increased, and high solar penetration creates challenges for power system operation. Accurate, near-real-time solar power forecasting over short time intervals is therefore important for supporting stable grid operation and efficient energy...

Muhammad Imran Munawar, J. Kilby, K. Prasad · 0 citations
Open access 2026

Multi-Scale Short-Term Wind Power Forecasting with Graph-Based Super-Resolution of Numerical Weather Prediction

: Short-term wind power forecasting up to 72 h ahead is essential for reliable power system operation and efficient electricity market participation under high renewable penetration. However, operational numerical weather prediction (NWP) products are substantially coarser than the spatial and temporal scales governing...

Can Zhang, Jing-Wen Yu, Xian-Yong Xiao et al. · 0 citations
Open access Oct 2026

STGAF-Net: A Multi-Source Spatiotemporal Gated Adaptive Fusion Network for Wind Speed Nowcasting

STGAF-Net is proposed, a multi-source spatiotemporal gated adaptive fusion network that uses radar reflectivity and surface wind speed as complementary inputs, employs collaborative local–global encoding to capture local details and long-range spatial dependencies, and adaptively integrates multi-source features throug...

Huan-Tong Geng, Mei-Jing Shi, Shao-Qiang Ma et al. · 0 citations
#machine learning Preprint Oct 2026

LLM-enhanced spatio-temporal learning for grid-level docked bike sharing demand prediction

Short-term bike-sharing demand forecasting is complicated by spatial-temporal non-stationarity and the practical difficulty of incorporating unstructured external text into numerical pipelines. Conventional approaches rely on historical flow sequences and fixed graph structures, thereby constraining their accuracy when...

Xuxilu Zhang, Francesc Soriguera · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.