Hierarchical Temporal Diffusion for Dispatch-Oriented Multi-Energy Scenario Generation
Main Article Content
Abstract
Multi-energy scenario generation should capture intraday dynamics, cross-channel dependence, and dispatch-sensitive risks. This paper proposes Hierarchical Temporal Diffusion (HTD) for joint PV, electricity, cooling, and heat scenario generation. The framework contains three coupled components. A hierarchical temporal decomposer separates each channel into trend, periodic, and fluctuation components. A weather-conditioned coupling graph learns cross-channel dependencies from meteorological and temporal conditions. A bounded decision feedback module builds dispatch-risk pressure from electric-equivalent demand, import exposure, ramp stress, and PV surplus. It then corrects generated scenarios within anchor constraints and channel-wise trust regions. Experiments show that the TDD-WCG generator improves joint statistical quality, while HTD with BDF improves dispatch-related risk metrics with a controlled trade-off in marginal accuracy and coverage.
你可以做些什么？
你是怎样分析图片的？
你能帮我翻译这段英文吗？
Downloads
Article Details
Data Availability Statement
The data used in this study are publicly available from the HEEW multi-energy dataset. Additional materials generated during this study are available from the corresponding author upon reasonable request.