Improving Day-Ahead Grid Carbon Intensity Forecasting by Joint Modeling of Local-Temporal and Cross-Variable Dependencies Across Different Frequencies

2026-01-10

arXiv.org

AI Summary

This study addresses the challenges in day-ahead forecasting of grid carbon intensity factors (CIFs), which arise from local temporal dependencies, dynamic high-order variable interactions, and complex multi-frequency patterns. To tackle these issues, the authors propose a parallel dual-module joint modeling framework: one module employs multi-scale wavelet convolutional kernels to capture multi-frequency local temporal features, while the other models dynamic cross-variable dependencies in the time–frequency domain, enhanced by an interpretable attention mechanism that adaptively focuses on critical variables and time intervals during perturbation events. This approach is the first to jointly model multi-frequency temporal dynamics and dynamic cross-variable dependencies. Evaluated on electricity market data from four Australian regions with varying renewable energy penetration levels, the method consistently outperforms state-of-the-art models, with ablation studies confirming the complementary gains from the dual modules.

Technology Category

• Computer Vision: Multi-modal Vision
• Reasoning under Uncertainty: Relational Probabilistic Models
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Accurate forecasting of the grid carbon intensity factor (CIF) is critical for enabling demand-side management and reducing emissions in modern electricity systems. Leveraging multiple interrelated time series, CIF prediction is typically formulated as a multivariate time series forecasting problem. Despite advances in deep learning-based methods, it remains challenging to capture the fine-grained local-temporal dependencies, dynamic higher-order cross-variable dependencies, and complex multi-frequency patterns for CIF forecasting. To address these issues, we propose a novel model that integrates two parallel modules: 1) one enhances the extraction of local-temporal dependencies under multi-frequency by applying multiple wavelet-based convolutional kernels to overlapping patches of varying lengths; 2) the other captures dynamic cross-variable dependencies under multi-frequency to model how inter-variable relationships evolve across the time-frequency domain. Evaluations on four representative electricity markets from Australia, featuring varying levels of renewable penetration, demonstrate that the proposed method outperforms the state-of-the-art models. An ablation study further validates the complementary benefits of the two proposed modules. Designed with built-in interpretability, the proposed model also enables better understanding of its predictive behavior, as shown in a case study where it adaptively shifts attention to relevant variables and time intervals during a disruptive event.

Problem

• grid carbon intensity
• multivariate time series forecasting
• local-temporal dependencies
• cross-variable dependencies
• multi-frequency patterns

Innovation

• multi-frequency modeling
• local-temporal dependencies
• cross-variable dependencies
• wavelet-based convolution
• interpretable forecasting