Optimization of Real-Time Carbon Emission Monitoring on the Power Demand Side Based on System Dynamics

Unofficial AI-generated summary based on the public title and abstract. Not an official translation.

🔬研究者:Demonstrates a novel integration of system dynamics with machine learning for high-resolution carbon monitoring, offering a replicable framework for demand-side emission tracking.

🏢実務担当者:Provides a technical pathway for real-time carbon monitoring at minute-level granularity, enabling better operational decisions for energy-intensive industries and grid operators.

🏛政策担当者:Highlights the importance of high-temporal-resolution monitoring for effective carbon market design and grid optimization, informing policy on data infrastructure for emission reduction.

Against the backdrop of China’s “dual carbon” strategic goals, the precise monitoring of carbon emissions on the grid’s demand side is crucial for advancing the transformation of the energy structure and achieving low-carbon development. Current mainstream hourly-level monitoring methods have insufficient temporal resolution, coarse information granularity, and weak localization data support. These issues compromise the effectiveness of carbon reduction efforts and may even lead to secondary risks and economic losses. As a major province with high energy-consuming industries, Henan Province experiences frequent short-term high-energy consumption events in its heavy industry sector, where carbon emission characteristics are complex, making traditional monitoring methods inadequate. To address these challenges, this study focuses on thedemand side of the power grid in Henan’s heavy industry. It specifically analyzes the carbon emission characteristics of high-energy-consuming equipment and industrial electricity usage patterns during short-term high-energy events. Innovatively, a system dynamics model integrating STL, LightGBM, and a stream computing subsystem is constructed. This model not only achieves a technical upgrade from hourly to minute-level carbon emission monitoring but also fully considers the internal structures and coupling relationships among subsystems. Empirical research demonstrates that the model possesses the capability for minute-level data collection and dynamic updates of carbon emission factors, significantly improving the temporal resolution and response efficiency of the monitoring system. It enables high-frequency dynamic tracking of instantaneous load during industrial electricity consumption. These results provide robust data support for grid dispatch optimization and carbon market trading. The model enhances the accuracy of carbon emission predictions and offers reliable foundations and new optimization directions for exploring technological pathways toward energy conservation and emission reduction. Its application holds significant theoretical value and practical importance for promoting the modernization of regional energy governance, optimizing industrial structure, and formulating scientifically sound policies.

• openalex https://doi.org/10.32996/jbms.2026.8.6.2first seen 2026-05-05 19:11:39