Research on Intelligent Carbon Emission Calculation Models for Enterprises Based on AI Large Model Data Driven Approach

CHEN Zhang; RUI Zai; CHENG Jin; BOJUN Kong; JINFENG Zhang

English Version Reader LoginJournal Login

您当前的位置：

首页 >

文章列表页 >

Research on Intelligent Carbon Emission Calculation Models for Enterprises Based on AI Large Model Data Driven Approach

更新时间：2025-09-25

- Research on Intelligent Carbon Emission Calculation Models for Enterprises Based on AI Large Model Data Driven Approach
- Issue 5, (2025)
- 作者机构：
  
  国网江苏省电力有限公司扬州供电分公司
- 作者简介：
- 基金信息：
- DOI：
  CLC：
- Published：2025
- 稿件说明：
移动端阅览
CHEN Zhang, RUI Zai, CHENG Jin, et al. Research on Intelligent Carbon Emission Calculation Models for Enterprises Based on AI Large Model Data Driven Approach[J]. 2025, (5).
DOI：

CHEN Zhang, RUI Zai, CHENG Jin, et al. Research on Intelligent Carbon Emission Calculation Models for Enterprises Based on AI Large Model Data Driven Approach[J]. 2025, (5). DOI：

摘要

随着我国"双碳"目标推进进入关键时期，建立精准、实时的碳排放监测体系成为实现碳达峰、碳中和的关键支撑。针对现有企业碳排放监测方法存在数据源异构、时效性差、精度有限等问题，本文提出基于AI大模型数据驱动的企业碳排放智能计算模型。该模型利用AI大模型的深度语义表示能力，构建多源异构数据智能融合框架，实现电力负荷、生产参数、气象数据等的统一处理和质量优化；基于信号时频分析建立单一总电表负荷智能分离模型，精确剥离静态基准、照明、温控和生产负荷；构建ARDL智能能耗推算模型，实现"以电折能"的高精度预测；结合动态电碳排放因子和绿电交易证书修正，建立企业碳排放量智能核算体系。以某大型汽车零部件制造企业为例验证模型有效性，结果表明AI数据驱动框架显著提升了数据质量，负荷分离识别精度良好，ARDL模型预测精度高，系统可实现分钟级实时碳排放计算，年度核算误差较小。该研究为企业碳排放智能监测提供了数据驱动的技术解决方案，对推动"双碳"目标实现具有重要意义。

Abstract

As China"s "dual carbon" goals enter a critical implementation period

establishing precise and real-time carbon emission monitoring systems has become crucial for achieving carbon peak and carbon neutrality. Addressing the challenges of heterogeneous data sources

poor timeliness

and limited accuracy in existing enterprise carbon emission monitoring methods

this paper proposes an intelligent carbon emission calculation model for enterprises based on AI large model data-driven approach. The model leverages the deep semantic representation capabilities of AI large models to construct an intelligent fusion framework for multi-source heterogeneous data

achieving unified processing and quality optimization of electricity loads

production parameters

and meteorological data. An intelligent load decomposition model based on signal time-frequency analysis is established for single master meters to precisely separate static baseline

lighting

temperature control

and production loads. An ARDL intelligent energy consumption estimation model is constructed to achieve high-precision "electricity-to-energy" prediction. Combined with dynamic electricity carbon emission factors and green electricity certificate corrections

an intelligent accounting system for enterprise carbon emissions is established. The model effectiveness is validated using a large automotive parts manufacturing enterprise as a case study. Results demonstrate that the AI data-driven framework significantly improves data quality

achieves good load separation identification accuracy

exhibits high ARDL model prediction accuracy

enables minute-level real-time carbon emission calculations

and maintains low annual accounting errors. This research provides a data-driven technical solution for intelligent enterprise carbon emission monitoring

which is of great significance for promoting the achievement of "dual carbon" goals.

关键词

Keywords

references

姚顺春, 支嘉琦, 付金杯, 等. 火电企业碳排放在线监测技术研究进展[J]. 华南理工大学学报(自然科学版), 2023, 51(6): 97–108.

Sur R, Sun K, Jeffries J B, et al. Scanned-wavelength-modulation-spectroscopy sensor for CO, CO2, CH4 and H2O in a high-pressure engineering-scale transport-reactor coal gasifier[J]. Fuel, 2015, 150: 102–111.

姚倩倩, 李新皓. 含高比例分布式资源的乡村配电系统有功-无功协同优化调控方法[J]. 电力大数据, 2024, 27(9): 19–30.

张智刚, 康重庆. 碳中和目标下构建新型电力系统的挑战与展望[J]. 中国电机工程学报, 2022, 42(8): 2806–2819.

贾彦鹏. 碳排放统计核算体系建设背景下完善电碳因子测算的措施研究[J]. 中国经贸导刊, 2024(17): 56–58.

崔杨, 邹新鹏, 赵钰婷, 等. 考虑动态电碳排放因子的新型电力系统电碳综合需求响应调度方法[J]. 电力自动化设备, 2024, 44(10): 1–7.

张倩. 环境监测大数据驱动的化工企业碳排放智能核算[J]. 中国轮胎资源综合利用, 2025(5): 162–164.

张东霞, 苗新, 刘丽平, 等. 智能电网大数据技术发展研究[J]. 中国电机工程学报, 2015, 35(1): 2–12.

胡波, 邹洪, 郭瑞鹏. 基于电力网络监测数据的大数据安全分析平台关键技术研究[J]. 电力大数据, 2021, 24(3): 51–58.

侯昝宇, 蔡亦浓, 刘博, 等. 基于电碳计算模型的区域与行业碳排放监测应用实践分析[J]. 数字技术与应用, 2024, 42(11): 85–87.

萧展辉, 柴雁欣, 李学龄, 等. 面向多源异构数据的语义对象识别技术研究[J]. 自动化技术与应用, 2024, 43(12): 103–106.

邹明明, 陈建军, 龚顺子. 基于大模型的多源异构数据集成与清洗方法[J]. 数字技术与应用, 2025, 43(3): 217–219.

闫玮丹, 齐冬莲, 闫云凤, 等. 面向电力领域的知识图谱与大模型融合关键技术及其典型应用[J]. 高电压技术, 2025, 51(4): 1747–1762.

刘天蔚, 边晓燕, 吴珊, 等. 电力系统碳排放核算综述与展望[J]. 电力系统保护与控制, 2024, 52(4): 176–187.

李峰, 王琦, 胡健雄, 等. 数据与知识联合驱动方法研究进展及其在电力系统中应用展望[J]. 中国电机工程学报, 2021, 41(13): 4377–4390.

邓晓平, 张桂青, 魏庆来, 等. 非侵入式负荷监测综述[J]. 自动化学报, 2022, 48(3): 644–663.

李青青, 苏颖, 尚丽, 等. 国际典型碳数据库对中国碳排放核算的对比分析[J]. 气候变化研究进展, 2018, 14(3): 275–280.

He H, Zhou S, Zhang L, et al. Dynamic Accounting Model and Method for Carbon Emissions on the Power Grid Side[J]. Energies, 2023, 16(13): 5016.

Tranberg B, Corradi O, Lajoie B, et al. Real-time carbon accounting method for the European electricity markets[J]. Energy Strategy Reviews, 2019, 26: 100367.

Chen X, Liao Z, Gao Z, et al. A Calculation Model of Carbon Emissions Based on Multi-Scenario Simulation Analysis of Electricity Consumption[J]. Sustainability, 2022, 14(14): 8765.

Gao B, Kong X, Liu G, et al. Monitoring high-carbon industry enterprise emission in carbon market: A multi-trusted approach using externally available big data[J]. Journal of Cleaner Production, 2024, 466: 142729.

Ling C, Yang Q, Wang Q, et al. A comprehensive consumption-based carbon accounting framework for power system towards low-carbon transition[J]. Renewable and Sustainable Energy Reviews, 2024, 206: 114866.

尹利, 于波, 马飞, 等. 基于深度学习算法的非侵入式工业负荷分解方法[J]. 电工电能新技术, 2022, 41(5): 61–70.

武昕, 于金莹, 彭林, 等. 基于用户边缘侧事件解析的工业电力负荷非侵入式感知辨识[J]. 电力系统自动化, 2021, 45(4): 29–37.

张逸, 姚文旭, 邵振国, 等. 基于电能质量监测数据的企业环保异常工况识别[J]. 电力系统自动化, 2023, 47(5): 180–189.

曾梓铭, 魏德样, 张文静. “双碳”目标下中美体育用品进出口贸易及对碳排放的影响[J]. 成都体育学院学报, 2024, 50(6): 79–91.

郑义, 秦炳涛. 能源消费、碳排放与经济增长的关系研究——基于中国(1970-2010年)ARDL模型的实证分析[J]. 科研管理, 2016, 37(8): 130–139.