Hourly Clean Electricity Tracing and Masked Computation Support Enterprise Carbon Certification
Annual and regional emission factors obscure hourly and nodal variation in clean electricity availability. We developed a consumer-side carbon accounting workflow that combines hydropower dispatch, proportional-sharing electricity tracing and masked certification. In a modified IEEE 39-bus system, nodal clean electricity fractions ranged from 2.95 percent to approximately 100 percent, while three constructed industrial consumers had period-total fractions of 1.96–23.70 percent. A fixed Gaussian mask was shared among trustees and removed before release; a smaller hourly perturbation remained. Across 100 Monte Carlo runs, the released estimates had a mean absolute error of 2.200 MW and a period-total absolute error of 36.08 MWh, compared with 9.272 MW and 240.78 MWh for the tested local-noise baseline. The hybrid scheme also had a 49.24 percent mean absolute percentage error and a 9.33 percent clipping rate. These results support finer spatial attribution and lower absolute error under the selected settings, but do not establish a formal differential-privacy guarantee.
Authors
- Yumin Chen (ORCID: https://orcid.org/0000-0002-8672-8910)
- Lingyu Chen (ORCID: https://orcid.org/0000-0003-2331-5876)
- tianze yu (ORCID: https://orcid.org/0009-0005-8825-587X)
- Junhua Zhao
- Xiangxi Duan
- Hongli Liu
- Han Zhang
- Yang Wei
Institutions
- Chinese University of Hong Kong, Shenzhen (CN)
- State Grid Sichuan Electric Power Corporation (China) (CN)
Publication Details
- Journal
- Research Square
- Published
- 2026-11-20
- DOI
- https://doi.org/10.21203/rs.3.rs-11499776/v1
- Primary Topic
- Smart Grid Security and Resilience
- Type
- preprint