Authors:
Dharni Patel, Urvi Deore, Y. Anantha Vishwa Priya, M. Mohamed Sameer Ali, Luigi Orlando Freire Martínez, Jaime Alfonso Flores Navas

Addresses:
Department of Artificial Intelligence, Faculty of Engineering and Technology, SRM Institute of Science and Technology, Ramapuram, Chennai, Tamil Nadu, India. Department of Computer Science and Engineering, Dhaanish Ahmed College of Engineering, Chennai, Tamil Nadu, India. School of Engineering and Applied Sciences, Technical University of Cotopaxi, Latacunga, Cotopaxi, Ecuador. Faculty of Sciences, National Autonomous University of Mexico, Mexico City, Ciudad de México, Mexico.

Abstract:

The economic development sector's industrial operations have struggled with carbon emission monitoring and regulatory compliance, especially as India prepares to implement its Carbon Credit Trading Scheme (CCTS) by October 2026. VayuCredit analyses Indian companies' carbon emissions in different CCTS-regulated regions using three machine learning models. The World in Data Global Carbon Project trained all three models: Scale-invariant classifiers predict emission trends and make CCTS-aligned recommendations, Copula-Based Outlier Detection models find unsupervised anomalies, and XGBoost regressors with log-transformed targets and Optuna hyperparameters. From 1990 to 2022, this collection has 7,748 records from 235 nations. Researchers avoid cross-country data leakage by generalising across unobserved national emission trends using GroupKFold cross-validation. VayuCredit's real-time CCTS compliance module meets MoEFCC's October 2025 and January 2026 GEI targets for cement, aluminium, chlor-alkali, pulp and paper, petroleum refining, petrochemicals, and textiles. GHG Protocol Tier 1 emission parameters for India are based on CEA 2023 grid data and BEE benchmark. The compliance engine calculates Indian Rupee profits and penalties for Carbon Credit Certificates (CCCs) at a market price of ₹830-1,000 per tonne CO₂e. Double market-rate BEE fines. Emission computation, ML prediction, anomaly detection, trend recommendation, and CCTS compliance evaluation are FastAPI REST API endpoints. All three multi-models combined national-scale training data with company-scale industrial inference to exceed accuracy standards in industrial carbon monitoring and compliance management testing.

Keywords: Anomaly Detection; Unsupervised Learning; Random Forest; Carbon Emission Prediction; Carbon Credit Trading Scheme (CCTS); Economic Development; Carbon Credit Certificates (CCCs).

Received on: 09/03/2025, Revised on: 16/05/2025, Accepted on: 05/08/2025, Published on: 03/03/2026

DOI: 10.69888/FTSESS.2026.000693

FMDB Transactions on Sustainable Environmental Sciences, 2026 Vol. 3 No. 1, Pages: 43-53

• Views : 28
• Downloads : 8

Download PDF