EcoChainAI: Multi-Facility Industrial Telemetry and Carbon Emission Benchmark Dataset Suite

The EcoChainAI Benchmark Dataset Suite is a multi-facility temporal telemetry and greenhouse gas (GHG) emission benchmark consisting of 100,000 continuous time-series samples across a 90-day simulated operational period (10-second sampling resolution). It captures three primary industrial domains:1. Thermal Energy Generation Facility (35,000 samples): Modeled on the UCI Gas Turbine and US EPA CAMPD CEMS benchmarks.2. Chemical Processing Facility (35,000 samples): Based on the Tennessee Eastman Process (TEP) benchmark across 52 process variables.3. Automated Manufacturing Facility (30,000 samples): Modeled on the Skoltech Anomaly Benchmark (SKAB) IoT testbed. Key Features:- Comprehensive feature attribution across Scope 1 (direct stack), Scope 2 (indirect electricity/steam), and Scope 3 (logistics/value chain) emissions as per the WRI GHG Protocol.- Physics-aware mass and energy balance conservation residuals.- 15% injected industrial cyber-physical attack scenarios (sensor drift, freeze, multiplicative spikes, and replay attacks).- 16,657 aggregated sliding window batches (Tw=60, Sw=6) with SHA-256 Merkle root hashes for blockchain oracle verification.