An IIoT-enabled real-time ESG carbon intelligence and MRV framework for industrial facilities: from sensor-level emission measurement to auditable carbon-credit quantification

Corporate greenhouse-gas (GHG) inventories for manufacturing sites are still assembled from monthly utility invoices and engineering estimates, which limits both the granularity and the timeliness of environmental, social and governance (ESG) disclosure. This paper develops and evaluates an Industrial Internet of Things (IIoT) architecture that couples 15-minute machine-level electrical and thermal metering to a GHG accounting engine, a measurement-reporting-verification (MRV) layer and a conservative carbon-credit quantification module. Two contributions are formalised: a provenance-tier traceability index that propagates data quality into the reported inventory, and a deduction rule that links the creditable share of an emission reduction to both measurement uncertainty and traceability. The framework is evaluated on a physics-based one-year simulation of a twelve-machine, three-line discrete manufacturing plant (420,480 machine-interval records) containing degradation, telemetry loss, grid outages, injected faults and on-site photovoltaic generation. Against conventional periodic reporting, continuous monitoring reduced machine-level attribution error from 11.29 percent to 0.33 percent and reporting latency from a median of 90.6 days to 8.0 minutes, both significant under paired Wilcoxon tests. Three findings qualify the benefit. First, the annual facility-level inventory improved only marginally, because emission-factor uncertainty contributes over 92 percent of the total variance and is unaffected by better activity data. Second, unsupervised outlier detection on raw telemetry was close to useless (F1 = 0.121); the same algorithm on context-conditioned residuals reached F1 = 0.340, indicating that feature construction rather than model complexity governs performance. Third, day-ahead carbon-intensity forecasting failed for every model tested. The quantified benefit of granularity appears instead in the credit module, where a daily rather than monthly adjusted baseline reduced the conservative deduction from 14.33 percent to 6.24 percent, raising the creditable quantity by 21.8 tCO2e. All results are simulated; no hardware measurements are claimed.