Prediction the Carbon Emission Intensity of Coal-fired Power Plant Based on Electricity Data

The coal consumption rate for power supply serves as a crucial metric for evaluating the efficacy in energy conservation and consumption reduction of coal-fired power plants. Within the strategic framework of China's Carbon Peaking and Carbon Neutrality Goals, and to better facilitate the formulation of low-carbon dispatch strategies for the power grid, a precise understanding of the carbon emission intensity of coal-fired power plants under diverse operational statuses is imperative. This study comprehensively analyzes a spectrum of factors influencing carbon emission intensity and, employing net electricity generation supplied to the grid as a principal input variable in synergy with historical operational data sourced from operational power plants, proposes a predictive model for the carbon emission intensity of coal-fired power plants based on Bayesian-optimized LightGBM. The proposed model's efficacy is subsequently validated using comprehensive annual operational and carbon emission datasets from six conventional coal-fired power plants at a thermal power plant in Guangdong Province. Empirical validation demonstrates that the developed model not only facilitates real-time forecasting of carbon emission intensity for coal-fired power plants but also elucidates an overall inverse correlation between the unit's carbon emission factor and its installed capacity; notably, for analogous levels of power supplied to the grid, units with greater installed capacities exhibit a higher carbon emission intensity. The outcomes of this investigation furnish a robust theoretical framework and essential technical support for the realization of low-carbon, sophisticated dispatch within power grid operations. Moreover, this work offers salient insights conducive to mitigating aggregate carbon emissions from coal-fired power generation and augmenting the grid integration of renewable energy resources.