A digital twin of the heat supply network in managing domestic energy demand

With its production of 440 million metric tons and exports of 211.2 million metric tons in 2025, Russia’s coal industry remains unprofitable; therefore, stability of the domestic demand for thermal coal is of critical importance. This article presents a quantitative assessment of the thermal component in the domestic coal market, as well as a systematic error of the current normative method for planning fuel requirements for heat supply, and the economic benefits of transitioning to adaptive forecasting of heat consumption based on a digital twin of the heating network and AI-simulated climate scenarios. Based on the assessment results, the domestic heat demand amounted to 30.2 million metric tons of coal equivalent, or 50.2 million metric tons of natural coal, which corresponds to 28.2% of the domestic market and 11.3% of the annual production. The weather-related variation in coal demand was calculated to be 2.4 million metric tons with a standard deviation, which is comparable to the effects of the price and logistical factors. It was found that the current normative planning method systematically overestimates the calculated load. The average absolute percentage error of the normative method was 8.2%, compared to 4.3% for the inertial method. Transition to adaptive forecasting based on a digital twin of the heating network will reduce the planning error by at least twofold and will provide savings in the heating fuel costs of approximately 3–6%. The results obtained are applicable at three levels: (1) for heat supply organizations; (2) for coal mining companies; (3) for regulatory authorities. Further research will focus on verifying the model coefficients using data from the system’s pilot operation, expanding the analysis to the municipal level, where climatic variations within a region are most pronounce, and with account of scenarios for provision of natural gas supply to coal-mining regions.