Abstract: Against the backdrop of the "dual-carbon" goals and the new power systems, coal-fired power generation is accelerating its transition toward deep and flexible peak-shaving. However, when coal-fired units operate at low-load conditions, prominent challenges including substantial degradation in safety and reliability as well as sharp growth in coal consumption remain to be addressed. To this end, we systematically elaborate the theoretical methodology of thermodynamic system reconstruction, analyzes its influences on boiler combustion stability, hydrodynamic characteristics, heating-surface metal safety, and emission control under low-load conditions, and discusses its improvement effects on heat consumption at low loads. By establishing a correlation mechanism of "key performance-characteristic parameters-key-component structures", the proposed theoretical methodology employs modular thermodynamic system topology control and flow-field reconstruction to raise characteristic parameters such as feed-water temperature, main-steam pressure, and steam flow-field under low-load operating conditions to levels close to their rated values. Research on a 1000 MW ultra-supercritical unit shows that, at 30%~50% rated load range, this technology can increase feed-water temperature, denitrification (SCR) flue gas temperature, and hot air temperature by approximately 65 ℃, 44 ℃, and 30 ℃, respectively, significantly improving key performance indicators including low-load combustion stability, hydrodynamic and heating-surface safety, and SCR reliability. Meanwhile, it reduces the net coal consumption for power supply by 8~27 g/(kW·h), offsetting about 55% of the coal-consumption increment of conventional units at low-load operation. Conducting research on turbine ultra-low load operation technology based on this theory represents an important development direction for flexible coal-fired power thermodynamic systems.