Under the “dual-carbon” goals, the construction of a new power system imposes requirements for deep peak regulation and flexible operation on conventional coal-fired power units. However, due to their inherent characteristics, coal-fired units face challenges such as degraded stability of thermal control systems when operating over a wide load range. Traditional control strategies struggle to adapt to complex operating conditions, necessitating a novel control paradigm. This paper proposes a digital twin–driven self-evolutionary optimization framework for wide-load thermal control systems of coal-fired power units. By deeply integrating digital twin technology, a high-fidelity virtual unit model is constructed to achieve real-time mapping and accurate prediction of the full life-cycle states of the physical entity. On this basis, a self-evolutionary mechanism is designed that integrates online identification and intelligent decision-making functions. The system can autonomously diagnose performance bottlenecks based on real-time data and historical experience, dynamically adjust control parameters, evaluate strategies through pre-simulation within the digital twin, and finally deploy optimized strategies to the physical system, forming a complete closed loop. This framework provides critical support for the efficient and stable wide-load operation of coal-fired power units and for the intelligent and green transformation of thermal power generation.
Wang Hong-Wei (Wed,) studied this question.