A flexible operation scheme for ultra-supercritical unit under wide load variations based on improved EADRC and modified northern goshawk optimizer

To enhance the grid's capability to accommodate large-scale renewable energy integration and accelerate the construction of the new power system, it is imperative to improve the flexible operation capability of the ultra-supercritical (USC) unit across wide-load ranges. In this paper, an operation scheme based on the improved error-based active disturbance rejection controller (IEADRC) and modified northern goshawk optimizer (MNGO) algorithm is designed from the three aspects of dynamic modeling, control strategy and operation optimization to improve the flexible operation capability of the USC unit under wide load variations. Firstly, the modeling problem is reformulated as a constrained optimization problem in multidimensional coefficient space and solved via the developed high-performance MNGO algorithm, yielding high-precision transfer function matrix models for four typical operating conditions. Secondly, an IEADRC strategy that perfectly integrates the reduced-order observer, the robust term and the actuator rate-limiting compensation structure is proposed. Building on this, a multi-objective optimization scheme is constructed by considering the physical constraints during the operation of the unit, tracking accuracy, smoothness of control variables, the economic cost of the power generation process and the carbon emission. The optimal parameters of the controller are then derived based on the MNGO algorithm. The integrated control strategy combines superior tracking performance, enhanced disturbance rejection, and strong robustness. Finally, extensive comparative simulation experiments on a 1000 MW USC unit across 30–100 % boiler rated load substantiate the effectiveness and practicality of the proposed integrated operation scheme of modeling, control, and optimization. The simulation results demonstrate that, relative to ADRC, the integrated control strategy reduces the average settling time of the unit during wide load variations from 290.9s to 220.3s and lowers the average steady-state error under the deep peak shaving condition from 0.1236 MW to 0.0248 MW. Consequently, the scheme successfully establishes an effective reference framework for the flexible operation of USC units to meet deep peak shaving demands.