<p>In order to improve the wind power consumption capacity of the power grid system and reduce the operating costs of the power grid, a hierarchical optimization strategy is proposed to integrate high-energy consumption loads as peak shaving resources into the power grid dispatch system. Firstly, based on the actual operational regulation characteristics of high energy consumption loads, a peak shaving regulation model for high energy consumption load power grids is constructed by comprehensively considering multiple variable constraints such as power balance, rotation reserve, and climbing speed. The regulation objectives and model constraints of the upper and lower layers are given. Secondly, this model is integrated with the deep peak shaving model of thermal power units, and a layered optimization calculation model is designed around the dual objectives of minimizing the curtailed wind energy (upper-level optimization) and operating costs (lower-level optimization) of the power grid system, focusing on the regulation of high-energy consumption load resources and deep integration of thermal power units. The experimental results show that the dual layer optimization model of source load coordination proposed in this study has a smoother overall output curve for thermal power units, which can significantly reduce the load peak valley difference. At the same time, the wind power plan has fewer limited output periods, lower maximum blocked power, and a reduced wind curtailment rate of 9.6%, which is 9.7% lower than the comparative scheme. The amount of abandoned wind power is reduced by half, and the wind power consumption level is significantly improved, which can improve the system’s operational economy.</p>