Abstract: Aiming at the peak load demand increase and the uncertainty of new energy grid-connection, this paper proposes a bi-level economic operation method of energy storage and thermal power unit joint peak load regulation. The upper layer of the model takes into account the energy storage configuration and the flexibility of thermal power units, and the objective function is the comprehensive income of the system. At the same time, in order to mobilize the enthusiasm of energy storage to participate in peak load regulation, a compensation strategy based on the energy storage life model is proposed. The lower level of the model takes typical daily operating conditions as decision variables, with the objective function being the system operating cost considering the integration of renewable energy consumption. Then, the spectral normalization generation adversarial network (SNGAN) and K-medoids algorithm are used to generate typical scenarios to improve the accuracy of the model. The advantages of spectral normalization generation adversarial network are highlighted by the time correlation analysis and the comparison with historical data. Finally, the simulation results show that the proposed method can effectively take into account the economy and the ability of new energy consumption.