Multi-timescale coordinated optimization of hydro-wind-solar-hydrogen integrated system operator considering prediction error

Uncertainties from source and load create challenges for the integrated system dispatch. To mitigate the adverse effects of prediction errors on dispatch operations, this study proposes a two-stage multi-timescale optimal dispatch model based on adjustable robust adaptive model predictive control (ARAMPC), thereby enhancing the dispatch accuracy. A Stackelberg game model is developed for the day-ahead stage to capture the conflicting interests of the integrated system operator (ISO) and energy user (EU) is established. Positioning the ISO in the leading role and the EU in the following role, the model seeks to maximize the economic benefits of both parties. Furthermore, an adjustable robust optimization approach is introduced to handle source-load uncertainty. During the intra-day stage, a double closed-loop MPC with adaptive step size (DCAS-MPC) model is established. The double closed-loop mechanism is formed by the forecast errors of stochastic variables, including load and renewable energy, as well as prediction errors of ISO operating revenue to adaptively adjust the rolling time step. This not only improves the dispatch efficiency but also corrects the day-ahead dispatch deviation of the ISO. Finally, a case study demonstrates that the proposed ARAMPC model enhances the economic performance by 7.47% compared with the single day-ahead dispatch model. Moreover, compared with the single-feedback adaptive variable-step MPC model and the traditional MPC, the economic efficiency of the DCAS-MPC proposed in this study is improved by 0.62% and 0.73% respectively. In addition, by setting an appropriate adjustable coefficient, the trade-off between security and economy can be realized.