Deriving adaptive long-term complementary operating rules for a large-scale hydro-photovoltaic hybrid power plant using ensemble Kalman filter

•An adaptive complementary operating rule with time-varying parameters was proposed.•State transition and observation from the same period of previous year are suitable.•The proposed method is still effective for nonstationary reservoir inflow and solar radiation. The complementary operating rules of hydro-photovoltaic (PV) hybrid power plants have been explored fully to operate reservoirs for complementing the PV output. However, prevailing operating rules were constant, which is incapable to be dynamically updated based on new observations. To address this issue, adaptive complementary operating rules were developed to deal with the future uncertainties of reservoir inflow and solar radiation using ensemble Kalman filter (EnKF). First, a long-term complementary optimization model was developed to derive the initial parameters of linear operating rules. Then, the optimal decisions under historical condition were used as the observation. Finally, six assimilation schemes were proposed, according to the two state transition equations (the same period of the previous year, the previous period) and three observations (the same period of the previous year, the previous period and their joint). China’s Longyangxia hydro-PV hybrid power plant was chosen as a case study. Results show that (1) the two state transition equations have comparative effects on operation results, (2) the choice of observation plays a decisive role. The assimilation scheme, in which the optimal operation decision of the same period in previous year is as observation, outperforms the stationary operating rules. Specifically, the total power generation and guaranteed rate are improved by 7.14% and 5.33%, respectively, (3) the deriving framework works effectively under the synthetic nonstationary reservoir inflow and solar radiation, and ensures the stability and security of energy system in the changing environment.