A novel day-ahead scheduling model to unlock hydropower flexibility limited by vibration zones in hydropower-variable renewable energy hybrid system

Giving full play to the flexibility of hydropower and integrating more variable renewable energy (VRE) is of great significance for accelerating the transformation of China's power energy system. For middle- to large-sized hydropower units, its irregular vibration zone (VZ) is a major factor affecting its flexibility, because VZs limits the power output and adjustment range. In the existing studies on hydropower unit commitment and flexibility scheduling, VZs only limits the output of hydropower units, while the impact on the flexibility adjustment ability of units is ignored. The hydropower flexibility in day-ahead scheme is overestimated, which increases the risk of system flexibility shortage causing power curtailment and load loss. In this study, a novel day-ahead scheduling model considering the flexibility limited by the VZs and the probability of flexibility shortage is constructed for the day-ahead generation scheme of hydropower-VRE hybrid generation system (HVHGS) to optimize the target load, VRE output and hydropower unit commitments. The impaction of the VZs on hydropower flexibility is firstly modeled by chance constraints and stochastic programming method. Moreover, a data-driven model based on machine learning and an efficient solving approach based on successive linear programming is carry out to describe the uncertainty of VRE output more realistically and ensure the timeliness of optimization scheme, respectively. The proposed model is applied to a real hydropower station in the Hongshui River Basin in China. In 16 representative scenarios, the proposed model can complete the optimization in an acceptable time, with a maximum of 444.57 s. Compared with the traditional interval optimization model, the proposed model effectively improves the flexibility supply capacity of hydropower in the day-ahead scheme. The maximum reduction value of flexibility shortage probability and expectation reach 98.77% and 442.66 MW, respectively. In particular, the flexibility of the model is most obvious under heavy load demand in flood season, and it is almost not at the cost of daily target load adjustment, which provides practical reference for decision makers.