A Gradient Boosting Decision Tree Algorithm for An Adjustable Capacity Market Clearing Model

With the proposal of carbon peak and carbon neutrality, it is of great importance to take effective methods to deal with the high penetration of renewable energy, which may pose a threat to the reliability of the power grid. One promising way is to establish the power capacity market where the generation units are compensated for their capacity characteristics. In this paper, an adjustable capacity market model is developed, where renewable energy is consumed with priority. Meanwhile, the thermal power units and hydropower units obtain capacity fees for flexibility upgrading. The paper also considers capacity demand, clearing method, and transaction method. In addition, considering the significance of predicting capacity demand with high precision, the paper applied Gradient Boosting Decision Tree (GBDT) algorithm to forecast the demand. A real case is studied and analyzed numerically, whose results show the validity of the proposed model. hope that this paper can provide some useful thinking for the construction of the power capacity market.