Surrogate-Model-Based Sequential Algorithm for Weather-Dependent Probabilistic Power Flow With High Calculation Efficiency

 Recently, a novel power flow model called weather dependent power flow (WDPF) has been proposed to incorporate conductor temperatures in power flow calculation. It is promising to adopt WDPF model in probabilistic analysis since it provides more accurate electrical and thermal states of a power system. However, calculation of the WDPF model is more computationally expensive than that of conventional power flow model, and hence repetitive calculation of the WDPF in probabilistic analysis is quite time-consuming. To overcome this challenge, a surrogate?model-based sequential algorithm is proposed for calculation of the WDPF model with high computational efficiency and good accuracy. Compared to the method that directly learns the whole WDPF model, the proposed method only learns the power flow sub-model whereas applies a linearization method to solve the thermal resistance sub-model. The construction of the proposed method considers the learning difficulty of the sub-model and the whole WDPF model, thus ensuring the high calculation accuracy. This study is the first to conduct a probabilistic power flow analysis based on the WDPF model. Numerical results in the IEEE 30-bus, 57-bus, and 118-bus systems demonstrate the importance of adopting the WDPF model in probabilistic analysis and the efficacy of the proposed method.