Variance-based global sensitivity analysis for power systems

Knowledge of the impact of uncertain inputs is valuable, especially in power systems with large amounts of stochastic renewable generations. A global sensitivity analysis (GSA) can determine the impact of input uncertainties on the output quantity of interest in a certain physical or mathematical model. The GSA has not been widely employed in power systems due to the prohibitively computational burden. In this paper, it is demonstrated that, via the implementation of a basis-adaptive sparse polynomial chaos expansion, a GSA can be applied to the power system with numerous uncertain inputs. The performance of the proposed method is tested on both the IEEE 13-bus test feeder and the IEEE 123-node test system, in presence of a large amount of independent or correlated uncertain inputs. The possible application of a GSA on the basis of the basis-adaptive sparse polynomial chaos expansion in power systems are discussed in terms of various sensitivities. The findings cannot only be used to rank the most influential input uncertainties with respect to a specific output, such as variances of the nodal power, but also to identify the most sensitive or robust electrical variables such as the bus voltage with respect to input uncertainties.