Exploring adaptive interpolation to mitigate non-linear impact on estimating dynamic states

Accurate estimation of dynamic states is important for monitoring and controlling transient stability. This paper proposes an adaptive interpolation approach to improve the performance of the Extended Kalman Filter (EKF) for estimating dynamic states of a synchronous machine. This approach consists of two major steps. First, the non-linearity of state transition function and measurement function is quantified. Second, when the non-linearity is severe, pseudo measurements are added through interpolation to mitigate the negative impact of non-linearity on the estimation accuracy. Using the 2-area model, it is shown that the proposed adaptive interpolation approach can make a good tradeoff between estimation accuracy and computation time when estimating the dynamic states of a synchronous machine.