A structural time series approach to modeling dynamic trends in power system data

Structural time series models provide a natural framework for modeling time-varying trends in measured data. In this paper, a statistical framework for analyzing and estimating time-varying trends in measured data is developed. In this model, temporal patterns in measured data are modeled within a stochastic state space setting. Estimates of the time-varying parameters are then obtained using an optimal estimation method based on Kalman filters and associated smoothers. Both, synthetic and observational data are used to assess the predictive capability of the model. Results are compared to other detrending techniques in order to assess the potential of the methodology.