In this paper, we propose a robust structural investment and dispatch model of electric systems including commitment and storage constraints under auto-correlated residual demand. We associate it to a novel approach to robust optimization focusing on uncertainty parameter trajectories. Using Principal Component Analysis, we approximate conditional order statistics for the differential distribution of components of residual demand using parametric polynomial regression. This flexible method allows us to derive a set of extreme trajectories maximizing the level and variability of residual demand. Finally, we apply our dynamic robust model to the electric system of the French region Auvergne Rhône-Alpes and discuss the implications in terms of investment decisions and cost performance.