Due to the increasing deployment of intermittent renewables, the residual load profile, as seen by the dispatchable generation units, becomes lower and more volatile. This paper introduces a new system planning model on a power plant resolution, taking into account technical operational constraints. The objective of this model is to determine the optimal set of generation units, able to serve a given demand. Two initial solutions are obtained; one from a classical screening curve model, and another from a model using mixed integer linear programming (MILP). These initial solutions are perturbed and combined with an operational model to validate and further improve the solution. The developed model complements other models available from the literature, in its level of detail (power plant level and full year -hourly time resolution) combined with a fast computation time (< 1 h). The evolution of the optimal amount of generation capacity as a function of the installed wind capacity is examined in a case study. As the share of wind power increases in the portfolio, a shift takes place from base load generation towards mid and peak load. This shift is triggered both by the lower demand and the increasing volatility. This demonstrates that operational constraints of power plants (individual basis) have an important impact on the configuration of the optimal generator set, and need being considered, especially at increasing rates of intermittent renewables.