This thesis investigates the potential use of the dynamic wake model FLORIDyn for Annual Energy Production (AEP) estimations in large-scale wind farm applications. Traditional low-fidelity wake models, based on steady-state assumptions, fall short in capturing the transient nature of wake interactions under varying wind conditions. FLORIDyn addresses this by incorporating time-dependent wake effects. This study includes a review of existing models, modifications to FLORIDyn, and calibration procedure using an inverse Bayesian approach with surrogate models. On the calibrated model, a validation study against SCADA data revealed that FLORIDyn's predictions did not outperform steady-state models, mainly due to biases in flow convection and wake accumulation inaccuracies. While FLORIDyn shows promise for small wind farms with large turbine spacings, it requires significant refinement for broader industry application. The thesis concludes with recommendations to address the current deficiencies in both the FLORIDyn model and validation methodology, aiming to enhance AEP estimation accuracy and provide more reliable wind farm performance assessments.