The main goal of this task is to develop a monitoring system of the AFP process with the objective to detect the most common defects like gaps, overlaps, twisted or folded tows and wrinkles and predict how these defects will affect the next manufacturing stage of the process and the properties of the final part. The next images depict the most typical defects of the AFP process. To predict how these defects will affect the permeability of the final preform and, in turn, the filling time during the resin transfer moulding (RTM) process injection, new monitoring systems were added to the AFP head. Specifically, a thermal camera to obtain the profile temperature of each tow and a profilometer to acquire a 3D point cloud representation of each tow of the laminate. Reduce order models (ROMs) have been deployed in an edge device to simulate the effect of these defects and predict the permeability of the preform and the filling time of the RTM process. These ROMs received real time data from the AFP process (including the new monitoring systems) as an input to carry out their simulations.