Numerical technique for the space discretization of resin infusion mould sensing with artificial vision

On-line control strategies are used to ensure full reliability of Resin Infusion processes. The top half of the mould is transparent or translucent, allowing the use of artificial vision as a sensor for mould filling processes. It has been recently introduced for Liquid Composite Molding processes [1], [2], [3]. In this sense, the CCD sensor can be considered as a matrix of nodes that produces a space discretizacion. It implies that it is possible to relate the regions defined by the finite elements used in the process numerical simulations and the image. Therefore, it is possible to define a single camera vision sensor as a space discretizer and to relate it directly with on-line Finite Element discretization. Also is possible to relate an off-line mesh with camera vision sensor. In both cases, a direct relationship between the FEM or proxy simulations and the real process evolution can be defined. The mould complexity for real 2.5D resin infusion process can be easily solved including multiple cameras and calibrated them with Stereovision techniques [4]. As a result, a 2.5D mould mesh is obtained where the camera vision acts as Finite Element mesher and process sensor simultaneously.