A product data model and computer aided manufacturing for the process chain of profiling

The process chain for the production of profiled strips, beams and panels includes a number of processes as mould design, fixing the number of cutting passes, tool grinding, tool measurement, moulding, sanding, profile-wrapping, coating and different types of quality assessment. This chain is still not consistently connected by means of information technologies. The design and the setup for each process in the chain are mostly done manually, using the experience of the operator. Computer aided systems (CAD/CAM) as widely used in the furniture industry are not in this field. The procedures are more complex and could not be automatized by the state of the art feature based algorithms. One attempt to solve the problem is the creation of a product data model that contains all information for the whole production process - not only geometrical data and tolerances but also semantic information (e.g. material, coating, batch size, etc.). Every process in the chain is now transforming the profile in a systematic and stochastic way. Both can be computed by the use of rules of artificial intelligence and tolerance calculations by means of statistical methods. Special CAM modules, which generate the specific process data, give also back the transformed profile and other necessary information to the superordinate product data model. The actual specific process data, for example for the process of profiling (e.g. setting parameters, pitch) can then be computed by a post processer using actual inputs as the geometry of the specified machine. The whole process chain can so be simulated before the real setup. The tools and processes can so be designed simultaneously in advance. The time for the production process can be reduced. This systematic is generating the same potentials as feature based CAD/CAM systems in the carcase furniture industry. The implementation of IT in the profiling process chain is accelerated and coordinated by the introduction of the linking product model.