Zero defect manufacturing in metal forming can be achieved by combining real-time measurements with fast and detailed models in process control. Variability of metal forming processes is caused by many distinct sources like scatter in material and lubrication properties. These variations can usually not be measured directly. Knowledge about the state of the process can be maximized by relating indirect measurements and process models in the framework of probabilistic state estimation. By doing so, valuable information becomes available for process control, which may lead to unprecedented increases in production accuracy.
Model-based state estimation in precision manufacturing has a strong potential from scientific as well as from industrial perspective. The nature of process variability has been investigated from statistical point of view, but has been barely understood at product-to-product level. Process models for control have been developed, but they lack the levels of detail needed for precision manufacturing. Development of real-time state estimation of complex manufacturing processes is a big step forward to be taken. From the perspective of industry, state estimation and control may lead to a reduction of scrap rate and make finishing steps redundant, reducing production costs. Detailed process models can be used to guarantee production accuracy of flexible processes, complying to the need for mass customization.