Katharina Lankers, R&D Scientist, Schott AG
Hot forming processes for glass depend on lots of parameters – a classical situation where statistical design of experiments (DOE) seems to be the appropriate tool for understanding correlations and finding best settings. However, in a production environment, planning and conduction of experiments has to be adjusted to available time slots and actual boundary conditions, which might change very quickly. As a consequence, DOE suitable for production requirements has to be flexible: Settings may have to be adapted quickly, and analyses have to be fast and easy if the results are used as a basis for the next test series. Moreover, the staff conducting the experiments might not be familiar with JMP and may want to use Excel sheets instead for entering results. This paper shows how we solved these problems for a glass hot forming process. We developed a couple of appropriate JSL scripts, allowing for rapid setup changes in experiments, compatibility with Excel templates for production staff and comfortable data visualisation and analyses. (And, by the way, we obtained good results for process optimisation!)
