In many production processes, a consistent high quality measure plays a crucial role. Due to natural machine and material-related fluctuations, a modern quality control can only be guaranteed with variable and highly adaptive categorization methods. This task often can not be displayed by classical statistical methods. To meet this problem IngB RT&S developed and implemented a special neural net based quality control system for one of the largest paper recycler. This software filters automatically high, medium and unacceptable complex structures from individual paper products.
Our software makes it not only possible to guarantee online a consistent quality in the production process, but also - if necessary - to monitot and readjust superior production parameters. Due to the fact that our software can be set at any time in it's conditioning mode, new product lines can be placed as well without any problems, such as construction-related changes in the production system and its sensors can be contained.