Overview

In the AICUT project, the ISSE is conducting research on the automated detection of process disturbances and quality deviation in machining production. Currently used methods such as envelope and trend analyses for process monitoring based on in-process variables (e.g. recorded mechanical loads on the machine spindle) do not allow comprehensive, precise and fine-grained quality assurance, making time-consuming and cost-intensive post-process measurements (e.g. of manufactured, geometric variables on the component) necessary in most cases. In this project, it will be researched whether quality monitoring can be improved by using machine learning. Thus, on the basis of in-process variables recorded by the spike measurement technology of the company pro-micron, it should be possible to draw conclusions on post-process variables without having to measure them. On the basis of a highly automated learning phase with coupled measurement of in-process and post-process variables, the system is to approximate this mapping so that measurement efforts can be reduced during its later use in production.

Funded by

Partners