GratKI – AI-supported deburring of components using a modular deburring machine and reinforcement learning

Motivation

In metalworking, deburring components is an essential process, but one that is often still manual, time-consuming, and prone to errors. Conventional automated solutions reach their limits because they either do not provide the required precision or are not economically viable for small and medium-sized production runs.

In addition, there is a lack of flexible systems that can be quickly adapted to different workpiece geometries and quality requirements. There is therefore an urgent need for an innovative solution that improves product quality while significantly reducing process times and costs.

Objective

The objective of GratKI is to develop a globally novel, modular, AI-supported deburring machine. At its core is an intelligent quality control system that uses reinforcement learning to dynamically adjust process parameters and continuously monitor quality.

An intuitive CAM software solution will also enable new deburring processes to be programmed in the shortest possible time with just a few mouse clicks. The solution combines maximum precision with the flexibility required in industry for a wide range of components and batch sizes, helping to reduce scrap rates and make production more sustainable.

Approach

The approach begins with a detailed analysis of the requirements for modern deburring processes. Based on this, a high-precision, 5-axis CNC deburring machine with an innovative automatic grinding system will be developed.

BIBA will implement the AI-supported quality control system and optical inspection for dynamic process guidance. In parallel, an easy-to-use software solution will be created that massively reduces programming effort.

Finally, the overall system will be tested and validated using selected practical examples in order to demonstrate efficiency gains and economic viability under real-world conditions.

Project Duration

04/2026 – 04/2028

Funded by: