Abrasive waterjet cutting is used to cut sheets from a wide range of materials, including metals such as aluminum, hardened tool steel, titanium, copper, and brass, as well as ceramics and carbon fiber composites. It is applied in many industries, including energy, electric motors, watchmaking, aerospace, and consumer goods. Waterjet cutting is a flexible, sustainable, and environmentally friendly manufacturing process.
In abrasive waterjet cutting, the nozzle, usually made of sapphire or ruby, forms the jet by transferring energy from high-pressure water to a high-velocity stream. This nozzle is one of the most critical components affecting the technical and economic performance of a waterjet system. Because sapphire nozzles produce a jet with a defined shape and droplet spectrum after a certain distance from the exit, this engineered jet can be mixed with abrasive particles in the cutting head at an optimal ratio.
The resulting abrasive jet can cut hard materials such as metals, ceramics, and composites. However, these nozzles are prone to unpredictable failures after limited operating time. Cracks and subsequent breakage change the jet shape immediately, reduce cutting quality, or stop the process. For this reason, sapphire nozzles are often replaced preventively before reaching their full technical lifetime to avoid damage to the workpiece. This leads to frequent nozzle changes, higher processing costs, lower efficiency, and an unsustainable aspect of the technology.
To maintain consistent cutting results, there is a need for a sensor system that can detect imminent nozzle failure during operation. Such a system would prevent damage to parts and reduce unwanted machine downtime.
The SmartCut project develops an intelligent cutting head equipped with sensors to detect upcoming nozzle failure and enable full and safe use of each nozzle. Machine learning methods are applied to identify microcrack growth in the sapphire, which eventually leads to failure. Anomaly detection algorithms for time-series data are designed and tested. Thresholds are learned to distinguish between normal behavior, early-stage failure, and complete failure. In addition, a new sensor based on AI-supported predictions of suitable polymer structures is being developed and evaluated.
In waterjet cutting, abrasive material is mixed into a high-velocity waterjet to improve cutting performance. The quality of the cut strongly depends on the waterjet, which is shaped by a sapphire nozzle before passing through the mixing chamber of the cutting head. The nozzle typically lasts about 40 hours before failure due to crack-induced breakage. A failure immediately changes the waterjet shape and its ability to carry abrasive particles, requiring the process to stop. Unexpected nozzle failures during unmanned shifts have caused high maintenance costs.
These failures lead to scrap parts, unplanned machine downtime, and replacement costs for sapphire nozzles. To avoid this, industry often replaces nozzles preventively, sometimes before the end of their actual usable life. This results in increased costs and sustainability issues, including:
Material waste due to broken waterjet nozzles
Wasted nozzles from early preventive replacement
Unplanned downtime from unexpected nozzle failures
The project aims to develop an intelligent waterjet cutting head, the SmartCut head, with integrated AI-based online monitoring to detect sapphire nozzle failure. This allows the cutting process to be stopped before defective waterjets produce scrap, improving process sustainability by:
Reducing scrap rates and associated waste
Reducing machine downtime and increasing productivity
Shortening nozzle replacement time and costs
This is achieved through the integration of sensors, AI-driven anomaly detection models, and the development and testing of an intelligent AI-based polymer sensor.
This project is a ZIM-Project funded by the Germany Federal Ministry of Economic Affairs and Energy (BMWE).
