Involved Organizations: scitis.io, ATEMAG, Fraunhofer IPA
Project Duration/Period: January – Dezember 2021
Brief Description: Optimization of tool usage time and process control with the help of virtual sensors
The Industrial Revolution is everywhere now. Small and medium-sized companies also have to consider how they can benefit from the developments and use the Industrial Internet of Things for themselves. The company ATEMAG has addressed this very question. ATEMAG develops, produces and sells aggregates for machining wood materials, aluminum and plastics on CNC machining centers. As a machine builder in the woodworking industry, ATEMAG’s goal is to continuously improve the uptime of its products. Since a holistic lifecycle management plays a central role in this, a powerful but also cost-effective solution is required.
In order to achieve a relatively cost-effective entry into Industry 4.0, low-resolution and thus low-cost sensors can be integrated into the production systems for process monitoring. – Why low-resolution sensors? High-resolution sensors cannot be integrated close to the process due to their high sampling rates and the associated power requirements, as they must be connected to the machine with a wired data and power line. In contrast, low-resolution sensors can be integrated close to the process using rechargeable batteries and transmit data wirelessly. One problem with low-resolution sensors, however, is that they do not have the data quality and quantity for strategies such as predictive maintenance or predictive quality. This is exactly where the ViSKI research project comes in.
For process monitoring in metal-cutting woodworking, a virtual sensor is being developed that can be supplied with data from a low-cost, low-resolution sensor system. By means of an AI it is possible to detect patterns even in low-resolution vibration data, which allow conclusions to be drawn about the condition of the tool and the machine. By constantly monitoring the condition of the mold, the mold can be changed precisely when the quality requirements can no longer be met. The data from the sensors is read out using a cloud plug and collected in the cloud. There, the data can be processed almost instantaneously. The AI automatically detects a machine stoppage, a normal processing operation, as well as machining with a blunt tool.
Extending the analysis to other sources of error such as defective bearings or incorrectly clamped tools is also conceivable.
Thanks to the low-cost, easy-to-integrate and effective recording of process parameters, the service life of the tools can be extended and, at the same time, rejects in production can be avoided. In addition, long-term monitoring enables predictive maintenance and quality control.