Adaptive process control during grinding with sensor integration close to the point of action

Generating grinding poses particular challenges and problems, as it involves working with a geometrically indeterminate cutting edge. This means that the abrasive grains are stochastically distributed on the grinding wheel, which leads to unpredictable machining effects. In addition, the abrasive grains change due to wear and conditioning, which affects the efficiency and accuracy of the process.

To ensure process reliability, high safety factors are often applied and the technological parameters are kept constant. However, this approach leads to reduced productivity as the possibilities for optimization are not fully exploited. At the same time, the reduced tool life of the grinding wheels leads to increased tool costs.

Process reliability is also impaired by the fact that there are no direct measured variables for monitoring the condition of the grinding wheel. This means that the exact condition of the grinding wheel remains unknown during the process, which further impairs efficiency as well as process reliability. This uncertainty prevents the potential for process optimization of the grinding wheel and machine from being fully exploited.

There is therefore still considerable potential for optimization in generating grinding, which could be realized through improved monitoring mechanisms and more precise control of the grinding process.

Sensor integration close to the point of action can make a significant contribution to optimizing processes in the challenges of generating grinding. As part of the Fraunhofer Cluster of Excellence Cognitive Internet Technologies (CCIT), Fraunhofer IWU and Fraunhofer IIS have developed a test setup with a sensor for acoustic emission (AE) measurements during continuous generating grinding.

The detection of AE, also known as structure-borne noise, offers an effective way of monitoring the generating grinding process. Using so-called piezo sensors, which operate in the MHz range, high-frequency signals can be recorded that are caused by mechanical influences such as the impact of the grinding wheel on the workpiece or collisions. The advantage of this method is that structure-borne sound can be measured practically anywhere on solid bodies, which provides flexible measuring points and a wide range of applications. The sensitive sensor technology enables continuous monitoring of the grinding process and registers faults, defects or, for example, breakouts on the grinding wheel throughout the entire process.

However, AE sensor technology requires high-rate data transmission and powerful measurement data processing in order to initially transmit the signals to be recorded in full bandwidth. Based on this raw data, algorithms for anomaly detection and fault detection can also be trained for a process on a machine and later contribute as sensor-related AI to data reduction and event detection directly at the sensor.

A special test setup has been developed for measuring structure-borne noise during continuous generating grinding, which ensures precise recording of the signals and their transmission.