Fraunhofer IZFP The Ear on the Machine

AcoustiX is already in industrial use at John Deere for permanent quality monitoring of cutting systems of combine harvesters and is currently being validated for series production.

Scientists at the Fraunhofer IZFP have developed a sensor system that uses acoustic sound evaluation - similar to human hearing - to detect errors or irregularities in moving systems and machines at an early stage.

During operation, machines generate characteristic vibrations and thus sounds. These provide information about the quality, because assembly errors or other defects often cause a change of the normal operating sounds. Oftentimes, the assembly personnel with a good hearing and long experience is entrusted with this inspection task. However, human hearing is subject to a certain subjectivity: it tires after a certain time or is negatively affected by ambient noise.

Currently available acoustic test systems often require adaptation or calibration, especially if the design of the equipment or machines to be tested needs to be changed even if they are ­ever so slightly. This is where the AcoustiX sensor system comes in, developed by scientists at the Fraunhofer Institute for Non-Destructive Testing (IZFP). It uses the data from individually adapted acoustic sensors, which are attached directly to the machine or record vibrations or noises from machines or systems via microphones without contact. Errors or irregularities are automatically analyzed and finally logged. Based on the signal evaluation, specific statements about the correct assembly and function of the system or machine can be made within a few minutes.

The Fraunhofer researchers' "hearing" sensor system is designed for a wide variety of industrial applications. The system can be used in all areas in which final assembly inspection or permanent operational monitoring is important, e.g., for monitoring large, autonomously operated machines and systems or for quality evaluation of individual assemblies, which are operated on test benches, among other things.

"The development focus in particular lies on our analysis methods, the algorithms. These can easily be integrated into existing test systems and customized," explains Matthias Heinrich, scientist at Fraunhofer IZFP. The medium-term goal of the researchers is exact fault localization and detailed determination of the type of fault using intelligent algorithms and analysis methods.