Ultrasonic predictive maintenance

As part of the development of its research activities on sensor networks and predictive maintenance, the Autonomy and Sensor Integration Laboratory (DSYS/SSCE/LAIC) of CEA-LETI in Grenoble, France, is offering a thesis on “Ultrasound-based predictive maintenance”. Ultrasound emission is one of the first signs of ageing in an industrial system, before the appearance of vibrations, noise or heat. The aim of the thesis will be to implement a network of ultrasonic sensors driven by a microcontroller, which, combined with analysis tools based on Embedded Artificial Intelligence (Edge-AI), will make it possible to detect premature ageing of an industrial machine.

The PhD student will have to implement a test bench around an industrial machine, which will be instrumented with ultrasound sensors. The data collected will then be processed by classification or regression algorithms (neural networks, SVM, random forest, etc.), for an accurate and reliable diagnosis of the state of the machines.

Facebook

Twitter