Predict potential engine failures simply by using sound

Predict potential engine failures simply by using sound

Arduino Team — December 16, 2022

Almost every manufacturer uses a machine at some point in their process, and every one of those machines is almost guaranteed to contain at least one motor. In order to maintain availability and efficiency, these engines must always operate properly, as even a small failure can have disastrous effects. Predictive maintenance aims to achieve this without going overboard by trying to prevent them entirely by combining sensors with predictive techniques that can schedule maintenance when failure is likely.

Shebin Jose Jacob's solution uses the Arduino Nano 33 BLE Sense, along with its built-in microphone, to capture audio and predict when a motor is about to fail. He achieved this by first creating a new Edge Impulse project and collecting samples for four classes of sound: OK, Anomaly 1, and Anomaly 2, as well as general background noise. After designing a pulse and training a classification model on the samples, he was able to achieve an impressive accuracy of around 95% on the test samples.

The final step was to deploy the model as firmware for the Arduino, which would allow it to classify sounds in real time by reading continuously from the microphone. Each time an anomaly is detected, a red LED at the top lights up.

You can read more about the project here in the Edge Impulse blog.