The acoustic method for diagnosing machines operating under variable conditions

The acoustic signal can be used for mechanical fault diagnosis in conditions where measurement with a vibration sensor cannot be realised. This problem often occurs on production lines, where it is necessary to stop the line to mount the sensor. However, the acoustic signal measured in the diagnosis machine's vicinity may contain noise from other sources under industrial conditions. This causes the signal to require more advanced analysis methods than methods based on vibration measurement. The paper proposes a method for diagnosing rotating machinery insensitive to acoustic interference from other machines based on the acoustic signal. The technique allows the diagnosis of machines operating under varying load, speed and temperature conditions. A laboratory experiment on a bench was carried out to analyse the functionality of the proposed method in gearbox diagnosis. The new approach was compared with a method based on vibration measurement.