Acoustic and Vibration Monitoring: Choosing the Right Signal
Industry Operations Note 3 of 6: acoustic and vibration signals
Teams starting rotating-equipment monitoring tend to hit the same fork first: acoustic or vibration? Neither wins in general. The right signal follows from where the expected change appears and what the maintenance team has to decide.
The measurements differ at the source. A microphone receives airborne sound. A vibration sensor measures motion at its mounting point. Acoustic emission concerns signals generated inside a material and carried through the structure. Treat the three as one measurement and both the installation and the interpretation go wrong.
The question picks the signal
| Question | Signal to consider first | Context to record |
|---|---|---|
| Has a new sound appeared around the machine? | Airborne acoustic data | Background noise, microphone position, operating state |
| Has motion changed at a structural point? | Vibration | Mounting point, direction, speed and load context |
| Is the task focused on transient events within a material or structure? | Acoustic emission | Transmission path, sensor coupling, analysis purpose |
The table is a starting point for choosing where to observe, not an automatic diagnosis. Confirming a fault still takes equipment structure, operating context, maintenance history, and a field check.
Figure 1. Select the observation point, compare a baseline and trend, then send the finding into a field inspection.
What the microphone hears is not what the accelerometer feels
Airborne acoustics can catch changes where mounting a sensor is impractical. The price is that nearby equipment and work noise arrive too, so microphone position and collection context have to stay consistent. Hearing an unfamiliar sound and identifying its cause remain two separate tasks.
Vibration runs the other way: the mounting point decides what you see. A bearing housing and a frame on the same asset can tell different stories. A comparison only holds when the location, direction, and operating state travel with the data.
Acoustic emission is not ordinary audio
Field conversations sometimes use acoustic monitoring and acoustic emission interchangeably. As measurements they are different jobs. Acoustic emission picks up structure-borne signals from events inside a material; its sensors, coupling, and analysis purpose have little in common with a microphone recording airborne sound.
Name the method precisely in the monitoring plan. Airborne audio should not inherit an acoustic-emission criterion, and an acoustic-emission result should not be read like a general noise measurement.
Trends only compare under matching conditions
ISO 20816-1 bases operational monitoring on vibration measured during normal operation. Speed, load, process state, and sensor placement all move the numbers, so that context belongs in the baseline.
An outlier is a reason to inspect, not a finished diagnosis. Check the prior trend and the maintenance history, then verify the machine itself.
Before the sensor order goes out
Two things belong on paper before any signal is chosen: the change being looked for, and the maintenance decision it would support. Record sensor position and direction in a repeatable form, keep background noise and operating state with the baseline, and re-baseline after maintenance or sensor changes. When several signals run together, give each one its own role; otherwise the program grows data instead of decisions.
For motors and gearboxes, Motor Sound Diagnostics shows one way to wire these signals into field alerting. The field question still comes first.