What Is Condition Monitoring?

Condition monitoring is the practice of measuring the physical condition of machinery, such as motors, pumps, fans, gearboxes and compressors, to spot changes that signal a developing fault. Every machine has a normal signature. Condition monitoring records it and watches for drift: a bearing that vibrates more, a motor that runs hotter or draws more current, oil that carries more wear particles. Catching those changes early turns a breakdown into a planned repair.

The process follows a simple loop, described in ISO 17359, the general guidance standard. Choose the machines that matter most, decide which failure modes to watch for, pick the measurements that reveal them, set baselines and alarm limits, then collect data, analyse trends and act. The data can be gathered by a technician on a regular route with a handheld instrument, or continuously by permanently fitted sensors connected to a monitoring system.

No single measurement catches every fault, so most programmes combine several techniques matched to the machine.

Vibration analysis is the most widely used. ISO 20816, which replaces the older ISO 10816 series, gives vibration severity zones for many machine types, so readings can be judged against a published reference as well as the machine’s own baseline.

The two terms are often used together, but they are not the same thing. Condition monitoring produces the data: measurements, trends and alarms. Condition-based maintenance acts on it, servicing a machine when its condition says so rather than on a fixed calendar. Predictive maintenance goes a step further, using trends and models, increasingly machine learning, to estimate when a fault will reach a critical level, so work can be scheduled before then. Good predictions depend on good data, so a reliable condition monitoring system comes first.

Most machine faults give warning long before they stop production. A bearing defect, for example, typically shows in ultrasound and vibration readings well before it becomes audible or causes heat. Condition monitoring is how a site hears that warning.

Programmes that succeed usually start small, prove their value on a few machines and then expand.

Continuous monitoring needs sensors on the machine, hardware to digitise their signals and a computer to analyse and forward the results. Vibration sensors produce fast analogue signals, often sampled thousands of times a second, so analogue I/O modules or dedicated vibration inputs capture them. Temperature, current and status signals can arrive over Modbus from existing drives and meters.

An edge computer or IIoT gateway then turns raw waveforms into features such as RMS velocity and frequency bands, applies alarm rules or AI models, and sends only the results onward. Our IoT edge gateway case study covers that relay on a factory floor, our industrial IoT gateways provide the computers, and our white paper on deploying rugged edge computers for AIoT covers running models on the data at the machine.

We build condition monitoring gateways and edge computers with the analogue inputs, serial ports and connectivity fitted, the operating system and data software configured, and the inputs tested as standard before they ship.