Condition monitoring lets machines tell you when something is wrong before they break. Condition monitoring sensors on a motor, a pump, a gearbox or a spindle measure how the machine behaves in real-time. When a bearing starts to wear or a shaft becomes misaligned, the readings change long before the machine stops. Maintenance teams can then act at the right time instead of waiting for a failure or replacing parts too early.
- Goal: detect faults early, reduce unplanned downtime and maintenance cost
- Typical measurements: vibration, temperature, oil condition, ultrasound, motor current
- Typical assets: rotating machinery such as motors, pumps, fans, gearboxes and compressors
- Standards: ISO 17359 (general guidelines), ISO 10816 and ISO 20816 (vibration)
How does machine condition monitoring work?
A condition monitoring system has three parts. Monitoring sensors capture data from the machine and its rotating equipment, such as vibration sensors on bearing housings or temperature sensors on windings. A data collection unit, often an edge device or a monitoring module, processes the signals and compares them with limits or reference patterns. Condition monitoring software visualizes machine health and equipment performance, triggers an alert when values exceed thresholds and, in larger setups, creates a work order in the computerized maintenance management system. Many plants centralize this data to compare machine conditions across sites. Monitoring can be continuous with permanently installed sensors, including battery-powered wireless sensors, or periodic with portable instruments on a route.
Main condition monitoring techniques
- Vibration analysis: the most widely used method for rotating machinery, usually combined with temperature in vibration and temperature sensors. Vibration monitoring detects imbalance, misalignment, looseness, gear damage and bearing wear, each with typical frequency patterns.
- Temperature monitoring and thermography: overheating bearings, motors or electrical connections become visible.
- Oil analysis: particles and chemical changes in lubricants reveal wear inside gearboxes and engines.
- Ultrasound and acoustic emission: early detection of friction, leaks and electrical discharges.
- Motor current analysis: electrical signatures reveal mechanical and electrical faults in motors and the driven equipment, supporting remote diagnostics.
Condition monitoring, preventive and predictive maintenance
Maintenance strategies range from reactive (fix it when it breaks) and preventive maintenance based on time (replace parts at fixed intervals) to condition-based maintenance (act when the measured condition requires it) and predictive maintenance (forecast when a failure will occur). Condition monitoring is the data foundation for the last two. Time-based maintenance often replaces parts that still had years of life, while reactive maintenance accepts equipment failures and expensive downtime. Condition-based monitoring avoids both by basing every decision on the actual state of the machine.
Benefits of condition monitoring for machines
- less unplanned downtime and fewer secondary damages
- lower maintenance cost through targeted interventions
- longer machine life and better spare parts planning
- higher safety, because dangerous failures are prevented
- data for process optimization, energy efficiency and higher productivity
Vendor neutral condition monitoring with open standards
Many condition monitoring solutions work well with one vendor’s sensors and software but are hard to combine with others. A plant with machines from different manufacturers ends up with several monitoring systems that do not talk to each other. Open standards solve this: OPC UA and MQTT transport sensor data to any application, and the Asset Administration Shell describes the machine, its sensors and their time series data in a standardized way.
OI4 and IDTA members have applied this in practice. Sensor manufacturer ifm, a member of both organizations, used their cooperation in a condition monitoring project in the food and beverage sector. In the OI4 energy monitoring project, ifm sensors for electrical energy and compressed air were installed on two pilot machines, with edge devices for data collection and preprocessing, to make consumption visible per workpiece. And in the onboarding project, the Time Series Data submodel delivers sensor values directly into the SCADA system for trend curves and lifecycle monitoring. Condition data in a standardized format can also feed the digital twin of a machine and, increasingly, industrial AI applications.
Frequently asked questions
What is a machine monitoring system?
A machine monitoring system collects data from machines, such as condition, state and output, and makes it visible to operators and maintenance teams. Condition monitoring is the part that focuses on the health of the equipment.
What is the difference between condition-based and time-based maintenance?
Time-based maintenance follows fixed intervals regardless of the machine’s state. Condition-based maintenance acts only when measurements such as vibration or temperature show that maintenance is needed.
Which sensors are used for condition monitoring?
The most common are vibration sensors and temperature sensors. Depending on the machine, oil quality sensors, ultrasound sensors, current sensors and pressure sensors are added, increasingly as wireless sensors for easy retrofitting.