Condition Monitoring

Accurate real-time detection of even the subtlest changes in machinery operation

Condition monitoring provides a vital early warning sign of potential problems whether this be with a production line, or a single machine, or a critical piece of infrastructure – picking up even the most subtle changes before a human could detect that something might be going wrong.

Parameters such as vibration, acoustics, temperature and humidity, to name but a few, can all be monitored in real time using Internet of Things (IoT) sensors attached to machines at key points of a production line, for example, or a remote piece of capital equipment or some critical hardware.

The aim is to detect any deviation from normal operation that might suggest maintenance or safety measures are required, with the data either streamed for real-time analysis, logged for regular reviews or automatically processed to generate alerts.

This kind of predictive maintenance can save time and money – cutting out unnecessary routine work and focusing efforts on the crucial maintenance that is required to address specific potential issues long before everything grinds to a halt. It also improves safety – detecting potential hazards early, or even predicting them in advance to prevent them from developing.


Real-world challenges

A world-leading healthcare product company

An increase in failures on an automated production line filling vials with drugs prompted a world-leading healthcare product company to turn to Plextek for help. In addition to delays to production while the problem was diagnosed and fixed, a costly re-sterilisation process needed to be performed each time maintenance was completed.

To improve monitoring of the filling process for early identification of components likely to fail, the Plextek team developed a custom sensing unit with an embedded processor and multiple sensors, supplied in a sterile encasement.

During a planned shutdown, multiple sensors were deployed on the line to collect real-time environmental data, and the line was set up to progress through a series of experimental states.

The data was then analysed to identify patterns associated with different simulated faults – and algorithms were developed which would be capable of detecting problems in real time and provide automatic alerts. These alerts could be used to identify potentially faulty components which could then be replaced during regular line maintenance, preventing unscheduled shutdowns.

Designing a sensor array

When underground electrical wires are damaged – causing sparks that ignite surrounding gases – the result can be explosions within service manholes.

That’s why a client wanted a system to detect and transmit early warnings about the build-up of gases at sites before they reached a critical state – so that preventative action could be taken.

Designing a sensor array capable of operating in such a harsh environment was no mean feat – with the solution even including a novel power harvesting technique using locally available ‘free’ energy sources to extend the sensor lifespan.

Defence Science and Technology Laboratory

Ceramic body armour can be damaged during training or operations – resulting in a considerable reduction in its ability to provide impact protection. Damage to the ceramic plate often takes the form of hairline fractures which are only visible during X-ray analysis.

When the Defence Science and Technology Laboratory asked Plextek for help to reduce the need for body armour to be returned for X-ray analysis, it was clear the solution would have to be low size, weight, power, and cost in addition to being easy to use.

The team developed a sensor system with fine-tuned accelerometers to detect the shock waves associated with the altered response of the armour to subsequent shocks if it was cracked – all with a five-year operating life from a coin cell battery.

A user can rapidly assess the integrity of their body armour through status information presented as a simple, single-screen graphic on a smartphone app or by a press-to-test button showing a simple pass/fail via an LED indicator.


Key skills

Image of a Condition Monitoring sensor
  • Sensor physics

    It is important to choose appropriate sensors to measure relevant parameters.

  • Time-frequency analysis

    Many failure modes produce a change in frequency of a periodic signal that varies as the condition worsens.

  • Machine Learning (ML)

    Using measured data, algorithms can learn to detect and classify anomalous operation.

  • Edge analytics

    Real-time analysis is often best performed on the sensor device itself rather than at a central server.

  • Low size weight and power electronic systems design

    Many condition monitoring systems must be small and last on battery power for extended periods.


It is costly to maintain more often than you need – but it is also costly to maintain less often than you need. Condition monitoring helps you maintain at the appropriate time dynamically, rather than having a fixed schedule. It can also help reduce maintenance costs because you have more targeted information that gives you a better idea of what needs maintaining.

Dr Damien Clarke, Lead Consultant
Dr Damien Clarke

Lead Consultant


What sets us apart when it comes to condition monitoring?

Developing a condition monitoring system from initial idea to a final deployed product can involve many steps. Plextek is well positioned to perform any – or all – of these steps, depending on the client’s needs:

  • Concept generation
  • Sensor selection
  • Preliminary data gathering
  • Wireless communication
  • Database management
  • Algorithm development
  • Live system testing
  • Custom hardware design
  • Manufacturing
  • Predictive maintenance
Contact Plextek

Contact Us

Got a question?

If you have got a question, or even just an idea, get in touch