BB13: Predictive Maintenance [FM] SOLD OUT!


The overall objective of the Monitoring Toolbox – Predictive Maintenance is to increase productivity through smart maintenance, reduce downtime and avoid catastrophic failures of machines, vehicles and production systems. It consists of an ecosystem of advanced signal processing and AI techniques for anomaly detection, fault diagnostics and prognostics, drivetrain modelling and cloud connectivity.

Through the usage of various sensing technologies, machine faults can be detected and predicted in bearings, gears, electric motors, clutches etc. Detection tools are available for bearing faults, gear faults, gear grinding faults, motor faults etc. Machine disturbances, such as gear noise in gearboxes, are mitigated through robust diagnostic and prognostic tools. These AI tools and models enable pre-emptive maintenance through in-time replacement of discrete components.

In a regular state, a machine exhibits a very specific vibration pattern/identity. If just a tiny fault is present, this vibration pattern changes. Sensors measure this vibration pattern up to high frequencies. The system filters out unwanted noise which is present in an industrial environment.

The validated Flanders Make advanced methodology highlights the modulation which exists in the spectrum and provides the user with accurate information in order to detect the fault early and accurately. Advanced algorithms continuously compare the vibration pattern from the machine when it was in perfect order (the reference model). These algorithms have an open character so they can be further modified and optimized on the application-specific requirements.

Example cases

  • Robust and early detection and diagnostic of bearing and gear faults:
    • issue: gear noise effects typically mask the effects of incipient bearing faults in gearboxes
    • solution: automated, computationally efficient, robust, and open algorithms for early fault detection running with embedded low-cost hardware.
  • Accurate prognostic methods for remaining useful life prediction of bearings with limited industrial data:
    • issue: limited amount of data available of failing machine components hampering the applicability of traditional AI tools
    • solution: hybrid remaining useful life methods
  • IoT architecture design for cloud-connected machine monitoring:
    • efficient local data reduction and transfer and cloud connectivity
    • transfer of classification models for easy deployment
    • efficient edge processing and metadata management

Links to Demo videos:

  • Condition Monitoring:
  • Smart Maintenance:
  • Data-efficient AI and digital-twin technologies for fault detection:

For more information, please contact:


Posted on

January 31, 2023