Powering Predict & PreventHumaware develop a range of data-driven predictive analytics tools that detect and diagnose defects to predict and provide actionable data to prevent asset failure. Backed by 30 years of experience in developing and fielding remote condition monitoring technology, we support both the technical and business challenges of implementing predict and prevent strategies.
The Humaware approach is to use the concept of condition indicators to identify the degradation of condition that is exhibited in the remote condition monitoring data. A condition indicator is a feature of the monitored data that directly relates to the defect, such as noise being present or frequency domain changes developing. Using this approach, our software is able to provide more detailed information about the defects to provide earlier, robust detections that can then be utilised to forecast maintenance intervention to prevent in-service failure of an asset.
Our toolset uses data-driven methods to provide the user with 100% surveillance of the asset for 100% of the time. There are no thresholds to set and the information is presented to the user via a novel dashboard that enables the root cause to be assessed without the need to analyse the monitored data and forecast the required changes to the maintenance schedule. This capability provides new, simpler protocols to enable the user to efficiently manage the information and accurately determine any intervention that is required.
Originally developed within aerospace, our toolset has been developed and validated in the rail sector in the following programmes:
- Innovate UK/RSSB funded Digital Rail Programme “Health & Prognostic Assessment of Railway Assets for Predictive Maintenance”. For escalators, a prognostics based dynamic maintenance scheduler was successfully demonstrated.
- Network Rail/RSSB funded Future Railway programme “Advanced Decision Support Tool”. For track circuits of all types, a robust automatic diagnostics technology was developed that removed the necessity to manually set and maintain fixed thresholds to detect anomalies.