Detect Faults Early.
Avoid Unnecessary Downtime.

Conventional threshold alarms react when a measured value crosses a preset limit. PrognostiX takes a different approach, combining physical fault signatures with AI-based analysis to recognize developing drivetrain problems and provide more meaningful diagnostic insights.

Our goal is to help operators detect emerging faults earlier, reduce unnecessary alarms, and make better-informed maintenance decisions.

PrognostiX focuses on three critical drivetrain components: main bearings, gearboxes, and generators.

Our approach targets developing mechanical faults that can lead to expensive repairs and unplanned downtime. The specific failure modes that can be evaluated depend on sensor coverage, vibration data quality, and turbine configuration.

The technology is being developed for both onshore and offshore wind turbines.

Not necessarily. PrognostiX is designed to work with vibration data from your existing condition-monitoring system (CMS).

If your current system provides suitable raw vibration waveforms, additional hardware may not be needed. We'll assess data compatibility before proposing a pilot.

PrognostiX primarily uses raw vibration waveforms rather than only summarized vibration measurements.

For an initial assessment, we're interested in the available vibration channels, sampling rates, sensor locations, and turbine configuration. Operating information such as rotational speed and load, along with historical maintenance records, can provide valuable additional context.

We'll help determine whether your existing datasets are suitable for evaluation.

Our current approach focuses on high-resolution vibration analysis.

SCADA data provides valuable operating information, including turbine speed, power, temperature, and load. However, conventional SCADA records generally lack the signal detail needed to analyze certain developing bearing and gearbox faults.

We therefore use vibration data as the primary diagnostic input, with SCADA data potentially providing additional context.

That's exactly what our approach is designed for.

PrognostiX aims to complement existing condition-monitoring infrastructure rather than replace it. By analyzing vibration data already collected by compatible systems, we aim to provide an additional layer of fault-detection and diagnostic intelligence.

Integration feasibility depends on data accessibility, format, quality, and the capabilities of your existing CMS.

Potentially, yes — historical fault data is a valuable starting point.

If you have vibration recordings associated with known component failures or maintenance events, we can explore a retrospective evaluation to assess whether PrognostiX can identify relevant fault signatures.

We'll first review data availability, quality, and event history to determine a suitable evaluation approach.

We believe the value of predictive maintenance should be measurable.

Before a pilot begins, we'll work with participating operators to agree on relevant evaluation criteria, which may include:

• Early-warning lead time before known failures.
• Fault-detection performance.
• False-alarm frequency.
• Diagnostic usefulness for maintenance planning.
• Compatibility with existing monitoring infrastructure.

Where suitable baseline data is available, we'll compare results against the operator's existing monitoring approach.

PrognostiX is currently at the pre-pilot stage.

We're actively seeking wind-farm operators and industry partners interested in evaluating our technology on real turbine data.

Our goal is to demonstrate measurable value through technical evaluations and collaborative pilot projects before broader commercial deployment.

We understand that turbine operational and condition-monitoring data can be commercially sensitive.

Before receiving non-public datasets, we would agree on data access, confidentiality, permitted use, processing arrangements, and retention requirements.

An initial discussion does not require sharing proprietary vibration recordings. We can start by reviewing your monitoring setup and available data types to determine the next steps.