Predictive Maintenance


Transform from monitor-and-react to predict-and-act

Problem

Unforeseen asset downtime is a high operational risk to manufacturers, requiring deeper insights into asset health that IoT sensors and real-time condition monitoring alone are not able to deliver.

Why?

The condition indicators, although useful, do not give enough time to plan. For e.g., overall vibration is a lagging indicator as the defect already exists.

Also, setting the alert thresholds is a challenge. The varying nature of the process and product recipe leads to many false alarms.

Solution

4PointX PdM solution uses AI to automate the analysis. By correlating signals from multiple sensors, the solution detects anomalies (aka leading indicators) that usually show up as early as 60 to 90 days prior to the actual failure.

A comprehensive AI-enabled predictive maintenance plan starts with business understanding.

- Gartner Research, Published on December 04, 2018

Our building blocks for an effective maintenance paradigm

Learn AI

Learns normal behavior of an asset

The Learn AI identifies the normal operating signatures of an asset by analysing the full spectrum of sensors - condition and process parameters.

  • Automatically selects the key tags that influence behavior
  • Works even for assets that have limited failure history

Predict AI

Detects unusual behavior when it occurs

The Predict AI scans every incoming asset data to compare and predict any unusual behavior that deviates from normal operating signatures.

  • Adapts to the unique behavior of asset
  • Scales easily to thousands of assets

Explain AI

Augmented decision making

Use our Explain AI to know what tags are causing the unusual behavior and compare with similar signatures, if any, in the past.

  • Assisted root cause analysis
  • Automated error code lookup from digitised user manual

Improve AI

Man-teach-bot. Bot-gets-better

As the user validates the abnormalities, our Improve AI incrementally learns further resulting in lesser and high-fidelity alarms.

  • No more false alarms
  • Continuous improvement

Want to learn more?