The offshore facility operated three turbine generators, normally running two in load-sharing mode with the third available as standby.
When one generator became unavailable, redundancy was significantly reduced. The remaining configuration created a high
operational risk: an issue affecting the active generator could result in loss of platform power, process disruption and costly
downtime.
The team needed continuous visibility into developing generator problems and sufficient warning to change operating configuration
before a failure occurred.
VROC trained predictive models using historical generator data that included five previous failure events. Soon after deployment, the
Time to Failure model predicted a defined undesirable event 5.6 days ahead. Contributor analysis highlighted increasing differential
pressure across the air intake filter, indicating developing airflow restriction.
Operations switched to the available backup generator and took the affected unit offline. When the generator was returned to service
three days later, VROC again predicted failure — this time in 3.7 days — indicating that the underlying condition remained unresolved.
The air intake filter was then thoroughly cleaned. Once the generator returned online, model predictions reflected stable, healthy
operation.
The Time to Failure model provided repeated advance warning while differential pressure across the air intake filter remained
abnormal. Following corrective cleaning, the model reflected the improved operating condition and returned to a healthy prediction.
Earlier intervention
A 5.6-day warning gave operations sufficient time to investigate the developing issue and perform a
controlled generator changeover.
Corrective action verified
When the initial intervention did not remove the underlying condition, the model continued to
predict failure. Following thorough filter cleaning, predicted performance returned to normal.
Critical power reliability maintained
Since deployment, the facility reports zero unplanned generator trips and no power
interruptions or plant shutdowns caused by generator failure.
VROC provides advance failure warning together with contributing-factor insights, helping teams intervene earlier and verify whether corrective action has restored healthy operation.
Explore Time to Failure Prediction | Talk to Our Team
Industry: Oil and Gas
Solution: Predictive Maintenance & Reliability
Use Case: Time to Failure Prediction
Facility: Offshore Platform
Asset: Turbine Generator
Related Case Study: Detecting a Turbine Generator Cooling Issue 18 Days Before Trip
Gain real-time visibility, predict failure earlier, optimize performance, and take control of your operations with VROC’s integrated solutions.