Refinery Use Case
Lube Oil Pump — Oil Pressure Stability
Detect lube oil pump degradation and bearing vibration anomalies before they cause unplanned shutdowns — using AI analytics on existing sensor and control data. No new instrumentation required.
Data Input
Analytics Findings
AI-ranked root causes of oil pressure and vibration instability with explanation and recommended action
Vibration is primarily driven by mechanical movement captured by paired sensors. Divergence between sensor readings indicates looseness, mounting integrity issues, or local resonance rather than true process excursions.
Implement a paired-vibration control limit-alarm on divergence between sensor 1 and sensor 2; trigger field checks for looseness, mounting integrity, or local resonance.
Lube-oil flow directly influences vibration outcomes. Unstable oil flow can amplify bearing vibration or create false excursions due to control valve hunting or pump recirculation issues.
Stabilize lube-oil flow by tightening control of the lube-oil flow sensor and confirming oil cleanliness and water content. Investigate control valve hunting and pump recirculation for root cause.
Sources: Feature importance values derived from AI model trained on lube oil pump sensor data. Methodology aligned with API 670 (Machinery Protection Systems) and ISO 10816 vibration severity standards for rotating equipment.
Expected Results
Bearing Vibration Excursions
−15–25%
Fewer false trips and unplanned shutdowns from sensor divergence
Unplanned Pump Downtime
−15–20%
Early detection of mechanical looseness and lube-oil degradation
Bearing & Seal Lifespan
+10–20%
Stable lube-oil flow reduces accelerated wear
Maintenance Cost per Pump
−10–18%
Condition-based intervention replaces fixed interval overhauls
Correlating paired vibration sensor divergence with lube-oil flow conditions catches mechanical degradation weeks earlier — using data the DCS already logs.
Monitored KPIs
Business Outcome
Lube oil pump failures in refineries are among the leading causes of unplanned unit shutdowns, with a single event costing $100,000–$500,000+ in lost production and repairs. By monitoring paired vibration sensor divergence and lube-oil flow quality in tandem, operators identify developing bearing faults and lubrication problems before they escalate — without adding instrumentation or replacing existing DCS infrastructure.
Sources: Rotating equipment reliability benchmarks from API 670 – Machinery Protection Systems (5th Ed.); vibration severity classification per ISO 10816-1 – Evaluation of machine vibration.
Data Privacy: Built for the Refinery Floor
On-Premises / Edge Deployment
All inference runs locally. No process data leaves the facility.
No Cloud Dependency
Fully air-gapped operation possible. Analytics continue even without internet.
Zero Data Sharing
Pump, vibration, and lube-oil data never transmitted externally.
Works with Existing Systems
Connects directly to DCS/SCADA. No new infrastructure required.
Your process data stays in your refinery. Full data sovereignty with AI-powered lube oil pump monitoring.
"The vibration and lube-oil data to predict pump failure already exists.
We just make it speak."