Feature 10 · Artificial Lift Reliability · Well B-12 · DEMO MODE
Explainable AI monitors intake stability, loading, speed and fluid conditions to predict representative pump-unsetting exposure.
Closed-loop AI workflow
Pump unsetting risk
Local representative model evaluates fluid pound / partial fillage precursors before pump unsetting occurs.
RISK CLASS
HIGH
60%
PUMP UNSETTING RISK
Equipment health
78.2%
Failure probability
67%
Predicted RUL
43 days
Explainable contribution ranking
Pump load
88% of modeled load envelope.
Speed profile
9.4 SPM at 48 Hz.
Fluid condition
720 cP at 63.5°C.
Rod load
91% peak rod-load utilization.
Low pump fillage
62% pump fillage reduces intake stability.
Low fluid level
34% modeled fluid level.
Modelled risk comparison
Representative model estimate; optimization combines intake stability with SRP speed, stroke and VFD corrections.
Current
60%
AI optimized
23%
Representative ESP32 / twin signals
12-hour risk model trend · DEMO MODE · representative data
Actions prioritized to improve pump intake stability.
Reduce pumping speed and align the VFD to the safer SRP envelope to restore modeled intake fillage.
Reduce stroke from 2.8 m to 2.45 m; verify the next dynamometer card for fluid-pound signatures.
Maintain the next CSS heating response above 70°C to lower representative heavy-oil viscosity and pump load.
Fillage
Fluid
Speed
The modeled risk is driven by partial pump fillage, elevated pump and rod loading, and a speed profile that exceeds the representative intake-stability envelope. Applying the AI action updates SRP stroke, SPM and VFD parameters locally, then immediately recalculates predicted improvement. All operational values are representative SIH prototype data.