Back to Insights

Optimizing ESP Performance in Unconventional Wells

The Challenge

A Permian Basin operator was experiencing frequent Electrical Submersible Pump (ESP) failures across a 12-well pad in a tight oil formation. Production downtime averaged 8 days per month per well, with ESP replacement costs exceeding $150,000 per failure — not including lost production revenue estimated at $40,000 per day per well.

Key pain points identified:

  • Rapid gas interference causing pump-off conditions and gas locking
  • Inconsistent fluid properties due to changing GOR (800-2,500 scf/bbl)
  • No real-time monitoring or automated control system
  • Reactive maintenance culture instead of predictive reliability
  • ESP run life averaging only 214 days (industry benchmark for Permian is 450+ days)

Our Approach

TerraQuint deployed a three-phase solution to address these challenges, combining real-time data analytics, physics-based modeling, and automated control.

Phase 1: Data Diagnostics & Sensor Installation

We installed real-time sensors and collected 90 days of high-frequency operational data including:

  • Pump intake pressure (downhole gauge)
  • Motor amperage (VFD data)
  • Discharge pressure and temperature
  • Fluid temperature and vibration signatures
  • Production rate and water cut (daily tests)

Phase 2: Nodal Analysis & System Modeling

Using Petroleum Experts IPM and WellFlo, we built a complete integrated system model from reservoir to sales:

Reservoir → Perforations → Tubing → ESP → Choke → Flowline → Separator

The model was history-matched against 90 days of production data and identified:

  • Optimal operating frequency range: 52-58 Hz (previously run at fixed 60 Hz)
  • Critical minimum intake pressure: 280 psi (below which gas breakout occurs)
  • Gas interference threshold: 15% free gas at pump intake
  • Underperforming wells requiring choke adjustments

Phase 3: Real-time Control & Predictive Analytics

We implemented an automated control system with machine learning-based anomaly detection that:

  • Adjusts VFD frequency dynamically based on intake pressure trends
  • Triggers real-time alarms when parameters deviate from optimal envelope
  • Provides daily optimization recommendations via dashboard
  • Predicts ESP failures 7-14 days in advance using pattern recognition

The Results

After 12 months of continuous operation, the improvements were significant and sustained:

Metric Before After Improvement
Production Uptime 73% 96% +23%
ESP Failure Rate 4 per year per well 1 per year -75%
Average Run Life 214 days 412 days +93%
Operating Cost (LOE) $42,000/month per well $31,000/month -26%
Unplanned Downtime 8 days/month 1.5 days/month -81%

Economic Impact

The financial benefits were substantial:

  • Annual savings from reduced ESP failures: $450,000 per well
  • Additional revenue from increased uptime: $1.2 million per well per year
  • Total annual benefit per well: $1.65 million
  • Payback period on automation investment: 3.5 months

Key Takeaways

  1. Real-time data is essential — ESP optimization requires high-frequency downhole and surface data
  2. Nodal analysis identifies root causes — surface symptoms often hide subsurface problems
  3. Automated control pays for itself — typical payback of 3-4 months
  4. Predictive maintenance prevents failures — machine learning models detected issues 1-2 weeks before failure
  5. One size does not fit all — optimal ESP operating parameters vary significantly between wells

Conclusion

This case study demonstrates that a data-driven, integrated approach to ESP optimization delivers substantial improvements in uptime, reliability, and operating economics. For operators facing similar challenges in unconventional assets, the investment in real-time monitoring, nodal analysis, and automated control typically generates returns within the first quarter of operation — while also extending ESP run life by 2-3x.

Struggling with ESP failures or lost uptime?

We help operators diagnose artificial-lift problems, tune their systems and extend run life, remotely or on site.

Get in touch → Free initial assessment call · We respond within 24 hours