Idea
ML workflow optimizing gas lift injection rates to boost production in unconventional oil fields without costly downhole data.
Research Paper
Core Innovation
This paper introduces a data-driven ML model that accurately forecasts the Gas Lift Performance Curve using only historical production time series, eliminating the need for downhole gauges or multi-rate well tests. It integrates Bayesian Optimization to determine optimal gas injection rates under facility constraints, enabling practical deployment in unconventional fields.
Why It Matters
Unconventional oil fields often lack downhole gauges or multi-rate well tests due to cost or facility limits, hindering gas lift optimization. This workflow improves production by over 5% on average using only historical surface data, enabling scalable, cost-effective optimization across many wells. It transforms operational efficiency and production economics in constrained environments.
Market Size (TAM)
$2–10B TAM for oilfield production optimization; $1–3B SAM from unconventional field operators. Driven by rising unconventional production and cost pressures to optimize lift methods.
Potential Customers & Pain Points
- Oil and gas operators – High cost and infeasibility of downhole data acquisition
- Field service companies – Need scalable optimization tools
- Asset managers – Desire improved production without expensive testing
Business Model
Subscription-based SaaS platform offering continuous gas lift optimization with tiered pricing by well count and feature set; consulting and integration services for deployment.
Competitive Landscape
- Schlumberger
- Halliburton
- Baker Hughes
- Weatherford
Implementation Challenges
- Integration with existing field operations and control systems
- Operator trust in ML-driven recommendations without traditional data
- Data quality and variability across different fields
Validation Strategy
- Expand pilot deployments across diverse unconventional fields
- Demonstrate consistent production uplift and cost savings
- Collect operator feedback to refine model and user interface
- Partner with service companies for broader market reach
Research Paper Overview
A Machine-Learning-Based Gas Lift Optimization Workflow for Unconventional Fields
Summary
This paper presents an automated ML-driven workflow that forecasts gas lift performance and optimizes gas injection rates without downhole data or multi-rate tests. Piloted on 30 wells in Bakken, it achieved over 5% production uplift and is now deployed across 200+ wells, offering a cost-effective solution for unconventional fields with facility constraints.