Optimizing WAG Injection in a Carbonate Field Using CRM & ML

Background

A major carbonate field with 53 producers and 45 injectors had been under waterflooding since the 1950s. In the mid-1990s, Water-Alternating-Gas (WAG) was introduced to improve sweep efficiency. Despite decades of operation, suboptimal injector performance and inefficient gas allocation limited oil recovery.

Challenges

Difficulty in identifying which injectors were truly effective in contributing to oil production.
Conventional reservoir simulation was slow, costly, and uncertain.
Need for a rapid, data-driven optimization tool to guide field decisions.

Solution – CRM-Driven Hybrid Workflow

Applied a Capacitance Resistance Model (CRM) to capture injector–producer connectivity.

Enhanced CRM with Machine Learning (ML) for higher confidence and noise reduction.

Integrated with a modified Power Law oil model to quantify injector impact on oil (not just fluids).

Optimized Gas-to-Water Ratios (GWR) per injector to maximize recovery.

Results

Identified 5 key injectors (out of 45) with the highest impact on oil production.

Delivered optimized recommendations within 4 months.

Production Uplift

Battery oil rate increased from 560 → 740 bopd (+32%).
Specific producers near optimized injectors saw oil gains up to 300%.

Achieved 1.4% incremental recovery in 2 years, at minimal cost.

Key Takeaways

Faster Decisions

Full-field injector ranking without lengthy simulations.

Targeted Optimization

Focus on injectors that matter most.

Proven Gains

Tangible production increase from quick wins.

Scalable

Approach can be replicated in other mature waterflood or WAG fields.