If you’ve ever taken a statistics class, you’ve likely encountered cross correlation. Cross correlation is essentially a measure of how two signals are related. A positive correlation means increases observed in one signal are also observed in the other signal. Cross correlation is an incredibly simple statistical model with applications across industries and subject areas—from pattern recognition to signal processing. Luckily for us, cross correlation also has a place in the oil and gas industry.

While waterflooding in mature fields is commonplace, the challenge of identifying effective injectors and determining the rates of water injection remains. How then can companies continue to optimize their oil recovery? A combination of qualitative (cross correlation) and quantitative analysis (CRM) for a waterflood field can help optimize injection to improve oil recovery (5-10%).

Mature fields are fields that have been producing for 10-15 years and have started observing declines. In these fields, waterflooding is implemented for pressure support and as a means of sweeping the remaining oil. Waterflood operations in mature fields are characterized by frequent and higher variation in production and injection rates.

When looking to optimize, we can use cross correlation analysis to determine the location and strength of connections between injection and production wells.

One method of combining qualitative and quantitative analysis is using cross correlation analysis in conjunction with capacitance resistance models (CRM). Capacitance resistance models are one of the most widely used reduced physics-based analytical models to determine the connectivity between injection and production wells. You can read more about CRM here .

This figure shows how cross correlations identified effective injectors A9A and A1 for this field in Egypt. Taken from “A New Continuous Waterflood Operations Optimization for a Mature Oil Field by using Analytical Workflows that Improve Reservoir Characterization” (Yadav and Malkov et al, 2019). 

While cross correlations may pick up some false positives, CRM can be both computationally expensive and time consuming for a field with multiple (more than 100) wells. This limits our ability to adapt these models directly for decision making. Using cross correlation as a precursor to using CRM approach helps reduce variables in the CRM model. This approach also allows us to identify effective injectors, thief zones and water recycling issues before validating these resulting with the CRM approach.

Cross correlation analysis, being a qualitative analysis, can be further verified using salinity data as shown in this JPT magazine article.  Changes in salinity levels can indicate the extent of water recycling and further supplement the observations obtained through a combination of qualitative and quantitative analysis. 

With the easy availability of data through digitization, using production and injection data through unique ways can help create hybrid models that can enable quicker decision making. How have you been using your data?

Share this Blog Post

Categories: Main