ISI Logo

Characterization for Unconventional Assets

Problem Statement

To gain competitive advantage, a large independent decided to identify the performance drivers in its unconventional asset in the DJ basin. The complexity of the operation that included large number of laterals with massive, multi-stage hydraulic fractures minimized the applicability of analytical, numerical and even traditional statistical analysis. Upon successful completion of this study the approach was implemented in other assets such as Eagle Ford Shale..

Methodology

Intelligent Solutions, Inc. used its IMprove technology (Data-Driven Analytics for Production Optimization) in order to identify the performance drivers in these assets. The part of the IMprove technology for identification of performance drivers incorporate fuzzy set theory.

Results

The hard data collected from this asset (field measurements) suggest that the collection of characteristics attributed to Rock properties (porosity, TOC, water saturation, reservoir thickness, reservoir pressure …) are the main driver behind production. However, degree of impact the rock properties on production changes as a function of time when compare to other collection of properties such as completion characteristics.

Rock properties have a higher impact on production than characteristics attributed to well construction and completion practices.

Furthermore, many hidden patterns and trends in the data was discovered using ISI's proprietary fuzzy pattern recognition technology. Cartesian cross plot of any of the rock, well or completion related parameters versus production indicators suggests no trends or patterns in the data.

Cartesian cross plot of TOC versus production indicator (90 Days Cum. BOE) for hundreds of wells.

When fuzzy set theory is used to classify the wells in this asset into "Poor", "Average", "and "Good" wells, and then average the parameter being investigated based on each well's fuzzy membership function and plot them as bar charts, patterns and trends start to emerge. Even after the granularity of the analysis is increased from three fuzzy sets to four and then to five fuzzy sets, the trend discovered in the data holds. Dominant trends in data will survive the scrutiny of granularity.

Wells in the asset are classified into fuzzy clusters of "Poor", "Average", and "Good" wells. Average TOC for each class of wells starts showing a clear trend. Better wells have been completed in the rock with higher TOC.

Increasing granularity of the analysis by classifying wells into fuzzy clusters of "Poor", "Average", "Good", and "Very Good" wells. The trend holds. Better wells have been completed in the rock with higher TOC.

Increasing granularity of the analysis by classifying wells into fuzzy clusters of "Poor", "Average", "Good", "Very Good", and "Excellent" wells. The trend holds again. Better wells have been completed in the rock with higher TOC.

Using a complex and proprietary descriptive data-driven analytics algorithm, the granularity of the analysis can be increased to the limit of total number of wells in the asset. This will result in discovering an unmistakable pattern in the data.

Comprehensive Fuzzy Pattern Recognition (Maximum Granularity). The complex, hidden pattern in data is discovered.