ISI Logo

Shale Production Optimization Technology (SPOT)

"Hard Data" vs. "Soft Data"

“Hard Data” refers to field measurements. This is data that can readily be, and usually is, measured during the operation. For example, in hydraulic fracturing variables such as fluid type and amount, proppant type and amount, injection, breakdown and closure pressure, and injection rates are considered to be “Hard Data”. In most shale assets “Hard Data” associated with hydraulic fracturing is measured and recorded in reasonable detail and are usually available.

In the context of hydraulic fracturing of shale wells, “Soft Data” refer to variables that are interpreted, estimated or guessed. Parameters such as hydraulic fracture half length, height, width and conductivity cannot be directly measured. Even when software applications for modeling of hydraulic fractures are used to estimate these parameters, the gross limiting and simplifying assumptions that are made, such as well-behaved penny like double wing fractures, renders the utilization of “Soft Data” in design and optimization of frac jobs irrelevant.

Another variable that is commonly used in the modeling of hydraulic fractures in shale is Stimulated Reservoir Volume (SRV). SRV is also “Soft Data” since its value cannot be directly measured. SRV is mainly used as a set of tweaking parameters (dimensions of the Stimulated Reservoir Volume as well as the permeability value or values that are assigned to different parts of the stimulated volume) to assist reservoir modelers in the history matching process.

The utility of micro-seismic events (as it is interpreted today from the raw data) to estimate Stimulated Reservoir Volume is at best inconclusive. While it has been shown that micro-seismic may provide some valuable information regarding the effectiveness of multi-stage hydraulic fractures in Eagle Ford Shale, the lack of correlation between recorded and interpreted micro-seismic data and the results of production logs in Marcellus Shale has been documented.

Due to its interpretive nature “Soft Data” cannot be used as optimization variables. In other words, one cannot expect to design a particular frac job that results in a well behaved induced fracture with a designed half length, height and conductivity by tweaking the amount of fluid and proppant that is injected. Similarly, designing SRV (size and permeability) by modifying the amount of fluid and proppant that is injected during a frac job or by modifying the injection rate and pressure is not an option . Therefore, although “Soft Data” may help engineers and modelers during the history matching process, it fails to provide a means for truly analyzing the impact of what is actually done during a frac job.

ISI's unique, proprietary and innovative Shale Production Management technology uses only "Hard Data" and is based completely on Data-Driven Analytics and Data Mining using state of the art Machine Learning. This technology includes three modules:

  • Pre-Modeling Data Mining
  • Data-Driven Predictive Modeling
  • Post-Modeling Analysis & Optimization

Pre-Modeling Data Mining

The complexity of oil and gas production from Shale assets is revealed rather quickly upon the use of conventional statistical analysis.

Example 1: Average Maximum Injection Pressure
Example 2: Injected Proppant per Stage of Hydraulic Fracturing

Frequency distribution of these two hydraulic fracturing parameters are shown below:

Let us examine the behavior of "Average Maximum Injection Pressure" and "Injected Proppant per Stage of Hydraulic Fracturing" against a production indicator such as the "Best 30 Days of Production":

Lack of any trends, patterns, or correlation between these "Hard Data" and prod cut ion from wells in this Shale asset is obvious.

Now lets use ISI's Fuzzy Pattern Recognition on the data shown above and see if there is actually a correlation between these "Hard Data" and prod cut ion that can be revealed once advance Machine Learning technology is applied.

Shale PM

Presence of a clear trend in these "Hard Data" and correlation with prod cut ion is unmistakable in the above figures. This trend is only revealed using ISI's Fuzzy Pattern Recognition technology.

ISI's unique and proprietary "Stairs" visualization technology uses Fuzzy Set Theory in order to categorize wells into "Poor", "Average", "Good", and "Excellent" wells based on their produ ction behavior. Using these new categories, "Hard Data" such as "Average Maximum Injection Pressure" and "Injected Proppant per Stage of Hydraulic Fracturing" are plotted.

Shale PM

Presence of a clear trend in these "Hard Data" and correlation with well qualities is unmistakable in the above figures. This trend is only revealed using ISI's "Stairs" visualization technology.

Data-Driven Predictive Modeling

A data-driven predictive model is developed upon completion of detail data mining. The data-driven model is validated using blind data (production from wells not used during the development of the model). It is ensured that the predictive model has the required generalization capabilities and honors the known physics of the problem.

Shale PM

Post-Modeling Analysis & Optimization

Upon completion and validation of the data-driven models, it is used for multiple purposes:

  • Predicting Production Behavior of New Wells
  • Generation of Type Curves
  • Uncertainty Analysis
  • Design of Optimum Frac Jobs for New Wells

Development of Type Curves provide valuable information about the impact of each parameter on production. Type Curves are developed for different regions of the asset to make sure that the impact of Shale heterogeneity on the production is accounted for.

Using Monte Carlo Simulation, uncertainty analysis is performed to identify the P90-P50-P10 for new wells.

ISI's Shale Production Management Technology has a small computational footprint, such that it can be deployed (upon completion) on a smart phone or a tablet computer for fast track and on-the-fly analysis.