The leadership of many oil & gas producers can significantly widen the application of predictive analytics to achieve value from the digital data their company collects and manages.
I’ll illustrate how predictive analytics achieves business value for oil & gas producers through the following five examples along the upstream oil & gas well life cycle:
Predictive analytics encompasses various statistical techniques, including predictive modelling, machine learning, and data mining to analyze current and historical facts to make predictions about future or otherwise unknown events.
Predictive analytics exploits patterns found in historical and transactional data to identify risks and opportunities. Models capture relationships among many variables to assess risk or potential associated with a particular set of conditions. The goal of predictive analytics is to guide decision-making and reduce the risk of making investments that turn out to be unprofitable.
First, I’ll clarify the relationship among some similar terms that are sometimes used as synonyms when they are not:
Reducing exploration risk requires an earth model for the area of interest. Explorationists use the earth model to predict:
Reducing exploration risk is often achieved with a cross-section data visualization such as this one.
We can quickly see how the wells interact with their surrounding lithology. The well tests, which are represented by ellipses along the wellbore, help us predict hydrocarbons in place and forecast likely initial production rates. A right-click on any of the well test ellipses will display information about that well test.
The cross-section data visualization helps to confirm or refute your hypothesis about the formation of the hydrocarbon reservoir.
Once we’ve forecasted the approximate hydrocarbon resources in place, we move on to creating development scenarios.
This predictive analytics application for development economics brings together a:
On this predictive dashboard, you can change a variable’s value, such as the oil or natural gas price forecast or the recovery factor for oil in place. Then all the charts are updated in real-time.
This predictive visualization is designed to:
Predictive analytics is used to aggregate a massive amount of data to produce multiple development scenarios.
Once we’ve forecasted the implications of alternative development scenarios, drilling engineers move on to drilling and completing wells.
This predictive visualization compares gas well production volumes per unit of time for selected operators. It illustrates that significant opportunities exist for some operators to improve drilling and completion design and execution. There’s a 4-fold difference in initial production rates from the best operator to the worst operator. This difference narrows to a 2-fold difference a few years later as the wells are produced.
In this predictive visualization, you can hover over every curve that represents an aggregation of all the wells drilled in the area by a specific operator to reveal the well identifiers and few items of drilling and completion information.
By double-clicking on a curve, you will create a visualization consisting of all the wells for the selected operator.
There are enough variables at play here that predictive analytics can make a valuable contribution to better understanding the specific opportunities to improve drilling and completion outcomes.
Once we’ve completed all the planned wells, production and reservoir engineers move on to improving production operations at the wells.
Most upstream oil & gas software packages include decline curve analysis with a predictive analytics component. Oil & gas professionals analyze decline curves to death.
In this predictive visualization, you can hover over every dot that represents monthly production volume for a specific well to reveal the well identifier and few items of well information. Typically, you can also choose from a variety of formulas for calculating the forecast.
It’s often useful to right-click on the outliers to see what’s happening at those wells. Most often, you’ll reveal a data problem. Occasionally you’ll stumble on an insight about well operations that will have wide-ranging implications.
Executives and perhaps some engineers are prone to forget the curves are predictions. The curve values are not a certainty. The uncertainty in the production volume prediction should trigger caution in deriving a cash flow estimate from these volumes. Unfortunately, that caution is sometimes not in evidence.
After producing the wells for a period, engineers and business development staff often evaluate acquisition and divestiture opportunities.
This budget vs. actual lease operating expense comparison and forecast dashboard compares operating expense for individual wells and various aggregation levels.
This predictive visualization immediately shows the more and less profitable wells, fields or areas. That data shows you what assets to divest and what to keep.
With a regional perspective, a visualization may also show you areas to consider for acquisition.
The predictive aspects of this comparison and forecast allow end-users to:
We can use these charts to support our potential acquisition and divestiture analysis.
Overall predictive analytics allow oil & gas producers to:
To explore additional oil & gas predictive analytics examples, please read Column: Oil & Gas predictive analytics applications.