Analytics & Big Data
Dec 22, 2015

How are oil well drilling companies using analytics to drive productivity and the return on assets

This is the final part in our three-part series about how big data and analytics are transforming operations across industries. In the first part, we focused on how analytics is driving operational excellence. In the second part, we provided our point of view on the use of analytics in the aviation industry. This part of the blog series provides our point of view on the use of analytics in the oil and gas industry.

Historically, well drilling for oil exploration using heavy duty drills was a time-consuming activity. Machines drilled down to specified depths and collected the soil-layer samples, which were examined in the lab for properties such as porosity, density, and gamma. The visual core analysis of these samples was analyzed by geo-scientists, who made recommendations on subsequent drilling—for example, to drill vertically, at an angle, or horizontally—in order to maximize the probability of hitting a shale deposit. The accuracy and expertise of geo-scientists greatly impacted the forecast.

Oil drilling is highly capital-intensive, involving machines that cost around $350-$500 million—or $400,000 a day. As only about 40% of these machines' time is spent drilling it is important to reduce their idle time to improve the return on asset (ROA). Companies are now combining digital technology and analytics to reduce costs and increase the accuracy of their operations.

The oil and gas industry collects massive amounts of data from exploration and drilling operations via sensors installed in oil wells, textual operational logs, and other measurement devices. Sensors fitted on drill bits capture and transmit these signals on a real-time basis where a sample set of signals for certain layers of soil are analyzed in real time to compute the soil parameter values. One raw seismic dataset is usually in the hundreds of gigabytes, which poses a classic big data problem to solve.

We developed fuzzy logic to capture and convert geologists' expert knowledge into an automated intelligent reasoning system using the Hadoop environment. Geo-scientists' knowledge is captured in a structured repository, which also has the history of the drilling decisions they have made based on the parameter values from soil samples and the accuracy of those decisions. This provides the foundations of the forecasting logic with a high level of accuracy.

The soil parameter value combination of certain soil layers are matched against historical inferences to predict the depth and direction that maximizes the probability of detecting oil deposits by identifying the accurate drilling action required. The solution achieved high prediction accuracy (90%), which maximized ROA for the drilling assets.

These are exciting times for manufacturing industries. Operations functions in particular are at a crossroads, as is analytics. Big data analytics is no longer a golf-course conversation for CIOs, for it has conquered the board room as well. Early movers are already benefiting, and the rest are catching up with astonishing pace. Following the hype, the results are becoming reality. From being a tool to identify customer-acquisition levers, credit scoring, or fraud detection, analytics is also making a huge difference in the manufacturing space by enhancing efficiencies and effectiveness, saving costs, and improving fulfillment. In this complex ecosystem, fostered by the open source era, everyone has a role to play. What remains to be seen is how quickly companies can figure out how to apply analytics in the context of their business challenges to drive true business impact.

About the author

Sudhanshu Singh

Sudhanshu Singh

Senior Vice President & Global Practice Leader, Analytics & Research

Follow Sudhanshu Singh on LinkedIn