Anadarko uses analytics applications to discover profitable business practices and avoid missteps.

Wouldn't it be a different world if engineers could eliminate hundreds of unworthy prospects at a glance? An emerging field called product analytics is introducing applications that do just that.
By creating dynamic visual representations of oil companies' exploration, drilling, production, economic and financial data, these new applications reveal the often hidden relationships that could help oil exploration and production companies make faster, smarter decisions about where to drill.
Data from local and disparate databases can be related using interactive images that show researchers and engineers trends and patterns that numeric tables and static images obscure. Product analytics empowers engineers to sift interactively through the deep, often scattered databases that exploration, drilling and production companies use. These software applications give engineers the hands-on ability to interact with the data through instant multidimensional pictures based on the parameters they choose. Engineers use these applications to test limitless "what if" scenarios by moving sliders across a visual representation of their data. The applications create a common interface for communicating vital information to multiple people across an organization, enabling them to make educated decisions about whether, where and how to drill for optimum business benefit.
When the engineers make decisions at Anadarko, the largest independent oil and gas company in the world, millions of dollars ride on their interpretations of the data used to determine whether to pursue potential drilling prospects. Managers generally don't have the luxury of taking the time to do everything twice to eliminate human error. Anadarko drills a well somewhere in the United States or one of 14 other countries every 5 minutes. At that pace, the ability to quickly uncover trends in data is essential. And there is no room for error; at an average cost of about US $2 million per well, the stakes are high, and delayed or misinformed decisions mean missed opportunities and costly mistakes.
Anadarko bases much of its information-driven approach on economic modeling applications. These applications let engineers test the various financial scenarios of a prospect based on various input parameters - all in the name of reducing risk while increasing profits. Anadarko uses Monte Carlo simulation methods to generate as many as 3,000 financial scenarios employing distributions for multiple reservoir variables, well parameters and costs. This particular technique tends to be computer intensive, with many problems taking hours to solve. Anadarko's data analysis tools display the results graphically and in detailed statistical reports that comprise more than 100,000 data points. Hidden in the data is information that's invaluable for people faced with the expensive proposition of deciding where and when to drill. Anadarko's predictive modeling tools help its engineers understand correlations and make preliminary hypotheses, but fully analyzing that much data to uncover the hidden trends and patterns that inform sound decisions poses considerable challenges.
A light in the darkness
Before it can use these predictive modeling tools as efficiently as possible, Anadarko needs to interpret the data, quickly understand key drivers and identify trends hidden within the numbers. Anadarko tried standard reporting applications and graphical packages to plot the data, but none would allow the company to fully understand its complex information interactions at one glance. As Bill Liddell, Anadarko's senior staff reservoir engineer, put it, "Identifying trends within a large tabular dataset is very difficult without advanced visualization and interaction capabilities."
Anadarko opted to enhance its analysis with a new analytics program from Spotfire Inc. DecisionSite is a configurable analytical application for asset life-cycle processes in any data-rich field, such as pharmaceuticals, genomics, manufacturing or semiconductor production. Traditional business intelligence applications are not well suited to the continual "what-iffing" needed to explore the myriad options and alternatives that drive oil and gas decision-making because they focus on static representations of data or on monitoring known performance metrics. Oil companies use DecisionSite to analyze data quickly, intuitively and visually.
Anadarko brought its simulation data into the DecisionSite environment. DecisionSite's dynamic query devices and visualization environment let Anadarko's engineers quickly eliminate scenarios with parameters that fall outside acceptable ranges. Based on the remaining scenarios, the engineers can easily determine whether a prospect will produce a large enough profit to be worth drilling - which also allows them to begin creatively changing a plethora of variables.
For example, an Anadarko engineer looking at 3,000 economic scenarios might begin by removing the scenarios with negative net present values. He can further reduce the dataset by examining such parameters as recovery efficiency, net pay, permeability and estimated total cost, and eliminating scenarios with values that are either too high or too low. Simply put, based on the number of scenarios that remain, he can decide whether one well is worth drilling or move on to the next one. DecisionSite relates data from deep and scattered sources into interactive images that show researchers trends and patterns that numerical tables and static images obscure. Users can move a mouse to ask their data endless streams of questions, then watch as the changes unfold on their computer screen. The combination of analytics, data access and visualization allows for limitless multidimensional, hypothetical data queries. It could, for example, tap geographical and financial databases in the same room or anywhere in the world for permeability and flow rates, zone thickness, temperature, distance and timing parameters, financial equations, quality control statistics or parameter probability distributions - in any combination. Engineers can move among as many parameters as they wish, performing dozens of data queries every time they nudge a graphical slider up or down and watch as the results take shape in the form of changing 3-D images onscreen.
Since rolling out the DecisionSite application, Anadarko has avoided several potentially disastrous drilling projects. In one such project, Anadarko partnered with another company in Alaska and was asked to participate in an exploration well. Anadarko's share of the projected cost would have totaled more than $9 million. Using DecisionSite, it quickly became evident the critical drivers in the project were zone thickness (net pay), permeability and the corresponding flow rates. Further evaluating these critical parameters revealed the attractive economic scenarios resulted from high-end - and hence financially risky - probability distributions. Anadarko chose not to participate, thereby freeing millions of dollars for investment in other projects with higher chances of success. Anadarko is using the technology in other parts of the organization to improve exploration, production and business decision-making, and plans are under way to expand the system to support other departments.
Upgrade released
Spotfire recently introduced DecisionSite 7.0, which supports offline analytics, enabling users to continue analysis tasks in disconnected mobile mode, then automatically resynchronize when returning to the network. Included in the new release is the ability to deploy the application to facilitate enterprisewide administration and management. Also included are the developer toolkit and a power library of dynamic statistics and data mining tools.
For additional information, visit www.spotfire.com, or send e-mail to info@spotfire.com.

Comments