Predictive Analytics Methods Begin to Take Hold in E&P



Similar documents
Collecting and Analyzing Big Data for O&G Exploration and Production Applications October 15, 2013 G&G Technology Seminar

EarthStudy 360. Full-Azimuth Angle Domain Imaging and Analysis

DecisionSpace. Prestack Calibration and Analysis Software. DecisionSpace Geosciences DATA SHEET

Cloud-based Data Management for Quick Asset Valuation, Opportunity Identification and Risk Quantification

Nexus. Reservoir Simulation Software DATA SHEET

Certificate Programs in. Program Requirements

Shale Field Development Workflow. Ron Dusterhoft

Data Mining and Exploratory Statistics to Visualize Fractures and Migration Paths in the WCBS*

Geothermal. . To reduce the CO 2 emissions a lot of effort is put in the development of large scale application of sustainable energy.

RESERVOIR GEOSCIENCE AND ENGINEERING

Unconventional Challenges: Integrated Analysis for Unconventional Resource Development Robert Gales VP Resource Development

Graduate Courses in Petroleum Engineering

DecisionSpace Earth Modeling Software

An Artesian Whitepaper

Quick Look Determination of Oil-in-Place in Oil Shale Resource Plays*

Bringing Oilfield Data into the Enterprise

Unlocking your Data to Improve Performance in the Shale. Fred Kunzinger Upstream Practice Lead ECIM 2014

Recommended Practices Associated with Hydraulic Fracturing Operations

Nautilus Global Schedule 2016

INDEPENDENT REPORTS CONFIRM SIGNIFICANT POTENTIAL FOR LINC ENERGY S SHALE OIL IN THE ARCKARINGA BASIN

Big Data analytics in oil and gas

WHITE PAPER. 2014, Vesmir Inc.

Understanding Tight Oil

The ever increasing importance of reservoir geomechanics

Building an Unconventional Gas Business in Australia:

BS PROGRAM IN PETROLEUM ENGINEERING (VERSION 2010) Course Descriptions

Lists of estimated quantities to be performed and prices Estimated quantities to be performed. Prices

Exploration. Exploration methods

PERSPECTIVE OF SHALE GAS PROSPECTION IN POLAND

EMBRACING ANALYTICS ACROSS THE WELL LIFECYCLE GREG PALMER SR. INDUSTRY CONSULTANT

DEPARTMENT OF PETROLEUM ENGINEERING Graduate Program (Version 2002)

On the Impact of Oil Extraction in North Orange County: Overview of Hydraulic Fracturing

Global Oil & Gas Suite

Master big data to optimize the oil and gas lifecycle

Exploring Success in Shale Zhiyong Zhao, Vice President, Hess Corp

Study Assesses Shale Decline Rates

Guidelines for the Estimation and Reporting of Australian Black Coal Resources and Reserves

EnerCom The Oil & Gas Conference. An Integrated Workflow for Unconventional Reservoirs

Integrated Reservoir Asset Management

In Development. Shale Liquids Production Analysis. Value. Key Deliverables. Principal Investigator: Investment per Sponsor $52K (USD)

Turning Data into Action: How Credit Card Programs Can Benefit from the World of Big Data

VII Seminario Estratégico - SPE Evaluación de las Perspectivas Energéticas de la Argentina

Optimization applications in finance, securities, banking and insurance

Infosys Oil and Gas Practice

GAS WELL/WATER WELL SUBSURFACE CONTAMINATION

Mineral rights ownership what is it and why is it so unique in the USA?

Midland Houston. In Alliance With

Appendix 25. Content of a Competent Person s Report for Petroleum Reserves and Resources

Monterey Shale Potential

Eagle Ford Shale Exploration Regional Geology meets Geophysical Technology. Galen Treadgold Bruce Campbell Bill McLain

The Economic Benefits of Oil and Natural Gas Production: An Analysis of Effects on the United States and Major Energy Producing States

FlairFlex. Real-time fluid logging and analysis service

Winning in Oil and Gas with Big Data Analytics

The Shale Gale Also Brings a Data Blizzard Author:

Technology Trends in Land Management Geographic Information Systems

Specialist Reservoir Engineering

The successful integration of 3D seismic into the mining process: Practical examples from Bowen Basin underground coal mines

Five Predictive Imperatives for Maximizing Customer Value

SEYMOUR SLOAN IDEAS THAT MATTER

Geoscience AT ITS best. Software solution. Consulting. International Oil & Gas Consultant and Software Solution Provider

14TH INTERNATIONAL CONGRESS OF THE BRAZILIAN GEOPHYSICAL SOCIETY AND EXPOGEF

Society of Petroleum Engineers SPE Global Training Committee Training Course Review Process

FIBER-OPTIC SENSING TECHNOLOGIES

PRODUCTION OPTIMIZATION CONSULTING SERVICES. Delivering comprehensive optimization solutions from a single well to a full field PRODUCTION

Business Analytics and the Nexus of Information

How big data is changing the oil & gas industry

How Did These Ocean Features and Continental Margins Form?

I N T E L L I G E N T S O L U T I O N S, I N C. DATA MINING IMPLEMENTING THE PARADIGM SHIFT IN ANALYSIS & MODELING OF THE OILFIELD

Master of Science in Marketing Analytics (MSMA)

Daniel T Shaw Principal Molten Consulting. 12Sep2014. Data - the Oil & Gas Asset that isn t managed like one. Presented by Daniel T Shaw

Implementing the Data Management Continuum A Practical Demonstration

HESS CORPORATION. Lean Leads the Way in a Low-Cost Environment

London STOCK EXCHANGE

Operations in the Arctic areas? New challenges: Exploration Development Production

An Integrated Rock Catalog for E&P Geotechnologists

Chartis RiskTech Quadrant for Model Risk Management Systems 2014

Figure 1. The only information we have between wells is the seismic velocity.

Solve your toughest challenges with data mining

Integration of Geological, Geophysical, and Historical Production Data in Geostatistical Reservoir Modelling

A Study on Software Metrics and Phase based Defect Removal Pattern Technique for Project Management

RESERVOIR EVALUATION. The volume of hydrocarbons in a reservoir can be calculated:

Chapter 1 Introduction

Degree/ Higher Education Jobs:

Energy Industry. Valuing the great shale play

sufilter was applied to the original data and the entire NB attribute volume was output to segy format and imported to SMT for further analysis.

Decisioning for Telecom Customer Intimacy. Experian Telecom Analytics

Figure 2-10: Seismic Well Ties for Correlation and Modelling. Table 2-2: Taglu Mapped Seismic Horizons

NSF Workshop: High Priority Research Areas on Integrated Sensor, Control and Platform Modeling for Smart Manufacturing

Unconventional Oil and Gas Production Drives Trends in Water Management and Treatment

Information-Driven Transformation in Retail with the Enterprise Data Hub Accelerator

Well-logging Correlation Analysis and correlation of well logs in Rio Grande do Norte basin wells

Transcription:

white paper Predictive Analytics Methods Begin to Take Hold in E&P A white paper collaboration between Transform and NEOS

TABLE OF CONTENTS Introduction 3 The Challenge 4 The Response 5 Applied Example #1 6 Applied Example #2 7 Summary 7

Predictive Analytics Methods Begin to Take Hold in E&P A white paper collaboration between Transform and NEOS by Murray Roth, Transform and Jim Hollis, NEOS Twenty years ago, predictive analytics was a buzz phrase little heard by those outside of academia. Ten years ago, the concept began to emerge prominently within the financial sector, as investment managers started to leverage powerful, software-driven analytical programs to assess investment alternatives given the interplay of a complex set of variables, including historical performance, asset correlations, relative risk, and projections of future economic conditions. More recently, companies ranging from Wal-Mart to United Airlines have been analyzing reams of product and consumer data to determine what products are selling, to whom, where, and why, and thereafter using the insights derived to make future decisions about offerings, pricing levels, and promotional packages. Thanks to Billy Beane, general manager of Major League Baseball s Oakland Athletics, and the book and movie, Moneyball, predictive analytics has erupted into the mainstream. Moneyball describes how Beane and his assistant, Paul Podesta, shackled with both a restricted budget and the limitations of a mid-level regional market, managed to assemble a competitive and highly successful baseball team capable of defeating opponents with deeper pockets, larger markets, and star-laden rosters. By using predictive analysis to interpret and integrate a vast pool of multi-variate statistics and performance attributes, the duo was able to quantify and leverage hidden reserves of talent and capability. For a time, they had access to asymmetric information that allowed them to field a winning team at a fraction of the cost and, in the process, change the face of modern sports. Today, it is becoming commonplace for companies across many industries to apply advanced mathematical methods to mine extremely large datasets, searching for distal correlations and associations. Pharmaceutical companies are mining patient histories and treatment regimens to determine which combinations of drugs are most effective. Online retailers are analyzing and influencing product purchase behavior given highly individualized web site presentations, direct marketing campaigns, and bundled promotions. A new battleground is being established, built upon the emerging foundations of advanced software and applications that enable deep, rapid, highly reliable insight into complex variables and conditions. The practice of information-driven decision making is establishing a new threshold for performance and driving companies to identify new strategies that capitalize upon previously unknown insights more readily than their competitors. Predictive Analytics Methods Begin to Take Hold in E&P 3

THE CHALLENGE Many of these new challenges are currently being identified and confronted within the E&P industry. In recent years, oil & gas leaders have increasingly leveraged predictive statistical methodologies to blend geophysical, geological and engineering data to estimate reservoir properties between well locations, This work is an extension of historical techniques that relied upon geological mapping and geostatistical computations to infer what occurs between drilled well locations; while these techniques integrate geophysical and geologic data, the predictive power of the outcome is largely based upon effective, or smart, interpolation and a fairly limited set of G&G inputs, namely seismic and well log data. E&P predictive analytics are extending beyond traditional G&G data to incorporate innovative multi-spectral and potential field measurements and a breadth of engineering information. To better understand the current needs and requirements, let s consider a pair of typical E&P scenarios. At the basin scale, explorers might be looking for indications of which areas are the most prospective to lease, either because they are more likely to contain commercially producible fields or, in the case of unconventional plays, the most productive portions of the reservoir for prioritized development. At the reservoir scale, explorers would typically be looking for localized sweet spots that contribute to the most productive wells. Let s take it a bit deeper with the case of an unconventional shale play. Successful exploitation depends on understanding the complex interplay among a host of subsurface variables including total organic content (TOC), thermal regimes, insitu hydrocarbon distributions, brittleness, fracture density, regional stress fields, reservoir thickness and proximity to faults. Understanding the interplay requires a rich analysis of numerous seismic and non-seismic measurements. A real world example includes liquid and gas producing reservoirs such as the Bakken, Eagle Ford, Permian and Barnett plays. These fields present variable profitability for current and prospective operators. To meet the challenge, they utilize quantitative analytics to churn through vast quantities of historic and new data to identify trends and best engineering practices. The data sets can include thousands of well log records for thousands of well bores, in conjunction with copious engineering measurements, 3D seismic volumes and, more recently, regional non-seismic measurements, attributes, and derivatives. In virtually any E&P scenario within today s oil and gas environment, the datasets are numerous and extremely large, often constituting multiple terabytes. And project data only continues to expand in both size and amount. It s an increasingly complex engagement that has begun to outstrip the processing capacity of the human brain and, in recent years, has mandated the development of deep algorithmic tools simply to keep pace. We can see how these datasets lend themselves to nothing short of the most advanced quantitative analysis, in order to achieve optimal interpretation and understanding. 4 Predictive Analytics Methods Begin to Take Hold in E&P

THE RESPONSE In many industries, predictive analytics is used in a behavioral context. That is, practitioners leverage analytically derived information to monitor and predict consumer behavior based on previous behaviors and tendencies. It makes sense. However, in the E&P world, and in particular seismic exploration, analytical tools, whether we classify them as predictive or otherwise, serve a more complex, workflow-driven purpose. They are used to not only improve and optimize subsurface interpretation and understanding, but equally to predict lifecycle performance and future variations in facies conditions as they relate to well planning, drilling and reservoir management. This response is predictable (no pun intended) given the emergence of advanced software and data processing technologies as extraordinarily valuable means of gaining even deeper and better insight into subsurface conditions. As the tools and methodologies have advanced, equally so has the competitive need to apply them in areas of critical business decision-making. The increasing complexity of hydrocarbon reservoirs places continuous strain on E&P workforces and technology processes; many reservoirs, particularly unconventional shale plays, require astute planning and subsurface visibility in order to achieve profitable (and optimally efficient) commercial-grade environments. Reservoir analytics can estimate the correlation of production and production declines with reservoir and engineering parameters, delivering very high prediction levels for identifying optimal engineering parameters and prospective sweet spot locations. There are also many existing wells and reservoirs that have exceeded prime or optimal production conditions and require meticulous and integrated analysis, including reliable forecasting, in order to achieve continuous profitability. In the past, most producers would routinely abandon these wells and reservoirs as production margins diminished below certain levels. Today, with the advantages provided by advanced interpretation and predictive analytics, it s no longer standard operating procedure to withdraw simply because existing configurations and reservoir conditions have reached certain thresholds or limitations. From a business standpoint, this type of decision-making can now be deemed perfunctory, premature and costly. Interpreters and operators are now faced with new issues of how to manage new and maturing assets and access harder to reach and harder to visualize regions in and around the reservoir. With a few adjustments, a sweet spot can continue to be a sweet spot. Or another previously undetermined sweet spot might emerge. Whatever the case, as we can all agree, it s a good problem to have. Currently, a small number of innovative companies are developing a de facto emerging methodology, of sorts, that provides E&P operators with deep integration of advanced analytics with modern interpretation approaches and methodologies. Visionary companies are establishing leadership roles in this area. NEOS GeoSolutions provides distinct expertise in multi-measurement interpretation, integrating a broad spectrum of geological and geophysical datasets (both seismic and non-seismic) including data available via public domain, owned by clients, or acquired using proprietary platforms to produce a highly constrained 3D model of the subsurface. Predictive Analytics Methods Begin to Take Hold in E&P 5

This enables interpreters and operators to determine which portions of a basin might be the most prospective and, at the lease level, what areas are most likely to contain commercial quantities of hydrocarbons or minerals. Transform focuses on reservoir analytic interpretation and modeling, an approach designed to improve the operator s understanding of reservoir characteristics and conditions. The company s innovative software suite simplifies interpretation challenges by automating the integration and visualization of microseismic, well log, seismic and other E&P data types. By combining geophysical, geological, and engineering workflows in a streamlined and quantified manner, asset teams are well positioned to better characterize unconventional oil and gas, heavy and enhanced oil, and most conventional reservoirs. Applied Example #1 Reservoir analytics seek a cause-and-effect relationship between the reservoir parameters that an interpreter or explorer inherits, and the engineering parameters that are applied in order to achieve a particular quantity or threshold of well production. Let s take a look at the Bakken unconventional play in North Dakota and Montana as an example of how analytical interpretation can provide remarkable benefits in today s E&P world. This predominantly oil and water producing play is notorious Bakken optimized production pattern for presenting substantial challenges when predicting, or attempting to forecast, well production from a myriad of engineering and geologic data. For oil producers, the Middle Bakken and Three Forks formations are key determinant objectives for effectively calibrating geologic conditions with variable hydraulic fracture effectiveness. For example, by analyzing microseismic data, interpreters can generate volumetric estimates of simulated reservoir volumes for each fracture stage, in addition to determining the degree of fracture zone overlap. Drilling and completions engineering decisions are a major determinant of relative well production in unconventional plays. Well performance in these plays is driven by: well length, fracture stage spacing, fluid and proppant rates and amounts, completion techniques and more. However, well engineering is not the complete story, as sweet spot locations in the Bakken and Three Forks formations are driven by many geologic factors. Foremost are the geochemistry of the Upper and Lower Bakken formations, specifically thermal maturity and organic content, coupled with depth and thickness of the Bakken and Three Forks reservoirs. And so, well production is dependent upon spatial location within the play and well construction and trajectory through reservoir rock. However, with increased proximity of wells, and the emerging multi-bench development of the Three Forks, interwell completion and production interference are increasing in importance. 6 Predictive Analytics Methods Begin to Take Hold in E&P

How do we begin to unravel the intertwined complexity of what creates a good and bad well in this play? Predictive analytics are undaunted by the size, amount and relationship of rock prospectivity and well parameters which is good, as leading operators are collating databases of 50 or more attributes for thousands of wells. Setting initial production or ultimate recovery estimates as prediction targets, non-linear multi-attribute statistics build a model to relate cause-and-effect. More than creating a prediction model, often with very high accuracy, reservoir analytics highlight the relative importance and relationship of geology and engineering on well production. As operators endeavor to optimize field development strategies and operations, reservoir analytics provide a roadmap as to what choices are more important and effective. Marcellus sweet spot map. Tioga County, Pennsylvania Appalachian Basin, USA. Applied Example #2 In Pennsylvania s Marcellus Shale, a predictive analytics methodology has been used to identify the most promising areas to lease and drill at the county scale (roughly 1,000 square miles). The technique involves comparing nearly 100 individual measurements, attributes, and derivatives from both legacy and newly acquired geophysical datasets - the latter generally acquired using airborne gravity, magnetic, electro-magnetic, radiometric, and hyperspectral sensors - to those associated with the most productive wells in the area. While the geostatistical techniques used are highly objective and algorithmic, the inputs to the predictive modeling effort generally all have high levels of explainable geological relevance. For instance, gravity data might feed regional subsurface models, resulting in horizon thickness isopach maps for targeted reservoir intervals. Magnetic data might provide indications about proximity to faults or fracture systems that enhance well deliverability. EM resistivity data might feed hydrocarbon distribution maps both within (and above) prospective reservoir intervals, providing an indication of the relative charge in different parts of the hydrocarbon system. And hyperspectral data might highlight direct (and indirect) hydrocarbon indicators on the surface, which could be associated with micro-seepage from nearby reservoirs within the subsurface. In the Marcellus example, twenty measurements, attributes, and derivatives were identified as highly correlative with the best wells in the county. These twenty measurements can be grouped in four categories: structural context; size and composition of the tank ; reservoir plumbing; and the halo above the reservoir intervals of interest. Predictive analytics allows the geoscientist to map the correlative anomalies throughout the county, with hot colors showcasing the sweet spots the methodology suggests would be most prospective for leasing and drilling or, at a minimum, for further investigation by the informed and experienced interpreter. To date, an analysis of drilling results suggests a strong correlation between the sweet spots that were predicted and those that were actually encountered by the bit. Predictive Analytics Methods Begin to Take Hold in E&P 7

Summary The E&P industry is emerging as a rapid adopter of predictive analytics methods to identify subtle patterns and correlations across multiple G&G and engineering datasets. The sheer number and sizes of these datasets creates an opportunity to apply analytically driven, multi-measurement interpretation methods to predict exploration and drilling sweet spots at both regional and local scales. The business advantages related to successfully leveraging these analytics include more efficient use of both human and capital resources; superior understanding of subsurface conditions and reservoir performance; and deeply integrated, highly-informed project decision-making. AUTHOR BIOS Murray Roth President, Transform Murray Roth is president and co-founder of Transform, headquartered in Littleton, Colorado. After graduating with a degree in astrophysics from the University of Calgary, he began his career at Geophysical Services Inc., working in seismic acquisition, seismic processing and special project groups. He joined Landmark Graphics in 1989, serving as executive vice president of research and development and marketing until he departed the company in 2004 to found Transform. Jim Hollis President and CEO, NEOS Prior to joining NEOS in 2010 as president and CEO, Jim Hollis served as president and COO of ION Geophysical where he established growth strategies and supervised financial execution of all ION business units. He also worked previously in management and technology roles for Landmark Graphics. He holds a BS in geophysics from the University of California, Santa Barbara, and an MS in geophysics from the University of Utah. 8 Predictive Analytics Methods Begin to Take Hold in E&P

Predictive Analytics is Changing the Game neosgeo.com transformsw.com