Oilfield Analytics: Optimize Exploration and Production ......Title: Oilfield Analytics: Optimize...
Transcript of Oilfield Analytics: Optimize Exploration and Production ......Title: Oilfield Analytics: Optimize...
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
ENERGY ANALYTICS SUMMIT 2014
Keith HoldawayPrincipal O&G Domain Expert, Global Energy Practices,
SAS Institute Inc.
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
OILFIELD ANALYTICS:
OPTIMIZE EXPLORATION AND PRODUCTION WITH
DATA-DRIVEN MODELS
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
AGENDA
• Data Mining Virtuous Cycle
• Data Mining: What is it?
• Data Mining: O&G Input Space
• Deterministic to Probabilistic
• SEMMA Process: Case Studies
• The Witch Methodology
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
DATA MINING
VIRTUOUS CYCLE
“Those who do not
learn from the past
are condemned to
repeat it.”
George Santayana
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
DATA MINING: WHAT IS IT?
• Data Mining Styles
• Hypothesis Testing
• Directed Data Mining
• Undirected Data Mining
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
DATA MINING: O&G INPUT SPACE
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
Data
•Historical
•Real-time
Deterministic analysis
•Experience
Outcomes
•Situation A
•Situation B
•Situation C
Data
•Historical
•Real-time
Probabilistic analysis
•Experience
•Variability
•Complex relationships
Predictive Outcomes
•Situation A 95%
•Situation B 22%
•Situation C 36%
Actionable workflows
•Workflow A
•Workflow B
•Workflow C
DETERMINISTIC TO PROBABILISTIC
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
SUBJECT MATTER EXPERTS
DATA SCIENTISTS
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
CASE STUDIES
ENERGY ANALYTICS SUMMIT 2014
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
Isolated the impact of significant variables Reconciled fragmented, unreliable, sparse data Analyzed well information from 211 wells and 2399 stages Developed accurate measure of poor vs. exceptional wells
Identified new models to include pressure depletion Identified 7 potential stages that could be eliminated Showed positive economic impact starting at $2.60 Mscf/d Identified additional opportunities to further economic benefit Identified optimization opportunities in 25% of the wells studied
Declining stimulation success in Pinedale Anticline Field Unable to understand critical factors impacting well performance High degree of uncertainty in recommending completion strategy Unable to understand interaction of multiple variables in complex geology
10
Value: 2-3% incremental annual production with lower cost
Problem
Benefits
Approach
PINEDALE: COMPLETION STRATEGY OPTIMIZATION SPE 135523
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
USING OUR NEURAL
NETWORKDERIVING VALUE FROM COMPLEXITY
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
EXPLANATION OF
PARAMETERSFROM DATA SET MESA_SAS
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
Value: 30% cost reduction in proppant with considerable savings
Identified key performance factors Analyzed data from over 11,000 wells Stratified wells by various key characteristics Developed statistical clusters based on significant variables
Reduced cycle time for job planning Increased annual well production over lifetime Developed consistent and repeatable workflows Identified significant opportunities for fracture cost reduction Enabled modeling methodology for application in de-risking new plays
Unable to understand impact of proppant volume on production Unable to isolate significant variables impacting fracking process Unable to understand interaction of multiple variables in complex geology
13
Problem
Benefits
Approach
MAJOR BARNETT OPERATOR: HYDRAULIC FRACTURING
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
Many work silos causing inefficiencies & slowing analytical efforts Inconsistent data quality, data verification, and analytics Limited toolsets and capabilities to analyze large numbers of wells Deterministic sampling process created uncertainty in forecasts Inconsistent results from well studies Inability to make best decisions based on available data
Background
Solution
Reduction in new well drilling Deferment reduced 25% Improved forecast accuracy from production wells and reservoirs 1/3 faster data collection Standardized and automated production surveillance
Problem
Reservoir maturity impacting performance creating increased need to change business process workflows for faster response to production deferments
14
Value: 5% improvement in forecast accuracy; 2% increased recovery factor
AUTOMATING WELL SURVEILLANCE: SPE 141110
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
Well Categorization
Production performance
Estimated EUR
Performance
Production forecast
Plan vs. Action
Data Integration
Integration
Standardization
Preparation
Well categorization
Well categorization
Good and Bad Wells
DCA Estimation
Event estimation
Analysis and distribution
TEXT
Well control and monitoring
Seismic data
Well Logs
Well Portfolio
DCA catalog
Production Reports
DCA Actual
Plan vs. Actual
Event quality control
Control analysis
Case Study OPTIMIZE PRODUCTION WITH
INTELLIGENT WELL
MANAGEMENT
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
OPTIMIZE PRODUCTION WITH
INTELLIGENT WELL
MANAGEMENT
Case Study
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
DATA
MANAGEMENTCase Study
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
EXPLORATORY DATA
ANALYSISCase Study
• Surface hidden patterns
• Identify trends and correlations
• Establish relationships among independent and
dependent variables [Factors and Targets]
• Data QC
• Reduce input space: Factor Analysis, PCA
• Identify key parameters for model building
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .Paper 167428•Data Mining Methodologies enhance Probabilistic Well Forecasting•K.R.Holdaway
• Traditional DCA
• Probabilistic methodology
• Well Forecasting Solution• Bootstrapping module
• Clustering module
• Data mining workflow
Case Study DATA
MODIFICATION
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .Paper 167428•Data Mining Methodologies enhance Probabilistic Well Forecasting•K.R.Holdaway
1. Cumulative liquid production
2. Cumulative oil or gas production
3. Water cut (Percentage determined
by water production/liquid
production)
4. B exponent (Decline type curve)
5. Initial rate of decline
6. Initial rate of production
7. Average liquid production
Case Study CLUSTER
ANALYSIS
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
METHODOLOGY
ASSESSMENTCase Study
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
ANALYTICAL CENTER OF
EXCELLENCE
“The fundamental idea of cross-functional
teams and goals appears to surface about
every 10 years with a new label. Usually,
attempts to implement this concept in the E&P
business ended with utter failure for a variety
of reasons.”
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
THE WITCH
METHODOLOGY
Which Witch?
Copyr i g ht © 2013, SAS Ins t i tu t e Inc . A l l r ights reser ve d .
Q&A
ENERGY ANALYTICS SUMMIT 2014