New sub-surface signals DOE SubTER Briefing Nov 17, 2015 Multi-Lab Working Group: Alain Bonneville (PNNL) & Robert Mellors (LLNL) 25 researchers from ANL, BNL, LANL, LBNL, LLNL, NETL, PNNL and SNL have contributed.

Motivations “…our inability to clearly characterize critical subsurface features and monitor processes in real time”

Build on results from other pillars  Long term well-bore behavior  Precise measurements of subsurface stress  Characterization of fracture flow paths

Enable adaptive control using new sensors and analysis tailored to specific problems

Objectives “…developing new approaches to sense the subsurface, analyze multiple datasets” Combine results from wellbore, stress, and Permeability pillars to develop adaptive control tools:

 New sensors to collect necessary data  New algorithms to image subsurface and forecast subsurface behavior.

Significant Benefits  Better understanding of the subsurface processes and our ability to control them in near-real time lead to more efficient and therefore lower cost operations.

 Likelihood of unexpected events will decrease  Inadvertent environmental impact will be reduced  Public confidence in the employed technologies will be increased.

Pillar Elements Develop new tools to characterize fracture distribution and behavior in situ as well as associated fluid flow and reactions with host rocks or other fluids. Incorporates advances in data mining, joint inversion, and analysis. This element is a necessary step in the definition of diagnostic signatures and then in the control of processes. Using laboratory and field data collected during abrupt transitions in subsurface system behavior, identify diagnostic signatures of critical system transitions. New acquisition software will be needed to autonomously trigger and co-acquire data from multiple sensors and to stream those datasets to computational centers

New Sensing Approaches Provide the next-generation tools and datasets required for monitoring subsurface processes

Sensors  Novel tracers for fracture system characterization and monitoring  Advanced fiber-optic monitoring tools for seismic, EM and ER detection  Large N-Arrays (seismic, deformation, gravity, ER and EM)  4-D muon density tomography

Key points  Leverage advances in material science, manufacturing, and massive data handling  Ranges from wellbore to reservoir and regional scale.

Sensors: activity plans Novel tracers for fracture system characterization and monitoring

Identify candidate intrinsic tracers, co-injected tracers, and natural fracture geophysical signatures suitable for pursuit. Demonstrate in field the use of improved tracers and natural signals to characterize a field fracture network.

Priority activities

Advanced Fiber-Optic Monitoring Tools for Seismic, EM and ER Detection

Design and construct a fiberoptic point EM vector sensor and distributed EM sensor.

Demonstrate the utility of the enhanced fiber-optic sensing systems for field scale real-time monitoring of fracture behavior.

adaptive control through identification of diagnostic signatures and critical thresholds by transformative collection and analysis of subsurface signals.

Novel tracers for fracture system characterization and monitoring Develop new ways to image subsurface flow.  Inject conductive and non-conductive tracers  Track using array of electrical resistivity sensors for time lapse monitoring

 Identification of isotopic/geochemical signatures associated with fracture stimulation  National Labs possess advanced experimental and tracers data analysis capabilities.

Advanced fiber-optic monitoring tools for seismic, EM and ER Detection Develop reservoir scale cheap monitoring of fractures and seismicity  Low-cost fiber optic enables largescale deployment for monitoring of wellbore conditions including flow, induced seismicity, and new signals.  Resilient to extreme conditions (high temperatures)  Leverages national lab capabilities in custom fiber manufacture, photonics, modeling, field deployment, and analysis.

State-of-the-art  Acoustic sensors available commercially (both distributed and point [e.g. Bragg] sensors)  Useful for wellbore environment (flow and temperature); less useful for reservoir.  Fiber based sensors capable of sensing electric or magnetic fields possible but not deployed in field  Little testing of optimized custom fiber design (rather than standard telecom type)  Modeling of fiber response and coupling poor at present.

Integration of multi-scale and multi-type datasets Improve understanding of the subsurface system at all scales.  Improved resolution of sub-surface permeability and flow from wellbore to reservoir  Use all data including time-lapse and ‘big data’

 Essential for adaptive control (needs to be fast)  National labs possess world-class modeling capabilities necessary for multidataset inversion in addition to high-performance computing.

State-of-the-art  Deterministic joint inversion (e.g. MT and seismic)  Stochastic using Monte Carlo type methods  Industry: generally reflection seismic plus gravity or MT

 Need to increase speed for adaptive control

Approaches    

Add additional datasets to joint inversion Monte Carlo stochastic for uncertainty resolution Increase capability for large datasets Reduced order models for increased speed necessary for adaptive control  Use all data and combine with modeling for validation  National Labs are experienced in analysis, modeling, and UQ

Log data

seismic

surface

THMC

Diagnostic Signatures and Critical Thresholds Identify dramatic transitions and anomalous events (sudden failure in wellbore integrity, induced seismicity on fault, breached caprock).

If successful signatures are found, this would be a game-changer.  Current state-of-the-art level: low (TRL < 3)  Nonlinear and need complexity-based systems  Adapt techniques from other fields (cybersecurity)

Triggered seismicity, Spain. Cesca et al., 2014. ~$1.8 billion loss.

Activities  Collect community datasets (esp. from industry)  Correlating near-surface and seismic signatures to map stress release

Adaptive Control Processes Real-time data acquisition combined with simultaneous modeling for prediction and optimization.    

Ultimate goal of SubTER Measure conditions, determine action, apply action, measure response Requires contributions from all pillars Expected emphasis in years 5-10  Adaptive sensor deployment  Identification of critical thresholds  Real-time modeling capability  Metrics:  Improved fracture design [metric: cost].  Reduction of unexpected events [with respect to plan].  Current state-of-art: basic

Metrics for success Element

Product

Possible metric

New sensors

Hardware and analysis

Cost; resolution; durability; environmental impact; adoption by industry

Multi-scale

Software algorithms

Resolution; speed; licensing

Critical thresholds

Technique

Resolution; speed; publications

Adaptive control

Hardware, software

Cost; effectiveness

New Subsurface Signal Activities important for Success of other Pillars Element

Activity

Wellbore

New sensing

Tracers

✔ ✔

Critical thresholds

Adaptive control

✔

✔ ✔

✔

✔

✔

Stochastic

✔

✔

Deterministic

✔

✔

✔

✔

Muon Multi-scale

Perm. ✔

✔

Arrays Fiber optic

Stress

Joint analysis

Community datasets

✔

✔

Stress release

✔

Interferometry

✔

✔

✔

✔

✔

✔

Data acquisition Adaptive sensor

✔

Potential Industry/University Involvement/Participation  Petroleum industry is heavily involved in new sensor development and permanent deployment for reservoir monitoring (e.g. BP, CGGVeritas, Schlumberger).  Industry is also engaged in efforts related to adaptive control (“smart field” approaches) as a way to increase efficiency and has access to thousands of wells for testing.  Several innovative monitoring techniques and big data approaches are developed in universities or research centers associated to Universities. Such university/industry consortia can greatly facilitate access to field testing of instruments and methods and to “Large-N’ type datasets.

Sapling project: Borehole Muon Detector for 4D Density Tomography Motivation & Challenge  Develop miniaturized muon tracking detectors capable of fitting in standard boreholes to perform 4D density tomography of geological structures.  Develop a rapid and efficient inversion method that will take into account not only the different muon paths, but also the data generated by other techniques, such as seismic and gravity. First result

The first prototype has been built and is operational with its power supply and computer interface. A series of tests will be conducted in laboratory and in tunnel during the next coming months.

Summary  Adaptive control requires improved real-time monitoring and rapid analysis which in turn depend on optimized data acquisition, rapid data assimilation, and innovative data management and fusion.  To reach this ultimate goal, new sensors and new algorithms to image subsurface and forecast subsurface behavior will be developed.  These new tools and methods will be focused on characterizing fracture distribution and behavior in situ as well as associated fluid flow and reactions with host rocks or other fluids.  These approaches will be multi-scale applying to both wellbore and reservoir scales.