Two Takeaways

  • Survey verification should not require expertise in surveying Error-model-based
  • QC should be possible using the Error model

Expertise Requirement

  • Most people who drive cars are not mechanics
    • There are warning signs when you need one
  • Most who drill and survey wellbores are not survey experts
  • Consumers of the data may be even less of an expert
  • How do they know when there is a problem?
  • Importance of error-model-based QC

Error-Model-Based QC: A Brief History

  • Pre-Error Model
    • Measure deviation from references
    • Many standards, usually fixed thresholds
  • SPE 103734, Ekseth, et al (2006)
    • Define weighting functions, Root-Sum-Square
    • Dynamic QC – Changes with orientation
  • Maus, et al (2017)
    • Account for error covariance
    • Compute “sigma distance”

Shortcomings of These Methods

  • Focus on single survey evaluation
    • User is interested in the set as a whole
    • Exception: MSE in 103734, but use and interpretation requires a knowledgeable user
  • Real-World workflows can lead to complacency
    • Once one survey fails, all the rest will likely fail
    • “Drill ahead, this always happens near vertical!”
  • Escalation procedures often assume some level of expertise
    • “If you identify interference from an offset well, notify town”
    • Assumes that they already know if the survey is good or bad

Decision Making With QC

  • When do I stop drilling?
  • When do I need to resurvey the well?

...

View the entire Presentation:

Survey QC, Decision Making, and a Modest Proposal for Error Models

Marc Willerth, MagVar

 

More from Meeting 47