SLB Limited · NYSE:SLB
SLB — Technical Presentation (2019)
12 slides from SLB's technical presentation, 2019
Slide 1: DELFI cloud native app for corporate cloud-based wellbore interpretation storage
DELFI cloud native app as a solution to build the corporate cloud-based storage with wellbore interpretation parameters and prepare the environment for advanced analytics and data mining Yngve Bolstad Johansen Chief Petrophysicist – Aker BP
Slide 2: DELFI cloud app project
DELFI Wellbore Interpretation Insights – Results from a Formation Evaluation Pilot Project DELFI cloud app project Ø Introduction Ø Todays status vs Preferred Status Ø Solutions – Cloud native app with AL algorithms Ø AkerBP Test of the DELFI Wellbore Interpretation Insights App Ø Conclusions and th
Slide 3: Introduction - Formation Evaluation
Introduction- Formation Evaluation Parametrization Sand parameter Shale parameter Measurements: Petrophysics Post Evaluation Results: Petrophysics: • Density • Net Volume • Mine data • Resistivity • Pore Volume • Digest data • Neutron counts • Oil & Water Saturation • Methods • Other.. • Permeabilit
Slide 4: Todays status vs Preferred status
Todays status vs Preferred status Ø Todays Formation Evaluation Workflow • Siloed data and expertise • Subjectivity in parametrization and answers • Experts like different data/wells • Experts like different methods • Time consuming - case by case data mining Ø Why is this unwanted? • Poor reproduci
Slide 5: Todays status vs Preferred status
Todays status vs Preferred status Ø Preferred Formation Evaluation Workflow • All data at fingertips (No silos) • Objectivity in answers • Efficient mining of data • Mine more data (not only favorite well) • Raw data • Expert behavior (Parametrization) • Use weighted averages from multiple experts Ø
Slide 6: Solutions - Why Augmented Learning?
Solutions- Why Augmented Learning ? Why a cloud native app with AL algorithms ? AkerBP, Oilfield Ex., Por vs Permeability Ø Machine Learning • Mine trends from a perfect database Producible Permeability • Requires good structure • Requires a perfect set of relevant meta data What explains 100 mD ove
Slide 7: Concept - Store method and parameter choices of Experts
Concept– Store method and parameter choices of Experts Easily find legacy interpretations Mine interpretation parameters Wells interpreted in the past in a Display smart dashboards, view simliar context, geological and histrorical records depositional environment Store interpretations in a cloud env
Slide 8: AkerBP Test of DELFI Wellbore Interpretation Insights App
AkerBP Test of DELFI Wellbore Interpretation Insights App
Slide 9: AkerBP Test of DELFI Wellbore Interpretation Insights App
AkerBP Test of DELFI Wellbore Interpretation Insights App
Slide 10: Augmented Learned Interpretation vs. AkerBP Test Reference Interpretation
Augmented Learned Interpretation vs. AkerBP Test Reference Interpretation (DELFI Wellbore Interpretation Insights App) Reference SWT Ref. Por Ref. SWT AL Por AL
Slide 11: AL based method vs AkerBP Expert, ML based method vs AkerBP Expert
AL based method vs AkerBP Expert ML based method vs AkerBP Expert DPHIT Jurassic Poor Net DSWT Jurassic Poor Net DPHIT Jurassic Poor Net DSWT Jurassic Poor Net -15 pu 15 pu 0.3 v/v 0.3 v/v -15 pu 15 pu 0.3 v/v 0.3 v/v DPHIT Jurassic Good Net DSWT Jurassic Good Net DPHIT Jurassic Good Net DSWT Jurass
Slide 12: Conclusions
Conclusions Mining Expert behavior Cloud Environment Adds valuable information on how It is perceived as useful to to do interpretations have been aggregate methods and done in the past, and identify parameters used by experts inconsistencies Augmented parameter advisor Advise parameters to be used