Skip to content

dataset_coverage_audit.md

Latest commit

 

History

History
54 lines (35 loc) · 2.44 KB

dataset_coverage_audit.md

File metadata and controls

54 lines (35 loc) · 2.44 KB

Dataset Coverage Audit (Volve)

Motivation

The public Equinor Volve dataset is a field archive release rather than a curated research dataset. Data availability varies across wells and disciplines. Before engineering analysis, a systematic audit of data availability was performed to identify which wells and data types are sufficiently represented for defensible research.

The audit is designed to answer practical questions:

  • Which wells contain drilling telemetry?
  • Which wells contain structured daily drilling reports?
  • Which wells contain wireline logs?
  • Which wells contain reservoir modeling files?
  • Which wells contain seismic or interpretation data?

Methodology

A frozen filesystem catalog (metadata-only) was used as the source inventory. The catalog enumerates file paths and tags without modifying the Volve dataset.

The audit steps were:

  1. Filter to file entries only (directories excluded).
  2. Remove system-level archive folders that dominate row counts but do not represent petroleum engineering content.
  3. Extract well identifiers from file paths using Volve well naming patterns (for example, 15/9-F-1 and wellbore suffixes).
  4. Use existing catalog tags to classify files into petroleum-relevant categories (for example: DDR_XML, WITSML, XML, DLIS, LAS, SEGY, WELL_TECH, DOCS, RMS_MODEL).

A well x data-category matrix was produced indicating presence or absence of each category per well. Category counts were also computed.

Results summary

After excluding system-level archives, the remaining dataset is dominated by:

  • Drilling telemetry and structured XML
  • Well technical documents and reports
  • Wireline logs (DLIS, LAS, LIS)
  • Reservoir modeling files (RMS model)
  • Seismic and interpretation data in smaller volumes

Well-level content was identified for a subset of wells. Data richness varies substantially between wells.

Implications for research design

This coverage audit enables defensible well selection and reduces selection bias. Subsequent engineering studies should select wells based on documented availability of:

  • continuous drilling telemetry and structured operational reporting
  • supporting logs and technical documentation

Wells lacking sufficient category breadth should be excluded from operational behavior studies and time-series based analyses.

Outputs produced

  • outputs/coverage_by_tag.csv
  • outputs/coverage_by_well.csv
  • outputs/coverage_matrix_binary.csv
  • outputs/coverage_matrix_counts.csv