Repeatability

Key Takeaways

• 4D seismic (time-lapse seismic) acquisition demands the highest standard of repeatability in seismic data collection because the entire interpretive value of a 4D survey depends on isolating genuine reservoir changes (from fluid substitution, pressure depletion, and compaction) from differences caused by inconsistent acquisition between the baseline and monitor surveys — if the source and receiver positions, source signatures, recording filters, ambient noise conditions, and near-surface conditions are not replicated within tight tolerances between survey vintages, the differences seen in the data may reflect acquisition inconsistency rather than reservoir change; 4D repeatability is measured by the Normalized Root Mean Square (NRMS) difference between baseline and monitor surveys in areas of the reservoir that are not expected to have changed, with NRMS values below 20% considered excellent and values above 40% indicating significant acquisition noise that will obscure subtle reservoir changes; achieving high 4D repeatability requires positioning receivers within 5-10 meters of baseline positions (or using permanently installed ocean-bottom nodes), matching shot positions within similar tolerances, and using identical acquisition parameters between vintages; the multi-billion dollar investment in permanent seabed receiver systems for 4D monitoring at major North Sea fields is justified entirely by the superior repeatability of permanent receivers compared to towed streamer resurveys.
• Wireline log repeat sections are the standard quality control check for wellbore measurement repeatability, in which the logging tool is run back over a 200-500 foot section of the wellbore a second time immediately after the main log pass and the two readings are compared to verify tool stability — a well-functioning tool should reproduce the main log response within the manufacturer-specified repeatability tolerance (typically 0.5% for density, 1 porosity unit for neutron, and 0.02 ohmm for resistivity in the moderate range); a repeat section that shows significant departure from the main log indicates a malfunctioning sensor, tool-to-borehole contact problems (tool spinning in a washed-out section), borehole rugosity effects, or electronic drift during the log run; when the repeat section fails to match the main log within tolerance, the well log data from the questionable tool is flagged as unreliable, and a decision is made whether to re-log the section (adding rig time and tool rental cost) or accept the data with documented uncertainty; the repeat section is routinely shown as a separate curve track on the delivered log print so formation evaluation geologists can visually assess the quality of the measurement at each depth.
• Core plug analysis repeatability is established by API RP 40 standards which specify that porosity measurements on replicate plugs from the same core sample should agree within 0.5 porosity units and that permeability measurements should agree within 5% for values above 1 millidarcy and within 15% for values in the microdarcy range — achieving these repeatability limits requires rigorous sample preparation (cleaning to remove all hydrocarbons and water, drying to a stable weight, and trimming end faces perpendicular to the plug axis), calibrated measurement apparatus (Boyle's law helium pycnometer for porosity, steady-state or pulse-decay permeameter for permeability), and consistent measurement protocols (same confining stress conditions for each plug, same fluid viscosity, same measurement temperature); core laboratories that submit to API RP 40 proficiency testing programs run the same blind samples multiple times and between multiple laboratories to verify that their repeatability and reproducibility meet industry standards; operators selecting a core laboratory for a major development program typically request the laboratory's proficiency testing results as part of the vendor qualification process, because core data quality directly affects the reservoir model used to justify the entire field development investment.
• Drilling fluid testing repeatability under API 13B-1 (water-based mud) and 13B-2 (oil-based mud) standards ensures that mud density, viscosity, gel strength, and filtration measurements made by different mud engineers on the same fluid will agree within acceptable limits — for example, API 13B-1 specifies that Marsh funnel viscosity measurements on the same fluid by two different technicians should agree within 2 seconds, and mud weight measurements with a pressure mud balance should agree within 0.05 lb/gal; these repeatability limits are important because drilling engineers make weight-on-bit adjustments, trip speed limits, and kill weight calculations based on mud weight readings, and inconsistent readings introduce safety risk; the same standards are used to demonstrate when a measurement device (funnel, balance, viscometer) needs recalibration, because measurements that fall outside the repeatability limits on identical samples indicate instrument error rather than sample variation; on offshore rigs where mud properties are measured multiple times per shift by different mud engineers, repeatability verification through cross-checks between shifts is a standard quality assurance step.
• Repeatability in PVT (pressure-volume-temperature) fluid analysis is demonstrated by running at least two representative samples from the same reservoir through the same PVT test program and verifying that key parameters (bubble point pressure, GOR at separator conditions, oil shrinkage factor, gas Z-factor) agree within published tolerance limits before the data are accepted for reservoir simulation use — discordant PVT results from two samples that should represent the same fluid indicate either sample contamination (drilling fluid filtrate invasion into the sample), phase separation during sampling (gas liberation before the sample reached the surface), or sampling from a geochemically different fluid compartment than expected; when PVT repeatability checks fail, the engineer must determine whether the discordance reflects real reservoir heterogeneity (the two samples came from different fluid contacts or different compartments) or sampling error (both samples were collected from the same interval but one is contaminated); reservoirs with unreliable PVT data require larger uncertainty ranges in the reservoir model and may need additional sampling wells before field development commitments are made.