Question 2 of 12: Characteristics that Affect Log Quality

Question 2: Characteristics that Affect Log Quality (6 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

Log quality — meaning the fidelity with which a recorded curve represents the undisturbed formation — is degraded by three broad families of influence: the borehole environment, the geometry of the beds relative to the tool, and the logging operation itself.

1. The borehole environment. Every log is recorded through a column of drilling fluid in a hole whose diameter and rugosity vary, and behind a mudcake, into a formation whose near-wellbore fluids have been displaced by filtrate. Hole enlargement puts extra mud between the sensor and the wall, which is severe for pad devices (density, SNP, microlog) because the pad loses contact, and material for mandrel devices (GNT, induction) because the borehole signal grows. Mud weight and type matter as well: a barite-loaded mud absorbs capture gamma rays and biases neutron–gamma readings, an oil-based mud kills the SP and the conventional resistivity measurement, and a very saline mud short-circuits the SP current. Invasion adds a third layer: shallow-reading tools see the flushed zone rather than the virgin formation, so $R_{xo}$ and $R_t$ must be separated by a multi-depth suite before saturation can be computed. Mudcake thickness and standoff enter the density log directly through the compensation ("spine and ribs") correction, which is trustworthy only up to about 20 mm of standoff.

2. Bed thickness and the surrounding beds. Every tool has a finite vertical resolution — roughly 300 to 600 mm for pad devices, a metre or more for induction and neutron tools — and a finite depth of investigation. A bed thinner than the resolution is never read at its true value: the curve is smeared, the peak amplitude is reduced, and the reading is contaminated by the adjacent shoulder beds. This is exactly why the SP opposite a thin sand is a pseudo-static SP smaller in magnitude than the $E_{SSP}$ computed in Question 1, and why bed-thickness correction charts exist for the SP, the induction log, and the laterolog. Formation dip relative to the borehole axis has the same effect, elongating the apparent bed and blurring the boundaries.

3. The logging operation and the tool itself. Nuclear measurements are counting measurements, so their statistical precision falls as the square root of the counting time: logging too fast, or with too short a time constant, produces a curve whose wiggles are radioactive noise rather than geology. Correct calibration (primary shop calibration, wellsite before-and-after checks) is what ties a count rate to a porosity at all, and a calibration drift is invisible on the curve. Tool position matters — a centralised device that becomes eccentred, or a pad device that rides off the wall in a rugose hole, reads the borehole instead of the formation. Finally depth control (cable stretch, tension changes, sticking and jarring loose) must be consistent between runs, otherwise curves from different tool strings cannot be crossplotted, which would silently corrupt an analysis such as the one required in Question 11. Repeat sections and the standard quality-control header exist precisely to expose these problems.