Core analysis · MunaiHub

Quick refresher

Core is a direct measurement, but only at points. A sample of a few tens of cubic centimetres represents the reservoir only together with hundreds of other samples and with the well logs.
The path of core: coring → preservation → scanning and depth matching → sawing and plugging (cutting out the samples) → cleaning and drying → measurements. An error at any step cannot be corrected by the later ones.
Routine core analysis (RCA) gives porosity, permeability, grain density and saturation on samples taken every 25–30 cm. It gives a profile of properties along the section. Special core analysis (SCAL) gives capillary pressure, relative permeabilities, wettability (the preference of the rock surface for water or for oil), and electrical and mechanical properties, on a few carefully chosen samples.
The state of the core is chosen for the task. For porosity and absolute permeability the sample is cleaned and dried. For wettability and relative permeabilities the core must be in its native state, or with restored wettability.
Laboratory conditions are not reservoir conditions: there is no overburden pressure (the weight of the overlying rocks), the fluid is different, and the clays are dry. The routine gas permeability is too high, and it is brought to reservoir conditions with corrections.
Core, well logs and well tests measure different volumes of rock. They can be compared only after depth matching, bringing them to the same conditions and correct averaging.

Key relationships

\text{core recovery} = \dfrac{L_{\text{recovered}}}{L_{\text{cored}}}\cdot 100\,\%

the share of the cored interval that is represented by core

z_{\text{log}} = z_{\text{core}} + \Delta z

depth matching of the core; the shift $\Delta z$ is found from the gamma-ray activity

P_s = \dfrac{200\,C_v}{\sqrt{N_s}}\ \%

accuracy of the mean over $N_s$ samples at a coefficient of variation $C_v$ (the spread of the values divided by their mean; confidence level 95%)

Common pitfalls

Comparing the gas permeability of core, measured without confining pressure (the pressure that squeezes the sample from all sides), directly with the result of a well test.
Measuring relative permeabilities on cleaned core without restoring the wettability.
Drying clay-rich samples in an ordinary oven. The structure of the clays is destroyed, and the porosity and permeability come out too high.
Taking the mean over the samples as the mean over the reservoir, without checking where the samples came from and how much core was lost.
Comparing core with well logs without a depth shift.

After this lesson you can

describe the path of core from coring to the laboratory report, and say which properties can be distorted at which step
tell fresh, preserved, cleaned and restored-state core apart, and choose the state of the core for the task
list what is included in routine and in special core analysis, and what these data are needed for
explain why core is scanned and how it is matched in depth to the well logs
name the differences between laboratory and reservoir conditions, and bring core values to reservoir conditions
compare core data with well logs and well tests, taking into account the scale of each measurement

What core gives and what it cannot give

Core is a column of rock cut out of the reservoir during drilling and brought to the surface. Only on core can the reservoir rock be seen as a whole and its properties measured directly. Cuttings, the small fragments of rock carried out by the drilling fluid (mud), give only the composition of the rock and an approximate depth. Well logging, seismic surveys and well tests do not measure the properties of the rock itself. They measure its response to an electric current, radiation, sound or the withdrawal of fluid. How to convert this response into properties is established on core.

Description of the rock: lithology (the type of rock), bedding, fractures and traces of living organisms. This is the basis for ideas about the conditions in which the sediment was laid down, and for the geological model.
Routine properties: porosity, permeability, grain density and saturation every 25–30 cm of the section.
Special properties: capillary pressure, relative permeabilities, wettability, and electrical, acoustic and mechanical characteristics, on a small number of samples.
Pore structure and composition: thin sections (slices of rock thin enough to look through under a microscope), electron microscopy, X-ray analysis of minerals, grain sizes.
Calibration of well logs: the relations “instrument reading — property of the rock” and “porosity — permeability”.

Core has three limitations. It is small: a sample of a few tens of cubic centimetres says nothing about fractures and non-uniformity at the scale of a metre. It is altered: during drilling the mud filtrate (the liquid part of the mud) enters it, while it is brought up the pressure and temperature fall, and the rock is unloaded (the stress on it is released). And it is rare: core is taken in few wells and not over the whole section, because this is expensive and slows the drilling. So we use core data together with well logs and well tests, not instead of them.

SchematicThe path of core from the well to the results: coring and preservation, scanning, sawing and plugging, measurements. All the samples go through routine analysis, and only a few chosen ones go through special analysis (a schematic, not to scale).

How core is taken

In ordinary drilling the bit breaks the rock over the whole bottom of the hole. To get core, the crew replaces the bit with a core bit with a hole in the centre. It cuts out a ring, and the untouched column of rock enters the inner tube inside the core barrel, the coring tool. The diameter of the core is usually from 67 to 133 mm, most often about 100 mm (4 inches). In one trip into the hole the crew cuts 9, 18 or 27 m (30, 60 or 90 ft) of core. At the surface the crew takes the core out, often right in a liner of aluminium or fibreglass (a tube that holds the core). They mark the depths and the “top–bottom” direction and cut the core into one-metre pieces.

Core recovery is the ratio of the length of core brought up to the length of the interval drilled. In strong, well-cemented rocks it is close to 100%. Weakly cemented and fractured rocks break up and wash out, and loose sand without a special tool simply falls apart at the surface. The weakest rocks are lost most often, and these are often the most porous and permeable ones. So when the recovery is low, the core that is left shows the reservoir as worse than it really is.

Full-diameter core: core of the full diameter. The laboratory studies its pieces whole when the rock is non-uniform at the scale of a small sample: fractured and vuggy carbonates (limestones and dolomites with holes and cavities in the rock), and conglomerates (rocks made of cemented pebbles).
Sidewall core: small samples taken from the wall of a well that has already been drilled, with a tool on a cable: a sidewall coring tool. A percussion (shooting) tool drives a hollow bullet into the wall and damages the rock. A rotary (drilling) tool cuts out the sample and damages the rock much less. It is quick and cheap, but the samples are small and are taken from the zone that the mud filtrate has invaded.
Pressure-retained core: the core barrel is closed at the bottom and holds the reservoir pressure, so that fluids do not leave the rock while it is lifted. At the surface the core is frozen, and it is thawed only in the laboratory, where the fluids that come out are collected. A simpler variant is a core barrel with an absorbent liner, which holds the oil that is pushed out of the core during the lift.
Oriented core: the crew marks longitudinal scribe lines on the core when it is cut, and an instrument records their position relative to the compass directions. This core shows the direction of fractures and the dip of the beds (how steeply the layers are tilted).

The drilling fluid also affects the properties of core. The filtrate of a water-based mud pushes the mobile oil out of the core, and the measured water saturation comes out too high. A hydrocarbon-based mud adds no water, but its surfactants (chemicals that change how fluids stick to a surface) change the wettability, which is the preference of the rock surface for water or for oil (Fluid saturation, Wettability and interfacial phenomena). So the crew chooses the type of mud for the main aim of the coring. Special core bits and fast drilling reduce the invasion of the filtrate. Sometimes the crew adds a tracer (a marker substance) to the mud, to estimate later how much filtrate has got into the core.

At the rig: how to preserve the core

Core starts to change as soon as it reaches the surface. Water and the light fractions of the oil evaporate, the oil is oxidised by the air, and clays dry out and crack. After only a few hours, unprotected core is no good either for measuring saturation or for studies that depend on wettability. So the crew preserves core for such studies at the rig:

They seal the ends of the liner in which the core was brought up.
They wrap the pieces in plastic film and foil and coat them with paraffin or wax.
They seal the pieces in multilayer bags that let no oxygen or moisture through.
They freeze loose core, or soak it in resin, so that it does not fall apart. Freezing can itself change the properties of the rock, so keep that in mind.

The laboratory report must say in what state the core was. The generally accepted classification is this:

Fresh core: just brought up and protected at once from drying and from air. The crew records the mud on which it was cut, because the mud decides what can be trusted in the core.
Preserved core: the same fresh core after storage in sealed packing.
Cleaned core: core from which the oil, water and salts have been removed with solvents. The laboratory records the method of cleaning, because solvents and heat can change the rock.
Restored-state core: cleaned core that has been saturated again with reservoir water and oil and held at reservoir temperature, to bring back the wettability that it had in the reservoir (Wettability and interfacial phenomena).
Pressure-retained core: core brought up without lowering the pressure. It is needed when saturation is the main aim.

Which core is needed for which measurement

Measurement: Porosity, grain density, absolute permeability
State of the core: cleaned and dried
Why: no fluids or salts must be left in the pores

Measurement: Saturation
State of the core: fresh, or better still with the pressure retained
Why: fluids must not be lost before the measurement

Measurement: Wettability, relative permeabilities, oil–water capillary pressure
State of the core: fresh with unchanged wettability, or with restored wettability
Why: wettability decides how the phases are distributed in the pores

Measurement: Electrical properties for the calculation of saturation
State of the core: with reservoir wettability
Why: the saturation exponent depends on where the water sits in the pores

Measurement: Mercury injection
State of the core: cleaned and dried
Why: mercury is forced into empty pores

Measurement: Strength and compressibility
State of the core: if possible not dried
Why: the strength of clay-rich rocks and chalk depends on moisture

Measurement	State of the core	Why
Porosity, grain density, absolute permeability	cleaned and dried	no fluids or salts must be left in the pores
Saturation	fresh, or better still with the pressure retained	fluids must not be lost before the measurement
Wettability, relative permeabilities, oil–water capillary pressure	fresh with unchanged wettability, or with restored wettability	wettability decides how the phases are distributed in the pores
Electrical properties for the calculation of saturation	with reservoir wettability	the saturation exponent depends on where the water sits in the pores
Mercury injection	cleaned and dried	mercury is forced into empty pores
Strength and compressibility	if possible not dried	the strength of clay-rich rocks and chalk depends on moisture

Fresh core whose wettability has been changed neither by the mud nor by storage is called core in the native state.

Scanning, depth matching and choosing the samples

While the core is whole, the laboratory studies it by methods that destroy nothing. These methods do two jobs: they tie the core to the well logs, and they show what is inside the core before anyone chooses the places for the samples.

Core gamma scan: the core on a conveyor passes by a detector in lead shielding, and the detector records the natural gamma-ray activity of the rock along the length of the core. The result is a curve that is compared with the gamma-ray log of the well.
X-ray computed tomography (CT): the rock weakens an X-ray beam the more, the higher its density and the heavier its atoms. The scanner passes X-rays through the core from different sides, and a computer builds a three-dimensional image from them. The image shows fractures, cavities, dense inclusions, bedding and traces of mud invasion. The detail is limited by the width of the beam and the size of one image element (the smallest dot of the image).
Magnetic resonance imaging: shows not the rock but the fluids: where the water and the oil are in the core.

The depth of core is known from the length of the drill string, and the depth of the well log from the length of the cable. These two depths differ by a metre or two, and sometimes more. In addition, when the recovery is incomplete, it is not known from which part of the interval the core is missing. To give each sample a depth on the well log, we shift the gamma-ray activity curve of the core until its characteristic peaks match the gamma-ray log curve.

CalculatedDepth matching of core for the data of the example below. Left: the gamma-ray log of the well. Right: the gamma-ray activity of the core, recorded against its own depth marks. The same peaks on the core are 2 m higher, and by this amount the core is shifted. The curves are assumed.

Worked exampleCore recovery and depth shift

Given (assumed data)

Cored interval by the length of the drill string: 2450.0–2468.0 m
Length of the core recovered: 17.1 m
Gamma-ray activity peaks by the core marks: 2452.4 m; 2458.9 m; 2465.1 m
The same peaks on the gamma-ray log: 2454.3 m; 2461.0 m; 2467.1 m
Sample: depth by the core marks 2460.25 m

Solution

Core recovery: $17.1/18.0\cdot100 = 95$ %. 0.9 m was not recovered.
Shift at each peak (log minus core): 1.9; 2.1; 2.0 m. The mean shift is $\Delta z = 2.0$ m.
Depth of the sample on the log: $2460.25 + 2.0 = 2462.25$ m.

Answer: Core recovery is 95%. The core must be shifted down by 2.0 m, and the sample falls at a depth of 2462.25 m on the log.

What it means. Without the shift, the sample would be compared with the log readings 2 m higher, which in a layered section can be quite a different rock. The shifts at the three peaks differ by 0.1–0.2 m. This is the accuracy of the depth matching, and it is enough for a log with a vertical resolution (the smallest thickness it can see) of a few decimetres. By default we assign the missing 0.9 m to the bottom of the interval, but we check this on the same curve. Here the shift at the lowest peak is the same as at the upper ones, so it is really the end of the interval that was lost. The data in this example are assumed for illustration.

After scanning, the laboratory saws the core lengthwise. It photographs the flat face in ordinary and in ultraviolet light (oil glows in ultraviolet), describes the face in detail, and keeps one part of the core as an archive. From the other part it drills core plugs, cylinders 25–38 mm in diameter (1–1.5 inches). Horizontal plugs follow the bedding (the layers of the rock), usually every 25–30 cm. Vertical plugs cross the bedding and are taken less often. Horizontal plugs give the horizontal permeability, and vertical plugs give the vertical permeability (Rock permeability). The offcuts from the ends go to thin sections and mineral analysis.

Cleaning and drying the samples

To measure porosity and permeability, everything must be removed from the pores: oil, water, and the salts left after the water evaporates. The laboratory washes the sample with solvents until they stop being coloured, and dries it to constant mass (until its mass stops changing).

Cleaning methods

Method: Soxhlet apparatus
How it works: the solvent boils, condenses, drips onto the sample and runs back, again and again
Features: gentle but slow: from a day to several weeks

Method: Dean–Stark apparatus (in Russian-language practice also called the Zaks apparatus)
How it works: the sample sits in the vapour of a boiling solvent; the water that is distilled off collects in a graduated trap
Features: gives the water saturation at the same time (Fluid saturation)

Method: Flow-through under pressure
How it works: the solvent is pumped through the sample in a core holder (a cell that grips the plug)
Features: fast; for tight rocks

Method: Centrifuge
How it works: the solvent is forced through the sample by centrifugal force
Features: fast; a risk of damaging weak rock

Method	How it works	Features
Soxhlet apparatus	the solvent boils, condenses, drips onto the sample and runs back, again and again	gentle but slow: from a day to several weeks
Dean–Stark apparatus (in Russian-language practice also called the Zaks apparatus)	the sample sits in the vapour of a boiling solvent; the water that is distilled off collects in a graduated trap	gives the water saturation at the same time (Fluid saturation)
Flow-through under pressure	the solvent is pumped through the sample in a core holder (a cell that grips the plug)	fast; for tight rocks
Centrifuge	the solvent is forced through the sample by centrifugal force	fast; a risk of damaging weak rock

There is no single recipe: the solvent is chosen for the oil and the rock. Usually the oil is removed with toluene, and the water and salts with methanol. For heavy oils and hydrophobic rocks (rocks that repel water) stronger mixtures are used.

Drying seems a simple operation, but it is the step that most often spoils clay-rich samples. Clay minerals hold water on their surface and between the layers of their crystal structure. If heat drives this water out, the clay particles shrink. Thin clay fibres, for example of illite (a clay mineral that covers the pores with a web), stick together and press against the pore walls. The pores are “freed”, and the measured porosity, and especially the permeability, come out higher than for the same rock in the reservoir.

Drying methods

Method: Ordinary oven
Conditions: about 100–115 °C
For which rocks: clean sandstones and carbonates without sensitive minerals

Method: Humidity-controlled oven
Conditions: about 60 °C at a relative humidity of about 40%
For which rocks: clay-rich rocks: a film of water stays on the clays, as in the reservoir

Method: Critical-point drying
Conditions: the liquid in the pores is replaced by liquid carbon dioxide, and this is heated above its critical point (the pressure and temperature above which liquid and gas are no longer different)
For which rocks: rocks with fragile fibrous clays

Method	Conditions	For which rocks
Ordinary oven	about 100–115 °C	clean sandstones and carbonates without sensitive minerals
Humidity-controlled oven	about 60 °C at a relative humidity of about 40%	clay-rich rocks: a film of water stays on the clays, as in the reservoir
Critical-point drying	the liquid in the pores is replaced by liquid carbon dioxide, and this is heated above its critical point (the pressure and temperature above which liquid and gas are no longer different)	rocks with fragile fibrous clays

The drying method decides what porosity you get. After an ordinary oven the pore volume includes the space that clay-bound water occupies in the reservoir; after drying at controlled humidity it does not.

Routine and special analysis

We divide the studies into two parts. Routine core analysis (RCA) means quick and inexpensive measurements on all the samples. Special core analysis (SCAL) means long and expensive tests on a few samples. These tests answer the specific questions of the project.

Routine and special core analysis

: What is measured
Routine (RCA): porosity, gas permeability, grain density, saturation
Special (SCAL): capillary pressure, relative permeabilities, wettability, electrical, acoustic and mechanical properties, pore compressibility

: Samples
Routine (RCA): all, every 25–30 cm
Special (SCAL): a few or a few tens, by rock type

: State of the core
Routine (RCA): cleaned and dried (fresh for saturation)
Special (SCAL): native state or restored wettability

: Test conditions
Routine (RCA): room temperature, low confining pressure
Special (SCAL): reservoir pressure, temperature and fluids, where possible

: Time
Routine (RCA): days to weeks
Special (SCAL): months

: Purpose
Routine (RCA): profile of properties along the section, calibration of well logs, relation between porosity and permeability
Special (SCAL): input data for the calculation of saturation, displacement and the reservoir simulation model

	Routine (RCA)	Special (SCAL)
What is measured	porosity, gas permeability, grain density, saturation	capillary pressure, relative permeabilities, wettability, electrical, acoustic and mechanical properties, pore compressibility
Samples	all, every 25–30 cm	a few or a few tens, by rock type
State of the core	cleaned and dried (fresh for saturation)	native state or restored wettability
Test conditions	room temperature, low confining pressure	reservoir pressure, temperature and fluids, where possible
Time	days to weeks	months
Purpose	profile of properties along the section, calibration of well logs, relation between porosity and permeability	input data for the calculation of saturation, displacement and the reservoir simulation model

Other lessons cover how each measurement is made and what it means: Porosity, Rock permeability, Fluid saturation, Capillary pressure and the transition zone, Relative permeability, Wettability and interfacial phenomena. Electrical measurements give the cementation and saturation exponents for Archie's equation (Archie's equation and saturation of clean reservoirs). Tests under load give the pore compressibility and the dependence of the properties on stress (Rock compressibility and the effect of stress on core properties).

Order of work

1
Questions
We draw up the programme before coring: which decisions depend on core, and which measurements they need. The mud, the method of coring and the method of preservation depend on this.
2
Non-destructive studies
Gamma-ray scan, tomography, photographs, description. We use them to mark the places for the samples.
3
Routine analysis
All the samples. From the “porosity–permeability” plot and the description, we divide the rock into types.
4
Choosing samples for special analysis
Take uniform samples from each rock type (check them by tomography), so as to cover the whole range of properties, not only the best reservoir rock.
5
Special analysis
First come the tests after which the sample can be used again. Destructive ones, such as mercury injection, thin sections and crushing, come last.

Laboratory conditions and reservoir conditions

Routine measurements are designed to be quick and reproducible (they give the same result when repeated), not to copy the reservoir. Before you use a result, understand how the conditions of the test differed from those in the reservoir.

How a routine test differs from the reservoir

Factor: Stress
In a routine test: a confining pressure of a few megapascals, only to stop the gas going round the sample
In the reservoir: effective stress of tens of megapascals
What happens to the result: porosity and permeability are too high

Factor: Fluid
In a routine test: dry gas: nitrogen, helium, air
In the reservoir: oil, water, gas under pressure
What happens to the result: gas slips at the pore walls: permeability is too high

Factor: Clays
In a routine test: dried
In the reservoir: contain water and take more space
What happens to the result: permeability is too high

Factor: Phases in the pores
In a routine test: one
In the reservoir: two or three
What happens to the result: absolute permeability was measured, but the flow needs the effective one

Factor	In a routine test	In the reservoir	What happens to the result
Stress	a confining pressure of a few megapascals, only to stop the gas going round the sample	effective stress of tens of megapascals	porosity and permeability are too high
Fluid	dry gas: nitrogen, helium, air	oil, water, gas under pressure	gas slips at the pore walls: permeability is too high
Clays	dried	contain water and take more space	permeability is too high
Phases in the pores	one	two or three	absolute permeability was measured, but the flow needs the effective one

All four differences act in one direction: they make the permeability too high, and the more so, the tighter and more clay-rich the rock is.

We find the corrections on some of the samples. We measure gas permeability at several pressures and convert it to liquid permeability (the Klinkenberg correction, Rock permeability). We measure porosity and permeability under a confining pressure that matches the mean effective stress in the reservoir (Rock compressibility and the effect of stress on core properties). We also measure permeability with water of the reservoir composition. Then we apply the relations we have found to the other samples of the same rock type.

Worked exampleFrom laboratory values to reservoir values

Given (assumed data)

Routine analysis of a sample: porosity 21.4%; permeability to nitrogen 85 mD
After the Klinkenberg correction: 78 mD
Under reservoir confining pressure (from tests on samples of this type): porosity × 0.96; permeability × 0.86
Permeability to reservoir water divided by permeability to gas under confining pressure: 0.90
Relative permeability to oil at irreducible water saturation (the water that cannot be pushed out of the pores): 0.80

Solution

Porosity in the reservoir: $21.4\cdot0.96 = 20.5$ %.
Gas slippage: 78 mD instead of 85, which is 8% less.
Confining pressure: $78\cdot0.86 = 67.1$ mD.
Water instead of gas: $67.1\cdot0.90 = 60.4$ mD, the absolute permeability at reservoir conditions.
For oil at irreducible water saturation: $60.4\cdot0.80 = 48$ mD, the effective permeability that a well test “sees” in the water-free zone.

Answer: Porosity at reservoir conditions is 20.5%, and the effective permeability to oil is about 48 mD instead of the 85 mD of routine analysis.

What it means. Only 57% of the laboratory number is left. If you compare the original 85 mD with a well test, people will blame the reservoir or the well for the discrepancy of a factor of 1.8, although the conditions of measurement explain it entirely. For tight, clay-rich rocks the same corrections reduce the permeability several times. The multipliers in the example are assumed. Find them by tests on your own rock, and do not take them from other people's data.

Special studies try to run at reservoir conditions from the start: under confining pressure, at reservoir temperature, with reservoir water and oil, sometimes with oil that has gas dissolved in it. This is expensive. But relative permeabilities and capillary pressure depend on wettability and on the interfacial tension (the tension at the boundary between the fluids), and they cannot be brought to reservoir conditions with simple multipliers.

Core, well logs and well tests: how to compare

The same property measured by different methods comes out different, and not only because of the conditions of the test. Each method averages the property over its own volume of rock.

SchematicThe volume of rock over which each measurement method averages the properties (orders of magnitude; logarithmic scale). Between a core plug and a well test there are about ten orders of magnitude (each order of magnitude is a factor of 10).

Before comparing

1
Match in depth
Shift the core to the depths of the well log and make sure that the same layers are being compared.
2
Bring to the same conditions
Bring the core values to reservoir stress and fluids, as in the example above.
3
Compare like with like
Core porosity after drying in an ordinary oven includes the space of clay-bound water, and it corresponds to the total porosity from well logs. Core gives the absolute permeability, and a well test gives the effective permeability for the fluid that is flowing.
4
Average over the volume of the measurement
Average the samples in a window equal to the resolution of the instrument or to the open thickness. Porosity is averaged arithmetically, and permeability in the way that fits the direction of flow (Rock permeability).
5
Allow for what the samples do not contain
Fractures, cavities, intervals of core that were lost.

Worked exampleThe mean over core and the result of a well test

Given (assumed data)

Open interval: 3.0 m; 12 samples every 0.25 m
Sample permeability brought to reservoir conditions, mD: 140; 95; 60; 180; 8; 45; 240; 55; 30; 110; 3; 160
Well test (on oil, with the skin factor, the effect of the zone around the well, taken into account): 88 mD converted to absolute permeability

Solution

Arithmetic mean (the ordinary average): 94 mD.
Geometric mean (the 12th root of the product of the 12 values): 57 mD.
Harmonic mean (12 divided by the sum of the inverse values, so the low values dominate it): 20 mD.
The test gave 88 mD, closest to the arithmetic mean: the difference is 6%.

Answer: The mean permeability over core is 94 mD (arithmetic), and it agrees with the well test.

What it means. Fluid flows to the well along the layers, and the layers work in parallel: their conductivities add up, and the two tight layers (8 and 3 mD) hardly interfere. So the arithmetic mean is the one compared with the test, provided that the layers continue over the whole drainage area. If the test had shown a value close to the geometric mean, 57 mD, it would have meant that the permeable layers are discontinuous. The harmonic mean, 20 mD, describes flow across the layers. It is 4.6 times smaller, and it is set by those same two tight layers. The data in this example are assumed for illustration.

Terms and symbols

Core: A column of rock cut out of the reservoir and brought to the surface.
Core recovery: The length of core brought up as a percentage of the length of the interval drilled.
Core plug: A cylinder 25–38 mm in diameter cut out of core for measurements.
Sidewall core: A small sample taken from the wall of the well with a tool on a cable.
Routine core analysis (RCA): Quick measurements of porosity, permeability, grain density and saturation on all the samples.
Special core analysis (SCAL): Long tests of capillary pressure, relative permeabilities, wettability, and electrical and mechanical properties on chosen samples.
Restored-state core: Cleaned core saturated again with reservoir fluids and held at reservoir temperature.
Depth matching of core: Shifting the depths of the core to the depths of the well log, so that the gamma-ray activity curves match.

Common mistakes and misconceptions

Wrong

“Core is a direct measurement, so core numbers should be trusted more than well logs.”

In fact

Core is measured directly, but on a tiny sample that was altered by coring, and in laboratory conditions. Core can be trusted when it is brought to reservoir conditions, matched in depth, and averaged over the volume with which it is compared.

Wrong

“The more carefully a sample is cleaned and dried, the more accurate the result.”

In fact

Harsh cleaning and hot drying destroy clays and change wettability. Choose the method of preparation for the rock and for the measurement. For relative permeabilities, cleaned core without restored wettability is unsuitable.

Wrong

“Routine gas permeability is the permeability of the reservoir.”

In fact

It was measured without reservoir stress, with dry gas and on dry clays. All three reasons make the result too high; for tight clay-rich rocks, several times too high.

Wrong

“Core recovery of 60% is no problem; we will find the properties from what was brought up.”

In fact

The weakest rock is usually lost, and often the most permeable, so the remaining core is a biased sample (it does not represent the whole reservoir fairly). We restore the lost intervals from the well log, calibrated on the core that was brought up.

In depth: advanced detail

How X-ray tomography of core works

A narrow X-ray beam passing through matter is weakened: some of the quanta (the particles of the beam) are scattered by electrons or absorbed by atoms. The intensity of the beam that gets through falls exponentially with thickness.

Attenuation of an X-ray beam

I = I_0\,\e^{-\mu h}

where:

$I_0,\ I$ intensity of the beam before and after the sample, the same units
$\mu$ attenuation coefficient of the material; it grows with density and with the atomic number of the elements, 1/cm
$h$ thickness of the sample along the beam, cm

Units: $\mu$ and $h$ in inverse units (for example 1/cm and cm)

Holds when

A narrow beam of one energy. In a real beam with a range of energies the soft (low-energy) part of the spectrum is absorbed first, and processing allows for this.
The material along the beam is uniform. For a non-uniform material the exponent holds the sum $\mu_i h_i$ over the sections of the path.

One scan gives only the sum of the attenuations along the beam. The scanner rotates the source and the detectors around the core and gets such sums in hundreds of directions. From them the computer reconstructs the value of $\mu$ at every point of the cross-section. Dense rock and heavy minerals (pyrite, and barite from the drilling mud) look light in the image, and pores and fractures look dark.

Density logging is based on a related effect, the scattering of gamma quanta by the electrons of the rock (Density log and photoelectric factor (Pe)). A tomograph for full-diameter core sees details of a fraction of a millimetre, which is enough for fractures, bedding and choosing the places for the samples. Micro-tomographs scan samples a few millimetres across with a resolution of micrometres and show individual pores. From such images we can calculate the flow at the pore level (Pore-scale flow modelling).

How many samples are needed for a reliable mean

The mean over a few samples is a random quantity: take other samples from the same layer and it comes out different. The more variable the property and the fewer the samples, the wider the scatter. We measure the variability by the coefficient of variation $C_v$ , the ratio of the standard deviation (a measure of how widely the values spread) to the mean (Descriptive statistics in geology).

Accuracy of the mean and the number of samples needed

P_s = \frac{200\,C_v}{\sqrt{N_s}}\ \%,\qquad N_0 = (10\,C_v)^2

where:

$P_s$ half-width of the interval in which the true mean lies with a probability of about 95%, as a percentage of the mean over the samples, %
$C_v$ coefficient of variation: the standard deviation divided by the mean, —
$N_s$ number of samples, —
$N_0$ number of samples at which $P_s = 20$ %, —

Units: dimensionless quantities

Holds when

The samples are independent and were taken without choosing “convenient” places.
The interval ±2 standard errors of the mean corresponds to a confidence level of about 95%. With a small number of samples it is somewhat wider.
The 20% threshold is a convention, not a law.

The standard error of the mean equals the standard deviation divided by $\sqrt{N_s}$ . Two such errors, expressed as a percentage of the mean, give $200\,C_v/\sqrt{N_s}$ . Porosity varies little ( $C_v$ about 0.1–0.2), and a few samples are enough for it. Permeability varies several-fold ( $C_v$ about 1 or more), and it needs tens or hundreds of measurements (Distributions and sample size).

Worked exampleAre twelve samples enough

Given (assumed data)

Sample permeability, mD: the same 12 values as in the example “The mean over core and the result of a well test”

Solution

The mean is 94 mD, the standard deviation is 74 mD, and $C_v = 0.79$ .
Accuracy of the mean over 12 samples: $P_s = 200\cdot0.79/\sqrt{12} = 46$ %, that is from 51 to 137 mD.
For an accuracy of 20% we need $N_0 = (10\cdot0.79)^2 \approx 62$ samples.

Answer: Twelve samples give the mean permeability with an accuracy of about ±46%. For ±20% about 62 are needed.

What it means. The result of the well test, 88 mD, lies deep inside the interval from 51 to 137 mD. A difference of a few per cent means nothing, and agreement is partly luck. To tell layers apart reliably by permeability, we need several times more samples than the routine spacing gives. Quick measurements of permeability on the sawn face of the core, and well logs, supply the missing points. The data in this example are assumed for illustration.

Connections to other topics

Porosity — how pore and grain volumes are measured and how the kinds of porosity differ.
Rock permeability — Darcy's law, the Klinkenberg correction and averaging over layers.
Fluid saturation — Dean–Stark extraction and the reasons why the saturation of core is distorted.
Wettability and interfacial phenomena — why the state of the core decides the fate of special studies.
Capillary pressure and the transition zone — the porous plate (semipermeable membrane), the centrifuge and mercury injection.
Relative permeability — steady-state and unsteady-state tests on core.
Rock compressibility and the effect of stress on core properties — measurements under load and bringing to reservoir stress.
Gamma ray log and natural gamma ray spectrometry — the curve by which core is tied to the well.
Porosity from the density log — porosity from well logs, which is calibrated with core.
Distributions and sample size — sampling, bias and the number of measurements.

Check yourself

1
Relative permeabilities are needed for a waterflood calculation. Which core is suitable?
1. A.Sidewall core taken with a percussion (shooting) tool
2. B.Cleaned and dried in an oven: it is the cleanest
3. C.In its native state, or with restored wettability
4. D.Any: relative permeabilities are a property of the rock
Show the answer
Correct answer: C. In its native state, or with restored wettability
Relative permeabilities depend on how the oil and water are distributed in the pores, that is, on wettability. Cleaning with solvents usually makes the rock artificially water-wet (hydrophilic), so core is needed that has kept its reservoir wettability, or cleaned core in which the wettability was restored by ageing in reservoir oil.
2
The routine gas permeability of a sample is 12 mD. What is the permeability of this rock in the reservoir likely to be, and why?
1. A.It cannot be predicted
2. B.Higher: in the reservoir the pores are held open by the reservoir pressure
3. C.The same: permeability is a property of the rock
4. D.Lower: gas slips at the pore walls, the sample is not under confining pressure, and the clays are dried
Show the answer
Correct answer: D. Lower: gas slips at the pore walls, the sample is not under confining pressure, and the clays are dried
All three differences between the routine test and the reservoir make the result too high: gas slippage at low pressure, the lack of reservoir effective stress, and dry, “shrunken” clays. For a rock with a permeability of about 10 mD the total correction is usually larger than for a high-permeability rock.
3
A gamma-ray activity peak on core is marked at a depth of 3120.4 m, and the same peak on the gamma-ray log is at 3118.9 m. What log depth should be given to a sample drilled from the mark of 3125.00 m on the core?
Show the answer
The shift is $3118.9 - 3120.4 = -1.5$ m: the core has to be raised by 1.5 m. The sample corresponds to a depth of $3125.00 - 1.5 = 3123.5$ m on the log. One peak gives the shift only roughly. Usually we use several peaks and check that the shifts are close.
4
Why are clay-rich samples dried at about 60 °C and about 40% humidity, and not in an ordinary oven?
1. A.So that a film of water stays on the clays, as in the reservoir, and their structure is not destroyed
2. B.So that the sample does not crack from the temperature difference
3. C.It is faster
4. D.So that the oil does not evaporate
Show the answer
Correct answer: A. So that a film of water stays on the clays, as in the reservoir, and their structure is not destroyed
Hot drying removes the water held by the clay minerals: the particles shrink, the fibres stick together, and the pores seem more spacious than in the reservoir. Drying at controlled humidity leaves a layer of water on the clays, and the measured porosity and permeability are closer to the reservoir ones.
5
In an interval of weakly cemented sandstone the core recovery was 60%. The mean porosity of the samples recovered is 17%. Can it be taken as the mean porosity of the interval?
Show the answer
No. What was not recovered is most likely the weakest parts, the least cemented and therefore the most porous and permeable. The mean over the core that was brought up is too low. The porosity of the lost parts is estimated from the well log calibrated on the recovered core, after the core has been matched in depth. The place of the losses is refined from the gamma-ray activity curve.
6
The mean permeability over ten samples from a layer is 60 mD, and the coefficient of variation is 0.9. How far can this mean be trusted?
Show the answer
The accuracy of the mean is $P_s = 200\cdot0.9/\sqrt{10} \approx 57$ %: the true mean lies, with a probability of about 95%, between about 26 and 94 mD. For an accuracy of 20% we need $(10\cdot0.9)^2 = 81$ samples. Ten samples are enough to say “tens of millidarcies”, but not to tell 45 from 75 mD.

Updated: 2 October 2026