Industry Report: Understanding the petrophysics of drilling mud and its effects on log data

GEOFFREY PAGE and STEPHEN VICKERS, Baker Hughes

Drilling fluid within the borehole must be accounted for during petrophysical interpretation. This was not a problem when muds were simple water-based or invert-oil emulsions, but with newer chemistry, complex modeling may be required to retain interpretation accuracy.

The potential effects of drilling fluid on the petrophysical measurements of formations are not always well understood. In addition to maintaining wellbore stability, removing cuttings and cooling the bit, drilling fluid should also allow the acquisition of accurate petrophysical information. As part of pre-well planning, the mud formulation should be reviewed to ensure its suitability for the required data acquisition strategy, especially in exploratory wells.

The physical properties of the drilling fluids are reported on a daily basis and should be reviewed by the petrophysicist for any major changes, particularly while penetrating the reservoir. The reported mud properties, however, tend to be limited to density, viscosity, chemicals used, etc. Additional properties are also important to the acquisition of high-quality log data, but are often difficult to ascertain, particularly at a later date.

Mud properties that affect data acquisition include:
• Mud weight, both of whole mud and of filtrate
• Viscosity, which can affect wireline tool running speeds and LWD pulsing capabilities
• Filtration control and expected depth of invasion
• Chemical composition of the makeup brine, including salinity, conductivity and reactivity
• Chemical composition of the base oil, and comparison to predicted native hydrocarbons
• Chemical composition of solids components, especially if any form of mineral logging is being undertaken.

Corrections for mud properties should be considered similar to detailed core analysis; they generally do not change the overall picture of a “quick look” analysis, but can help refine the data at a later date.

INVASION

An important consideration is how deeply liquids, or solids, invade the rock from the wellbore, in relation to the depth of the measurements. For deviated or horizontal wells, invasion may vary azimuthally around the wellbore due to anisotropic permeability variations and/or gravity.

Correctly sized solids should never invade more than a few pore depths from the wellbore. This requires the correct mix of small and larger particles in the mud mix, Fig. 1. Mud filtrate invasion is dependent on time, overpressure, permeability and mud cake development. For LWD data recorded shortly after drilling, this may be only a few inches; it may be considerably deeper if recorded days or weeks later.

Most petrophysical measurements have the bulk of their response within a few inches of the wellbore, and are affected by filtrate properties. A confusing factor is that for most resistivity logs, the “depth of investigation” is defined as 50% of the response coming from within that depth, whereas nuclear measurements are usually quoted at 90% response.

A density measurement is usually quoted as having a depth of investigation around 8 in. However, 50% of the signal comes from within 3 in. Neutron measurements are deeper, at around 12 in., which should be considered when combining density and neutron, as they may be seeing different fluids.

Acoustic measurements follow the shortest path, and invading fluids may force the measurement deeper or shallower depending on relative slowness of filtrate and formation fluids, but they generally stay within a foot of the borehole. Magnetic resonance has a depth of investigation between 1 in. and 4 in.

All of these measurements are usually within the (partially) invaded zone, and mud filtrate properties should be considered. The exceptions are the resistivity logs, which are designed to either see a lot deeper, sometimes 10 ft or more, or specifically to measure very shallow to define the flushed-zone resistivity (R_xo ). Frequently, a range of depths of investigation is recorded to define the invading fluid depth profile. Resistivity image logs are generally shallow reading, but the system as a whole can respond to resistivity of 15–20-in. depth, providing a mix of Rxo and true formation resistivity (R_t) responses.

Many methods can derive an apparent depth of invasion from at least three resistivity measurements of different depths; however, ideally one of these should be an accurate shallow R_xo such as from a pad-type device reading within the first few inches. Although R_xo and depth of invasion can be derived from many sonde-type tools, their shallowest raw data comes from around 10 in., which is beyond the first few inches where the porosity logs have most of their responses.

Alternative methods for defining apparent invasion depths include using the photoelectric effect (PE) for fluids with a high PE such as cesium formate brines,1 magnetic resonance fluid analysis in the 2–4-in. depth range, and acoustic, which can be used to define variations in the interval transit time (DT) in the first few feet away from the wellbore.

Fig 1. Correctly sized solids should not enter the formation, where they could cause damage and affect log responses. If solids are too small, they cannot bridge the pore space and whole mud will invade the formation. Solid particles that are too large can prevent a bridge, allowing smaller particles to invade. With the right mix of particle sizes, only clean filtrate will enter the formation.

DATA CORRECTION

Correcting petrophysical measurements for mud chemistry consists of borehole corrections based on hole size and mud properties to correct measurements back to calibration conditions. This provides a “true” value free from wellbore artifacts, but true bulk formation properties still contain effects from any invading fluids (or solids) that need to be accounted for. This includes fluid properties for porosity and saturation calculations, wettability, permeability and/or any mineralogical changes.

Since temperature and pressure may be different within the wellbore and formation, borehole parameters should be used when applying borehole corrections, and formation parameters when correcting for filtrate effects. Only after accounting for both of these factors will the virgin formation properties be determined.

AFFECTED MEASUREMENTS

Drilling fluid bulk density and the absorption of gamma rays have a small effect on formation density measurements due to backscatter within the borehole, and will also reduce total natural gamma ray (GR) counts. This may be exacerbated by high-PE components, such as barite and cesium, which preferentially absorb the lower energies, adding considerably to the PE curve. Mud cake density is generally compensated for by the design of multi-detector tools, but may still introduce high-PE effects. The liquid components that may enter the formation as filtrate will need to be considered in density porosity calculations. Acoustic measurements fully compensate for any variations in borehole fluid slowness, but invading filtrate slowness needs to be accounted for in porosity calculations.

Gamma density response of fluids is slightly different from true density, but a correction can be estimated (see Appendix 1).

Some mud system components may significantly reduce the hydrogen index (HI) of the mud, affecting neutron borehole corrections and porosity measurements. Nuclear magnetic resonance (NMR) measurements will also be affected by the HI, with additional changes in the relaxation constants T1 and T2 and in diffusivity spectra.

The HI of fluids can be estimated (see Appendix 2).

The mud sigma, or macroscopic thermal neutron capture cross-section, affects both the neutron borehole corrections and porosity. Modern tools can minimize borehole chloride/salinity corrections (essentially a sigma correction). Some mud/filtrate components that have a very high sigma, such as cesium, will require additional corrections.

Mud radioactivity can significantly add to natural GR measurements and distort spectral measurements, but it is rarely high enough to have a noticeable effect on density measurements. If the filtrate is radioactive, there may be a permeability-related component to the apparent formation GR signal. In general, spectral GR measurements are useful to identify and correct these responses.

Whole-mud resistivity will affect the choice of resistivity logging tools, and may restrict NMR acquisition capabilities in highly conductive muds. Filtrate samples should be taken and their resistivity measured to calculate flushed zone saturations and correct all measurements for invasion. Magnetic minerals in the mud, including swarf from milling, may affect orientation devices, resulting in bad orientation data. In these cases, running NMR after any orientated services should be considered.

Surfactants used to mix the oil and water in an oil-based mud (OBM) may affect the wettability of the flushed zone and, thus, calculated irreducible flushed-zone fluid saturations. This may also affect the data recorded by NMR tools, which generally read within the flushed zone. NMR core measurements, rarely done, can identify the magnitude of any such effects.

Lost-circulation materials (LCMs) contain polymers, asphalt or other materials designed to block pores and fractures. These may or may not be removable later under flowing conditions, and can affect the pre-testing pressure communication, permeability and sampling from formation testing tools. They may also affect other permeability measurements derived from acoustic measurements.

When cleaning up formation tester fluid samples, a good contrast between filtrate properties and native fluids will ensure that contamination levels can be identified and monitored. This may be a consideration when designing mud systems for exploration wells.

Ultimately, many of these effects can be identified and corrected if log data is calibrated to core. However, coring is expensive, not available in many wells, and rarely over entire reservoir intervals.

MUD SYSTEM COMPONENTS

Fresh water has a density of 1.00 g/cc and an HI of 1.00 at surface conditions. Increasing temperature causes water to expand. At 200°C and 225 psi (the lowest pressure at which it remains liquid at this temperature), fresh water has a density of 0.87 g/cc and HI of 0.87. Pressure alone at 10,000 psi compresses water, increasing its density and HI to 1.02, less than the temperature effect. Combined, these effects mean that fresh water at downhole conditions has a density and HI less than 1.0. At 200°C and 10,000 psi, the density and HI of an invading freshwater mud filtrate would be 0.92.

The addition of salts to water increases its density and reduces HI. Sodium chloride- (NaCl-) saturated water (250,000 ppm) at surface conditions has a density of 1.19 g/cc and an HI of 0.90. Increasing pressure and temperature change these values in a similar way to fresh water.

For a reservoir at 120°C and 6,000 psi, a 100,000-ppm NaCl filtrate would have a density of 1.04 g/cc and an HI of 0.94. If a value of 1.00 were mistakenly used for both, then porosity calculations would be reduced by around 0.5% (at 20%). At 200°C and 10,000 psi, the true density and HI for the same saline filtrate are 1.12 g/cc and 0.91, respectively, resulting in a loss of 1.5% porosity if 1.00 were used.

Fig. 2. The invert emulsion of oil and water in an oil-based mud.

Oil-based mud (OBM) consists of base oil and brine, with surfactants and emulsifiers to form an invert emulsion. Oil is the continuous phase, Fig. 2. These muds do not conduct electrical currents at the normal voltages and frequencies of logging measurements. It is also assumed (sometimes wrongly) that, as the continuous phase, only the oil invades through a mudcake into the formation. It is the properties of this filtrate that affect log measurements most.

Base oils vary widely in density, typically 0.75–0.85 g/cc, with an HI around 1.0. The lower the density of the base oil, the more solid weighting materials must be used to achieve the same mud weight. At surface conditions, typical base oil will expand with increasing temperature, similar to water, and contract under pressure. The pressure contraction is of a similar magnitude to the temperature expansion, but in the opposite direction. This results in downhole properties similar to those at surface. Oil with a 0.8-g/cc density and 1.01 HI at surface would have a 0.78-g/cc density and 0.99 HI at 120°C and 6,000 psi. These values assume no absorption of gas. An oil of this density at downhole conditions can potentially absorb up to 1.7 Mcf/bbl, resulting in a density of only 0.6 g/cc and an HI of 0.89.

Ester-based muds tend to be more aggressive toward elastomers than mineral OBMs. Esters are also a wetting agent and can affect near-wellbore wettability. They generally have a slightly lower HI than mineral OBM, which may be visible as a porosity reduction on neutron or NMR measurements. Esters can absorb gas at elevated temperature and pressure, and have been known to appear as a false gas effect on density/neutron logs.

Low-density mud for underbalanced drilling is created either using a standard mud aerated with air or nitrogen or using lightweight chemical foams. Most logging tools can be run in them, with some caveats. First, LWD data pulsing relies on a low-compressibility fluid column to transmit pressure pulses of data to the surface, and therefore is not practical for these mud systems. Electromagnetic transmission systems are available as a replacement data route. Acoustic measurements can’t be used, as they rely on a similar borehole fluid to couple the signals in and out of the formations. Borehole corrections for density and neutron measurements will be larger than for regular mud systems. The lack of a radiation-absorbing medium will result in higher count rates and lower densities and neutron porosities. Wireline instruments tend to be affected more than LWD due to the smaller fraction of the borehole that they occupy.

Formation tester pre-tests will exhibit draw-ups, as formation pressure will be higher than hydrostatic. This may require analysis logic reprogramming to interpret when pressures have stabilized and for determining permeability. The higher formation pressure tends to push the sealing pad off the wall, resulting in lost seals.

Induction and propagation resistivity logs should work effectively. However, shallow and lateral-type logs may not work due to increased mud resistivities, and it may be necessary to use an oil-based imager tool even in water-based fluids.

Brines other than NaCl solutions allow mud “base water” with density ranging from 1.0 g/cc up to more than 2.0 g/cc for the pure liquid phase alone. As these brines get denser, their HI decreases. Saturated sodium bromide brine has a density of 1.52 g/cc and an HI of only 0.8 at surface conditions. This is due to the high weight fraction of salts displacing the hydrogen in the dissolving water fraction. The heavier brines and formates may undergo gravity segregation with formation fluids, thus showing unusual invasion profiles.

One of the most common brines uses potassium chloride, which helps to stabilize clay structures and prevent swelling. When clay structures contain an interlayer of calcium or magnesium ions, they can bind several layers of water molecules within the structure, Fig. 3a. Sodium is particularly problematic, and can bind up to 10 water layers, causing clay swelling and borehole collapse. In contrast, potassium can form stable interlayers without water, resulting in clay stability, Fig. 3b. One result is that potassium in mud is “used up,” and additional potassium may be added constantly to maintain salinity. If mud potassium content is used to correct log data, then mud reports should be checked throughout the drilled interval, not just the final report.

Potassium acetate (CH3CO2K) has been used as a more environmentally friendly alternative to potassium chloride (no chlorine). From a logging perspective, this is just another brine. The HI of the potassium acetate itself is 0.14, compared to 0 for the chloride, so there may be a slight increase in the HI of the filtrate.

When calculating resistivities of salt solutions, the charts used are frequently those for sodium chloride. However, other salts can have significantly different resistivities. Charts exist to allow other salt/ion concentrations to be converted to “NaCl-equivalent” values. At very high temperatures (over 200°C), the resistivities may also start to deviate from the standard (Arps type) formulas, and above 300°C resistivities may actually increase with increasing temperature.1

Formate brines are a special case. The formate ion (HCOO−) can be combined with sodium, potassium or cesium to create very dense brines. Due to the very high percentage (by weight) of the formate and low water content, these brines have significantly different log responses and require special modeling to determine the required instrument corrections. In particular, the low HI will significantly reduce the neutron response and may create the appearance of a gas effect on the density/neutron log.

The resistivities of formate solutions also do not behave as expected. As the formate concentrations increase, the resistivity initially decreases, but then at very high concentrations the resistivity increases again as the solution becomes so overloaded with ions that they start to affect each other, reducing conductivity. The mixing of different formate brines also generates a non-linear mixing rule. In the case of saturated formate brine invading a freshwater-filled formation, as the saturated formate dilutes by mixing with the formation water, its resistivity may decrease.

Saturated cesium formate has a PE absorption of over 200 b/e, and very high log values of 30 b/e or more may be observed. This high PE absorption may also affect GR counts, as low-energy gamma rays will be absorbed. The high PE makes normal analysis of the PE curve impossible for mineral identification. However it can be useful for identifying and quantifying invasion, and for a net sand count.2

Glycol is a common constituent of water-based muds that helps reduce invasion and improve lubricity. At low temperatures, glycol is fully miscible in water. As the temperature is increased, the glycol “clouds out” into droplets and becomes separated from the water. This cloud-point temperature is designed around the expected wellbore temperature. The glycol droplets then coat the rock surface, block pores and inhibit invasion. Glycol also coats cuttings, preventing their dispersion. As the cuttings-laden mud returns to surface, the glycol cools, becomes miscible again, and is released, Fig. 4.

Glycols are hydrocarbon chains similar to alcohols but with two or three OH groups. They usually have a density of 1.0–1.1 g/cc and an HI close to 1.0, are non-toxic and have minimal effect on log data, as they should not invade more than a few pores’ depth. Resistivities of the whole mud may be different above and below the cloud-point temperature, but the invading filtrate should only be the water/brine fraction. Glycols have the potential for slightly altering wettability.

Chemical solids are added primarily to increase mud density and provide filtration control by forming a “cake.” If sized correctly, these solids should not invade the formation, but may add to borehole corrections of logging instruments.

Weighting agents are high-density solids that are insoluble in water. Of particular interest are barite, which is very common and has a very high PE response (267 b/e); galena, which has been used despite the toxicity of lead, with a PE response of 1,631 b/e; and minerals containing iron, which will be magnetic. In addition, mud solids will include dispersed rock flour from drilled formations.

Typical particle sizes of mud solids range 10–100 microns. However, when high quantities of solids are required to generate a heavy mud, micro-sized particles as small as 1 micron can be used. These are less likely to settle out of suspension, preventing mud “sag” and varying mud densities.

Filtration control additives can be either solid or dispersible in the liquid phase of the mud (such as glycol). The most common solids are barite, bentonite and calcium carbonate. If sized correctly, these solids should not invade the formation, but may add to borehole corrections of logging instruments. Other common chemicals include starch, carboxymethyl cellulose, polyanionic cellulose and asphaltene, especially at high temperatures. Most of these products are long chains of carbon, oxygen and hydrogen atoms, and will have minimal log effects. If these products are removable, then they will not impede formation testing and sampling, although an increased drawdown may be seen for a short time before the mud cake pops off. If the barrier is permanent, then permeability and potentially pressure measurement will be impeded.

Silicate muds are solutions of water-soluble glasses. The silicate starts out as monosilicate, which polymerizes rapidly to form negatively charged oligomers at a high pH, at which the mud is formulated. After entering a pore network, the pH drops, and the oligomers overcome their mutual repulsion and coagulate, forming 3D gel networks. Within the pore fluids and at clay surfaces, there are free polyvalent ions (calcium and magnesium), which then react instantaneously with the oligomers to form insoluble precipitates. This reaction is irreversible and occurs very rapidly before significant fluid loss and pressure invasion has occurred. These muds are run overbalanced to encourage invasion, the net result of which is a permanent glass barrier within the borehole wall.

These muds may significantly affect formation testing and sampling. Crystalline solids may precipitate on moving surfaces and valves, causing blockages and seals to fail if cut. The alkaline environment is also detrimental to many elastomer seal compounds. Any instruments run in these muds should be thoroughly washed inside and out immediately on return to the surface, as the precipitate may encase them in a thin permanent glass film, jamming threads and moving parts. The precipitate may affect the operation of mud turbines and pulsers for LWD, and acoustic data has been reported to be affected by the hard fast layer within the rock wall and/or plugging of the instruments. Shallow measurements such as resistivity and density will also be affected, and the density correction curve may show a dense mud cake. Unusual invasion profiles may develop, which require resistivities at multiple depths of investigation to solve for a true Rt.

Aluminum salts perform a similar function to silicates, but the process is reversible, lowering the risk of permanent formation damage. Aluminum salts stay in solution at pH greater than 10. As the mud filtrate penetrates formations, the pH drops below 10 and the aluminum precipitates. This blocks the pores and prevents additional filtrate entry. The process is fully reversible by increasing the pH again. Aluminum salts do not cause cuttings to accrete. The precipitate may cause some reduction in permeability at the time of formation testing, but the precipitate is not “stuck” to the rock in the same way as silicate. 

ACKNOWLEDGMENT
This article was prepared from Paper GG presented at the SPWLA 52nd Annual Logging Symposium held in Colorado Springs, Colorado, May 14–18, 2011.

LITERERATURE CITED
1 Ucok, H., Ershaghi, I., Olhoeft, G. R. and L. L. Handy, “Resistivity of brine-saturated rock samples at elevated temperatures,” Journal of Petroleum Technology, April 1980, pp. 717–727.
2 Kukal, G. C. and R. E. Hill, “Improved shaly sand analysis in heavy drilling muds: A simple technique for using the photoelectric measurement,” Paper U, SPWLA 26th Annual Logging Symposium, June 17–20, 1985.