PEH:Fluid Sampling

Publication Information

Petroleum Engineering Handbook

Larry W. Lake, Editor-in-Chief

Volume I – General Engineering

John R. Fanchi, Editor

Chapter 4 – Fluid Sampling

John M. Williams* and Sunil L. Kokal, Saudi Aramco

Pgs. 173-215

ISBN 978-1-55563-108-6
Get permission for reuse

Many general petroleum engineering texts have sections covering the measurement of phase behavior or pressure/volume/temperature (PVT) analysis, but few have detailed descriptions of fluid-sampling practices. This chapter covers the sampling of all produced reservoir fluids. It is intended to provide an overview of sampling methods, guidelines for selecting suitable methods, and detailed procedures for the most common practices.

An enormous range of reservoir fluids exists, and this means that the limited measurements of produced oil and gas properties that can be made in the field are far from adequate to provide the detailed characterization that modern petroleum engineering requires. In addition to PVT analysis, of fundamental importance to reservoir management, measurements relating to corrosion potential, solids formation, and nonhydrocarbon constituents have the potential to produce serious effects on the design of production facilities, on compatibility with pipeline transport, on product sales value, on refinery maintenance costs, and on reservoir asset values in general. The lack of such data could easily represent more risk than that tolerated when the decision to perform sampling and laboratory studies is taken. Examples of the financial impact of errors in fluid-property measurements are given elsewhere.^[1] Fluid samples are thus required to enable advanced physical and chemical analyses to be carried out in specialized laboratories. Samples must be collected from a wide range of locations, including separators, pipelines, tanks, wellbores, and the formation itself. This chapter primarily targets the sampling of fluids under pressures above atmospheric, where numerous tools and procedures have been developed that are essentially specific to the petroleum industry. Best practices are proposed for fluid sampling, reporting of data, and quality control of samples.

Reservoir-fluid-property measurements derive from a complicated series of processes relying both on the operation of equipment and the performance of people, so the scope for errors is very significant. The overriding challenge in fluid sampling is that of ensuring that the fluid entering the sample container is representative of the bulk fluid being sampled. It is equally important that the sample remains representative during handling and storage, until all required measurements have been completed. Although thorough sample-checking procedures can identify some of the most obvious problems, there is never absolute certainty that the fluid under study is truly representative of the reservoir fluid. On occasion, laboratory measurements can show that a fluid is definitely not representative (e.g., saturation pressure is significantly higher than reservoir pressure), but even here the problem could lie with errors in field measurement data rather than with the samples themselves. Thus, it is essential that all the necessary precautions are taken to prevent poor samples from leading to erroneous physical-property measurements.

Fig. 4.1 is a schematic diagram illustrating some of the most common sources of error in relation to the collection of production samples and data in the field. Perhaps the most important, yet often misunderstood, phase of any sampling program is that of well conditioning. A poorly conditioned well may still be producing drilling-mud filtrate, workover fluids, or reaction products and, in extreme cases, such materials may remain even after months of production. A conflicting aim of well conditioning is to avoid excessive pressure drawdown and the creation of a large region of two-phase reservoir fluid around the wellbore, which may be difficult to remove. This is especially important in the case of gas/condensate reservoirs, of which many are found at their saturation pressures. The sampling program must ensure that appropriate procedures are used to ensure that samples are taken under the best conditions.

Fig. 4.1 – A schematic view of wellsite sampling and measurement errors.^[1]

Measurements of reactive or nonhydrocarbon components of reservoir fluids are complicated by the potential for loss through reaction or adsorption in contact with the production tubing or with sample-bottle walls, especially during long storage periods. On-site measurements can be very important if performed and recorded properly.

The schematic in Fig. 4.1 emphasizes sampling activities in cased-hole wells, but pressurized samples are also obtained with formation-test tools in openhole wells. Here, contamination by mud filtrate or excessive pressure decrease (drawdown) during sampling means that it may not be possible to obtain quality PVT samples. Contamination by oil-based mud (OBM) is especially problematic.^[2] Sampling from tanks or pipelines also requires that care be taken to ensure that the fluid is representative of the location or condition required to be studied.

Not only may errors in the field mean that samples are not fully representative of the reservoir fluid, but even good fluid samples may be studied under invalid conditions. Pressure and temperature errors can influence measurements and their interpretation, but it is especially errors in gas/oil ratio (GOR) that can have a major influence on a PVT study. Even basic data, such as sampling date and time, if not recorded or erroneous, can reduce the value of samples, even to the point of making measurements meaningless.

*

Now with The Petroleum Inst., Abu Dhabi, UAE.

General Guidelines for Setting Up a Sampling Program

The specific requirements for samples and laboratory studies naturally will depend on the state of knowledge about a prospect. Thus, it may be advisable to perform extensive sampling and a complete suite of laboratory measurements on a wildcat well when nothing is previously known about the reservoir; this may provide the only fluid data on which to base future exploration work. However, the early wells in a field may not provide the best samples because drilling and workover practices will not have been optimized, and the wells’ response to testing programs may require changes that are detrimental to fluid sampling. It may then be necessary to repeat some analyses during the appraisal stage, typically when wells will yield samples that are more representative of likely production. In contrast, sampling late in the appraisal phase may be needed only on occasions when surface measurements indicate unexpected fluid character.

The composition of subsurface water commonly changes laterally, as well as with depth, in the same aquifer. Changes may be brought about by the intrusion of other waters and by discharge from and recharge to the aquifer. It is thus difficult to obtain a representative sample of a given subsurface body of water. Any one sample is a very small part of the total mass, which may vary widely in composition. Therefore, it is generally necessary to obtain and analyze many samples. Also, the samples may change with time as gases come out of solution and supersaturated solutions produce precipitates. Sampling sites should be selected, if possible, to fit into a comprehensive network to cover an oil-productive geologic basin. There is a tendency for some oilfield waters to become more diluted as the oil reservoir is produced. Such dilution may result from the movement of water from adjacent compacting clay beds into the petroleum reservoir as pressure declines with the continued removal of oil and brine. The composition of oilfield water varies with the position within the geologic structure from which it is obtained. In some cases, the salinity will increase up-structure to a maximum at the point of oil/water contact.

The first priority in developing a sampling program, whether extensive or limited, is to establish exactly what measurements are required. Table 4.1 gives a wide range of the measurements that are typically considered for exploration wells. This can be used as a checklist, together with direct contacts with users in other functions, to identify specific requirements for sampling and on-site measurements. Generally, it is advisable to plan to perform all applicable measurements unless sufficient information is already available from earlier tests of other wells. The fact that a measurement proves to be "normal," or an unwanted component is not detected, should not be regarded as a waste of resources because it can still provide essential information, especially if data are different on other wells or changes are identified during production. On-site measurements are recommended for all reactive components because concentrations may change with time (e.g., during a well test), and losses frequently occur during sample transport and storage. Table 4.1 is not a comprehensive list, and other measurements will be required in certain locations and for specific purposes.

Table 4.1

Having decided which fluid measurements are required, it is necessary to set up a suitable sampling program, taking into account the cost of the work, the quality and quantity of samples and subsequent measurements, the urgency with which data are required, and the application of safe practices. The program should specify who has overall responsibility if a change is required in the program, as often occurs. Sampling programs should not be developed in isolation from the other objectives of a well test because there is direct conflict in some cases, such as when the well test requires large drawdowns for gas/condensate fluids as part of flow-capacity tests. Thus, in the case of a well test, the overall plan should include the following: (a) establish the production potential, (b) determine the permeability, (c) determine the skin, and (d) collect fluid samples. Each objective should be defined in sufficient detail so that all parties involved are fully aware of their obligations, thus increasing the likelihood of achieving the objective. Objectives must be realistic and must allow for possible changes. In the case of fluid sampling and on-site analyses, the following sorts of questions should be considered in deciding the detailed sampling objectives: (1) How much information is available on the likely reservoir fluid? (2) What types of fluid sampling will be best? (3) What is the most suitable well-cleanup and -conditioning procedure, and how can this be integrated with other well-test objectives? (4) How many samples are needed, and do partners need duplicates? (5) When is the ideal time to take samples? (6) Will on-site analyses be required? (7) Who will perform sampling and analysis duties?

Fluid-sampling operations are often left to service-company personnel, but because significant variation in levels of competence exists within the industry and within service companies themselves, it is recommended either to use specialist laboratory personnel or to supervise the service-company operations closely.

General guidelines for choosing reservoir-fluid-sampling methods and sample quantities required are summarized in Table 4.2. Regardless of the actual volumes mentioned, you should collect at least two separate samples of each fluid, referred to as duplicate or replicate samples. This reduces the chance of losing information if one of the samples leaks or is accidentally damaged during laboratory operations, and it allows a comparison between the samples as part of the quality-control procedures.

Table 4.2

Surface-separator sampling is the most common technique, but the reservoir-fluid sample recombined in the laboratory is subject to errors in the measured GOR and any imprecision in the laboratory recombination procedure. Downhole samples (or wellhead samples) are not affected by such inaccuracies but require the fluid to be in monophasic condition when sampled; this can be confirmed definitively only afterward in the laboratory. Also, there is general reluctance to attempt downhole sampling in gas/condensate reservoirs because many are saturated, and the phases are likely to segregate in the wellbore. The ideal situation for a laboratory is to receive both surface and downhole samples because a choice is then available, and a good idea can be obtained of how representative the resulting fluid is.

In certain circumstances, it can be good practice to collect "backup" fluid samples at the earliest opportunity during a production test, even if a well has not cleaned up properly. If the test has to be aborted for some reason [well bridging, unexpected levels of hydrogen sulfide (H₂S), etc.], the backup samples may be of great value, even if they are not 100% representative. If the test is completed successfully, the backup samples can be discarded to avoid the cost of unnecessary shipment and testing.

If sampling is part of a long-term monitoring program, such as those required by government authorities or those forming part of custody-transfer contracts, the methods defined in the appropriate documentation or contracts must be followed as closely as possible, even if this constitutes differences with the procedures or recommendations in this text or in the industry standards cited here. Full use of this text and appropriate industry standards should, of course, be made in setting up new procedures and contracts that require long-term sampling and measurement programs.

If there is concern about whether the fluid is homogeneous in a flow line or tank, the best approach is to take samples from different locations and compare them. In a liquid flow line, take samples from the top and bottom; in a tank, take samples at different depths. If samples are indeed different, it is advisable to locate a better sampling point (e.g., where there is sufficient turbulence to homogenize the fluid). Failing this, the only solution may be to mix the samples together in an attempt to provide a representative average fluid. If, however, the purpose of the sampling is to study the nonhomogeneity, then separate samples should be taken accordingly.

When samples are collected from drillstem tests (DSTs), which do not involve surface production, the limited volume of fluid produced from the reservoir may be insufficient to remove mud filtrate or other contaminated or changed fluid. Thus, even samples collected from the last fluid that enters the drillstem may not be truly representative. This is especially the case for formation-water samples, which are more widely susceptible to contamination from drilling fluids, well-completion fluids, cements, tracing fluids, and acids, which contain many different chemicals. The most representative formation-water samples are usually those obtained after the oil well has produced for a period of time and all extraneous fluids adjacent to the wellbore have been flushed out.

In some cases, fluid sampling may be made on short notice in response to a problem, with the intention of identifying the cause and preventing any recurrence. Here, it is essential to record all the operating conditions and any changes that may have contributed to the problem. Also, it can be useful to collect a reference sample when operation is normal, if this is possible (e.g., a sporadic problem or a similar installation not affected), to allow comparisons. Laboratory personnel also should be contacted regarding the sample needs and the types of analyses that could be performed.

Reservoir-Fluid Type

One of the principal variables in reservoir-fluid sampling is the type of reservoir fluid present. This is rarely known with certainty and, in exploration wells, may be completely unknown at the start of testing. Determining the exact nature of a reservoir fluid is, of course, a key objective of sampling and laboratory study. Fig. 4.2 shows the relation between the major classes of hydrocarbon reservoir fluid in terms of a generalized phase diagram. Although the shape of the phase diagram is specific to the actual fluid composition, it is the reservoir temperature compared to the temperature T_c of the critical point (T_c determines if the fluid is an oil or a gas). When the reservoir temperature is lower than T_c, the fluid is an oil and will exhibit a bubblepoint when pressure is reduced into the two-phase region. If the reservoir-fluid temperature is above T_c, the fluid is a gas and will either show gas/condensate behavior and a dewpoint on pressure reduction or, if the reservoir temperature is also above the cricondentherm T_t, the fluid will behave as a one-phase gas with no liquid formation in the reservoir on pressure reduction. If the fluid exists in the reservoir at or close to its critical temperature, it is classified as a critical or near-critical fluid. These fluids exhibit neither bubblepoint nor dewpoint, but on pressure reduction into the two-phase region, they immediately form a system comprising large proportions of both gas and liquid (e.g., 60% gas and 40% liquid by volume).

Fig. 4.2 – A generalized phase diagram for reservoir fluids.

The reservoir pressure determines whether the fluid is at the boundary of the two-phase region (and referred to as saturated) or at a higher pressure than the two-phase region (and referred to as undersaturated). Saturated fluids will immediately enter the two-phase region when a well produces fluid because of the reduction of pressure in the well and near-wellbore region. More details on phase diagrams are available in the General Engineering section of this Handbook.

Both the reservoir-fluid type and the saturation condition influence the way fluid samples must be collected, yet this information can be estimated only at the time of the sampling program and is especially uncertain when fluids are close to the boundaries between the different types. Numerous correlations are available for estimating reservoir-fluid type and condition from produced-fluid flow rates and properties measured at the wellsite (such as those developed by Standing^[3] and those given elsewhere in this Handbook), but you should be careful in using these methods, especially when the fluid properties differ significantly from those used to develop the correlation. For this reason, it is good practice to allow for significant error in the reservoir-fluid character when designing and implementing sampling programs.

Well Conditioning

The best way to prepare a well for sampling is dependent on the reservoir-fluid type, as indicated earlier. With the exception of one-phase gas reservoirs, prolonged production will cause all reservoirs to reach saturation conditions, thus bringing about changes in the fluid composition throughout the reservoir. When this happens, there is no longer any possibility of obtaining truly representative fluid samples. Thus, although in one-phase gas reservoirs (and for a certain length of time in undersaturated reservoirs), the fluid will remain unchanged during pressure depletion—the true nature of the fluid will be unknown until samples actually have been analyzed in a laboratory—it is strongly recommended to take samples at the earliest opportunity in the life of a well.

Both in openhole and in cased-hole completions, the best depth or production interval for sampling will be as far away as possible from gas/oil, gas/water, and oil/water transition zones to reduce the chances of coning. Every attempt should be made to test zones individually because commingled production may be difficult to detect and is impossible to correct in the laboratory.

The possible influence of any matrix-treatment chemicals on sampling programs should be evaluated, and treatment schedules should be modified accordingly. Problems such as the liberation of carbon dioxide (CO₂) or H₂S after acid treatments are possible, as is the release of other components such as metal ions, and these could affect analyses. On the other hand, sampling after an acid treatment has been properly cleaned up has the probable advantage of reduced drawdown in the near-wellbore region.

Because of the enormous variety of constraints, there can be no definitive guidelines for well conditioning. The first phase of conditioning involves the cleanup, in which the well is flowed to the surface to remove any solids resulting from perforating activities, drilling mud or completion fluids in the well, and mud filtrate or workover fluids that may remain in the formation near the wellbore. Here, the production rate must provide a sufficient flow velocity in the production string to lift solids, hydrocarbon liquids, and water to the surface, but conditioning is typically performed at the maximum rate, as this reduces the total length of the cleanup period.

The cleanup period typically lasts from a few hours to 24 hours, and progress is monitored by regular measurements of flowing wellhead pressure, basic sediment and water (BS&W), and other brmeters. At the end of the cleanup period, production may be diverted through the separator to check its operation. This is an ideal moment to take backup samples.

Depending on the fluid type, significant differences can exist in conditioning procedures, which attempt to control or eliminate any modified reservoir fluid so that fluid entering the well is identical to that in the reservoir. For an unknown fluid, one of the most important considerations is the need to interpret the response of the well to different flow conditions and then develop the final sampling program during the test itself on the basis of this information. If initial production indicates an oil, the best approach is to evaluate the response of surface GOR to changes in production rate. As long as representative reservoir fluid enters the wellbore and is carried to the surface, and the same separator operating conditions of temperature and pressure are maintained, the GOR should remain stable for different choke sizes. A GOR that changes significantly between choke sizes is indicative of nonrepresentative production either caused by two-phase flow effects in the near-wellbore region or possibly by commingled production of more than one reservoir fluid (e.g., oil zone and gas cap), and production should be choked back until GOR no longer changes with choke size.

If, however the reservoir contains a saturated gas condensate, extra flow periods will simply compound the condensate-buildup problem, so sampling at the earliest stable rate is probably advisable. In fact, some modeling work^[4] has shown that when an important ring or bank of condensate has built up in the reservoir, it may be possible to produce at reasonably high rates with an apparently stable GOR while producing fluid that is not representative of the original reservoir fluid. At very high flow rates, the GOR may appear to increase as a result of liquid carry-over in the separator gas stream.

Gas wells that have a small flow velocity will exhibit liquid "slippage" in the tubing and heading or unstable flow rates at the surface. Several methods of establishing the required minimum flow rate are available, of which an industry nomogram^[5] has been used extensively. In very-low-permeability, saturated gas/condensate reservoirs, it may not be possible to lift condensate from the well without creating a major pressure drawdown in the reservoir and causing nonrepresentative fluid to enter the wellbore. The best approach here is to select a small tubing diameter before the test so that the minimum lift velocity can be achieved with a low flow rate and, thus, reduced drawdown.

In highly undersaturated reservoirs, it is possible to take downhole samples while the reservoir is producing, provided that the downhole flowing pressure is greater than the fluid saturation pressure. In many cases, however, the saturation pressure will be unknown or cannot be estimated with sufficient accuracy; then, the best recommendation for downhole sampling in an oil reservoir is to sample when shut in, as for reservoirs that are at or close to saturation pressure. Before downhole sampling with the well shut in, it is necessary to allow pressure to build up near to static and then to purge fresh reservoir fluid at a low rate to replace any "changed" fluid in the wellbore or in the near-wellbore region. The most suitable time for downhole sampling during the well test is probably after the initial cleanup and buildup, but an alternative is at the very end of the test if a long buildup is part of the test plan.

On a different aspect of well conditioning, the use of OBM during drilling operations can lead to contamination of the near-wellbore region, and any subsequent contamination of fluid samples by base oil may not be identified. This is unlikely to pose a problem if the well is properly cleaned up, but it can result in significant contamination if samples are collected using an openhole formation tester in which only small volumes are purged. It is thus best if lost circulation can be minimized during drilling and ideal if only water-based muds are used.

If surface-separator sampling is planned, another form of conditioning can be necessary if chemicals are in use. Injection of methanol or glycol upstream of the separator can be used to prevent gas-hydrate formation, and the injection of antifoaming agents and demulsifiers may be required in oil reservoir fluids. If possible, any such injection should be stopped before separator samples are taken, and enough time should be allowed for such potential contaminants to be purged from the separator (e.g., by waiting at least five times the residence time). Residence times can be derived from the nomogram given in the chapter on Design of Two- and Three-Phase Separators in the Facilities and Construction Engineering section of this Handbook.

Separator conditions themselves also have an influence on sampling operations. Separator temperature can be controlled only by changing the production rate or by the use of a heater, but there is generally more flexibility in the separator pressure, which can be set at any value not exceeding the choke downstream pressure limit for critical flow (or the working pressure of the separator, if it is lower). The advantages of using the highest-acceptable separator pressure include more intermediate components being in the liquid (increasing the liquid flow rate somewhat), more gas in sample bottles because of the increased pressure, and generally a leaner gas stream with less condensation on cooling. Although service companies may be reluctant to operate separators at higher pressures, these benefits can be important for the quality of fluid measurements. For low-GOR oil production, a lower separator pressure may be advisable because it can significantly increase the gas flow rate and improve its measurement accuracy. Separator liquid levels also can be adjusted in many separators; lower levels increase gas residence time and thus can reduce carry-over problems for gas wells, whereas higher levels increase oil residence time, which can reduce emulsion or foaming problems.

In view of the concerns presented above, the current best practices for conditioning a well for sampling should include the following steps. Complementary details of guidelines can be found elsewhere.^[1]^[6]

For oil reservoirs:

Clean up the well until wellhead pressure and BS&W stabilize.
For surface sampling, reduce the flow rate in steps until the separator GOR does not change between choke sizes, then stabilize separator conditions and take separator samples.
For downhole sampling, shut in the well and build up to static pressure; produce at a low rate for long enough to remove all changed material in the near-wellbore region, and briefly shut in.
Finally, purge fluid past the sampler in the well at a bleed rate and shut in before sampling (this step may be omitted if downhole fluid is known to be monophasic).
For gas/condensate reservoirs: (This seems like it should be the start of a new bullet point list)
Clean up the well until wellhead pressure and BS&W stabilize, then flow the well at the lowest flow rate that will lift liquids up the tubing.
For surface sampling, stabilize separator conditions and take separator samples.
For downhole sampling (undersaturated reservoir with monophasic flow downhole), collect downhole samples.

Recommendations for well conditioning in a near-critical-fluid reservoir are not widely available, both because these reservoirs are fairly rare and because there is no sure way of identifying the situation from surface measurements. A rule of thumb is that reservoirs with near-critical fluids often exhibit separator GORs in the region of 4,000 scf/bbl (700 m³/m³), but there are no well-established GOR ranges for the near-critical region, and correlations are rarely applicable in this area. Because pressure drawdown can result in major changes in the reservoir fluid, it is advisable to perform sampling at the earliest moment (a good practice for all reservoir types) and to condition the well by flowing it at successively slower rates to remove all nonrepresentative hydrocarbon phases as far as possible. However, it may be difficult to establish when the well is adequately conditioned because the surface GOR may change only slightly with different quantities of downhole phases, and expert advice should be sought to evaluate all the observations.

Selection and Preparation of Sampling Equipment

Equipment planning must start as soon as the sampling program is defined to ensure that all necessary equipment will be available and checked well in advance of the sampling operation. For pressurized sampling operations, metal cylinders are invariably used, almost always with valves at each end that facilitate filling, transfer, and cleaning operations. For downhole samples and separator liquid samples, the sampling procedure requires maintaining pressure on the sample while the sample cylinder is filled. Achieving this by draining mercury from a full cylinder^[7] has been largely discontinued owing to the safety and environmental concerns with mercury. The industry is now using piston cylinders, which have an internal piston to separate the sample part of the chamber from the hydraulic fluid (commonly a mixture of water and ethylene glycol). Maintenance of the piston cylinders is more complicated because the piston seals must be in excellent condition to prevent the occurrence of internal leaks.

Cylinders are commonly made from stainless steel or titanium, the latter being significantly lighter and offering better resistance to H₂S (while being incompatible with mercury). Aluminum cylinders are in common use because larger volumes of sample are generally required, and the cylinder weight is a concern. Cylinders with internal coatings, such as Teflon^® * , are also used occasionally. The wide variety of materials used for pressurized sampling demonstrates the fact that there is no perfect solution to the problems of resistance to corrosion, sample preservation, volume, and weight. All cylinders must be provided with plugs for the valves, endcaps to protect the valves, and storage containers to facilitate handling and to protect cylinders during transport. It is good practice to plug and label cylinders when they have been prepared for sampling to avoid any confusion when at the job site. Trademark of DuPont Corp., Wilmington, Delaware

For atmospheric samples, containers that are used for water samples include polyethylene, other plastics, hard rubber, metal cans, and borosilicate glass, but the choice should be based on the measurements that will be performed on the samples. For stock-tank hydrocarbon liquids, glass or plastic containers are more resistant to corrosive components and are generally preferred for smaller volumes. Although plastic containers are more robust, they can contaminate samples with plasticizers, and glass bottles must be used for storing stock-tank samples intended for geochemical analysis. Glass containers must be well protected against the risk of breakage.

For water samples, glass will adsorb various ions such as iron and manganese and may contribute boron or silica to the aqueous sample, so its use should be avoided if ionic analyses are required. Plastic and hard rubber containers are not suitable if the sample is to be analyzed to determine its organic content, and a metal container is often used if the sample is to be analyzed for dissolved hydrocarbons such as benzene. Otherwise, a polyethylene bottle is probably the most satisfactory container, especially if the sample is to be stored for some time before analysis. Also, a plastic container is less likely to break than is glass if a water sample is transported in freezing temperatures. However, not all polyethylenes are acceptable because some contain relatively high amounts of metal contributed by catalysts in their manufacture. The approximate metal content of the plastic can be determined by a qualitative emission spectrographic technique.

In addition to sample containers, sampling equipment must include fittings, valves, gauges, and lines to enable samples to be recovered safely from the required location. Trained personnel must be assigned to collect samples. They should have copies of the sampling program and field procedures, the prepared forms for recording data, and a supply of labels. All sampling equipment must have been previously pressure tested and be clean and dry. Downhole-sampling equipment such as production samplers and formation-test samplers are extremely sophisticated and must be prepared by specialists. Great care must be taken with maintenance, cleaning, and assembly to ensure the maximum chance of correct operation and the minimum chance of contamination or other nonrepresentative sampling.

*

Trademark of DuPont Corp., Wilmington, Delaware

Pressurized Hydrocarbon Fluid-Sampling Procedures

The procedures covered here apply to reservoir fluids or production streams above ambient pressure, and they are highly specific to the petroleum industry. The American Petroleum Inst. publishes a detailed recommended practice,^[6] which is the most complete industry standard covering the sampling of pressurized hydrocarbon fluids. It should be consulted for additional information to that presented here. The choice of sampling method depends on the reservoir-fluid type; this has been explained in the guidelines mentioned above. Here, the various methods have been subdivided into downhole- or surface-sampling methods. The former obviously apply to a specific well, whereas the latter can be used for wells, gathering stations, or other surface facilities.

Although here, "hydrocarbon" is intended to mean fluids containing hydrocarbons and nonhydrocarbons but no (or only small) quantities of water, techniques in this section also can be applied to the sampling of pressurized water fluids, though this is not very common.

Production Downhole Sampling

Production downhole sampling, also referred to as bottomhole sampling, involves running a special sampling tool into the well on wireline so that a sample of the fluid in the well can be collected under the increased pressure of the fluid column. Careful well conditioning is necessary, as described earlier, to ensure that the fluid is in monophasic condition. Modern samplers are triggered either by a timer or a mechanical clock in the tool itself, or by an electric signal conveyed by electric line. The former system is more common, being able to be run with any wireline unit, but it has the inconvenience of needing a preset delay to allow the tool and well to be set up for sampling. The sampler should be lowered into the well until it is a short distance above the upper limit of the perforated interval (unless there are mechanical limitations that prevent the tool from reaching this depth) to collect a sample that is representative of all the produced intervals. Various drillstem and tubing-conveyed installations are available for downhole samplers, which allow them to be operated without the use of wireline. These can allow samples to be collected downhole in high-risk wells in which wireline operations are not permitted.

One advantage of downhole sampling is that it can be performed without a separator at the well. There are several problems that can occur in downhole sampling: the fluid around the sampler may be in two-phase condition, or it may have segregated in the wellbore; a mechanical problem can lead to incorrect opening or closing of the device; the fluid may be contaminated with water or drilling mud; or the sample may not be made fully homogeneous before transfer into a shipping bottle. Use of a pressure survey may help check the whereabouts of any interfaces in the wellbore, but lack of an interface does not guarantee that the fluid present has not lost any material in the form of condensation or wax or asphaltene precipitation.

It is common practice for downhole samples to be transferred at the wellsite, as this allows a measure of the quality of the sample to be obtained, and can allow additional sampling runs to be made in most cases (if needed) while still at the wellsite. This approach also can reduce rental charges for the downhole samplers if supplied by a service company. Because it is difficult either to transfer the entire downhole sample (such that it need not be homogeneous) or to make it fully homogeneous (and just transfer a portion), the best practice is to try to achieve both objectives. Samplers with moving metal parts to facilitate mixing are now fairly common and are preferred. General recommendations to be followed for downhole sampling are given in Table 4.3.

Table 4.3

Downhole samples are commonly transferred to shipping bottles at the wellsite, and the following subsection describes a method suitable for most production downhole samples and many formation-test samples. This step-by-step method is reproduced from RP 44 (where it appears as Section 6.2.5)^[6] by kind permission of the American Petroleum Inst. (API). It may need to be modified according to the actual type of transfer equipment available.

API RP 44 Method for Sample Transfer to Shipping Container. If the sampler itself is not used to transport the sample to the laboratory, the sample must be transferred to a transfer container for shipping or transport. Whatever vessel is used, it must have an adequate pressure rating and be certified to meet all applicable shipping regulations. Further, the shipping cylinders must be cleaned thoroughly; this is particularly important to avoid contamination of the sample from trace amounts of heavy components remaining in the cylinder from previous use.

The primary concern in transferring a downhole sample to a shipping container is to maintain the integrity of the sample during the transfer operation. This requires that the fluid in the sampler be maintained in a single-phase condition during the entire sample-transfer process or, if the fluid is in a two-phase condition, that the entire contents of the sampler be transferred. (The sampler should be heated if wax or asphaltenes are present.) If only a portion of a two-phase sample is transferred, the fluid transferred to the shipping container will differ from the original sample because the two phases in the sampler almost certainly cannot be transferred in the proportions that exist in the sampler. Because valid transfer is crucial to sample quality, the preferred procedure is to maintain the fluid in a single-phase state and transfer it in its entirety. An important consideration is that pressurizing the sample may produce a single-phase condition but may not homogenize the sample; thus, thorough agitation (by rocking the cylinder) during the process is important.

In addition, the sample composition must not be altered either by (a) leaks of hydraulic fluid across the piston of piston-type samplers or (b) by selective absorption of components from the sample into a transfer fluid (e.g., water or glycol) in cases in which the transfer fluid is in direct contact with the sample. The latter is a particular problem in samples containing CO₂ or H₂S, which are very soluble in the transfer fluid.

At all stages of the transfer process, the pressure must be maintained substantially higher than the sample saturation pressure. Fig. 4.3 shows a schematic diagram of a transfer apparatus for piston-type samplers and transport containers. The 1966 Edition 1 of API RP 44^[7] should be consulted for transfer apparatus involving direct contact between the sample and mercury (as the hydraulic fluid).

Fig. 4.3 – Diagram of sample transfer apparatus.

The transfer procedure is as follows.

Use the pump to fill all lines between valves B and F with hydraulic fluid (refer to Fig. 4.3). This can be done by loosening the fittings at these valves and pumping until hydraulic fluid appears, then tightening each fitting. Note: Valves A and B and F, G, and H may be integral parts of the sampler and transfer container, respectively, depending on the design of these vessels. Also, valve H and its line may be arranged somewhat differently from Fig. 4.3 so that valve H simply "tees" into the line from valve A to valve G.
With valves A, D, E, and F closed and valve C open, slightly open valve B and note the opening pressure of the sampler. Valve B is often hydraulically or spring-actuated in cases in which it is part of the sampler; if so, use the pump to raise the pressure until valve B just opens, and record the opening pressure.
Open valve G and evacuate through valve H the line between valves G and A, including the sample side of the transfer container.
Close valve H.
Open valve F and use the pump to bring the hydraulic oil pressure in both the sampler and the shipping container to a pressure well above the saturation pressure of the sample.
Slightly open valve A and fill the line between that valve and the upper face of the piston in the shipping container with sample fluid, using the pump to keep the pressure on the gauge well above the saturation pressure during this transfer process.
Close valve C, then slightly open valve D, allowing hydraulic fluid to drain slowly into the hydraulic oil reservoir (open to atmospheric pressure) as fluid flows from the sampler to the shipping container. Use the calibrated pump to (a) keep the pressure in the sampler above the saturation pressure and to (b) keep track of the amount of sample transferred. When the desired amount of sample has been transferred, close valve D, then close valves A and G.
Before the transferred sample can be shipped, a vapor space must be created in the shipping container. To do this, slightly open valve E and allow hydraulic oil to drain from the shipping container into an open calibrated receiver. Close valve E, then valve F, when the volume of hydraulic fluid in the receiver equals 10% of the volume of the shipping container. This will result in a 10% vapor space ("ullage" or "outage") in the shipping container. Such a void volume is required for safety because very high pressures can result if the temperature increases even slightly in a totally liquid-filled, closed vessel. Note: Special sample cylinders with an auxiliary gas cap are available for samples that must be retained in single-phase (monophasic) condition.
Close valve B if it is not self-sealing. Open valve C, then valve D, to relieve pressure in the pump. At this point, the sampler and shipping (transfer) vessel can be disconnected from the transfer apparatus.

Downhole Sampling With Formation Testers

The collection of reservoir-fluid samples by formation-test tools was originally a secondary benefit of their use for the measurement of pore pressures. Formation-test tools can obtain reservoir-fluid samples without any production to the surface. The tool is typically run into an openhole well containing drilling mud or completion fluid to a specific depth, and a probe is forced against the formation, providing a seal and allowing formation fluid to flow into the tool. Modern formation testers generally can be equipped with numerous devices designed specifically to enable samples of reservoir fluid to be collected in a series of sample chambers. These tools offer the advantage of the ability to collect samples without performing a DST with fluid flow to the surface, and they are especially useful in obtaining fluids from a number of discrete depths, thus helping identify possible fluid gradients. However, the principal disadvantage is the limited cleanup that is possible, resulting in various levels of contamination by drilling-mud filtrate. These problems have been reduced by developments allowing the pumping of significant volumes of fluid into the well and the monitoring of the quality of the fluid flowing into the tool. Nevertheless, samples collected almost always contain some contamination (both from mud filtrate and from small quantities of water used to fill connecting lines when the tool is prepared), but when the drilling mud is water-based, such contamination can be separated in the laboratory, and it is a significant concern only when sampled fluids contain soluble components such as H₂S and CO₂. When sampling in wells that have been drilled with OBM, contamination is more difficult to detect and impossible to remove physically. Advanced spectroscopic detection systems have been developed for formation-test tools, but the industry is now beginning to accept that there always will be problems with formation-test sample contamination where OBMs have been used, and laboratories have developed various methods for evaluating the level of contamination and for estimating the true physical properties of uncontaminated reservoir fluid.^[8] In fact, this problem is not limited to formation-test samples because in some cases, production testing may not fully clean up OBM filtrate, especially if there have been significant losses during drilling.

Formation-test tools are extremely sophisticated and must be run by specially trained engineers and wireline operators. In addition, significant differences exist between the tools available from the various service companies, and technological developments are occurring all the time, so specific operating details will not be given here. However, in addition to the well preparation described earlier, a number of recommendations can be made for the sampling process:

Planning must optimize the match between tool capability and sampling and analysis needs.
Sample sizes collected should be compatible with storage containers so that individual samples can be transferred in their entirety.
Sample chambers containing mixing devices are to be preferred because they facilitate sample homogenization before transfer; where possible, duplicate samples should be taken from each depth sampled.
To determine depth gradients, samples should be collected from at least three different depths spanning the reservoir interval; when available, fluid-quality monitors should be used to evaluate cleanup of the fluid entering the tool.
The fluid-sampling rate should be adjusted where possible to minimize pressure drawdown, unless downhole bubblepoint measurement or estimation are available that allow higher sampling rates to be used with confidence.
If OBM was used in drilling, collect a sample of the mud that has been used most recently, and contact a laboratory that will analyze the samples to establish which fluid samples are needed; for some correction techniques, samples are required from the same depth with different levels of filtrate contamination.
Use of the formation-tester pump to compress collected samples (sometimes referred to as "overpressuring") may help reduce the effects of cooling, but it should not be used if final pressures are to be used as a measure of sample quality. If phase segregation on cooling must be avoided, single-phase sample chambers should be selected as described below.
If fluid pumpout into the well is not possible (e.g., for safety reasons—H₂S, low overbalance, etc.), large sample chambers should be used at the start of sampling to serve as "dump" chambers, allowing better-quality samples to be collected afterward.
The depth and sampling time must be recorded together with the serial number of each chamber.
If possible, avoid using OBM when drilling, or switch to water-based mud for probable hydrocarbon-bearing intervals. Handling and transfer of formation-test samples should be along the lines described above for production-test downhole samples.

Single-Phase Sampling

Downhole samples cool down as they are pulled out of the well, and the associated fall in pressure will usually result in the sample entering the two-phase region, thus necessitating homogenization before transfer. Special versions of downhole samplers now available, known as single-phase or monophasic samplers, use the release of gas pressure behind an additional piston to maintain a downhole sample above reservoir pressure while it is brought to the surface. This design of sampler is especially used for reservoir fluids likely to precipitate asphaltenes, which are very susceptible to pressure reduction and difficult to homogenize. For other fluids, single-phase samplers facilitate sample transfer and reduce the chance of the transferred fluid not being representative of the fluid in the sampler. One disadvantage of the one-phase sampler is that a "bubblepoint check" cannot be performed on site because the gas buffer will mask sample behavior. One solution to this limitation is to run a conventional sampler in tandem to permit a quality check on one of the samples in the field.

Although the single-phase sampler will prevent the formation of a gas phase in most cases, it does not prevent the formation of a wax phase in waxy reservoir fluids, which commonly occurs with cooling. A sampler with a heated chamber is available but has not been used widely. Also, gas/condensate fluids undergo significant shrinkage on cooling, and single-phase samplers may not prevent the formation of condensate in the sample chamber. Single-phase versions of formation-tester sample chambers are also available.

Other Downhole-Sampling Tools

Various other tools can be used to collect downhole fluid samples, such as DST chambers, but thought must be given to the problems in recovering a valid sample from the tool, and preference must be given to configurations that allow samples to be homogenized, transferred under pressure, and preferably contained in a single storage cylinder. Industry practice now favors the use of standard wireline samplers conveyed into the well as part of the DST tool.

Separator Sampling

Surface sampling primarily involves sampling individual gas and liquid streams from a production separator or similar installation, and it is by far the most common method of sampling pressurized hydrocarbon fluids. The operation of oil and gas separators is covered in detail elsewhere in this Handbook. Usually, the objective of separator sampling is to obtain a fluid representative of the production of one well that enters the separator in its entirety, but the method also can be used to obtain a fluid representing commingled production from a number of wells into a single gas/oil separation plant. In either case, the objective is to collect separate samples of the gas and liquid exiting the separator and to measure the separate flow rates of the two phases and obtain the GOR. Although the two phases are never in perfect equilibrium, providing that the two samples are representative of the separate flows, it is possible to mix the two samples together in the same proportion in which they are produced to obtain a recombined sample that represents the fluid entering the separator.

Some of the biggest errors affecting fluid samples are related to the measurement of separator-gas and -liquid flow rates, which are crucial for the recombination process in the laboratory. Good accuracy is often considered to be in the region of 5%, but the figure can be much worse, for example, if there is carry-over of liquid in the gas exit stream (or carry-under of gas in an oil with foaming tendencies). Problems are especially common for gas-well production tests, where very high flow rates can be used, and special techniques are available for trying to measure liquid carry-over in such situations. However, the best approach involves proper sizing and adjustment of the separator for the production rate. Another important source of error in this domain involves confusion over whether liquid flow rates are reported at separator conditions or at tank conditions; this has serious implications for gas/condensate fluids in which the separator-liquid shrinkage is typically much larger than in the case of an oil.

Although broad guidelines were given above concerning the volumes of samples that should be collected, special attention should be given when collecting gas samples from separators operating at low pressures because the lower density may result in the collection of insufficient weight of gas. Fig. 4.4 enables the required volume of gas to be estimated simply as a function of separator pressure, GOR, and the volume of liquid that is required. This chart is reproduced with the kind permission of Saudi Aramco.

Fig. 4.4 – Separator-gas volume to be collected as a function of pressure, GOR, and oil volume.

In line with current trends in the oil industry, this work recommends using evacuated bottles for gases and piston bottles for liquids and avoiding any use of mercury in sampling operations. At extremely low temperatures, piston bottles have been reported to leak past the piston seal, so sampling under such conditions should be avoided if possible. If these methods cannot be used, then repeated purging (a minimum of five times) should be used for gas samples, and the displacement of brine should be used for liquid samples unless high H₂S or CO₂ levels are present, in which case it is preferable to use separator water saturated with gas if it is available. The principal guidelines to be followed for surface sampling are given in Table 4.4.

Table 4.4

The following two subsections describe the two most common separator-sampling methods in detail.

API RP 44 Gas Method No. 1: Filling an Evacuated Container. The following step-by-step method is reproduced from RP 44 (where it appears as Section 6.3.7.1)^[6] by kind permission of API.

This method is especially simple and accurate. The principal undesirable feature of the method is the requirement that the vessel be evacuated before its transport to the sampling point (with possible loss of vacuum during transport), or that a vacuum pump be provided at the wellsite. Testing pre-evacuated vessels for adequate vacuum at the time of sampling should be done only by personnel well trained in vacuum-testing procedures because improper testing often leads to loss of vacuum or introduction of air into the sample vessel. (Collecting an additional sample may be preferable to vacuum testing.) A clean, evacuated container should never be purged with separator gas and re-evacuated in the field because any liquid that condenses in the container during the purge may not totally re-evaporate during evacuation in the field.

Sample collection is accomplished by the following steps:

Locate an appropriate sample source valve A (see Fig. 4.5 ) on the separator from which the desired sample can be collected. Clean any debris from valve A; open the valve briefly to blow it out, and then close it.
Connect the fitting on the flexible tubing of the sampling rig securely to valve A on the separator. Open the line valve B, and open the purge valve C.
If a vacuum pump is available and personnel are qualified in vacuum techniques, connect the sample inlet valve D on the sample container to the fitting on the sampling rig, as shown in Fig. 4.5. Connect the vacuum pump to valve C, open valve C and valve B to evacuate the sampling rig, then close valve C and disconnect the pump. Slowly reopen valve A completely to establish full separator pressure on the entire sampling rig from valve A to valve D, and proceed to Step 6.
If a vacuum pump is not available, open valves B and C, then open and close valve A in one quick burst to purge air from the sampling rig, and quickly close valve B. Slowly reopen valve A completely to establish full separator pressure on the entire sampling rig from valve A to valve B.
Connect the sample inlet valve D on the sample container to the fitting on the sampling rig, as shown in Fig. 4.5. Open valve C, then open and close valve B in one quick burst to purge air from the line connecting valves B and D, and close valve C promptly. Note: Use a long vent line on valve C if H₂S is present. Reopen valve B to establish full separator pressure on the entire sampling rig from valve A to valve D.
Cautiously crack open valve D, while carefully monitoring the pressure gauge, and fill the container slowly. Continuously adjust valve D as needed to keep full pressure on the pressure gauge. Filling a large container can take as long as 20 minutes. The progress of the filling process can be monitored by listening for a hissing sound at valve D (and in the container) and by monitoring the pressure gauge. When you think that the container is full, open valve D further while listening to the container and monitoring the pressure gauge.
When the container is full, close valve D, and then close valve B.
Slightly open valve C to bleed the connections between valves B and D to atmospheric pressure. Note: The line from valve A to valve B, including the pressure gauge, is still under full pressure. Use a long vent line on valve C if H₂S is present.
Disconnect the sample container. This is the last step in collecting the first sample. The apparatus is now ready for collecting additional samples by repeating Steps 5–8.
Following collection of the last sample, close valve A securely, then open valve B (and valve C, if it is not already open) to bleed pressure from all parts of the line and sampling rig before disconnecting the line from valve A. Note: Use a long vent line on valve C if H₂S is present.
Insert sealing plugs into the valves on each sample container; then check the valves for leaks by immersing them in water or painting them with soap solution. Before inserting the sealing plugs, the threads should be lubricated by stretching Teflon^® tape into the threads or by applying pipe dope. After a container is determined to be leak-free, it should be tagged and otherwise prepared for storage or transit.

Fig. 4.5 – Diagram of a sampling apparatus.

API RP 44 Oil Method No. 3: Filling a Piston-Type Container. The following step-by-step method is reproduced from RP 44 (where it appears as Section 6.3.8.3 )^[6] by kind permission of API. It refers to the same sampling rig as that used for the gas-sampling method above, though the sample cylinder will contain a piston, and valve E will represent the hydraulic-fluid connection, as indicated in Fig. 4.6. Some steps in this procedure may need modification depending on exact equipment design; this is notable for sample cylinders, which have an additional purge valve at the sample inlet end of the cylinder.

Fig. 4.6 – Diagram of a liquid-sampling apparatus.

This is a preferred method for nonmercury liquid-sample collection. It has the advantage that the liquid sample can be kept at the saturation pressure throughout the collection process, which avoids gas breakout from the sample. In addition, the sample does not come into contact with any other fluids during sampling or during transfer in the laboratory. The undesirable feature of the method is that with sample containers, the potential for contamination with hydraulic fluid exists if the seal on the piston leaks. (Water can be used as the hydraulic fluid to minimize the possibility of contamination, but the operator should first check with the manufacturer to ensure that water will not damage the container.)

If a piston-type container is being used, hydraulic fluid must be preloaded behind the piston so that the piston position is fully toward the sampling end. A danger is that inexperienced personnel may not know this and may attempt to use this type of container without a proper fill of hydraulic fluid and without proper hydraulic-pressure support on the piston seal. In such a case, full pressure will not be maintained on the separator oil during sampling, and the process essentially will be the same as filling an empty container, except that the seal on the piston might leak. The manufacturer’s instructions should be consulted to ensure that the operation of the piston-type container is completely understood before commencing the sampling operation.

The procedure is as follows:

Locate an appropriate sample source valve A on the separator (see Fig. 4.6) from which the desired oil sample can be collected. Clean any debris from valve A, hold a rag over the valve (or attach a temporary purge line connected to a suitable container), open valve A slowly, purge sufficient oil through the valve, and then close valve A. Remove the rag or temporary purge line. Note: Use a long vent line if H₂S is present.
Connect the fitting on the flexible tubing of the sampling rig (see Fig. 4.6) securely to valve A on the separator. Open the line valve B, and open the purge valve C.
If a vacuum pump is available and personnel are qualified in vacuum techniques, connect the sample inlet valve D on the sample container to the fitting on the sampling rig, as shown in Fig. 4.6. Connect the vacuum pump to valve C, open valve C and valve B to evacuate the sampling rig, and then close valve B. Slowly reopen valve A completely to establish full separator pressure on the entire sampling rig from valve A to valve B. Open valve D to evacuate the connection and the small dead volume in the container (the internal volume between valve D and the face of the piston when the piston position is at the sampling end), then close valve C and disconnect the pump. Slowly reopen valve B completely to establish full separator pressure on the entire system from valve A through valve D to the face of the piston in the container, and proceed to Step 6. Be sure that valve D is completely open.
If a vacuum pump is not available, close valve B and open valve A slowly (the pressure on the gauge should rise to the separator pressure). Close valve A, attach a purge line at the end of the rig below valve C, close valve C, and open valve B to let the pressure deplete to atmospheric. Close valve B, then slowly reopen valve A completely. Slightly open valve B, and slowly purge a volume of oil equivalent to several times the volume of the sampling rig, collecting the purged oil in a suitable container (maintain full separator pressure on the pressure gauge during this purge). Close valve B, and remove the purge line. Full separator pressure should now be on the entire sampling rig from valve A to valve B.
Connect the sample inlet valve D on the sample container to the fitting on the sampling rig, as shown in Fig. 4.6, and attach a purge line at the end of valve C. Open valve D, close valve C, and open valve B slowly to pressure up the connection with the container and any dead volume in the sample container. Close valve B, and open valve C to let the pressure deplete to atmospheric. Close valve C, then slowly reopen valve B completely. Slightly open valve C, and slowly purge a volume of oil equivalent to several times the volume of the connection, collecting the purged oil in a suitable container (maintain full separator pressure on the pressure gauge during this purge). Close valve C, and remove the purge line. Full separator pressure should now be on the entire sampling rig from valve A through valve D to the face of the piston in the sample container. Be sure that valve D is completely open. Note: This method is not perfect because the oil in the dead volume in the sample container has not been purged under pressure. However, if the piston position is fully toward the sampling end of the container, the amount of oil in the dead volume will be negligible.
Cautiously crack open sample outlet valve E while carefully monitoring the pressure gauge, and allow the sample fluid to slowly displace the preload hydraulic oil into a suitable collection vessel. Continuously adjust valve E as needed to be sure that the rate of sample collection is sufficiently slow so that full separator pressure is maintained on the sample side of the piston (as indicated by the pressure gauge). The sampling operation can be ended when a desired volume of sample is collected (as indicated by a given volume of hydraulic fluid being displaced to the collection vessel). The operation must be stopped with at least enough preload liquid left in the container to provide the "outage" required in Step 7. Close valves E, D, and B, in that order. (If the container has a magnetic indicator to show the position of the piston, then nitrogen gas can be used as the hydraulic fluid behind the piston, and Step 7 can be eliminated so long as approximately 10% volume of nitrogen remains on the hydraulic side of the piston.)
Open valve E slightly (with valve D closed), and drain into the collection vessel a volume of hydraulic oil equal to approximately 10% of the container volume. This will create the necessary vapor space in the container without altering the overall composition of the oil sample. (Be sure to leave at least some hydraulic oil behind the piston so that there is pressure support on the seal and very little pressure drop across the seal). Close valve E securely.
Slightly open valve C to bleed the connections between valves B and D to atmospheric pressure. Note: The line from valve A to B, including the pressure gauge, is still under pressure. Use a long vent line if H₂S is present.
Disconnect the sample container. This is the last step for the first sample and leaves the apparatus ready for collection of additional samples by repeating Steps 5–8.
Following collection of the last sample, close valve A securely, then open valve B (and valve C, if it is not already open) to bleed pressure from all parts of the line and sampling rig before disconnecting the line from source valve A.
Wipe the valves on the sample container clean and inspect for any signs of leakage. After a container is determined to be leak-free, insert plugs in the valves, then tag the container and otherwise prepare it for storage or transit. Before inserting the sealing plugs, the threads should be lubricated by stretching Teflon ^® tape into the threads or by applying pipe dope.

Wellhead Sampling

Wellhead sampling, more commonly known as flowline sampling, involves the collection of a fluid sample at the surface from the wellhead itself or from the flowline or upstream side of the choke manifold, provided that the fluid is still in one-phase condition. This option is restricted to wells producing dry gas, very-low-GOR oils, and some high-pressure/high-temperature reservoir fluids. Dry-gas wellhead samples can be collected as for gas sampling from a separator, whereas wellhead sampling of other or unknown fluids should be performed as for separator liquids. However, all equipment must be compatible with maximum wellhead pressure, and as the state of the fluid is not usually known with certainty, separator sampling also should be performed if possible, as a backup.

Isokinetic sampling, also known as split-stream sampling, involves collecting samples from well production in two-phase flow, using a small side stream to allow the two-phase fluid to be collected and measured in laboratory scale equipment at the wellsite. There are two principal challenges in this approach: controlling the side stream so that it is removed from the main flow at identical velocity (hence the term isokinetic) to avoid disproportionate sampling of the two phases, and ensuring that the flow is turbulent upstream of the sampling probe so that the minor phase is finely distributed in the major phase. Although this special type of sampling has been used for more than 60 years, mainly for sampling gas/condensate production, many still consider it to be at the development stage,^[2] and it has never achieved wide acceptance. A more recent development of isokinetic sampling involves sampling of the exit gas stream from a separator and calculation of a figure for separatorefficiency. This efficiency is then used to modify the GOR used for recombining separator samples, but it should be compared to the separator efficiencies reported elsewhere in this Handbook.

Nonpressurized Hydrocarbon Fluid-Sampling Procedures

The sampling of nonpressurized or atmospheric-pressure hydrocarbon fluids from lines is relatively simple to perform, but attention must be paid to the need to purge sampling lines and pipework with at least three times their volume of fresh fluid before each sampling session. This is especially important in some installations and processing facilities, where the sampling point may be at the end of a "dead-leg" or trap in which fluid has collected or stagnated over a long period of time. In general, oil or condensate samples should be collected from a sample tap on the side of the line or the top of the line to avoid any water or sediment that may have accumulated at the bottom of the line. Atmospheric gas samples are rarely collected, but if they are required, they should be collected in evacuated chambers to minimize contamination by air.

Atmospheric hydrocarbon samples also may be collected from pressurized lines or from samples collected in pressurized chambers, such as downhole samples. Usually, this will involve the release of gas and the collection of oil or condensate. Because the separation procedure that releases gas is dependent on the temperature and pressure (which may be above atmospheric if the liquid is collected in a closed trap), the properties of samples collected in this way may vary. Also, because the fluid in a sample chamber may already be in two-phase condition (or may have segregated), liquid from the entire sample should be collected to minimize uncertainty in the sample quality.

Sampling from tanks is complicated by the need to collect samples from various depths to allow for any property changes or segregation that may exist. The procedure given next is a traditional method used for measuring and testing a field tank of crude oil, frequently referred to as "running" when related to custody-transfer transactions. It was published as API Standard 2500 but is no longer available. The method is reproduced here with the permission of the American Petroleum Inst. It is intended to support operations still using this method, or methods derived from it, and serve as a guideline to engineers setting up similar methods. Detailed descriptions of individual calibration and measurement methods are available in the API Manual of Petroleum Measurement Standards (MPMS),^[9] which represents all branches of the petroleum industry and is the recognized standard for downstream measurement methods.

Procedure for Typical Measuring, Sampling, and Testing of a Tank of Oil

The tank is vertical, nonpressurized, and has a fixed roof with side outlets; it is to be gauged by the innage method (a process to determine the depth of liquid in a tank, which is measured from the surface of the liquid to the tank bottom or to a fixed datum plate).
The oil viscosity is less than 100 Saybolt seconds at 100°F and is a liquid at atmospheric temperature and pressure.
A cup-case thermometer is used to read the temperature of the oil in the tank.
A thief is used to obtain fluid samples from the tank. (A "thief" is an industry term for a bottom-closure, core-type sampler used to secure samples from chosen depths in storage tanks.)
The API gravity scale hydrometer test method is used to determine the API gravity of the oil; the temperature of the oil has to be near 60°F (±5°F).
The water and sediment in the oil are to be determined by the centrifuge method with a 203-mm (8-in.) cone-shaped tube.

The following outline gives the sequence of steps to be taken and the key points to be noted at each step.

Isolate the tank to be checked.
Use safety precautions and fresh air bottles if an H₂S hazard exists.
Ground yourself to a stair railing or tank shell before reaching the top. This prevents static-electrical discharge in a hazardous area.
Stand to the side of the hatch when opening it to permit wind to blow gas away from you.
Measure the temperature: suspend a thermometer in the oil tank. The thermometer should be 12 in. or more from the tank shell and must be immersed in oil for 5 minutes.

Use an American Soc. for Testing and Materials (ASTM)-approved, wood-back or corrosion-resistant metal cup case. If atmospheric temperature differs by more than 20°F from that of the liquid in the tank, the cup case should be given at least two preliminary immersions. Empty the cup case after each immersion.

Rapidly withdraw the thermometer and read and record the temperature to the nearest 1°F. Note: The number of temperature measurements varies with the depth of the liquid.

In a tank containing more than 15 ft of liquid, three measurements should be taken: (1) 3 ft below the top surface of the liquid, (2) in the middle of the liquid, and (3) 3 ft above the bottom of the liquid.

In a tank containing 10 to 15 ft of liquid, two measurements should be taken: (1) 3 ft below the top surface of the liquid, and (2) 3 ft above the bottom surface of the liquid. In a tank containing less than 10 ft of liquid, one measurement should be taken in the middle of the liquid. For tanks over 10 ft high with a capacity of less than 5,000 bbl, one measurement in the middle of the liquid should be taken.

With a thief, take sample(s) for a BS&W centrifuge test. Note: The number of samples to be taken for BS&W determination varies.

In tanks larger than 1,000-bbl capacity that contain more than 15 ft of liquid, equal-volume samples should be taken (1) 6 in. below the top of the liquid, (2) at the middle of the liquid, and (3) at the outlet connection of the merchantable oil, in the order named. This method also may be used in tanks up to and including a capacity of 1,000 bbl.

In a tank larger than 1,000-bbl capacity that contains more than 10 ft and up to 15 ft of liquid, equal-volume samples should be taken (1) 6 in. below the top surface of the liquid and (2) at the outlet connection of the merchantable oil, in the order named. This method may be used on tanks up to and including a capacity of 1,000 bbl.

In a tank larger than 1,000-bbl capacity that contains 10 ft or less of liquid, one sample may be taken in the middle of the column of liquid.

Place the BS&W composite sample in a sample container. The sample should be a blend of the upper, middle, and lower samples (if three samples were required), containing equal parts from the samples taken.
Seal the sample container. In the lower 48 states, with the exception of California, the sample is ready to be tested for BS&W, as described in Step 17. In California, the container should be labeled and delivered to the laboratory for BS&W determination. ( Note: These U.S. state references were part of the original standard .)
With a thief, take a sample for gravity determination. The sample should be taken midway between the oil surface and the pipeline connection. Hang the thief in the hatch. Remove bubbles, and place the hydrometer in the oil sample.
Determine and record the sample temperature to the nearest 0.5°F. The hydrometer must float away from the wall of the cylinder; the temperature of the surrounding media should not change more than 5°F.

Depress the hydrometer two scale divisions and release. Read the hydrometer to the nearest 0.05°API on a scale at which surface or liquid cuts scale.

Read and record the sample temperature to the nearest 0.5°F. Repeat the gravity reading if the temperature of the sample before and after the gravity reading has changed more than 1°F. Apply any relevant correction to the observed hydrometer reading (correction scale on bulb) to the nearest 0.1°API. Record the mean temperature reading observed before and after the final hydrometer reading to the nearest 1°F.

Note: Hydrometer scale readings at temperatures other than calibration temperatures (60°F) should not be considered more than scale readings because the hydrometer bulb changes with temperature.

Convert the relevant corrected value to standard temperatures. Use API MPMS Chapter 11.1 (Table 5A)^[9] for crude oils.
Take the bottom thief sample for BS&W height. Lower the clean, dry thief slowly into the oil to the desired depth, trip the thief, and raise it slowly to avoid agitation. When the sample is taken, the top of the thief must be 2 in. above the bottom of the pipeline connections.
Determine and record BS&W height in the tank. Pour the remaining thief sample over a test glass until contamination appears. Measure and record (as the BS&W height) the distance from the bottom of the thief to the top of the contamination in the thief. If BS&W height is less than 4 in. from the bottom of the pipeline connection, do not run the tank.
Gauge the tank. Do not gauge a boiling or foaming tank. Use steel innage tape with an innage plumb bob. Always make contact between the gauge line and the hatch while running tape into the tank.

Gauge the tank only at the reference point on the gauging hatch. On tanks of 1,000-bbl capacity or less, read to the nearest 1/4 in. On tanks of 1,000 bbl or more, read to the nearest 1/8 in. Record the reading immediately; repeat until two identical gauges are obtained.

Saturate solvent with water. Toluene is approved solvent; it is flammable and toxic. Care should be taken when using toluene.

Fill a 1-qt or 1-L glass bottle with a screw top with 700 to 800 mL of toluene. Add 25 mL of either distilled or tap water. Screw the cap on; shake vigorously for 30 seconds. Loosen the cap; place the bottle in a bath for 30 minutes. Maintain the bath at a constant temperature of 140 ±5°F. Remove, tighten the cap, and shake vigorously for 30 seconds. Repeat three times.

Allow the bottle of water/toluene mixture to sit in the bath for 48 hours before using. Be sure that no free water is left in the bottle.

Shake the sample container until the sample is well mixed. Fill two 203-mm (8-in.) cone-shaped centrifuge tubes with 50 mL of sample. Use a pipette to add 50 mL of toluene. Toluene should be water saturated at 140°F. Read the top of the meniscus at both the 50- and 100-mL marks. Add a 0.2-mL demulsifier if necessary for a clean break in the oil/water contact.

Stopper the tube tightly; invert the tube 10 times to ensure that oil and solvent are uniformly mixed.

Loosen the stopper slightly. Immerse the tube to the 100-mL mark in a bath for 15 minutes. The bath must maintain a temperature of 140 ±5°F; by contract agreement, the bath temperature may be 120 ±5°F.

Remove the tube from the bath and tighten the stopper. Invert the tube 10 times to ensure that oil and solvent are uniformly mixed.

Place the tubes in trunnion cups on opposite sides of the centrifuge. Spin for 10 minutes while maintaining minimum relative centrifuge force of 600.

Following the spinning, read and record the combined volume of water and sediment at the bottom of each tube. Read to the nearest 0.05 mL for oil from 0.1- to 1-mL graduation. Read to the nearest 0.1 mL above 1-mL graduation. Estimate to the nearest 0.025 mL below 0.1-mL graduation.

Return the tube to the centrifuge without agitation. Spin for 10 minutes at the same rate. Repeat this operation until the combined volume of water and sediment remains constant on two consecutive readings.

Record the final volume of water and sediment in each tube. The sum of the two admissible readings is the vol% of water and sediment in the sample.

After the tank has been run, the following closing data should be obtained.

Closing oil temperature: no closing temperature is necessary on tanks of 5,000 bbl or less; on tanks of 5,000 bbl or more, always read to the nearest 1°F.
Obtain a closing gauge reading at the same point and in the same manner as the opening gauge reading.
Obtain the bottom thief. If the BS&W level is lower than the opening gauge, report this to a supervisor.

More information concerning the specific measurement methods referred to here can be found in the API MPMS.^[9] The manual is being updated continually, and care should be taken that the current standard or chapter is used. Identification of the appropriate section can be made using the publication catalog on the API website (http://www.api.org/cat/).

Oilfield Waters

Oilfield waters are often referred to as brines, especially when they contain significant quantities of dissolved salts. They also frequently contain dissolved gases (more details are available elsewhere in this Handbook) and may contain small quantities of the heavier hydrocarbons found in oils. Water can be present in a surface separator during production, either from liquid water in the zone being tested or by condensation from water vapor in the produced gas, or possibly from both. Water from aquifers or seawater may also need to be analyzed in connection with water-injection activities.

The analysis of oilfield waters has a wide range of applications, including identifying the origin of produced water, characterizing aquifer properties, interpreting wireline-log measurements, predicting formation damage from water incompatibility, investigating scaling tendencies in surface and downhole equipment, monitoring fluid movement in reservoirs, identifying the presence of bacteria, evaluating disposal options and environmental compliance, and predicting and monitoring corrosion. Water analyses also can be useful in diagnosing and correcting numerous oilfield operating problems.

API publishes Recommended Practice 45,^[10] which contains information on the applications of oilfield-water analyses and gives recommendations for the proper collection, preservation, and labeling of oilfield-water samples. RP 45 also gives a description of numerous analytical methods and recommends appropriate reporting formats for analytical results. This publication should be consulted for more information about specific analytical methods and any special sampling or storage requirements linked to such methods. Numerous analytical methods are also available as ASTM standards.^[11]

When sampling and analysis are part of a long-term monitoring program, such as those required by government authorities or those forming part of custody-transfer contracts, the methods defined in the appropriate documentation or contracts must be followed as closely as possible, even if this constitutes differences with the procedures or recommendations in this text or with the industry standards cited here. However, the guidelines provided here should be taken into consideration before contracts are drafted or when existing contracts are renewed.

If samples are to be collected for the measurement of trace components, biological species, or reactive chemicals that are likely to be affected by storage, container material, or ambient conditions, on-site analyses should be considered. API RP 45 lists the following measurements that should be carried out immediately in the field after sampling and filtering oilfield waters: (1) pH, (2) temperature, (3) alkalinity, (4) dissolved oxygen, (5) CO₂, (6) H₂S, and (7) total and soluble iron. Other measurements or preparations to be performed in the field include (8) turbidity on an unfiltered sample, (9) total suspended solids with at least primary filtration and washing performed in the field, (10) bacteria with filtering and/or culturing in the field, and incubation and counting performed in the laboratory. Biological determinations are outside the scope of this document but are covered in detail elsewhere.^[12]^[13]

For many other analyses, special preparation and preservation measures are required to be performed in the field. This can involve acidification with various acids, addition of other chemicals, refrigeration [ideally 39°F (4°C)], and storage in the dark. If there is any uncertainty concerning sample storage conditions, the laboratory that will perform the analyses should be consulted for advice. If no information is available, it is advisable to keep samples cool and out of the sunlight.

For a DST that does not flow to the surface, great care must be taken to determine if the test has flowed sufficient fluid to allow representative reservoir brine into the tool. The best practice is to sample the water after each stand of pipe is removed. Normally, the total-dissolved-solids content will increase downward and become constant when pure formation water is obtained. A test that flows water will give even higher assurance of an uncontaminated sample. If only one DST water sample is taken for analysis, it should be taken just above the tool because this is the last water to enter the tool and is least likely to show contamination.

Surface sampling is commonly used to obtain a sample of formation water from a sampling valve at the wellhead or another sampling point. A plastic or rubber tube can be used to transfer the sample from the sample valve into the container. Fig. 4.7 shows a simple method of excluding air when sampling water in this way. After purging the sample valve and line to remove any foreign material, water is delivered to the bottom of the sample bottle, which is placed in a large, much taller beaker until the water fills the beaker and overflows. Then, the cap is immersed in the beaker and inverted to eliminate air bubbles before removing the delivery tube and closing the sample bottle under water. This technique cannot be used when acid or other preservatives must be added to the sample.

Fig. 4.7 – Water-sampling method to avoid air.

An alternative sampling technique for use when a clean source of water is available is shown in Fig. 4.8. Here, once the sample point and line have been purged, the sample is collected in the sample cylinder by closing the two valves. This system should not be used to collect pressurized water samples.

Fig. 4.8 – An alternative method for collecting air-free water samples.

In many producing wells, it may be impossible to locate a suitable sampling point free from oil or gas, such as for pumping wellheads in which the brine will surge out in heads and be mixed with oil. In such situations, a larger container equipped with a sampling tap near the bottom can be used as a surge tank or oil/water separator. Such a device is shown in Fig. 4.9. This method will serve to obtain samples that are relatively oil-free.

Fig. 4.9 – Oil/water separator and water-sample filtering system.

For some measurements, it is necessary to obtain a field-filtered sample. The filtering system shown in Fig. 4.9 is simple and economical and can be used for various applications. It consists of a 50-mL disposable syringe, two check valves, and an inline disk-filter holder. The filter holder takes size 47-mm-diameter, 0.45-μm pore-size filters, with the option of including various prefilters. The syringe fill line should be connected to a source of brine free from oil or gas, either directly to a suitable sample point or to the brine outlet from a suitable separation vessel, as shown in Fig. 4.9. The brine is drawn through the inlet line into the syringe and then forced through the filter into the collection bottle. The check valves allow the syringe to be used as a pump for filling the collection bottle without needing to open and close valves. If the filter becomes clogged, it can be replaced in a few minutes. Approximately 2 minutes are required to collect 250 mL of sample. Usually two samples are taken, with the one being acidified to pH 3 or less with concentrated hydrochloric or nitric acid. The system can be cleaned easily or flushed with brine to prevent contamination.

If pressurized water samples are required, most of the procedures described previously for pressurized hydrocarbon fluids can be used, including downhole sampling. Piston sample bottles are essential because the sample cannot be collected by the displacement of water or brine.

Sampling-Data Measurement and Recording

In the same way that laboratory measurements require representative samples to be meaningful, the samples themselves must be supported by accurate data to provide a unique identification and to record all important production and sampling brmeters that will be used in checking the sample and (in many cases) in determining the exact measurements that will be performed. This section reviews the importance of data measurement and provides guidelines for recording and validating the necessary data.

Provided that flowmeters and pressure gauges are properly sized for a measurement, so that readings are not made at the low end of the measurement range, random errors are generally small. Systematic errors are a major concern, however, for all measurements, deriving from sensor malfunction, poor (or lack of) calibration, and human error in general; the latter item can include both errors in recording and reporting data and those deriving from the use of computer-based acquisition systems (e.g., entry of erroneous calibration data, incorrect sensor connections, and even software bugs). Although systematic errors are comparatively rare, their magnitude can be significant. In fact, on some occasions, errors are identified only when measured values are so large that the values become ridiculous.

The GOR is considered to be the most important measurement for separator samples, and it is dependent on errors in both the gas flow rate and the oil flow rate, which are measured separately. New techniques have seen limited application to reduce errors in GOR, such as the injection of a standard marker chemical upstream of the separator and measurements of the concentrations in the separated gas and liquid streams. Also, use of various carry-over measurement techniques has been made, such as the isokinetic approach described briefly earlier. However, significant improvements can be achieved simply by proper sizing, calibration, and recording.

In production testing, gas flow rate itself is widely measured by the orifice meter. This system has been in use for a long time, but new standards have been issued more recently that improve gas rate calculations.^[14] The orifice meter relies on a range of coefficients or factors to calculate the flow rate from the differential pressure measured across the orifice. Many of these factors are derived from on-site measurements of the gas. The measurement accuracy can be improved by ensuring that the orifice plate has been sized correctly for the flow so that it falls within 30 to 70% of full range (or higher, if there is no chance of going off scale). Likely additional sources of error come from what could be considered mechanical factors, such as the physical condition of the orifice plate itself; waxy deposits or damage will change the flow performance and can lead to significant errors. An obvious but commonly overlooked potential error concerns not the condition of the orifice, but the recording of the orifice size. An oversight here can have serious implications not only for fluid analysis but for well-test interpretation. Errors in the differential-pressure measurement derive primarily from poor calibration of the recording instrument or from liquid buildup in the lines that have not been purged. The orifice meter pressure-base factor is a common source of errors because variations do exist between the reference pressure and temperature used for gas measurements (e.g., a variation between 100 kPa and 14.7 psia is an increase of 1.4%). Thus, it is essential that reference conditions are quoted correctly. The actual source of gas gravity and supercompressibility factors (F_g and F_pv) is usually not important for fluid studies because accurate values are commonly calculated in the laboratory on the basis of compositional analysis of the gas sample, but it is necessary to know exactly which values were used to correct gas flow rate to the new values. To ensure the highest accuracy in gas flow-rate measurement, a check should be made that all the meter factors are determined and used correctly. Approximate flow rates can be derived from the choke setting, and a comparison should identify any major error in the orifice meter calculation.

Condensate or oil flow rates are normally measured by a positive displacement meter that is placed in the outlet line from the separator upstream of the flow control valve. The most common error derives from incomplete reporting of the measurement conditions for the oil rate, especially whether the oil flow is measured at separator or at stock-tank conditions, and the meter factors and shrinkage values that should be applied if stock-tank rates are reported. The most likely causes of error in the measurement itself are poor calibration, worn seals (allowing liquid to bypass the measuring element), or the release of gas leading to high-volume, two-phase measurement. The latter problem can be treated by the installation of a "gas eliminator," which is effectively a tiny separator before the meter. Gas breakout in the meter may be signaled by sudden flow-rate fluctuations, whereas stable foams with some oils (occasionally referred to as "carry-under") may be less obvious and may require antifoaming agents to overcome. Any water and sediment in the oil flow should be determined by the BS&W measurement and corrected for accordingly. It is good practice to size the flowmeter according to the expected flow rate, as recommended for gas flows. Flow rates also should be checked by gauging the stock tank regularly.

BS&W measurement is performed by centrifuging a sample of liquid mixed with solvent; although relative error in the measurement can be very important at low BS&W, measurement accuracy is generally adequate for the purposes of flow-rate correction. Of more concern is whether the sample used for the measurement is representative, so samples should be taken from the top and bottom of the liquid flowline, and a comparison should be made.

The shrinkage factor, used to relate separator-liquid volumes to stock-tank conditions, depends on a differential liberation of gas and may give different values from the true flash process as separator liquid enters the tank stage. In normal circumstances, it is thus much better to rely on a separator flow rate measured with a calibrated meter than to use the tank flow rate corrected according to the shrinkage tester. In the worst case, with no reliable liquid flow rates at separator conditions, an experimental shrinkage factor must be determined in the laboratory and used with the average tank flow rate to obtain the necessary rate.

Further details of proper oil- and gas-measurement practices are available elsewhere in this Handbook and in other sources.^[1]^[2] Table 4.5 provides a checklist that can ensure that surface-measurement data are as reliable as possible. Other surface measurements should be validated in similar fashion; for example, wellhead pressures should be measured with a dead-weight tester or with a pressure gauge that has been calibrated recently.

Table 4.5

Among the downhole measurements, it is the reservoir temperature that is the most important for fluid studies because this is the temperature at which reservoir-fluid-property measurements will be made. In addition, pressures, gradients (density, pressure, and temperature), and, indeed, the depth at which these measurements are made are all important in validating samples and in interpreting laboratory measurements. Downhole temperature and pressure gauges should be calibrated, under well conditions if possible, and adequate time allowed for temperatures to stabilize if fluid production or injection has influenced downhole temperatures. Good knowledge of temperatures in a reservoir may only be available once measurements have been made in several wells.

The data listed in Tables 4.6 and 4.7 must be considered essential if fluid samples are to be studied properly in the laboratory. These data are the absolute minimum needed for valid laboratory studies. Every attempt should be made to provide all the information requested on sampling sheets. An independent check at the wellsite is advisable to ensure that sampling personnel have achieved this need. Many additional measurements are of value in sample validation, and measurement trends with time are important in monitoring well behavior (such as during cleanup or when evaluating the effect of changes in production). To enable a proper check to be made of well conditioning, separator stability, and data recorded on the sampling sheets, it is recommended that a full copy of the well-test report (or records of production data for production facilities) be sent to the laboratory that will be working on the samples.

Table 4.6
Vol1 Page 0206 Image 0001.png

Table 47

Water can be produced in a surface separator—either from liquid water in the zone being tested or by condensation from water vapor in the produced gas, or possibly from both—and can affect measurement accuracy. The effect of water on gas gravity (and, thus, the gas flow rate) is currently ignored because it is not routinely measured either in the field or in the laboratory. In most cases, this is an acceptable approach, but in separators operating at high temperatures and low pressures, the water content of the gas stream can reach significant proportions (for further details, refer to the nomogram^[15] "Water content of hydrocarbon gas"). Water production can have more serious consequences if separator-liquid flow rates are not properly corrected for BS&W measurements.

The data in Table 4.8 should be recorded for the sampling of water from a well. Similar data should be recorded for water samples taken from other installations or facilities.

Table 4.8

Quality Control of Samples

Selecting Samples for Study

This is an area in which there have been significant improvements in recent years, with significantly more details of quality-control tests being reported by laboratories, yet only limited information has been published on the aspect of fluid sampling.^[16] This section highlights the principal controls that should be performed and gives guidelines for selecting which samples are most likely to be representative. Newer concerns involve the quality of formation-test samples from wells drilled with OBM. Because of the wide range of fluids and sampling conditions, comparison of duplicate (or, more correctly, replicate) samples is generally the best method of evaluating whether the sample is representative.

The primary objectives must be selection of a fluid that is most representative of the reservoir fluid and identification of any serious quality problems related to the samples or the sampling data; these problems must be communicated to the client before proceeding with the fluid study.

Poor sample quality can arise from such sources as sampling nonrepresentative fluid, human error during sampling or field transfers, contaminated sample containers, and leaks during shipment.

For separator-gas samples, the quality checks that should be made when the sample bottles have been heated to, or slightly above, separator temperature are (1) determination of opening pressure; (2) compositional analysis, including air content; and (3) determination of residual liquids, possibly from carry-over.

Separator liquids transported with a gas cap must be homogenized by pressurization and agitation. In this instance, the controls that must be performed are (1) determination of initial opening pressure; (2) determination of bubblepoint pressure at ambient or (preferably) separator temperature; (3) a check for presence of sediments or an aqueous phase; and, when feasible, (4) flash separation to give GOR, shrinkage, gas gravity, or composition.

Downhole samples should be checked in the same way, except that bubblepoint pressure can be measured at either ambient or reservoir temperature. Measurement at reservoir temperature takes longer but is preferable for comparisons with downhole static or flowing pressures. Ambient bubblepoint-pressure estimates are often available from field transfers, but they should be used only as a guide because thorough mixing of the sample may not be achieved during recompression, and temperatures may be unstable. In downhole samples of a highly volatile oil or a gas condensate, no "break point" will be seen on the recompression curve, and a saturation pressure must be determined in a windowed PVT cell.

The following are brmeters that should be used, in order of preference, when a sample is selected on the basis of sample quality alone (i.e., when samples are essentially duplicates collected at the same time and under the same conditions): (1) an adequate sample volume or pressure; (2) a downhole sample bubblepoint pressure lower than downhole pressure during sampling; (3) contamination levels lower than, or similar to, duplicate samples; (4) bottle opening pressures that agree with sampling data (i.e., leaks are unlikely); (5) surface sample bubblepoint pressures that agree with separator data; (6) a close correlation between laboratory measurements on duplicate samples; and (7) one sample that represents "average" properties of duplicates.

For gas and oil samples collected from a separator, if at all possible, production test reports or other documentation should be studied in addition to the sampling sheets that normally accompany samples because a high proportion of sampling sheets contain inconsistencies. Data should be studied with the following objectives: (1) to identify what well or plant conditioning has been performed; (2) to look for the stability of gas and liquid rates when the surface samples were taken and, possibly, to calculate averages at the time the samples were taken; (3) to ensure that the GOR is based on oil flow rate at separator conditions; (4) to determine which gas gravity and nonideality ( Z ) factors were used, as well as the reference pressure and temperature; and (5) to verify reservoir temperature and static pressure.

If all samples meet the quality criteria, the choice can be made on the basis of field data alone, although the selection tends to be a compromise in some cases. Both operator and laboratory personnel must be involved in these choices. Primary emphasis should be given to (1) samples collected after proper well conditioning, (2) surface oil and gas samples taken simultaneously or close together, (3) a downhole sample that was collected above its bubblepoint that compares well with the bubblepoint pressure for duplicate samples, (4) a good downhole sample in preference to a recombined surface sample, and (5) a recombined surface sample if doubt exists about the quality of downhole samples. In cases in which downhole samples have been backed up by surface samples (an excellent practice), creation of a recombined surface sample from the best surface samples might be worthwhile, especially if there are only one or two downhole samples that appear to be valid. This allows comparison of the recombined surface sample with the downhole samples. In fact, in important wells, complete analyses of the two types of reservoir-fluid samples might be useful. Such an approach can give a high level of confidence in the data and could provide a crosscheck of separator GORs.

All validation data and analyses do not need to be included in the report when laboratory measurements are reported, but it is good practice to use a minimum of one page to explain sample selection and to detail any quality or field-data problems. This information can be of major value to engineers using the measured data.

Wells drilled with OBM are particularly problematic for formation-test-tool sample quality, and many research and development (R&D) centers worldwide are working on correction techniques. Figs. 4.10 and 4.11 demonstrate the sort of contamination that can occur; because the base oil is miscible with reservoir oil, it is impossible to remove this contamination from samples. It is essential to perform "fingerprint"-type gas chromatography (GC) analyses as a minimum quality control on formation-test samples to provide a qualitative indication of contamination, and even on production-test samples for which thorough cleanup may not have been achieved.

Fig. 4.10 – Chromatogram showing broad OBM peak.
Fig. 4.11 – Chromatogram showing narrow OBM peak.

Special correction techniques are increasingly used within the oil industry, and because these techniques vary between organizations and laboratories, sample selection should be done only after considering which method to use. Many companies are forced to use oil-based drilling muds to manage drilling costs in water-sensitive formations, and the added expense of handling contaminated samples (and the risk associated with poorer-quality data) must be used to evaluate the overall economic balance.

For water samples, comparisons of duplicates also give a good indication of quality. Where fluid concentration may be stabilizing (e.g., at the end of a cleanup), sequential samples should be used to look for compositional trends and thus to help decide if representative fluid has been sampled. For some sampling procedures involving trapping or precipitation of particular components, it is highly recommended to use blank "samples," which undergo exactly the same treatment and storage as the actual sample and provide a reference measurement to assist with the interpretation of laboratory measurements. More details are available in API RP 45.^[10]

Although this chapter concentrates on sampling rather than on analytical measurements, it is worth providing simple quality-control guidelines for GC here. This is because of the importance of GC analytical techniques both in the quality-control procedures described above and, increasingly, in on-site measurements and because simple guidelines of the sort given in Table 4.9 are not widely reported.

Table 4.9

Selecting Fluid Samples for Storage

Decisions concerning sample storage involve the following constraints: (1) discarding samples may prevent future measurements or checking of dubious results; (2) long-term storage of pressurized samples may incur very high rental costs; (3) sample-bottle purchase involves higher "short-term" cost (instead of long-term rental charges) but may sidestep the issue of deciding on a long-term storage policy; and (4) long-term storage requires a safe storage area and a catalog to be maintained (subcontracting is an option to be considered by producers).

One policy could be to keep duplicate samples for a short length of time on a rental basis and then to transfer minimum sample quantities into bottles purchased specifically for long-term storage. In all cases, it is advisable to budget for long-term storage within the project costs.

A good approach is for the laboratory report to recommend which samples should be kept or discarded on the basis of the quality checks and the study itself. An initial selection can be made at the end of the study, or even at the quality-control stage, if useless samples are identified. Then, the "customer" can respond on the basis of this information and the other needs of the project.

Further measurements may be warranted for a number of reasons: doubts about initial measurements; measurements required at different temperature or pressure conditions; advanced PVT measurements that are deemed necessary (interfacial tension, phase diagram, etc.); new analytical techniques that are developed before the reservoir is developed; or an asset purchase or joint venture that changes needs or requires independent measurements.

Because of the complexity of the heavy components in reservoir fluids, it is impossible to make an adequate synthetic mixture based on the liquid composition determined. It is thus good practice to store at least 1L of stock-tank liquid. This can be stored cheaply in a low-pressure closed container, but it also could be blended with a synthetic gas mixture (based on separator-gas and stock-tank gas analyses) to recreate a sample close to the original fluid composition, if further work was eventually required on the reservoir fluid itself.

A small volume (e.g., 100 cm³) of separator gas should be stored under pressure in case more-accurate or new analytical measurements are required on the gas later. For a downhole sample, a flash separation can be made at a convenient pressure to generate a suitable gas sample for storage.

Note that the quantities specified above are recommended in the case that no further laboratory measurements are anticipated. If further work appears likely, appropriate quantities of samples should be stored in addition to these minimum volumes.

This work recommends that a suitable sample-storage policy be identified and implemented in operating companies.

Hazards

No text covering fluid sampling would be complete without reference to the numerous hazards that must be considered when establishing safe working practices and sampling programs. Like many areas of the petroleum industry, if not managed properly, hazards can lead to equipment damage or loss, personal injury, and even death. Common hazards include the following (though this list is not exhaustive and may not cover unusual locations or special operating practices).

Hydrogen Sulfide (H₂S). This poisonous chemical is present in numerous hydrocarbon reservoirs and can be present both in gas streams and dissolved in hydrocarbon liquids. Although H₂S is recognizable by its smell at the low parts per million (ppm) level, above approximately 100 ppm the human nose becomes insensitive to the gas, and personnel could easily be exposed to lethal levels of H₂S (700 ppm can lead to instant death) if proper safety equipment is not in use. Safety measures should range from automatic alarm systems, personal monitors, and evacuation equipment to positive-pressure breathing systems, depending on the exact nature of the risk.

High Pressures. Fluid sampling frequently involves pressures up to 10,000 psi (700 bar), and even higher pressures are becoming increasingly common. Basic precautions should involve careful checking that equipment has a working pressure rating compatible with the maximum pressure that can be encountered at a sampling point (beware that flowing streams can produce a "hammer" effect when valves are closed suddenly), routine wearing of eye protection, and releasing of pressure before tightening leaking connections and attaching the ends of lines used to vent pressure.

Flammable Materials. Reservoir-fluid samples contain combustible hydrocarbons, so care must be taken to eliminate all sources of ignition from areas in which samples are collected or stored, especially where hydrocarbons are released during the purging of lines. Equipment must never be pressurized with oxygen or air (e.g., to clear blockages), as this can result in autoignition of heavy hydrocarbons (the "diesel" effect).

Solvents. Cleaning agents may contain dangerous compounds such as chlorinated solvents, and indeed, produced fluids may contain benzene. Breathing of vapors and skin contact with solvents should be avoided as much as possible. Solvents should be used as efficiently as possible, and all waste materials should be stored in closed containers before proper disposal.

Transport and Storage. Physical shocks are common during transport, so sample containers should be shipped with connecting ports plugged and exposed valves protected by endcaps. Liquid-filled containers are at risk of developing high pressures when heated; the best protection is to collect samples so that the liquid is in two-phase condition, with a gas cap representing approximately 10% of the capacity, or to use sample containers with a special separate gas cap. Rupture disks can be used to provide similar protection, but there is an increased risk of sample loss and venting of hazardous material.

Other Hazards. Examples of other dangers that sampling personnel must be aware of include offshore operations (special survival training is available), lack of oxygen in enclosed areas where large volumes of gases can be vented (notably nitrogen, but other gases such as hydrocarbons are an equal danger), and the toxicity of mercury, still used in some sampling operations.

Personnel who are not trained to work safely in the presence of these or other hazards must not undertake fluid-sampling activities. Assistance in properly managing all hazards should be obtained from qualified safety specialists.

Special Topics in Fluid Sampling

Sampling of Crude-Oil Emulsions

Samples of emulsions may be required for several reasons, including verifying crude specifications, evaluating the performance of emulsion-treating systems, or for laboratory testing such as choice of demulsifiers and optimum concentration. Emulsions frequently must be sampled under pressure, and special procedures must be used to obtain representative samples. For crude specification testing, it is not important to maintain the integrity of the water droplets; however, the sample location point may be critical. In general, samples should not be withdrawn from the bottom of the pipe or vessel, where free water may accumulate, affecting the BS&W reading and, thus, the validity of the sample. In addition, the sample should not be withdrawn from the top of the vessel or pipe, as it is likely to contain primarily oil. The best position in the pipe to take an emulsion sample is from the side at approximately midheight, preferably with a sampling probe (often known as a "quill"). Choosing a sampling point at which there is turbulence and high fluid velocity in the pipe may also avoid problems caused by segregation and ensure that the sample is homogeneous.

As for all sampling activities, every effort should be made to obtain representative samples. When sampling from pressurized lines and vessels, care should be taken to ensure that emulsification does not occur during the sampling process itself. For example, samples obtained at the wellhead or production headers may show a high percentage of emulsion (as a consequence of the sampling as the sample was depressurized into the sample container), whereas the actual oil and water inside the piping may or may not be in the form of an emulsion. Also, emulsions exhibit a wide range of stability, so samples of emulsions collected in the field may separate partially or even totally during shipment to the laboratory.

The best sampling procedure to use for samples from pressurized sources, without further emulsification of the liquids, is the technique based on a floating-piston cylinder, as described in Section 4.6.5. A setup similar to that in Fig. 4.6 is used, with the hydraulic section of the cylinder filled with a pressurizing fluid (e.g., a glycol/water mixture or a synthetic oil) and the top of the cylinder evacuated. Purging of the sample line can be made but should be carried out at a bleed rate. The sample collection should be performed slowly to obtain the sample with a minimum pressure drop between the cylinder and the sampling point. Alternative methods use a simple sample cylinder (without any floating piston), which is initially filled with water (or mercury). Once the pressurized sample is captured, the cylinder can be depressurized, if required, by removing further quantities of hydraulic fluid extremely slowly with little effect on the sample.

In situations in which sampling into a pressurized container is not possible, the best method to take an emulsion sample is to bleed the sample line very slowly into the sample container. The idea is to minimize shear and reduce emulsification that may be caused by the sampling procedures. Emulsions are covered in more detail elsewhere in this Handbook.

Waxy and Asphaltenic Fluids

Great care should be taken in sampling fluids that have potential for the precipitation of wax or asphaltenes because loss of a solid or flocculated phase during sampling or handling will produce fluids that are no longer representative. Because these tendencies may not be recognized until the samples are being studied, the same precautions are advisable for all fluids that have not been characterized previously in detail.

Asphaltene problems in particular are difficult to predict owing to a poor correlation between asphaltene concentration and flocculation tendency. Because of the difficulty of homogenizing fluids containing flocculated asphaltenes, it is highly recommended that downhole samples be collected using single-phase samplers with the pressure raised well above reservoir pressure by the nitrogen charge. Separator oils and atmospheric oils generally suffer from limited asphaltene deposition problems because they contain very few of the light hydrocarbons that contribute to flocculation.

Paraffinic or waxy crude oils can be extremely difficult to sample, with separator liquids occasionally solidifying in sampling lines and equipment, and the use of heated, short, large-diameter sampling lines is recommended. At downhole temperatures, sampling is generally easier, and limited availability of heated sampling tools exists. Single-phase sampling tools can be used, especially if asphaltene and wax problems may occur, but pressure maintenance alone will not prevent wax precipitation on cooling, as this is strongly dependent on temperature. Though rarer, gas/condensate reservoirs can also produce liquids that show wax-forming tendencies, which require special handling procedures.

Sample bottles containing movable mixing devices are recommended in all these cases, as returning samples to original sampling conditions and agitating for a lengthy period (e.g., overnight) gives the highest chance of recovering representative samples from samplers or sampling cylinders.

On-Site Measurements

It is worth giving some details of common on-site measurements because they must be performed on representative samples or sample streams and are frequently included in sampling programs.

The most common on-site gas analysis method is the "length-of-stain" detector tubes (often called "sniffer" tubes or Drager tubes, after one of the suppliers) used primarily for H₂S but available for CO₂ and a wide range of other gases and vapors. This method is relatively simple to use, and principal errors derive from incorrect use of response factors or stroke counts. The ASTM has a number of standards that apply to this method,^[17]^[18]^[19] and all propose a sampling system based on a modified polythene wash bottle (or equivalent setup) to ensure that measurements are performed on a representative sample stream. Flexible tubing is connected from a control valve at the sample point to the wash-bottle delivery tube, and the screw cap is removed (or perforated). Then, the control valve is opened slightly to allow gas to flow into the bottom of the wash bottle, which purges air out of the top. After purging for at least 3 minutes, the length-of-stain detector can be inserted through the top of the wash bottle and the pump operated to perform the measurement. This arrangement ensures that the gas sampled is at atmospheric pressure. Detector tubes have a limited life and should not be used beyond the date limit. Care must be taken to avoid contacting any liquids with the end of the tube. An alternative approach used a gas bag, which must be made from an inert material and purged completely at least five times immediately before the measurement. In some circumstances, gas concentrations above the detector-tube limit can be estimated by using fewer or fractional pump strokes, but such practices must be recorded clearly to help interpret measurements. For reactive species like H₂S, there is significant justification for making on-site concentration measurements by two independent techniques, as this provides on-site quality assurance.^[20]

Portable gas chromatographs are becoming more common at the wellsite, and they bring the advantage of early characterization of gas composition, together with an identification of most nonhydrocarbons present (depending on the carrier gas used). However, such instruments are accurate only when operated by trained personnel and when properly calibrated. Also, the additional cost of the service must be justified, though this could occur in the following cases: (1) high nitrogen or helium is anticipated, (2) early decisions must be made on the basis of gas sales value, (3) variable nonhydrocarbon concentrations occur within a field and will contribute to property mapping or be used to determine if fluid sampling is required, or (4) sample transport logistics mean that laboratory analyses may take a very long time.

The possibility of sulfate-reducing bacteria (SRB) contaminating completion fluids and even souring the reservoir itself represents a significant risk, and tests on site or sampling for SRB are recommended to identify and address potential problems. Tests that give a negative result can be particularly helpful in identifying the origin of SRB development later in the life of a field.

Many additional analytical measurements can now be performed on site because there is a wide variety of portable chemical test kits available, especially for water analysis. However, suitably trained operators are essential.

Drilling-Mud Gas

During drilling operations, returning drilling mud is commonly monitored for the presence of hydrocarbons, both for formation-evaluation purposes and for safety concerns. Generally, hydrocarbons are extracted from the mud to provide an air sample containing hydrocarbons, and as the extraction technique is dependent on equipment design and installation, measured compositions are rarely quantitatively representative of the concentrations in the drilling mud. This limitation is commonly accounted for by the use of hydrocarbon ratios when interpreting drilling-mud hydrocarbon analyses and logs, but it has been shown^[21] that effective quantitative measurements can be obtained with careful location of the mud-sampling point, the use of a special extraction device, and care to account for losses of hydrocarbon gas from the return line before the mud-sampling point. Mud logging is covered in detail elsewhere in this Handbook.

Water-Cut Measurements

Downhole sampling can be used as a means to measure water cut in producing oil wells, especially when there are no separation facilities or suitable measuring instruments at the wellsite or when measurements with depth are required as an alternative to (or validation for) production logs. In these cases, the well should be flowing at normal producing conditions, unless the purpose of the sampling is to investigate the water cut at other production conditions. As in all cases in which sampling is attempted from two-phase flow, there is the potential for preferential collection of one of the phases; this is especially likely if the two phases are not well distributed. It may be advisable to set the sampler to collect the sample in the shortest time to minimize any segregation. Care also should be taken in high-angle well sections, where the sampler will lie in the lowest part and is likely to sample water preferentially.

For a measurement of total water cut, it is advisable to sample from a depth a moderate distance [e.g., 20 ft (6 m)] above the top of the perforated interval. Also, it is good practice to take a number of separate samples under identical flow conditions to allow evaluation of the repeatability of water-cut measurements. Good agreement between replicate samples, although not absolute proof, gives high confidence in the reliability of the measurements, whereas significant variation is a clear sign of unrepresentative sampling or of variable water cut in the fluids entering the wellbore. Sampling from a range of depths both above and within the perforated interval in a well can provide more-detailed information on water production, but it does not give production-rate information on its own.

Volumetric measurement of water cut should ideally be made on site so that repeat measurements can be made as required. If emulsion is present in the sampled fluid, this should be given time to break, or suitable chemicals should be used to demulsify the sample before the definitive water-cut measurement.

Conclusions

Optimizing costs in all petroleum-industry activities continues to have a major effect on sampling operations, with competition between production testing and formation-test-tool operations leading to widespread industry acceptance of lower-quality fluid samples from the latter. A key challenge at present is to get better-quality formation-test-tool samples not simply by advanced tool capability but by better planning and preparation for the test. Increasing efforts to obtain reservoir information from short well cleanups is also putting pressure on fluid-sampling operations, but in some cases (such as on-site measurements), this change in emphasis may provide opportunities for measurements that otherwise would not be available.

Multiphase metering is likely to have an increasing impact on sampling operations because the accuracy of flow-rate measurement is seen to be improving, and the economic benefits of avoiding the use of separators will be major. However, even with significant development of special sampling approaches (such as isokinetic sampling), it seems unlikely that the same quality of samples will be available as with traditional separator methods.

Among the major developments in the past 10 to 15 years is the progress toward the worldwide elimination of the use of mercury in fluid-sampling operations, producing significant improvements in personnel safety and environmental protection. Efforts must continue to achieve better management of sampling programs and cost-efficient sample storage, despite the difficult challenge of trying to assign monetary values to fluid samples and the measurements made on them.

Specific technical developments that can be anticipated are a greater use of automatic surface-sampling systems, introduction of equipment better designed to preserve reactive samples such as fluids containing H₂S, an increase in gas/condensate downhole sampling and in use of downhole sampler technology (such as heated chambers for sampling waxy fluids), and downhole measurement of simplified composition and physical properties such as bubblepoint.

With the tremendous pace of the development and functionality of new downhole tools, there is a need for speedy updating of industry standards documentation to allow broad dissemination of new sampling knowledge and practices; however, it is clear that this is not a role readily filled by traditional standards organizations such as API. Working groups and forums involving service companies and operators may provide the best prospect of developing standards updates, which can be published in peer-reviewed petroleum engineering journals.

Nomenclature

F_g	=	gas gravity meter factor defined as
F_pv	=	gas supercompressibility meter factor defined as
G	=	gas gravity
P_sep	=	separator pressure
T_c	=	critical temperature
T_t	=	cricondentherm
Z	=	gas nonideality (or compressibility) factor

References

↑ ^1.0 ^1.1 ^1.2 ^1.3 _
↑ ^2.0 ^2.1 ^2.2 _
↑ _
↑ _
↑ _
↑ ^6.0 ^6.1 ^6.2 ^6.3 ^6.4 _
↑ ^7.0 ^7.1 _
↑ _
↑ ^9.0 ^9.1 ^9.2 _
↑ ^10.0 ^10.1 _
↑ _
↑ _
↑ _
↑ _
↑ _
↑ _
↑ _
↑ _
↑ _
↑ _
↑ _

SI Metric Conversion Factors

°API		141.5/(131.5 + °API)		=	g/cm³
bar	×	1.0*	E + 05	=	Pa
bbl	×	1.589 873	E – 01	=	m³
ft	×	3.048*	E – 01	=	m
ft³	×	2.831 685	E – 02	=	m³
°F		(°F – 32)/1.8		=	°C
in.³	×	1.638 706	E + 01	=	cm³
psi	×	6.894 757	E + 00	=	kPa

*

Conversion factor is exact.

[r1-1] 1.0 ^1.1 ^1.2 ^1.3 _

[r2-2] 2.0 ^2.1 ^2.2 _

[r3-3] _

[r4-4] _

[r5-5] _

[r6-6] 6.0 ^6.1 ^6.2 ^6.3 ^6.4 _

[r7-7] 7.0 ^7.1 _

[r8-8] _

[r9-9] 9.0 ^9.1 ^9.2 _

[r10-10] 10.0 ^10.1 _

[r11-11] _

[r12-12] _

[r13-13] _

[r14-14] _

[r15-15] _

[r16-16] _

[r17-17] _

[r18-18] _

[r19-19] _

[r20-20] _

[r21-21] _

[1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

PEH:Fluid Sampling

Contents

General Guidelines for Setting Up a Sampling Program

Reservoir-Fluid Type

Well Conditioning

Selection and Preparation of Sampling Equipment