ISO 3171 Petroleum liquids - Automatic pipeline sampling
4 Principles
4.1 Purpose
This clause defines the principles which it is essential to observe during sampling operations in order that the representativity of the sample taken corresponds to the specifications of this International Standard, and meets the acceptability criteria given in 4.4.

4.2 Principles to be observed
In order to determine the oil composition, quality and total water content of a batch of crude oil, samples that are representative of the batch are taken and analysed. The batch may be either a discrete pipeline transfer over a given period of time, or the whole or part of the cargo of a tanker, either loading or unloading.

Representativity depends on four conditions, all of which should be observed, since failure to comply with any one of them could affect the quality of the final result.

4.2.1 The first condition is that the samples that are taken from the pipeline should have the same composition as the average composition of the crude oil over the whole cross-section of the pipeline at the location and time of sampling. lt is not easy to comply with this condition, because of the possibility of a variable concentration gradient existing across the section.

This condition requires that at the sampling location:
a) the distribution or concentration of the water in the crude oil should be uniform across the section of the pipeline within the acceptance limits given in 4.4;
b) the diameter of the entry port of the sampling probe should be large in relation to the maximum water droplet size. The port opening should not be smaller than 6 mm (sec 7.3).

4.2.2 The second condition is that representativity should persist throughout the period of transfer of the batch, the composition of which may change between the start and finish of sampling. The rate of sampling, whether it is continuous or intermittent, should be in proportion to the flow rate in the pipe. When an intermittent sampler is used the sampling frequency and grab size should both be sufficient to guarantee acceptable representativity.

Furthermore, the representativity of the sample should be maintained in the automatic sampler from the sampling probe up to the final receiver. Samples should be taken with an appliance that camplies with the recommendations in clauses 7, 8, 9 and 10.

4.2.3 The third condition is that the sample should be maintained in the same condition as at the point of extraction, without loss from it of liquid, solids or gases and without contamination.

Storage and transfer of samples should comply with the recommendations in clause 11.

4.2.4 The fourth condition concerns division of a sample into a number of sub-samples in such a way as to ensure that each of them has exactly the same composition as the original sample.

The procedure for dividing each sample into sub-samples, and for transferring them to laboratory appliances, is given in clause 12.

NOTE - lt should be emphasized that this fourth condition concerns a critical activity and any error introduced is capable of destroying the representativity achieved by the first three.

4.3 Sampling tolerantes and validation
In order to ensure that each sample sent to the laboratory for analysis is representative of the whole batch, the composition of the sample should not differ from the composition of the batch by more than the tolerantes given in table 4 and as applied in 15.5.

In order to ensure that any departure from the conditions given above (see 4.2) does not result in a sample representativity which exceeds the tolerantes given in table 4, each step of the sampling operation should be validated as shown in figure 1.

4.4 General principles for sampling
Hydraulic laws governing the behaviour of heterogeneous liquids which will mix or will not mix in the pipe show that for stream conditioning a sufficiently high energy dissipation rate should be provided to keep drops of water and heavier solid particles suspended in the crude oil. Such an energy dissipation rate may be provided either from the velocity in an unobstructed pipe, or from a mixing device immediately upstream of the sampling location.

In considering the distribution of water over the cross-section, the acceptable limits of the values found in the profile test (clause 6) should be relative to the mean concentration of water in that plane and should be within +/- 0.05 g/100 g for samples having a water content up to 1 g/100 g and should be +/- 5 % of the mean concentration (relative) for samples having a water content greater than 1 g/100 g [but see also case 2 (4.4.2)].

NOTE - Although the concentrations above are quoted in mass units, they also apply to volume units.

In a horizontal pipe, three cases may be used to describe the ways in which the concentration of the different phases may vary over the cross-section of the pipe depending on hydraulic conditions (flow rate, product density and viscosity, dispersed Phase composition, interfacial tension modifiers, etc.).

4.4.1 Case 1 (see figure 2, profile type 1)
In this case, the concentration is the same, within the acceptable limits as defined above, across the entire cross-section of the Pipeline for all concentrations of water. The existing conditions are acceptable for sampling, since water is evenly distributed over the pipe cross-section. A representative sample consequently exists at the inlet of the sampling probe, which tan be placed at any point on the diameter although it is essential that care should be taken not to place the sampling probe too near the wall in order to minimize wall effects.

4.4.2 Case 2 (see figure 2, profile type 2)
In this case, the concentration varies from one point to another in the cross-section, but with a uniform gradient such that there is at least one point where the concentration is equal to the mean concentration across the entire section. Based on a theoretical analysis, this point is generally found between 0.4 and 0.5 of the vertical diameter from the bottom of a horizontal pipeline.

Sampling at this location will only be acceptable if the water concentration found at the sampling puint under wurst-case conditions is equal to the mean concentration within the acceptable limits as defined above.

4.4.3 Case 3 (see figure 2, profile types 3a and 3b)
In this case, the concentration across the cross-section of the pipe is non-linear, indicating the presence of segregation (type 3a). Some erratic profiles may also be encountered (type 3b).

As there is a range of concentrations at different points in the cross-section, sampling at a predetermined point is unacceptable, and it will be necessary to install a mixing device (see 5.3).

NOTE - If there is any free water, or an emulsion having a high concentration of water, at the bottom of the pipe, representative sampling is not possible.

4.5 Dispersed Phase - variations with time
lt is unlikely that the concentration of a dispersed phase component in the bulk liquid will be constant with time. For example in the discharge of crude oil from a marine tanker, in addition to more gradual changes in the base water content, there may also be periods when peaks of relatively high concentrations of water travel down the pipeline. Experimental observations indicate that these "transients" may contain over 50 % water and may be shorter than a minute in duration. Depending on the unloading procedures, the significance of the water discharged in the form of transients may vary relative to the base level carried with the bulk of the discharged cargo.

lt is apparent that the representativity of samples taken in such applications will be dependent upon the ability of the automatic sampler system to reflect, both accurately and proportionally, the integrated water content of these peaks in the total collected sample volume.

With intermittent-type samplers, accuracy will depend on the type of equipment and its frequency of operation in relation to the frequency and duration of the transients. With continuous-type samplers, accuracy will depend on the external collection and mixing arrangements and on the rate of secondary sampling if applied. For both types of sampler, the overall duration of the oil transfer, the duration of any water transients and the sampling frequency have statistical significance in the determination of sampling accuracy.

Theoretical analysis of the effect of dispersed phase transients on the performance of the different types of samplers leads to the following general conclusions:
a) in short-term transfers in which there is a possibility of frequent, short-duration transients, the accuracy of the continuous type of automatic sampler is least affected by the transients;
b) in short-term transfers in which there may be relatively few long-duration transients, the accuracy of the intermittent-type sampler approaches that of the continuous type of automatic sampler;
c) for long-term transfers, the mean error introduced by transients of any duration, using either type of sampler, is within the limits of acceptability defined in this International Standard.

4.6 Low water content
Attention is drawn to the fact that, if the concentration of water is about 0.1 % (m/m), i.e. near the level at which water is soluble in crude oil, the concentration profiles will show a good uniformity of water distribution under all hydraulic conditions.