ISO 3171 Production Hydrocarbons Allocation Sampling for challenging “Tie-ins”
with pressures close to RVP breakout.
By Mark Jiskoot, Managing Director, Jiskoot Ltd.
Summary
Introduction
There are an increasing number of applications where sampling is required for crude oils
and condensates at close to vapour breakout in production environments.
These are typified where the quality measurement must be extracted/made in the “oil” leg
of a separator. The exertion of back pressure (pressure loss) on the liquid leg to expand
the operating envelope is generally not acceptable due to the effect that this will have on
the separator efficiency and this frequently precludes the use of normal metering
technologies.
These conditions represent some of the hardest for quality determination.
In conventional separators, velocities are low to promote separation of the gas and liquids
and the therefore the gas pressure within the liquid phase is close to breakout. Frequently
there is an envelope of less than 0.5 bar below the operating pressure at which the fluid
will start to cavitate. Concurrent with the low velocity is the expectation of some free
water that would require mixing to ensure representivity.
There are frequently also severe restrictions on space and the piping arrangement in
which to achieve a representative off-take. The “rules” that are applied to representative
sampling are equally applicable to the use of water monitors (watercut, OWD) and to
densitometers.
An integrated quality measurement system that prevents any impact on the process and
also provides the required mixing to allow extraction of a representative offtake that can
be used for physical samplers as well as process instrumentation such as densitometers,
viscometers and water cut monitors can be achieved by careful design. Jiskoot has
adapted its CoJetix technology to meet these goals.
Sampling Requirements
Conventional sampling requires that four steps be met :
! Adequate dispersion and distribution of the pipeline content cross section
! Representative sampling of the batch
! Sampling handling and mixing
! Laboratory analysis
Our industry should never overlook any of the above steps, whether the objective be
extracting a physical sample or to provide a representative flow to a process analyzer; but
we are frequently surprised by how often the overall requirements are overlooked.
Page 1
199
Dispersion and Distribution
A cursory reading of the current standards provides a table informing the user of the
potential problems with poor mixing at low velocities. Unfortunately many miss the
small print and assume that if they meet the requirements in the “table”, they have
adequate mixing. Those that do this fail to take the opportunity to assess their process
against a much more detailed and accurate methodology which also accounts for
viscosity and density. For close on 20 years models representing the dispersion and
distribution of oil/water mixes have been published both in the ISO (3171 – Annex A),
API (8.2 – Appendix B) and the IP (6.2 – Appendix B)standards and its all the more
surprising that they are not used as these calculations are easily accessible on the internet.
The models allow the calculation of dispersion quality in many configurations including
downstream of a variety of devices including meters, valves as well as the more common
elbows, reducers, pumps etc.
There are many specialised situations where a deeper understanding is required, for
example :
It is not commonly understood that although a vertical pipe reduces the gravitational
segregation of oil/water mixes, over a short length it also promotes slug formation and
can spin the water towards the circumference of the pipe, i.e. make distribution worse.
Or, for example, that the use of a metering system as a mixing device may not always be
successful due to the possibility that the inlet manifold can act as a slug creator (if there is
a free water phase at inlet). Under such circumstances the mixing of water into oil in
each stream may be acceptable, but the commingling of the streams in the outlet header
can yield a poor cross sectional distribution. Given that these methods are available, why
do some process engineers still overlook the actual process and make subjective
judgments that should fail an easily applied audit of the process? Both engineers and
auditors are at fault here.
Representative Sampling of the Batch
This is all about getting a sample to be physically extracted from the pipeline and
deposited into a sample receiver that fully represents the batch in question. It remains a
constant surprise that many engineers still fail to read the standards in regard to the actual
requirements for sample frequency and many appeared blinkered by the de-facto marine
discharge requirement of 10,000 samples per batch. Collecting 10,000x1ml samples of a
high RVP product and either transporting it under pressure, or mixing and decanting it
into a smaller pressurised receiver for transportation to a laboratory is frequently at a
detriment to the overall measurement uncertainty. Poor handling and mixing can often
result. It should be under this heading that we discuss the sampling mechanism and
“Isokinetic” as applied to both in-line and fast loop sampling systems. Isokinetic -
Implicitly an “in-line” sample probe with a large opening through the head may have a
velocity closely matching that of the process through the sample chamber. (Though this
may not be true when the opening size is relatively small in comparison to the bluff body
of the stem that supports it!)
Page 2
200
However over time two clarifications of the desirability of Isokinetic sampling must be
qualified:
1. As the size of the opening to the sampling device increases relative to the
expected droplet size, then Isokinetic sampling becomes irrelevant. This
was first highlighted in qualification notes from testing within the IP
Section 6 part 2
2. If a “Fast Loop” is driven isokinetically i.e. matching the velocity at the
inlet to the velocity in the main pipeline, then there is a significant
problem at low flow velocities that water and sediment can fall out in the
sample loop this is in contradiction to the objective of the “Fast Loop”
which is to deliver a representative and well mixed stream to the sampler
or analyser.
Sound engineering design should recognise the use of large sampler or sample loop
openings and abandon the use or application of the work Isokinetic to oil sampling.
Sample Handling and Mixing
Consideration of the volume of sample that is actually required for analysis and retention
and the required sampling frequency is valuable. A 4 litre, high pressure, sample will
normally be more than adequate to represent a batch in a production environment.
We also often find demand for high pressure sample receivers to be used on stabilised
crude i.e. with an RVP below a bar, this is simply a waste of money.
Even with higher RVP products the sample used for water determination will need to be
depressurised for analysis.
Density and On-line water content measurement
We realize that there is growth in demand for using on-line water monitors particularly
for allocation (sometimes called Watercut or OWD’s – but NOT correctly described as
BS&W monitors), indeed there is a paper later this morning dedicated to this subject. But
without adequate pre-conditioning of the process stream (pipeline mixing) and suitable
and regular re-calibration these devices render little more than trending information.
Sampling technology continues to move from the old style and less accurate “in-line”
samplers to fast loop systems. By fast loop, we tend to classify them as a loop with a
piping size of 1” or larger nominal diameter and typically a loop flowrate of 4-10 m3/hr.
These rates and velocities concur nicely with the demands of OWD’s and densitometers.
There are many reasons for this slow migration, the general comprehension by engineers
that fast loop systems offer easier installation in a congested process environment and
simplified maintenance. Also since a loop is required for density measurement which,
according to the standards “should meet the same requirements as for sampling” and
more recently the body of evidence that fast loop technology is simply more
representative.
Page 3
201
It is possible to classify sampling systems into various categories and further to consider
the mixing. Since the 1980’s there have been significant water injection tests of all types
of systems. From the “DIY” samplers constructed by sites, to in-line systems with
“Static Mixers” or Power and Jet Mixers.
There have also been a considerable number of tests of fast loop systems with the same
types of preconditioning and more recently with the direct integration of fast loops with
JetMix systems so-called CoJetix.
One statement of fact :
NO sampling system, unless installed in a biased pipeline profile will measure the
absolute water content, ALL sampling systems tend to under measure water content, the
only question is by how much. Call this uncertainty or call this systematic bias, gathering
real data from over 200 water injection of various systems yields the following summary.
Water recovery from water injection testing of automatic sampling systems
0.20%
0.00%
(-ve number indicates undermeasurement of water
-0.20%
Absolute error in water recovery
-0.40%
-0.60%
-0.80%
CoJetix Sampling System
-1.00%
In-Line sampling system
with poor mixing
-1.20%
in-line & JetMix sampling
system
-1.40%
Number of water injection tests
The reason for these results are somewhat obvious:
The under measurement will depend in greatest part in the dispersion and distribution of
the water in the pipeline cross section. To minimise these influences a balance between
the mixture quality and the size of the opening and flow profile through sampling device
seeking to collect a sample are paramount. A fast loop is implicitly more competent than
an inline probe by virtue of its larger opening extracting water droplets from a pipeline of
equivalent dispersion quality.
Whatever you do for mixing, you MUST pay the price in energy terms, be it through
pumping power (pressure loss) in your pipeline or by adding energy such as with a
JetMix.
Back to back comparisons of an in-line (“Isokinetic”!) probe behind a static mixer and a
CoJetix system indicate a significant average offset (0.15%)
Page 4
202
6,000.00
CoJetix Sampler
Original In Line sampler (w/static mixer)
5,000.00
4,000.00
3,000.00
2,000.00
1,000.00
0.00
MAYA
BASRAH LT
Maya
Oriente
BASH LT
BASH LT
ARAB HVY
ARAB LT
BASRAH LT
BASRAH LT
ANS
ANS
ANS
ANS
ANS
ANS
ANS
ANS
ANS
ANS
Escalante
Escalante
Escalante
We have also been comparing the results of a variety of water cut monitors back to back
against a marine unloading environment, our conclusions are that provided the users
rigorously follow the calibration procedures and understands the limitation of the
technology, that they are capable of providing excellent trending information.
The summary below indicates the scatter between two different manufacturers units and a
proven (physical) sampling system. The two units are in the same loop with a sampler,
both density compensated – running batches of the same “blended” crude, sulphur is
plotted as a possible cause for the scatter.
ANS Error Relative to Lab Result
Sulphur
1.2000
1.0000
0.8000
0.6000
Water%
0.4000
0.2000
0.0000 ANS ANS ANS ANS ANS ANS ANS ANS ANS ANS ANS ANS ANS ANS ANS ANS ANS ANS ANS
6/10 6/11 6/13 6/21 6/23 6/26 6/29 6/30 7/7 7/13 7/14 7/17 7/30 8/02 8/07 8/12 8/16 8/19 8/24
-0.2000
-0.4000
Page 5
203
Design Study
Process Values
Liquid Process Data
Line Size 8” 8” 8”
Pipeline Fluid Condensate Condensate Oil
Flowrate Range 2696 to 12750 6785 to 84800 kg/h 8818 to 422700
bbls/d (10.6 to 122.1 m3/h) kg/h
(11.4 to 540.5 m3/h)
Density 645 to 654 kg/m3 643.1 to 694.1 kg/ 746.8 to 782 kg/ m3
m3
Viscosity 0.5112 cP 1.2 to 1.6 cP 2.0 to 3.75 cP
Line Temperature 55 to 75 °C 30 to 50 °C 42 to 67 °C
Operating Pressure 28 to 63 Barg 11 to 15 Barg 7 to 63 Barg
Water Content Unavailable 0 to 2% 0 to 2%
Pipeline Mixing
Given that the RVP is within 6psi (0.5 bar) of the process pressure any reduction in
process pressure will cause gas breakout.
Since the flow velocities are low, some form of mixing will be required as the separator
is expected to allow up to 2 % carryover of water in the oil leg.
The use of a static mixer in this environment would not have worked, (in one case a 50:1
turndown – we would typically suggest, even if pressure losses are acceptable that 5:1 is
the attainable limit). The maximum allowable backpressure was 0.5 bar at maximum
flowrate and this should have been the static mixer back pressure AT MINIMUM
FLOWRATE (and therefore significantly higher at higher flowrates.)
Some form of power (Jet) mixing was therefore desirable. In seeking to employ a pump
to extract flow from the pipeline, considerations of available pressure losses within the
suction pipeline and the NPSHa vs NPSHr (and specific engineering constraints applied
by the buyers) require due care in selection of and positioning of the pumps. If suction
head can be increased through elevation, this should be implemented.
A careful design requires consideration of a variety of issues, these include:
! Droplet dispersion and distribution vs. sample system inlet sizing.
! Configuration of the inlet and outlet connections, all the intermediate piping.
! Elevation.
! The obstruction potentially caused by the insertion of an offtake quill and a re-
injection jet.
Nominal calculation of the initial distribution yielded the following result.
Page 6
204
Pipeline Mixing Quality
1.00
0.90
0.80
0.70
0.60
C1/C2 Ratio
JetMix
0.50
Static Mixer
0.40
Straight Pipe
0.30
0.20
0.10
0.00
0 20 40 60 80 100 120 140
Flowrate (m3/hr)
The energy dissipation required for mixing to allow extraction of a sample with a
conventional in-line sample probe or fast loop provided an energy addition requirement
and pump sizing that was untenable, therefore the energy addition was curtailed in favor
of increasing the suction inlet profiling capability of the offtake quill and changing the
system into a CoJetix, this allows the mixing energy requirement and therefore demand
pumping rates to be reduced.
JetMix nozzle
A further constraint on the design is the compromise between volume and jet velocity
(pump outlet head requirements). Jet velocity in conjunction with the jet diameter
provides the means for both dispersion and distribution. Small diameter jets at high
velocity provide good local energy addition but limited distribution capability due to their
faster diffusion. Larger jets provide longer term bulk movement of the flow.
The pipeline section in to which the offtake quill and jet were inserted was increased to
ensure that the installation would not impact the overall system pressure losses through
obstruction.
The use of a CoJetix also allowed the implementation of the densitometer loop and the
use of a water cut monitor, samples are collected into a constant pressure receiver.
Page 7
205
Co-Jetix fiscal condensate sampling and analysis system
Page 8
206
Due to the high RVP of the product the system was designed to be able to use plain and
compact (constant pressure) sample collection receivers which can be mixed in a
laboratory. These can be mixed with a ShearMix (pressurised shearing vessel) into which
the sample is decanted for homogenization. The ShearMix can be used with almost any
vendors constant pressure cylinders and allows samples to be extracted from the top and
bottom of the vessel to ensure convergence in the analysis results.
CPC sample receiver ShearMix - CPC mixer
Water Separation Top to Bottom Vs Mixing Time
1.9000
1.8000
1.7000
1.6000
1.5000
1.4000
1.3000
1.2000
1.1000
% Water
1.0000
0.9000 Bottom Sample
0.8000
0.7000 Top Sample
0.6000
0.5000
0.4000 Absolute
0.3000 variance
0.2000 Absolute water
0.1000 content
0.0000
-0.1000
-0.2000 0 3 6 9 12 45
Mixing time (Mins)
A more recent application of this CoJetix style design is to a system where no sampling is
implemented but the problem of providing representivity to a watercut monitor remains,
under these circumstances the same approach is taken.
Page 9
207
Sampling system 3D design model
This style of design, whether it be for a physical extraction of sample, density
measurement OR water cut measurement singularly or concurrently can be used in almost
any application where measurement uncertainty is important to the overall process audit.
Your income stream in either joint venture environments or where production is taxed on
the basis of your measurement depends upon the best measurement achievable.
Page 10
208
References
[1] Paper presented at the North Sea Flow Measurement Workshop, a workshop
arranged by NFOGM & TUV-NEL
Note that this reference was not part of the original paper, but has been added
subsequently to make the paper searchable in Google Scholar.
