A123 – Petroleum-Testing Guidance/Interpretation
Revision 1.1 – February 2009
A121 – Petroleum-Testing Guidance/Interpretation
Introduction
This guidance/interpretation document is for use in conjunction with A120 – Petroleum-
Testing Checklist, when assessing methods specifically designed for testing in the petroleum
industry. Where there are differences between the checklist, this guidance/interpretation
document, the reference method, and ISO/IEC 17025, the requirements in the reference
method and ISO/IEC 17025 shall prevail.
The item numbers noted below correspond to the item numbers in A120 - Petroleum-Testing
Checklist.
Guidance/Interpretation
01.01 (Authorized test method)
Some common petroleum-testing references include, but are not limited to,
ASTM, GPA, and UOP.
02.01 (Method validation)
ASTM methods normally state expected method precision on two bases:
“Reproducibility” and “repeatability”. Reproducibility is the expected inter-
laboratory variance.
ASTM methods include precision statements to provide a guideline on the
variability expected between results obtained under prescribed conditions
and assess, in general terms, its validity for specific applications. The
repeatability (within lab variability) should normally be the basis for the lab
repeatability in method validation/verification. However, these performance
levels may not always be achievable unless using the same materials (e.g.
performance levels determined from refined products may not be achievable
for crude oils. If is not necessary that the lab achieve the repeatability
performance specified in the reference, but it should use this as a guideline for
expected variability.
Reproducibility (between lab variability) is probably useful in test method
selection and service validation.
02.01 (Measurement Uncertainty)
P19 (CALA Policy on the Estimation of Uncertainty of Measurement in
Environmental Testing) covers estimation of MU for most petroleum-testing
Rev 1.1 Page 1 of 6
A121 – Petroleum-Testing Guidance/Interpretation
methods. For methods used in petroleum-testing that are more or less
empirical, PT variability provide a reliable, comprehensive, quick and easy
estimate.
Testing for compliance to a specification is common for many petroleum-
testing methods. Gases and liquids transported in pipelines must meet
certain specifications (this may be done to protect the pipeline or other
vessels from corrosion etc.) There may be any number of specifications
around custody transfer points. These are usually contractual and can vary
from simple heating value / moisture content to more exotic concerns such as
specific contaminants such as PCBs, metals, benzene. Per the standard, this
has implications related to requirements for measurement uncertainty (client
knowledge related to uncertainty is, as with most industries, limited). The lab
can, as service, have policy and procedure related to notifying clients of
nonconformance to product specification and using specifications for data
verification / validation.
03.01 (Test method details)
Sample pretreatment is important for many petroleum methods (e.g. cleaning
oil, single phasing HP liquids). Most tests on crude oil are performed on oil
cleaned by centrifugation and are reported on an “after cleaning” basis.
Pressurized hydrocarbon samples may be anything from light gas mixtures
consisting mainly of methane to gasified crude oils with everything from C1 to
C30+ hydrocarbons. The latter are problematic to analyze since the
components represent a very wide range of boiling points. Labs may have
various approaches including separating the sample into gas and liquid
phases stable at ambient conditions, measuring each phase for composition
and volume or mass, then recombining these mathematically back into the
original mixture.
Identification and measurement of individual hydrocarbon components is
usually impractical in chromatographic analysis of hydrocarbon mixtures.
Analysis usually involves grouping of peaks by retention time that
approximates boiling point, which approximates carbon number.
Measurement is usually made on a mass basis but reported on mass, volume
and mole bases. Conversion requires molecular weight and density that, in
principle, can be estimated from the chromatographic analysis. However,
because of the approximations indicated in the previous paragraph, and the
error in assigning properties to a group based on carbon / boiling point these
estimates can seriously misrepresent the “true” result. These parameters can
be measured, but normally only on the stabilized phase. It is important that
the lab use a valid approach to calculation and that the calculations be
Rev 1.1 Page 2 of 6
A121 – Petroleum-Testing Guidance/Interpretation
compatible with the bases of all measurements made.
Data reduction is critical in some cases. The most important results for
valuation of natural gas (density and heating value) are not measured directly,
but are calculated from composition. If the scope is to include these, then the
data reduction must be detailed.
Quite a bit of data validation normally goes on with pressurized samples for
hydrocarbon composition (e.g. comparison to pipeline specs, equilibrium k-
plots, calculated vs. measured properties such as density of gas, comparison
to historical data, methane vs. pressure on liquids).
Normally for production testing:
Equilibrium k-plots (or equivalent) should be used i) to verify phase
separation (if this is done) and ii) to verify sample integrity from associated
gases and liquids (e.g. from an inlet separator). K-plots are thermodynamic
diagrams based on the Clausius-Clapeyron equation indicate whether the gas
and liquid are at equilibrium and can verify if results are mutually consistent.
It is recommended that historical data should be used where available and
where applicable.
The lab should have protocols for comparing related parameters for
compatibility / validity / expected behaviour. These can be such things as:
o A reasonably smooth variation of concentration vs. carbon number
in natural materials.
o Historical consistency (where applicable)
o Geological and geographical consistency (fluids from the same
geological pool in the same area can be expected to be broadly
consistent).
o Measured properties vs. the same property calculated from
composition (e.g., gas density).
04.02 (Sampling and sub-sampling)
The sampling process is critical for pressurized samples. There are reference
methods for gas sampling (GPA has this, so does ISO). They are generally
applicable but may not be applicable in all situations depending on the source
composition, source conditions, ambient conditions etc. Sampling procedures
that might work for lean gases may bias rich gases.
Rev 1.1 Page 3 of 6
A121 – Petroleum-Testing Guidance/Interpretation
If accreditation is going to cover sample collection, the lab should be
following good practices and have validated their sampling protocols for the
types of samples collected. Samplers would have to be appropriately trained,
be able to recognize potential problems, etc.
Hydrocarbons are generally sampled based on availability at agreed upon
points using techniques based on sound physical principles and good
practice. Usually samples will be collected from a single specific point along a
flowing stream, variation is normally with time. “Statistical” sampling plans
would therefore require a model of variation in the source over time and flow.
“Spot” samples are collected at a single point in time and may be taken to
represent the source over an assumed interval (this is quite reasonable if the
stream is consistent). “Proportional” samples use special sampling equipment
to collect sample in proportion with time or with flow rate. This technique
more costly, labour intensive and complex, but is more representative than
spot samples or statistical techniques. Either way, this is normally based on
agreement between lab and client; clients normally manage their own
proportional sampling.
Generating representative samples is not a problem if dealing with a stable
fluid. With high-pressure liquids, phase changes can occur which can result in
bias if not controlled.
The validity of sampling and sub-sampling methods can be critical if dealing
with multiphase samples. Sample integrity must especially be evaluated when
the samples are gas and high-pressure liquids because opening pressures, air
contents, bubble points etc. establish whether the sample can be
representative of the source.
The lab should be checking opening pressures / temperature of gas and liquid
samples and should have nonconformance criteria and actions to taken for
nonconformance based on source temperature and pressure.
The lab should measure O2 content of gas samples. O2 does not occur in
natural gas and is normally assumed an artifact of sampling (due to air
contamination). The lab should have a protocol for dealing with O2.
Normally, small quantities of air can be corrected out, O2 and an amount of N2
in proportion to that found in air, is taken from the result and the remainder
renormalized - air free basis. Larger amounts of air indicate a seriously
compromised sample and should normally rejected or – at minimum – a note
added on the report.
Care must be taken when measuring and correcting for O2 as Ar may co-elute.
There are ways of dealing with this (Ar carrier, cryo temperatures, suitable
Rev 1.1 Page 4 of 6
A121 – Petroleum-Testing Guidance/Interpretation
accounting in calculation…) the lab should have the means clearly defined.
Bubble points (essentially measuring the conditions required to return a
sample to single – liquid – phase) are useful as well in determining sample
integrity.
Many sample containers used in the oil and gas sector are NOT disposable
and get cleaned and reused. Therefore, the lab must have a procedure for
this step.
Auto samplers in use must meet requirements of ASTM D4177, or the
frequency and volume of manual samples is varied to meet ASTM 4057
requirements to avoid product representation issues.
05.01 (Calibration)
In many situations, calibration is more that of a piece of measurement
equipment than the method. This may be a basic physical measurement such
as temperature or pressure, or equipment such as viscometers, which may
require external/manufacturer calibration.
Calibration for hydrocarbon composition by GC may involve determining
relative response factors (RRF). A multilevel calibration may be redundant in
situations where RRF’s are used. RRF’s may even be stipulated in reference
methods.
09.01 (Equipment)
Of relevance beyond device calibration is the device itself. Some of these
methods are empirical. Most (e.g. ASTM) are quite detailed in measurement
equipment requirements and configuration. If there are any modifications to
equipment, these need to be both documented (clearly stated in the lab SOP)
and verified that the changes are within the allowance specified in the
referenced method.
10.01 (Supplies)
The containers themselves should be stainless steel and capable of
withstanding the sample pressure. They should also be suitable to the
sampling technique (certain liquid displacement sampling methods require
special equipment configuration). Carbon steel should not be used, it absorbs
CO2 and it corrodes so the integrity of the vessel can fail. Even stainless steel
is inappropriate for sampling gases for trace sulfur species as these may be
absorbed. Tedlar sample bags are typically used for reduced sulfur species.
Rev 1.1 Page 5 of 6
A121 – Petroleum-Testing Guidance/Interpretation
One the other hand, samples in Tedlar are generally not suitable for
hydrocarbon composition analysis due to permeability to some constituents,
and other effects that may result in bias.
11.01 (Record-keeping)
When measuring gases, it is important that conditions (temperature &
pressure) be appropriately measured and compensated.
Volume correction to reference conditions can be critical, especially for gases.
If necessary, the lab should have an appropriate means of measuring Boiling
Point.
With gases, it is critical to report the reference conditions and the basis of the
measurement. There are at least a dozen different sets of “standard
conditions”. They should be stated directly and unambiguously (The
Electricity and Gas Inspection Act of Canada specify 101.325 kPa and 15 °C
which is slightly different from the conditions used by some US bodies). Use
of “STP” or “NTP” is not acceptable.
Although gases are normally measured “as received”, properties can be
reported on different bases such as “moisture free” or “moisture and acid gas
free”. Both are valid, but these bases must be clearly indicated.
There are a few tests that may only be valid if done directly at the source (e.g.
dew points by the chilled mirror method ASTM D1142).
Rev 1.1 Page 6 of 6
