CAC – Reducing Errors & Uncertainties In Gas Quality Measurement

Natural gas and LNG are traded on total energy delivered. Energy delivered is the product of volume flow and gas quality. It follows that errors and uncertainties in the measurement of gas quality or flow will affect the total energy calculation. Due to the large volumes of natural gas flowing across fiscal custody transfer* points even a very small error in either measurement can lead to large financial discrepancies.

Natural Gas is made up of several component gases and is therefore subject to natural variation. This inconsistency affects the energy contained within a given volume of gas. Gas quality is most commonly described based on the measurements taken of the heating value, also known as calorific value, Wobbe Index, and relative density and others.

*A custody transfer point does not necessarily imply a change of ownership. It may be that at this point a measurement of the natural gas is taken to ensure that the contractual obligation between buyer and seller is being met. The obligation may require adherence to accuracy, linearity, repeatability or uncertainty standards as defined by the measurement standards they have agreed to operate under.

How do we measure gas quality?

Gas Quality Measurement is most commonly measured with the use of a gas chromatograph. A gas chromatograph is used to separate the components of a natural gas so that each major component can be quantified. The internal process consists of subsystems that inject the sample, separate the sample, detect the components and report the results.

In both LNG and natural gas measurement, the GC is essentially a cash register. The data obtained from the instrument along with flow or volume data will determine the financial value of the product. Considering the vast financial sums changing hands in these transactions, it is extremely important to ensure the most accurate measurement is made.

Where are error and uncertainty introduced?

• Sampling
• Conditioning
• Calibration Gas
• Analysis
• Staff
• Quality Control
• Maintenance

Sampling and Conditioning (natural gas)

When sampling natural gas it is imperative that the sample collected is the most complete representation of the gas as possible. The best method involves the use of a probe and certain considerations should be taken in regards to the probe assembly.

Installation of the probe is recommended on the top of the pipe with the probe extending into the middle 1/3 of the pipe from which the sample is being gathered. Probes should also be placed in an area of the process pipe in which turbulence is at a minimum, as turbulence can cause contaminants at the bottom of the pipe to be stirred up and included in the sample.

Turbulence is caused by bends, valves, headers and any other items that restrict, change, or impact the “normal” flow of gas. ISO10715 indicates that a sampling probe should be more than 20 diameters in distance from any flow measurement.

A conditioning system is then used to remove water, particulates and in some cases liquid hydrocarbons. It’s designed to prevent pressure fluctuation, flow fluctuations and retrograde hydrocarbon condensation.

Issues with sampling LNG

ISO8943 describes procedures for sampling of LNG by vaporising it to make it amenable to GC analysis.

LNG vaporisers are required to achieve a phase transition; (liquid to gas) a process which can be challenging. Pre-vapourisation and enrichment occur when the LNG sample is not completely vapourised in a single step. Partial vapourisation leads to the most volatile components vapourising first, leaving an enriched sample in the liquid phase. Both phases are not representative of the true sample composition.

Only by vapourising the sample in a single step will the true composition be achieved which then can be analysed.

Instrumentation and Calibration Gas

In its simplest form, the GC is simply a comparator. Similar to a high quality wrist watch, there is no point in it being precise if it is set to the wrong time!

For fiscal measurement of natural gas or LNG, the GC needs to be both precise and accurate, but the accuracy is primarily derived from the reference material (calibration gas) only. The calibration gas is the principle driver of the accuracy of the instrument.

When discussing the quality of a calibration gas, the relationship between accuracy, precision and tolerance is often confused.

In general terms, accuracy refers to how close the measured value is to a known value. Precision refers to how close two or more measurements are to each other. Precision is independent of accuracy. You can be very precise but inaccurate and you can also be very accurate but imprecise.

When talking about the accuracy of components within a calibration gas standard, the term tolerance is often used.

Tolerance is a term used to describe a manufacturing method’s ability to meet the end user’s requested nominal value for each component. The manufacturer will aim for the requested value, and the tolerance will determine the range into which this value will fall.

If it is a good quality gas standard, each component will then be analysed with a GC to verify this value.  Inaccuracies in the calibration gas will produce biased stream gas measurements. The uncertainties of both reference material and instrument will compound and can affect the results even further.

What can go wrong with calibration gas?

Low temperatures (transport, storage, usage)

A calibration gas standard is a homogeneous mixture of components that have the same proportions of its components throughout a given sample. A good analogy is like a cup of tea. If you were to add sugar to a cup of tea and not stir it, the sugar would sink to the bottom. As you drink the tea, it would be bitter at the beginning and get continually sweeter as you get to the bottom. In this case, the cup of tea would not be a homogenous mixture.

The same situation can occur with a calibration gas mixture. If not mixed correctly, the lighter components of a natural gas mixture (Methane, Ethane etc.) would be at the top and exit first when drawing a sample and heavier components would stay inside therefore not providing a clear representation of the sample.

When a calibration gas is prepared, it is rolled for a period of time after mixing to ensure it is homogenous. But these proportions can be altered when exposed to low temperatures during transport and storage.

If a hydrocarbon component calibration gas is exposed to temperatures below the minimum usage temperature. The heavier hydrocarbons can begin to move into the liquid phase, leaving the gas mixture as a two phase, non-homogenous gas standard. If used for calibration the instrument will be biased for higher hydrocarbons and a higher heating value.

Large or no uncertainties

It is a common misconception that measurement is an exact science. All measurements are merely estimates of the true value being measured and the true value can never be known. No matter how careful or accurate, every measurement result contains an independent amount of uncertainty.

Therefore, if measurement is important, then measurement uncertainty is equally important.

If the calibration gas has no uncertainties at all or an even 1% across all components (scientifically very unlikely) then accurate calculation of this uncertainty is impossible.

Methane reported as ‘Balance’

Considering natural gas is made up mostly of methane, it is the most vital component in terms of calculating the heating value, and thus the financial value.

A good quality calibration gas manufacturer will be aware that the reference value and uncertainty for methane is necessary to determine the uncertainty of the overall gas quality measurement.

How can we reduce these issues?

Inspections

Regular inspections will highlight the problems with the measurement process. Gas quality metering inspections provide independent assessment of your system, review of system versus industry best practice and confidence in metering system and data.

Analyser Performance

Evaluation of the analyser according to ISO 10723 will prove or disprove the analysers fitness for purpose for the measurement requirement (e.g. fiscal, custody transfer or emissions trading)

Analyser bias and uncertainty

The ISO 10723 performance evaluation will set a benchmark in time for the bias and uncertainty. Regular evaluations will identify deterioration before unacceptable bias and uncertainty are reached.

Analysis function parameters to reduce bias (errors)

Analysis function parameters are calculated during the data processing following an ISO10723 evaluation. ISO6974:2012 refers to these instruments using analysis function parameters operating in type 1 mode. Assuming the instrument has the capabilities to operate in type 1 mode, these parameters can be programmed into the instrument to reduce analytical bias. Type 2 instruments, as defined by ISO6974, operate with a single point calibration and are subject to ISO10723 evaluation to establish an inherent non linearity in the detector response for each of the components.

The diagram below shows a typical performance evaluation dataset containing 30,000 simulated gas compositions. The red markers indicate the spread of results assuming the instrument is operated in a type 2 mode and has a linear response function for all components. The blue markers show the same instrument operated in a type 1 mode, demonstrating significant less spread of results and almost zero error.

Data, Quality, Staff and Maintenance

Set alarm limits

It is important not to rely on un-normalised total alarms. Instead set low level limits (<50ppm) as this will ensure invalid data is highlighted.

Quality control measures

A Shewhart control chart can monitor GC performance between evaluations and highlight problems before failure occurs.

Staff and Procedures

Formally documented site procedures will ensure tasks are performed correctly and to best practice. Adequately trained staff can deal with a wider range of problems and know when to call in the vendors.

---