Unintended Introduction of Oxygen into Gas Export Systems as a Result of the Use of Flare Gas Recovery Systems

Summary
In the oil and gas industry, oxygen is not a naturally occurring constituent of the produced hydrocarbon as, if present in the original formation, it would have been consumed by oxidative reaction with the hydrocarbons during the millions of years of storage[1]. The presence of oxygen is undesirable as it promotes corrosion; for this reason, specification limits on gas in transport pipelines are often set at < 10 ppm. Whilst oxygen presence has major implications, it is not routinely monitored as it is not expected to be present; however, Qa3 has identified a source of oxygen from recent technology introduced to reduce emissions.

Background

There are a number of global initiatives aimed at reducing gas flaring which, for 2021, was estimated by the world bank[2] as being 144 billion cubic meters of gas across the globe. These initiatives include:
•   The Zero Routine Flaring (ZRF) initiative launched by Global Gas Flaring Reduction Partnership (GGFRP) that commits governments and oil companies to end routine flaring no later than 2030.
•   The Oil and Gas Climate Initiative (OGCI) commitment to finding viable solutions to eliminate routine flaring in existing fields by 2030 at the latest[3].
•   The UK’s commitment to net zero carbon emissions by 2050.
For compliance, modifications to the processes at most existing facilities will be required, for example, incorporation of gas reinjection back into the reservoir and / or the integration of Flare Gas Recovery (FGR) systems to prevent normal day to day flaring. By incorporating an FGR system, all gas that would routinely go to flare is diverted back into the gas production system; however, an unintentional consequence of this, is the introduction of oxygen into the natural gas export.
When systems are re-engineered to incorporate FGR systems, the introduction of oxygen via the nitrogen package will not necessarily be taken into consideration. Since the introduction of nitrogen into the exported product gas was never intended, nearly all nitrogen packages will contain oxygen at low % concentrations. If remedial actions are not put in place this could lead to extensive local corrosion in the FGR system and oxygen concentrations in the exported gas that exceed pipeline specifications and ultimately accelerate corrosion of the pipeline / process infrastructure.

Flare Gas Recovery Systems

Traditionally, gases that are sent to flare include production gas from off gases such as water vessels or from low-pressure oil / condensate separators where the gas volume produced is considered negligible. Other gases include process gases such as nitrogen, which are used in maintenance operations, purging or as safety blanket gases in units such as flare headers.
After FGR systems are installed, all gases normally sent to flare will be redirected, compressed and commingled into the production gas for export. One of the main areas of concern is that plant nitrogen will make up a significant proportion of this ‘FGR gas’ and, as oxygen is a natural component of such nitrogen, will result in elevated oxygen concentrations in the export gas which, where the FGR gas is not sufficiently diluted by the produced gas, may exceed the pipeline specification.

Case Studies

Case Study 1
A gas terminal receives gas from four offshore facilities. During a surveillance visit in 2020, analysis showed that the  commingled gas contained no oxygen; however,  18 months later  ~ 50 ppm v/v oxygen was found which exceeded the pipeline limit of 9 ppm v/v.
The oxygen presence raised safety concerns as such levels increase the potential for corrosion both at the terminal and upstream. An immediate investigation to determine the source(s) of oxygen ingress involved visiting each gas entrant to identify whether oxygen was present in the export gas and if present, identify the point(s) of entry across the facility.
•   1st Entrant – Export gas was free of oxygen.
•   2nd Entrant – Export gas was free of oxygen. An FGR was in use but not introducing nitrogen into the process.
•   3rd Entrant – Export gas was free of oxygen. An FGR was installed but not commissioned.
•   4th Entrant – Export gas contained 50 – 90 ppm oxygen. An FGR was in routine use which redirected all gases destined for low pressure (LP) flare back into production. This included plant nitrogen, used as a blanket for storage tanks and vessels across the process, that contained up to 2% v/v oxygen.  
A review of historical flow data showed that the fourth entrant was offline at the time of the first surveillance visit in 2020 which explains why no oxygen was present in the gas arriving at the terminal.

Case Study 2
An offshore facility needed to provide data for target contaminants each time a new field was brought online. A total of twelve on-site visits were conducted during a period of three years and with the exception of the second survey, no oxygen was found in the export gas until survey ten. At the time of the second survey a new well was brought online and the positive oxygen concentration was attributed by the operations team to be due to completions activities that may have introduced air into the system. As no oxygen was found in the export gas for the next seven surveys the positive concentrations found during the second visit were initially thought to be an outlier, however, positive oxygen concentrations were observed again during surveys 10 and 11 and 12.
A thorough investigation of the process found that an FGR system had been installed but, up until survey 10 had been used only periodically and was online during the second survey. Again, the consequence of the FGR system being in operation is that plant nitrogen, which was found to have percentage concentrations of oxygen, made up a significant volume of the FGR gas.
A detailed investigation to evaluate the effects of intermittently running the FGR system was carried out during survey 11.
As expected, when the FGR system was online the oxygen concentration in the export gas was found to be above the pipeline specification limit and when offline the oxygen concentration fell to below the specification limit; however, when the FGR was offline for the six hours under study, the oxygen concentration in the export gas did not fall to below the reporting limit (< 1 ppm v/v). This was expected as a proportion of the export gas is used for gas lift to aid production and therefore oxygen is introduced into the wellbore, which is then returned to the surface with the produced gas.

Sampling and Analysis to Determine Trace Oxygen

It is essential for operators to assess the oxygen concentration in (i) the plant nitrogen (ii) the LP flare gas and (iii) export gas to calculate the expected oxygen concentration in production gas prior to the modification of the production infrastructure to limit routine flaring. This data will be valuable in the forward planning of the design as it may instigate the need to provide a purer plant nitrogen or incorporate an oxygen removal system to eliminate oxygen from the plant nitrogen or LP flare gas.
It is of paramount importance that when monitoring oxygen concentrations sampling should be such that no contamination from air influences the results produced. Thus, as a QC check, the instrument calibration should be checked using a matrix matched certified calibration standard at or around the pipeline specifications before and after each period of monitoring.
To avoid interference from air ingress, the preferred method for quantification of oxygen in production and process gas is to utilise an analytical technique that can be operated at-line. This allows an extended blanking period with an oxygen free inert gas to ensure that a background measurement below the reporting limit of the instrument is achieved just prior to analysing live samples.  
When using conventional, commercially available oxygen analysers to quantify oxygen in accordance with ASTM D7607/D7607M-19, a lengthy equilibrium time requiring significant volumes of product gas is usually needed to be confident that no air contamination is introduced into the analytical instrument; however, Qa3 have designed an at-line system that utilises a commercial analyser but incorporates some in-house design modifications to the sampling apparatus and analysis protocol in order to (i) prevent contamination from atmospheric oxygen (ii) reduce equilibrium time and (iii) achieve a reporting limit commensurate with client requirements. Typically, a limit of detection of 0.3 ppm v/v is achievable, however, due to the variation seen when blanking the instrument, the reporting limit for the at-line method is set at 1 ppm v/v.
At-line measurement affords monitoring over an extended period which (i) provides a better weighted average oxygen concentration and (ii) allows the data to be trended over an extended period to observe any changes to the oxygen concentration due to variation in day-to-day production conditions.

Summary

A summary of the issue and how to check if an oxygen problem does or may exist?
•   As the oil and gas industry strives towards its climate objective in reducing carbon emissions, many operators are now investing in FGR systems to limit the amount of routine flaring from their facilities; however, this comes with the currently unconsidered potential for introducing undesirable oxygen into the exported natural gas.
•   Historically, any plant nitrogen used in process gases has been flared via the low pressure flare system. Therefore, it is very common to find nitrogen purities in the range ≥ 95%, which means that without any remedial actions, low % concentrations of oxygen are invariably present in the plant nitrogen that may be reintroduced into export natural gas via FGR systems.
•   The introduction of oxygen into the exported production gas may (i) result in a product that is above the pipeline specification limit for oxygen which may have fiscal implications and (ii) pose a serious corrosion risk to the steel infrastructure.

•   An assessment of the potential risk for oxygen introduction into the export gas should ideally be conducted prior to commissioning an FGR system. This requires analysis of the plant nitrogen, low pressure flare gas and export gas to calculate accurately the expected oxygen concentration in the export once a flare gas recovery unit is brought online.
•     If nitrogen is present in the FGR system, the oxygen concentration would be expected to be high (potentially %). An assessment of the corrosion risk to the FGR and export gas systems should be sought.
•    To minimise the risk of introducing air into the sample under test an at-line analytical instrument should be employed. In addition, blanking with an oxygen free gas and appropriate QC checks spanning each monitoring period should be incorporated into the procedure to provide an assessment of analyser performance and calibration integrity.  

What if there is or will be positive concentrations of oxygen in the export gas?
•   As oxygen has the potential to cause corrosion via a number of mechanisms it would be prudent to consult a materials engineer to assess the potential rate of corrosion of the steel infrastructure. It may highlight the additional need to quantify H2S and moisture in the export gas to provide the data necessary for optimum modelling of the potential corrosion risk.
•   The manufacturer of the plant nitrogen system should be consulted to see if, with modifications, nitrogen with a lower oxygen concentration can be produced.
•   If nitrogen with a lower oxygen concentration cannot be produced, it may be necessary to incorporate an oxygen removal system[11].
Offshore installations form part of a dynamic system where production rates often vary and new wells may be brought into production. The volume of plant nitrogen and off-gases sent to LP flare also change on a frequent basis. For this reason, it is likely that the oxygen concentrations within the gases will fluctuate over time and should therefore continue to be monitored on a regular basis.

References

1.  McMahon A.J., Groves. S. (1995). A practical guide to the selection and deployment of corrosion inhibitors in oil and gas production facilities. bp Corrosion Inhibitor Guidelines, Sunbury Report No. ESR.95.ER.050 (95).
2. The World Bank – 2022 Global Gas Flaring Tracker Report www.worldbank.org/en/topic/extractiveindustries/publication/2022-global-gas-flaring-tracker-report
3. OGCI – Talking Transition: Putting a stop to flaring (2021) www.ogci.com/talking-transition-putting-a-stop-to-flaring/
4. Weimin Z et al. (2018). Corrosion Failure Mechanism of Associated Gas Transmission Pipeline. Materials. 11(10).
5. Rubin A., Garcia C. (2020). Corrosion – Effects on Metals & Electronic Equipment, Processes & Prevention. AREPA www.arepa.com/media/pkfbmjhx/arepa-whitepaper-corrosion.pdf
6. Ajayi F. (2015). Mitigating Corrosion Risks in Oil and Gas Equipment by Electrochemical Protection: Top of The Line Corrosion. PhD Thesis, The University of Manchester, Manchester UK.
7. Morell L., Park N. (2010). Review: The Effect of Methanol on the Corrosion of Carbon Steel in Sweet and Sour Environments.NACE.
8. Howard M., Sargent A. (2001). Texas gas plant faces ongoing battle with oxygen contamination. Oil & Gas Journal, 99(52-59).
9. Francois M. (2020). Stability of dehydration glycols MEG and TEG. MSc Thesis, Norwegian University of Science and Technology, Norway.
10. Tiedtke D et al. (2001). Chemicals Influencing the Activity of Palladium-Based Catalysts for the Selective Hydrogenation of Acetylene to Ethylene in Acetylene Converters. Ethylene Producers’ Conference, Volume 10 (2001).
11. Jones R., McIntush K. (2010). Oxygen Removal in Natural Gas Systems. Trimeric Corporation. www.trimeric.com/assets/oxygen-removal-in-natural-gas-systems-lrgcc-paper.pdf