In the second of a series of blog posts about the validation of Shell FRED, we review how it has been proven for quick and easy screening of carbon dioxide releases.
Carbon Capture and Storage (CCS) or Carbon Capture, Usage and Storage (CCUS) has gone through several phases of research and investigation.
The current focus is on energy transition, so-called ‘blue hydrogen’ where hydrogen is generated via SMR (Steam Methane Reforming) from methane and the CO2 is stored rather than released to the environment, and in the capture of CO2 from power generation and other industrial processes.
All of these processes involve capturing, storing, processing and transporting CO2 in large volumes at pressure. Accidental releases are likely to be cold, dense phase gas clouds which will remain close to the ground for potentially extended periods of time resulting in a major accident hazard potential because of the asphyxiant and toxic nature of CO2.
In the late 2000s and early 2010s the major hazard potential of CO2 was investigated through research projects, Joint Industry Projects and experimental programmes.
In 2010 – 2011 Shell ran a series of large scale CO2 release experiments gathering data representative of CCS projects (this was presented at Hazards XXII).
In 2011, the UK’s Health and Safety Executive (HSE) published its assessment of the major hazard potential of CO2, and concluded that major accidents are possible in line with many other substances that are regulated by the HSE https://www.hse.gov.uk/carboncapture/assets/docs/major-hazard-potential-carbon-dioxide.pdf
Subsequently, Shell in conjunction with the HSE and the Health and Safety Laboratory carried out a validation exercise, comparing the application of the dispersion model in FRED for CO2 to experimental data and two CFD (Computational Fluid Dynamics) models. The findings were presented at the Hazards 23 conference that was held in 2012. https://www.icheme.org/media/9162/paper21-hazards-23.pdf.
Dispersion Modelling in Shell FRED
The Shell FRED software tool comprises a suite of interlinked hazard modeling codes. Of the models available, the primary scenario is the Pressurised Release (PR) model. There are several key components to this calculation:
- Calculation of thermodynamic equilibrium (calculation of the phase parameters of the stored substance)
- Outflow model (the rate at which material is ejected from the hole)
- Expansion to atmospheric pressure (since the stored material is at pressure, it will expand many times upon release to atmospheric pressure)
- Dispersion Modelling. This comprises two main sub steps in the case of dense phase CO2:
- Momentum jet dispersion using the AEROPLUME model
- Dense gas dispersion using the HEGADAS model, if required
This process is essentially the same as for any other dense gas release (such as LNG). Neutral or buoyant gas releases are passed to a different model, PGPLUME, following the momentum jet phase of the release.
The Pressurised Release model in FRED was compared to experimental data from DNV GL’s Spadeadam test facility, and to CFD results calculated using the ANSYS CFX and OpenFOAM codes.
For unobstructed releases, Shell FRED was shown to perform well.
Figure 1: Measured and Predicted centreline mole fraction (left) and temperatures (right) for test case. Reproduced from Dixon et al. 2012 https://www.icheme.org/media/9162/paper21-hazards-23.pdf
In the very near field, Shell FRED underpredicted the CO2 concentration slightly due to limitations in the model (it is unable to model solid-phase CO2), but within only a few metres from the release, the model was in good agreement with the test data. Further discussion and data was given in the referenced paper. The paper concluded that the Shell FRED model performed adequately for the purposes intended (releases in the free field).
This exercise showed that Shell FRED is an appropriate tool for CO2 hazard assessment in many cases, particularly due to the speed at which calculations can be set up and run. The particular drawback of the CFD models that were assessed in the same study is the time taken to set up and run the scenarios, which is quite a detailed process. This means that Shell Fred can be used for modelling large numbers of scenarios for screening or coarse assessments, where high risk or consequence scenarios can then be selected for further assessment when alternative tools such as FLACS-CFD could be used.
Development of Shell FRED
Shell FRED has been in constant development since the early 1990’s with a very extensive library of validation to support it.
As an operating company Shell understands the importance of accurate, validated models, and continues to invest in R&D projects to further increase the accuracy and usefulness of the tool, and we look forward to sharing these with our users.
For more information about Shell FRED, contact FRED@Gexcon.com.