Community Research and Development Information Service - CORDIS

Periodic Report Summary 1 - SAFELNG (Numerical characterization and simulation of the complex physics underpinning the Safe handling of Liquefied Natural Gas)

The main hazard of Liquefied Natural Gas (LNG) is the flammable vapour which can extend to kilometres as a greenhouse gas, or be ignited resulting in fire and explosions. SafeLNG focuses on six specific areas which are most relevant to facility risk management but for which both theoretical insight and predictive tools are lacking.
- To characterize different LNG release scenarios and develop robust source term models;
- To gain insight of the complex physics in LNG/fuel cascades and flammable cloud formation, and develop robust predictive tools;
- To develop a robust model for accurate prediction of rollover.
- To develop modelling strategies for assessing the environmental impact of large LNG spill by coupling micro scale dispersion models with mesoscale atmospheric models;
- To develop and validate LES based predictive tools for large LNG pool and jet fires; and
- To validate and improve models for explosions in non-uniform LNG vapour mixtures.

This Innovative Doctoral Training programme is hosted by Kingston University London as only legal participant. It is supported by six Associated partners including the University of Warwick (UK), the Health and Safety Laboratory (UK), GexCon AS (Norway), The European Centre for Research and Advanced Training in Scientific Computation (France) and the Universitat Politècnica de Catalunya (Spain).

SafeLNG officially started on 1 February 2014. Since then, six Fellows in their early stage of academic career have been recruited. All six have registered for PhD study at Kingston University and have been making excellent progress in their PhD research. All six research topics are primarily based on numerical study using computational fluid dynamics (CFD) techniques. The open source CFD code OpenFOAM is used by all six fellows as the basic numerical frame for model development. The object orientated structure of the code makes it easy for the fellows to work as an integrated team, sharing development which is common to all. It also renders it easy to consolidate the individual contributions into one single version of the code for future exploitation. After the researchers have learned to use the code as well as assembling new solvers for their specific study by addressing the underlying physics, a dedicated training course has been organised through Open CFD Limited, the vendor of OpenFOAM to address a list of topics compiled by the fellows concerning OpenFOAM as well as essential C++ required for programming within the frame of OpenFOAM. This is in addition to a number of other training course in which external experts from both industry and academia have delivered lectures on a range of topics covering LNG production, transport and utilization, LNG safety, CFD modelling for the oil and gas industry and Radiation modelling in OpenFOAM. The fellows have also been provided access to additional online training on Fire and Blast Issues Related to LNG and LNG production, storage, transportation and utilization. Furthermore, the Fellows have also been encouraged to share knowledge among themselves, e.g. one Fellows who already gained ample experience in mesh generation with OpenFOAM delivered a training course on this to other fellows. The fellows have also received complimentary skills training on a wide range of topics.
Excellent progress has been achieved towards the scientific objectives. A summary is provided for the state of play for each of the six topics below:

Rollover: The study has focused on developing a robust computational fluid dynamics (CFD) model which would be able to prevent any risk of rollover occurrence from the safety perspective. The initial development has completed. For code verification and validation, the numerical results have been compared with experimental data as well as numerical predictions of previous researchers and historic accidents. As there is virtually very little data available in the open literature for model validation, parallel experimental work has also been set up at premise of the Associate Partner, University of Warwick. The preliminary results have also served to provide meaningful comparison with the predictions to guide the development of the model.

LNG vapour cloud explosion: Flame acceleration (FA) and deflagration to detonation transition (DDT) in gasesous explosions have been widely studied both experimentally and numerically. However, in practice, the combustible mixtures are usually inhomogeneous and subject to a vertical concentration gradient due to gravitational effects. DDTFOAM, a density based solver has been assembled within the frame of OpenFOAM for FA and DDT in flammable mixtures with concentration gradients. As experimental data is lacking for the explosions of LNG vapour could, the initial numerical tests and validations have been conducted with a horizontal obstructed channel filled with inhomogeneous hydrogen/air mixture and achieved reasonably good agreement with the experimental data. The study has also examined the effects of Baroclinic torque (based on density-gradient and reflected-shock wave from the obstacles and walls), Richtmyer–Meshkov (RM) instability and Kelvin Helmholtz (KH) instability on FA and DDT.

LNG jet release and jet fire: Immediate ignition of pressurised LNG release can result in a LNG jet fire. If the pressure is relatively low and the source diameter is large, the jets will normally rain out burning liquid, but higher pressure and smaller jets are likely to consume all the liquid with no significant rainout. A jet fire may cause severe damage and be confined to a localised area. Apart from some very much simplified studies conducted in industry, there is no report of study on this topic in the open literature. The initial effort has led to the development of the basic approach to treat the two phase jets with the homogeneous relaxation method, which is now being implemented into OpenFOAM. The researcher also conducted predictions of gas jet flames for cases previously studied by others numerically.

LNG/fuel cascades and flammable cloud formation: CascadeFOAM, a new solver for modelling fuel cascades has been developed within the frame of the OpenFOAM toolbox. Both the buoyancy terms in the momentum equations, the film and splashing models have been modified. The predictions have achieved reasonably good agreement with the experimental data but identified the need to improve the existing splashing model, which is based on a published model developed and tuned for internal combustion engine applications in which the size and properties of the droplets as well as the physical domain are very different from the current application. As there lacks data in the literature in relation to droplet impact on a flat dry/wet surface for droplet diameters in the range of the current applications, investigations are underway to calibrate/modify the existing splashing model through numerical tests.
LNG pool fires: A combined CFD fire model with full coupling between gaseous and liquid phases to predict burning rates of liquid fuel pool fires has been developed within the frame of FireFOAM, the large eddy simulation (LES) based fire simulation solver within OpenFOAM. This work is dedicated to simulate the entire process of liquid fuel pool fire; buoyancy driven flow with combustion, soot formation and oxidation, radiative heat transfer, evaporation of liquid fuel and prediction of fuel burning rate. For validation, comparison has been made with published experimental data for burning rate histories versus time and achieved reasonably good agreement. This has laid the foundation for simulating LNG fires without relying on the crude assumption of mass burning rates used in previous numerical studies.

LNG pool spread, evaporation and dispersion: This project has been progressing in two fronts. Firstly, a CFD approach has been implemented in OpenFOAM for LNG dispersion. A source model was used to define the pool radius as well as evaporation rate, which was then coupled to predict dispersion. To verify the model three standard trials of the Maplin Sands series of experiments were taken into account. The model is capable of incorporating the effect of humidity, and generates atmospheric boundary conditions over the domain. Lack of enough information on time varying release rate and pool formation rate led investigations to an arcwise study, where verifications showed the model is in good agreement with the experiment, and the detection of low flammability limits is acceptable. On the second front, work is underway to develop a modelling approach for LNG spill with the ultimate aim to couple the spill and spreading process with the cloud dispersion. As part of this process, modifications more accurate evaluations of LNG properties are being implemented using the NIST Chemistry web-book.

Website: A dedicated project website is up and running http://www.safelng.org.uk.

Reported by

KINGSTON UNIVERSITY HIGHER EDUCATION CORPORATION
United Kingdom

Subjects

Life Sciences
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top