Search for ignition key

Technical data from BP explosion experiments is being used by a UK university as part of its research on the safety implications surrounding the growing use of hydrogen fuel.

The Centre for Fire and Explosion Studies (CFES) at Kingston University will use the information to develop computational fluid dynamics (CFD) techniques that accurately model hydrogen combustion in the context of fire and explosion safety for the EU-funded HYFIRE project. Also being used is data compiled from explosion experiments carried out by the Health and Safety Laboratory, Buxton.

The scientists, led by Prof Jennifer Wen, will carry out numerical investigations in four main areas: hydrogen jet flames from very high-pressure release; how flames impinge on surfaces and affect hydrogen storage vessels; how a hydrogen spill spreads and the conditions under which the fuel can ignite.

'Hydrogen is only a small part of BP's business — about one per cent. But because of the move to hydrogen economy, its has commissioned explosion tests as well as high-pressure hydrogen release trials. Those are larger in scale than universities can do, so BP has carried out some very expensive tests with commercial organisations and released that data for this project to use. We will run the data model against their data,' said Wen.

CFES has also compiled an accident database for hydrogen and hydrocarbon fuels. A total of 3,100 accident cases were collected, which includes 850 hydrogen cases.

Wen said: 'Analysis of this database shows that the majority of hydrogen releases have led to ignition, but the reason for ignition was unknown. There has been little research into why the probability of ignition for accidental hydrogen release is higher than other fuels.'

Since the start of the project last September, CFES successfully modelled the auto-ignition phenomenon in high-pressure hydrogen release (hydrogen ignites spontaneously on release). According to Wen, this is particularly significant for the safe operation of fuel cells in vehicles.

'Our biggest achievement is carrying out accurate numerical predictions of the auto-ignition which were previously observed in some laboratory high-pressure hydrogen release experiments. With the numerical tool developed, we have conducted further parametric studies to illustrate that it is a combination of the release pressure and size of the rupture in the cylinder which determine auto-ignition occurrence,' she said.

While the scientists are confident about these predictions, Wen said there was still ongoing work to predict how the ignited hydrogendevelops into a stable flame, which is made complex by a large number of variables, such as chemical reactions.

'The spontaneous auto-ignition is reallydifficult to predict because the underlying equations of the chemistry are really stiff, so you not only need the actual time state, but also really high order numerical accuracy because there are so many variables and so many reactions with the shockwaves,' she said.

As well as investigating the conditions for auto-ignition, the scientists hope to use CFD modelling techniques, such as large eddy simulation (LES) to predict the chances of ignition from a spillage of liquid hydrogen.

'We want to develop models for how the spill will spread and how the flammable cloud would form,' said Wen. 'We want to use a CFD model to look at that because it will be more accurate and also allows you to take into account the 3D effect, as your cloud will be 3D. LES techniques will solve the equation for the flow of the liquid or gas so it captures everything.'

The scientists will also use LES models to predict conditions that could lead to detonation in confined explosions (an explosion inside an enclosure with some obstacles) or semi-confined explosions (which could be in a car park, which has limited openings).

'We are using an open-source CFD code called OPENFOAM. That is the basic tool we are using for the explosion work. It is a generalcomputational fluidic dynamics code, so it can model complex flow as well as the reactive flow where there is combustion,' said Wen.

Finally, some modelling of jet, or spray, fires (caused by the combustion of fuel continuously released at speed in a particular direction or directions) has already been carried out by CFES, and follow-up research will involvelooking at the impact of this type of fires on storage vessels and making improvements to existing modelling software.

'Ideally we would like to find out what the consequences will be on the fuel that is stored inside. There are technologies to model the impingement, but they are not sufficientlyaccurate — they have not been validated.

'So we are developing and validating a model which can then be used as a tool by industry to make predictions,' said Wen.