Project ETANKFIRE
Published: 05 August, 2011
Biofuel usage is increasing worldwide and yet large-scale fires involving ethanol have yet to be successfully extinguished. Jose Maria Sanchez de Muniain reports from a project that aims to do just that.
Hosted by BP, the European ETANKFIRE workshop took place 28th June, and fire industry experts came together near Ascot (UK) to discuss how to drive forward the ambitious project.
The goal is to develop and validate a methodology for tackling tank fires containing ethanol fuels and determine the large scale burning behaviour of ethanol fuels.
Per Brännström, Swedish Petroleum Institute
Per Brännström, health and safety environment manager for the Swedish Petroleum Institute and managing director of a task force that deals with large fires (SMC) outlined some of his observations and concerns regarding ethanol tank fires.
Firstly, ethanol use is increasing and the European Renewable Energy Directive that is to be implemented in 2011 will define renewable fuels – of which ethanol is one – and each EU member state will be expected to replace 5.75% of all its petrol and diesels for transportation issues with this type of fuel. ‘In terms of volumes, we have seen this year an increase from 5% to 10% in several countries around Europe, and in the US there is a plan to increase from 10% to 15%.’
In the US, ethanol fuel production has increased from 175m gallons in 1980 to 13,000m in 2010 (source: Renewable Fuels Association). In Sweden alone, usage of E85/ED95 has increased from 31,225m3 in 2005 to 210,083m3 in 2010.
Per commented that discussions are already taking place with authorities as regard emergency response, fire protection systems, and adequate risk assessments. These discussions revolve particularly around ethanol stored in large tanks, and whether the heat given by ethanol in such a tank is more heat intensive than gasoline. ‘In small fire scenarios they are fairly similar but when you increase the scale there are indications that ethanol has a much larger heat flux than gasoline or petroleum. If you want to design a good extinguishing system or if you are an emergency response organisation you will need to tackle this problem.’
Per Brannstrom suggested that if current theoretical calculations are correct, then ethanol tank fires would result in a heat flux between 35 and 50% higher than gasoline. ‘And if this is correct you will have a problem with your depots, because it means you are only a few minutes away until very high temperatures are reached in adjacent tanks of petrol, which could catch fire very quickly. And today’s systems and tank layouts are designed for petrol – not ethanol.’
In short, the only way to find the answer to the ethanol question is to conduct large-scale fire tests.
Niall Ramsden, Resource Protection International (RPI), UK
Following on from a potted history of LASTFIRE and lessons learned, Niall Ramsden of Resource Protection International talked about some of the research carried out on storage tank fire hazard management including boilover tests, new extinguishing techniques, and radiant heat effects on exposed adjacent tankage– all issues that could be applicable to ethanol.
Niall pointed out that some biofuels have been shown to exhibit boilover type characteristics similar but generally less severe to crude, and he reiterated the main question that is yet to be answered: how long does a responder have of safe operation once a tank is on fire; and what is the best way to tackle this type of emergency? ‘One of the things we’ve learned about boilovers is that there are a lot of strange theories out there. I’ve heard one about pumping fuel into the bottom of the tank while it’s burning – which I think is not really a guaranteed risk reduction option!, I’ve heard pumping in compressed air, I’ve heard all sorts of strange things.’ The test work the Lastfire Group has done – which now exceeds more than 150 experiments – clearly shows that such theories are exactly that – theories – and they cannot be relied upon in a real situation.
Niall showed some data related to crude oil boilover testing that had been carried out with Loughborough University. Even on small scale testing with test pans and crude oil there was a massive seven-fold increase of radiant heat during boilover.
A film was showed of a boilover test in Abu Dhabi using a 4 metre diameter test pan: ‘If you look at the old 1920s description of these types of incidents, they talk about a “flaming Niagara”; which is spot on as you can see here. There was an initial boilover, and several hours later there were more boilovers.’
Niall invited the audience to imagine similar incidents with tanks that were 115 metres in diameter. Obviously these would be major incidents with massive escalation and risk to firefighters.
Other work that could be relevant to ethanol carried out by the LASTFIRE Group has involved the use of alternative extinguishing media, such as intumescent balls, and glass based materials.
LASTFIRE is also doing some initial work on ethanol, and referring to the foam manufacturers at the workshop Niall remarked: ‘However good you claim your alcohol-resistant foam is, the plunging application into ethanol is not very effective. Perhaps industry ignores that factor but it’s true.’
Niall accepted that bouncing foam off the tank walls resulted in a more gentle application, but he doubted this could be carried out in most real situations. A video clip was shown demonstrating the problem ‘This is using a good quality foam, with two nozzles at double the application rate that had previously worked on an ethanol fire with gentle application at similar solution rates to those used on hydrocarbons. When applied forcefully on ethanol it doesn’t work – the foam just breaks up straightaway. Eventually it would work through dilution but it would take a long time to do that and is not apractical solution.’
Ultimately the problem is that there is a shortage of data from ethanol fires and foam application to them. Hopefully ETANKFIRE will correct this situation.
Tank fire fighting research – previous work & knowledge gaps, Henry Persson
Henry began his presentation by pointing out some important differences between ethanol and petroleum products, including flammability, burning behaviour and extinguishment.
When comparing ethanol to gasoline in an open spill, gasoline is the worse fuel because it has a very low flashpoint of -40 C while pure ethanol is +12 C.
While gasoline’s heavier fuel vapour increases the risk of ignition due to its spreading on the ground, ethanol vapour has higher electric conductivity, reducing the likelihood for an ignition caused by static electricity. All together, this means in an open spill situation both fuels have similar risk levels.
Inside an enclosure such as a tank, however, ethanol is the more dangerous because it has a flammable range between +12 and +40 degrees. While gasoline’s flammable range is from -40 C up to about -25 C, once the temperature rises above this temperature the vapours produced quickly result in a ‘rich’ mixture that is difficult to ignite.
With E85 (summer quality) the flammability range is from about -16 C to +5 C, which means at ‘normal’ temperatures the resulting fuel has a lower risk of ignition than pure ethanol.
Henry pointed out that ethanol is normally stored in converted gasoline tanks with internal floaters, which aids in eliminating the problem of lean mixtures inside a tank. But he said that filling or emptying a tank results in increased fuel vapours (eg due to ‘wet’ tank walls) which could lead to flammable conditions. ‘The question is if you have flame exposure, bund fire or spill fire, a hot spot can very quickly reach auto ignition temperature close to flammable vapours inside the internal floater. Or the spills could result in increased fuel vapour mix inside the tank that could go to flammable conditions, and you could end up with an internal explosion.’
Henry demonstrated a number of incidents involving ethanol tank fires, including a fire test carried out by a Swedish insurance company that went spectacularly off-plan and which resulted in an explosion – the demonstration should have resulted in the release of vapour by the pressure vent on the top of the tank. ‘In Sweden we are talking about having flame arresters on venting equipment, and that operators should keep in mind that thermal exposure could go to ignition, and that the addition of gasoline could have an effect from a fire perspective.’
Moving on to differences in burning behaviour between ethanol and gasoline, a small-scale test was shown using a 1.7m2 pan for a spill scenario at a filling station. The gasoline flames reached 7m in height while E85 was less severe. Heat release measurements taken at 3m distances were higher for gasoline than ethanol too. Going up scale however and the results were different. A video comparing the burning behaviour of gasoline and acetone/ethanol mix showed that the polar solvent’s flames and radiant heat were much higher – in essence a reversal of the small scale experiment.
‘Now we don’t know the acetone/ethanol mix is the same as ethanol – I would say it is, but the fact that we have seen this difference means it is important it should be investigated.’
A modelling program used by Henry predicts flame height for gasoline reaching 25m in comparison with 40m for ethanol. Current recommended distances between gasoline tanks are 15m but if the data suggested by the modelling is to be believed, the distance between an ethanol tank and the next tank should be doubled to 30m. ‘The problem is we are lacking the verifying data on a large scale.’
The next big issue is extinguishment, and some work carried out in Sweden with methanol (Sweden’s initial biofuel of choice in the 1980s) demonstrated that the method of foam application was of paramount importance. Extinguishment couldn’t be achieved through foam plunging, and that for foam to work it had to be applied gently eg through bouncing it on the ground in front of a spill.
Tests carried out by the Ethanol Emergency Response Coalition in the US – a group that aims to provide educational material and training for emergency responders – concluded that the only agent that could successfully extinguish ethanol spill fires (with an emphasis on transportation incidents) was AR-AFFF. But these tests, as well as most standard test methods, are only representing spill fires with fuel layers of 35-70 mm, and the foam had to be applied by a back wall. In addition, the allowed extinguishing time of 3-7 minutes meant that the dilution effect was up to 30%, which is quite significant in comparison to tank fire applications.
Data on ethanol tank fires is virtually non existent. Out of 480 tank fires identified by Henry for research purposes, only 8 have contained water-miscible fuels. But the main point is that the conditions for tank fire fighting are completely different to spill fires – large diameters, large fuel depth (dilution would take a long time): ‘And you could say that backboard applications are not practical.’
In the small number of cases that have been documented the fire has not been extinguished even after large volumes of foam have been applied. At Port Kembla in Australia an explosion in a 32m diameter tank with 4,000m3 of ethanol swallowed up 50,000 litres of AR-AFFF without success. In the end the fire was allowed to die out and the remaining solution only contained an estimated 10% of ethanol.
As for extinguishment, Henry remained sceptical as to whether foam monitor application was likely to work, even using a large-scale monitor and backboard. ‘And that is why we think ETANKFIRE is important. There are many aspects of ethanol we don’t have the answer to. We can’t treat an ethanol tank as a gasoline tank and if we continue to do so then we will continue to have burnout situations and our investment will be a waste of money. So I would say that the growth of the ETANKFIRE project is about finding a way of extinguishing these tank fires and verifying it to a large scale.’
Regulations and tank fires, Niall Ramsden, Resource Protection International
Niall Ramsden commented that legislation varied from country to country – in India the regime was prescriptive as regards fire protection on tank farms, while typical European legislation (under SEVESO) was based on risk: ‘So it does not say, “you shall use this foam system or provide this number of responders”. The onus is on the operator to look at the safety aspect and for fire engineers to consider the risk – which is defined as a product of probability and consequences. One of the reasons we don’t know a lot about fighting these types of fires is that we are very good at preventing them.’
In essence, legislators are unlikely to overly mind if an ethanol fire burns down, as long as there is no risk to life or the environment. While in theory there should be no conflict between the different stakeholders – eg the operator wants to protect his facility and the legislator wants to protect life safety – in the real world there can be conflicts in interest between the operator, local authority and the safety executive. The operator will want to keep costs down and may rely on a local brigade for support, but then there are the instances where a brigade will believe it has the capability to respond effectively but doesn’t. Other factors may come into play, such as national interest in the case where a country’s main income is reliant on a single facility. Another factor is public image: a relatively small incident on a Staten Island oil storage facility (US) in February 2003 resulted in disproportionate media attention including press conferences by New York mayor Michael Bloomberg.
There have been incidents where tank fires have been allowed to burn out – Shell in New Jersey and Magellan in Kansas City are two examples. Where this is not an option, the solutions tend to be either fixed systems (which can be expensive in farms comprising of as many as 100 tanks), or mobile and flexible solutions such as monitors. ‘But people buy these sorts of things and forget about the back up. The worst case I saw was in the US, where a large refinery had a large-scale monitor. I noticed it was very clean and enquired when it was last tested. They said they couldn’t test it because they didn’t have the water supply.’
Niall pointed out that the UK’s Environment Agency had published a document called the Environmental Impact of Controlled Burns (Technical Report P388), which said that in some cases from an environmental point of view it could be better to allow a controlled burn than attempt to extinguish. Another publication, Pollution Prevention Guidance (Controlled Burn: PPG28, July 2007), highlighted how operators needed to think about public concerns and the fact that a controlled burn could take several days – a long time to be in the media spotlight.
‘My message is legislation can let operators burn if there is no harm to anybody and it is possible to do so while cooling [adjacent tanks]. But ultimately it is up to operators whether it is acceptable to them as a company – it may not be the way forward for their public image.’
Project structure
The suggested structure of the ETANKFIRE project involves six work-packages as shown below.
In order to optimize the test design in the project, the intention is to start with test series on a laboratory scale to investigate the relative influence of various extinguishing media and tactics. Based on the results of laboratory tests, the most promising extinguishing methods/media will then be selected for further evaluation and verification in a larger scale or scales.
In total, four work packages related to extinguishment of ethanol fires are proposed. WP1-2 involves testing in laboratory scale while WP3 (“medium” scale) and WP4 (“full” scale) need to be conducted outdoors. The burning behaviour and heat radiation from an ethanol fire in large scale conditions, will be handled in a separate work package (WP5). This WP will involve large scale ethanol pool fire tests to provide accurate data concerning the burning behaviour. Further, WP1-4 will to some extent provide further input to WP5 based on the results from the pre-burn periods in the small and medium scale fires. The test program will be focused on ethanol based fuels but will provide important information that will also be applicable to other water miscible products.
ETANKFIRE is looking for funding and support. If you would like to get involved with the project, contact Henry Persson on +46 10 516 51 98, henry.persson@sp.se, or Margaret Simonson McNamee on +46 10 516 5219, margaret.mcnamee@sp.se
A second workshop will be hosted tentatively in central USA later this year. More details can be found here as they emerge.
