Cassandra Osvatics, Hydrogen and Fuel Cell Technologies Office (HFTO): Hello, and welcome to this month's H2IQ hour for an update on changes to the National Fire Protection Association hydrogen technologies code, NFPA 2. My name is Cassandra Osvatics with the Department of Energy's Hydrogen Fuel Cell Technologies Office, or H2IQ, according stakeholder engagement and other outreach activities.
Please be aware that this Webex webinar is being recorded and will be published online on our H2IQ webinar archives. If you experience technical issues today, please check your audio settings under the audio tab. If you continue experiencing issues, please send me a direct message. Next slide, please. There will be a Q&A session at the end of the presentations and attendees will have the opportunity to submit questions in the question box. Next slide.
You can access the Q&A feature by clicking on the panel options or more options button depending on your operating system. That is the button with three dots at the bottom right corner of your window. To ensure that your questions are answered, please utilize the Q&A feature and do not add your questions to the chat. If you have trouble using the Q&A feature, please view the link in the chat for help with this issue. You can stop screen sharing. And now I'd like to turn it over to Laura Hill. Laura is a technology manager for safety codes and standards within HFTO.
Laura Hill, HFTO: Thanks, Cassie. Brian, Ethan, if you guys want to bring up your slides while I give a little bit of an introduction. I want to thank everybody for joining us today. We're excited to have two speakers from Sandia National Labs. The first speaker, although they will be tag teaming, but the first speaker is Brian Ehrhart. He's a chemical engineer at Sandia.
He has a bachelor of science in chemical engineering firm Rensselaer Polytechnic Institute and a master and Ph.D. in chemical engineering from the University of Colorado at Boulder. Since 2017, he has worked to support the technical analysis for safety codes and standards for alternative fuels, particularly hydrogen, and he leads the hydrogen safety codes and standards project at Sandia.
Ethan Hecht, our second speaker is also at Sandia National Labs. He join Sandia in 2005 after receiving a bachelor's science in engineering physics and a master's of science in mechanical engineering from the Colorado School of Mines. Ethan participated in a special degree program through Sandia and received a Ph.D. in chemical engineering from the University of Utah in 2013.
Ethan has been leading Sandia's research into hydrogen behavior as it relates to safety codes and standards since 2014. Today, Brian and Ethan are going to talk with us a little bit about the new changes to NFPA 2. So I will stop talking and allow them to expand on that. Thank you, Brian and Ethan.
Brian D. Ehrhart, Sandia National Laboratories: Great. Thanks, Laura. So, yeah, like Laura said, my name is Brian Ehrhart and Ethan and I are going to be talking today about some of the basis technical justification and calculations that provided revised set back distances for bulk liquid hydrogen storage in the NFPA 2 code. And I apologize about my voice but we'll just deal with it.
All right, so first, I wanted to give just a brief introduction to the NFPA 2 hydrogen technologies code specifically. So this is a fire code put out by the National Fire Protection Association that was founded in 2006. The NFPA 2 document was established in 2006 as a way to bring together all hydrogen-specific requirements for anything related to stationary systems that produce, store, transfer, or use hydrogen. So this code really does apply to those kind of stationary systems. So refueling stations, storage facilities, those sorts of things.
It does not apply to onboard vehicle storage or components, and is meant to apply to near pure hydrogen gas and liquid. Not necessarily other hydrogen containing mixtures or materials. The most recent 2023 edition of NFPA 2 was just published recently in December of 2022. So just a couple of months ago. And it is now the current edition of the NFPA 2 code.
This code, this document, is fairly widely and commonly adopted here in the U.S., although the specifics of whether or not this particular document is adopted and specifically what year or what edition of the code is adopted can vary by state or local jurisdiction. So the specific legal requirements in your particular jurisdiction, especially internationally but even here in the US in different states or counties can vary case by case. But NFPA 2 is still a pretty commonly used document. Even when it's not legally required, it still can be referred to and used as a basis for requirements for these types of systems.
So today we're going to be talking primarily about setback distances. And so I wanted to talk briefly about what those are, what they mean, and what they are intended to do. So a setback distance is a prescribed distance. So it's a specific distance that's listed in the code. So it's a particular requirement that is meant to provide physical separation from a potentially hazardous system, in this case a hydrogen storage system, and provide distance away from people, buildings, or other hazardous materials.
This can help protect those people, buildings, or other hazardous materials from potential damage should a hazardous condition develop. And it can also work in reverse actually. So external damage, like external fires to the system, that same distance could provide protection to the hydrogen system in reverse, even though that sometimes isn't always explicitly considered in the development of the setback distances. But it's still worth mentioning.
Setback distances do not completely eliminate risk. They are not necessarily based on an absolute worst case scenario. And so these very rare high consequence events may sometimes not provide sufficient protection just due to the setback distance alone. That's because these setback distances are meant to work in conjunction and together with other safety requirements, including design requirements, installation requirements, and operational requirements in the fire code. So those other types of requirements are really help to ensure the system is designed, installed, and operated safely.
And these setback distances provide an additional layer of protection, especially for more common leaks or releases that might occur due to the system operation. These setback distances, because they are prescribed in regulations, codes, and standards, in this case, the NFPA 2 code that we'll be talking about today, can really have a big impact on system location and siting.
This can include both the physical footprint a hydrogen storage system can take up, and how much space around the system needs to be kind of reserved or set aside. It can also affect where a hydrogen system can be located within a larger property that may contain other types of systems or other people, roadways, parking facilities, that sort of thing. So therefore, for all these reasons, it's really important that these setback distances have a strong technical justification. So that they can promote and improve safety and ensure these systems are operated safely without being unnecessarily onerous or restrictive
So previously, in previous editions of the NFPA 2 document, the bulk liquid hydrogen storage setback distances tended to be somewhat large, somewhat arbitrary and based on the storage volume capacity of the liquid hydrogen storage tank. And they also tended to vary quite a bit by individual exposures. So different types of buildings or different types of other materials tended to have different setback distances. And this made it somewhat complicated to apply these types of systems.
And as well, there are different systems with larger storage capacities were potentially subject to very large setback distances. And so what we wanted to do in conjunction with the NFPA 2 Technical Committee and Storage Task Group was take a look at the basis of these setback distances, see if we could come up with a revised basis to replace a poorly documented and somewhat unknown basis for the previous setback distances. And specifically try to identify were there places in which the setback distances could be reduced in which safety wouldn't really be adversely affected.
So to do that, we wanted to develop, validate, and verify some necessary numerical models that could be used to predict hydrogen release behavior because that's what we can base these setback distances on. We wanted to group all these various different exposure types into fewer groups that were more similar to make things a little bit more consistent. We wanted to use quantitative risk assessment tools and methodologies in order to determine a representative or relevant leak size basis to determine the setback distances, and then calculate these distances using those models and based on the basis determined.
This was done previously in previous iterations of the NFPA 2 document for gaseous hydrogen setback distances. And so we wanted to take a similar approach here that was done for gaseous setback distances and applied the same overall methodology or approach to the liquid hydrogen setback distances. Specifically, the approach, as described here and I kind of described on the last slide, used risk assessment tools to quantify the risk on a representative system design. That was used to inform but not directly generate the leak size basis.
There were different physical harm criteria that would be used in conjunction with the calculated model results in order to determine the actual setback distances. And the leak basis was changed from a storage capacity metric to a fractional leak size, which therefore depended on both the system pressure and pipe size, which we felt was a more representative and more useful basis for these setback distance calculations because it would allow the distances to vary more with the system characteristics.
There was also some passive mitigation such as fire barrier walls that allowed for reduction of the calculated setback distances, both for gaseous and liquid hydrogen setback distances in previous versions of the NFPA 2 code. And we'll be mentioning that as well because those will be used for these setback distances as well. So, yeah, overall, the same type of approach used for liquid hydrogen as was previously done for gaseous hydrogen setback distances.
Ethan S. Hecht, Sandia National Laboratories: Thanks, Brian, for the great introduction. This is Ethan Hecht now presenting a couple of sections here. So we're going to walk you through, basically, the process that we took to come up with these science-based setback distances for bulk liquid hydrogen systems in the 2023 edition of NFPA 2. As Brian pointed out, the first step was to figure out which models we were going to use and make sure that those models were validated. So we landed on using the HyRAM plus tool kit for the calculations.
You can see that that's free and open source software that's available from hyram.sandia.gov as a Windows gooey program that you can run on your own personal computer. The back end models are all based on Python and there's also individual Python packages that are -- the Python module is available from both PyPI and conda-forge. So you can run the back end models on your PC as well if you're familiar with the Python programming language.
What's contained within HyRAM+ our fast running reduced order models for physical phenomena associated with hydrogen and other alternative fuels. The specific models of interest for this work were for unignited dispersion. And that contour plot over on the right-hand side of this graph shows what those model results look like. So there's contours of different mole fraction levels in two dimensional space for a horizontal leak that is shown in that contour plot.
It can also simulate flames. And that's what the three frame plot shows in the middle of the slide there. So that it shows both the trajectory of the flame by the black line as well as different heat flux contour levels shown by the different colored contour bubbles there in that center plot.
There's also a model for unconfined overpressure. So the pressure that would be released should delay ignition of a plume occur. So if the system is leaking for a while and then you ignite it, there will be some overpressure associated with that. And that can cause hazard to people or other infrastructure and that was considered in this analysis.
So the behavior models can be used standalone as shown by these plots, or they're also tied into a quantitative risk assessment. And we also use the quantitative risk assessment piece, as Brian briefly mentioned, to come up with the characteristic leak size for the system. And Brian will be going into more details on that later. We did use the Python backend for all these calculations because it enables more flexibility in terms of the modeling. The results you can pull out of it. But similar models are available from the windows gooey front end.
This is important that these calculations are open source and freely available and somebody can repeat these analyzes and change the assumptions and come up with different setback distances or critique our approach here on these setback distances as well. So we tied it to a specific version of HyRAM+, specifically version 4.1, the development of which occurred as we were going through this work. If you can move on to the next slide, please.
So we wanted to ensure that the models were valid for liquid hydrogen systems specifically. We'd done a lot of previous work with HyRAM for gaseous systems but not a whole lot of validation or work using HyRAM for liquid hydrogen systems. And so one of the first models that we looked at was the flow through an orifice. This is important because this will be the amount of mass of hydrogen that's flowing into the system.
And we found an issue in version 4.0 of the flow model through an orifice. And that's depicted in that graph on the left-hand side where you see that dashed blue line provides a mass flux for a given pressure that's a bit below the solid black line, which is the implementation in version 4.1. And the difference between these two models was the way that the speed of sound was used in version 4.0, which is a sort of a difficult calculation and uncertain calculation for two-phase flows, which will occur during liquid hydrogen flows through an orifice.
And so we came up with another algorithm which relies on a search for the maximum mass flux through an orifice for a given pressure, and more certain parameters such as the enthalpy and entropy of two-phase flows rather than relying on that speed of sound calculation. And so the black line, the increased mass flux through the orifice, was used in version 4.1 and under all versions of HyRAM+ moving forward. And used for this work here, which gives a slightly higher mass flux than before.
We also considered some other types of modeling of mass flux of liquid hydrogen through orifices, such as the metastable liquid model which is shown by the dashed orange line at the top. But we decided that was a bit too conservative and unrealistic and would give too much mass flow of hydrogen through a system. We compared these to other implementations by Air Products, which is the AP symbols, as well as implementations by Chart who were members of this subtest committee who was really critiquing all the models and all the assumptions that went into this analysis.
And so we took that black line, that model that's shown on the left, and we compared that to data. And that's what's shown in the graphic on the right-hand side, where the data is shown by the black bars and the model results are shown by the blue bars. And there's two sets of data, one from the Presley project labeled Pressley and one from some experiments that were performed by DNV GL shown on the right-hand side.
You can see for the first three, the Presley results, we're doing a pretty reasonable job of predicting what that mass flow is. It's slightly under-predicted for that five-bar release, that 6 millimeter orifice, but there's a lot of uncertainty shown by that red line on top of that black bar in that measurement. So you can see that this updated flow model is doing a pretty good job there.
The experiments by DNV GL shown on the right-hand side of that plot, the right, 1, 2, 3, 4, 5, 6, 7, 8, data points there, you can see that the HyRAM+ model is over predicting that mass flow through the orifice pretty greatly. So it's a pretty accurate, if not conservative, estimation of liquid hydrogen flows through an orifice. You can move on to the next slide, please.
So then I briefly mentioned the other types of modeling that are done within HyRAM+. And these are directly related to what those setback distances need to be. And so we compared these other models that are downstream of the leak. So what happens when we compare the dispersion, the overpressure, and the heat flux to data as well.
And so the top plot up there shows some experimental data by the solid lines that are quite noisy and bouncing up and down due to unsteady wind during the experiments, as well as the model predictions which are the straight dashed lines that go across those two plots there. You can see it close in, well, reasonably close to the release point, so 30 meters away from the release point, the model's predicting slightly above the average of the highly fluctuating experimental data in terms of what that average mole fraction will be and the model's predicting the average mole fraction.
As we move further downstream and the mole fractions get lower at 50 meters and then 100 meters, the model's over predicting by quite a bit what that unignited concentration or unignited small fraction will be in those downwind regions. And those are really the regions of concern for these setback distances. We're concerned with specific mole fractions that are near the lower flammability limit. I'll get into more details on that in later slides.
The upper right-hand plot is a contour plot showing the overpressure. So the overpressure that would occur from delayed ignition of a plume. The data in this case is shown by the six dots that are there, and those are labeled by the amount of overpressure in PSI that was observed during the experiments. You can see that those overpressure values were quite low in the experiments, maxing out at 0.44 PSI.
The predictions on the other hand are up above 10 PSI. So between somewhere between 3 and 10 PSI within that three PSI contour that you can see towards the right-hand side of that plot. So the overpressure is quite greatly over-predicted using the model in HyRAM+ relative to this one set of experimental data.
And finally, the bottom right-hand plot shows the predicted heat flux by the shading and the contours. And that's once again compared to data. And the contours, the bubbles, the circles on there, are meant to show the setback distances. So that's really the setback distances that would be predicted by HyRAM+ using the specific flame in this experimental case.
And you can see that within that 20 kilowatts per meter squared contour bubble, or outside that 20 kilowatts per meter squared contour bubble, there was never a heat flux observed that was greater than that 20 kilowatts per meter squared. You can see that 22 kilowatt per meter squared is near that 20 kilowatt per meter squared maximum heat flux that's shown on the color bar on the right-hand side. But for all cases, for these specific contours, 29 and 5 kilowatts per meter squared, you can see that there's never a heat flux that's greater than that specific contour outside that contour bubble. Move on to the next slide, please.
So the next consideration was really what these specific criteria should be for the different setback distances. And so part of that was grouping the different exposures into separate groups so that we knew what we were trying to avoid. And so that's shown on the right-hand side of this chart here. And so for all of these exposures in group 1, the test group decided that these criteria should avoid harm to the general public, should avoid damage from heat flux, damage from overpressure, and any flammable concentration that would be affected by an ignition source.
For group 2, the avoidance criteria was harm to people aware of the risk, so people at the fueling station or people who knew they were near a bulk liquid hydrogen storage tank, avoiding significant damage to buildings, avoiding fire spread to ordinary combustibles. And then finally, for group 3, that should avoid escalation of events. So the event. So fire spread either to or from the hydrogen system. Next slide, please.
And so by trying to avoid these different criteria, we had to come up with what the specific values were for those criteria. And so if you remember on the last slide, group 1 was avoiding unignited concentration. And we based that unignited concentration level on the ability to form a jet flame. So we've done quite a few experiments at Sandia where we've looked at these diffusion jets of cryogenic as well as gaseous hydrogen releases.
And we found that there is a boundary where you can ignite things up into a jet flame but that's not necessarily the lower flammability limit. That there are areas of these diffusion jets where you can ignite up into ignition and ignition kernel, but there's so much turbulence in momentum in these jets that kernel will quickly blow off. And that's what's meant to be depicted in that video that's playing towards the center of the screen, where there might be an ignition kernel at a value that's below an 8% by volume average concentration but that kernel will quickly blow out.
On the other hand, if you're within this 8% by volume mole fraction contour, there's a pretty high probability that will light up into a jet flame. The actual value can be actually a bit above that 8%. It can be up to 12% or 15% by volume average concentration. But we found 8% to be a good criteria for the ability to ignite these diffusion jets up into a jet flame. So that's what we used for group 1 exposures, this ignited concentration. You can move on to the next slide, please.
So we also considered the criteria for heat flux. And this giant arrow at the top of your screen is meant to depict different heat flux values and the types of damage that these different heat flux levels can actually cause. And so for group 1 exposures, we used the criteria of 4.7 kilowatts per square meter. This is the same value that's used for group 1 and group 2 exposures for gaseous hydrogen setback distances and NFPA 2. And is also used in the International Fire code as the exposure for employees for 3 minutes.
For group 2, we used a level that's a bit above the group 2 criteria for NFPA 2 for bulk gaseous setback distances and NFPA 2. We use nine kilowatts per square meter. This is the same value that's used in NFPA 59A is the fatality. The risk of fatality to a person outdoor without PPE. We thought that was a good value to use for group 2
And for group 3, we use 20 kilowatts per square meter, which is the same as used in the, again, for the bulk gaseous setback distances in NFPA 2. And it's also in the International Fire code for combustibles as a good level to avoid fire spread between different combustible materials. Next slide, please.
And then the final criteria was coming up with different criteria for peak over pressures. And now this arrow is showing what damage those different overpressure values can cause. So for the group 1 level, which is 6.9 kilo pascals or one PSI, that has the ability to knock a person over or a chance of broken glass or minor damage to structures. At the group 2 level, which is 2 PSI or 13.7 kilo pascals, that can affect a person by causing skin lacerations or light injuries and partial collapse of structures.
And for group 3 that's just below the criteria where serious injuries are common and fatalities may occur. You can see that there's a very small chance, maybe a 1% chance of fatality at some pressures just below 3 PSI. But that was the criteria that we used for group 3 exposures. So 1, 2, and 3 PSIs for groups 1, 2, and 3. And if we can move on to the next slide, please. It's back to Brian.
Brian: Thanks, Ethan. So when considering how to calculate these setback distances, the models used in the HyRAM+ toolkit that Ethan described will be used to actually calculate the distance to the various harm and hazard criteria that Ethan was just talking about. But it's also important to consider the basis, the source of the leak that drives those particular setback distances.
Previously in NFPA 2, as I mentioned, this was based on storage capacity. But using a whole size rather than storage capacity allows setback distances to grow or shrink as a system operating pressure changes. And even system pipe size can change. And so if a-- one option would be to use a fixed hole size, say, a quarter of an inch hole size as the leak basis. But the downside of this is pipe sizes change the whole size of the leak basis and therefore, the resulting setback distances wouldn't change with those system design changes.
So instead, we use a fractional leak size. In this case, this is a fraction of the respective cross-sectional circular areas of the leak size as compared to the cross-sectional flow area of the interconnecting piping for the hydrogen storage system. The benefit to this is that this sort of constant fractional hole size allows systems that use smaller pipes that can inherently limit the consequences and take the benefit, take that credit of using smaller setback distances.
Whereas systems that use larger pipes that could potentially result in higher consequence releases, larger releases, would result in larger, longer setback distances. And the same with system pressure. Higher pressure systems are going to result in larger, longer setback distances. Lower pressure systems are going to result in shorter distances.
It's also worth noting that these rather simplified models do tend to assume steady state flow. In reality, a tank that is a leaking hydrogen is going to reduce the flow rate because the system pressure reduces over time as more and more hydrogen empties out of the tank. But in this case, we always assumed steady state flow. And so we're effectively always using the maximum flow rate at the highest system pressure, even though in reality the flow rate and system pressure would decrease over time as the leak went on. But we're assuming that steady state. So a somewhat conservative assumption there.
And so in order to consider what leak size basis to use, what fractional hole size we should actually use, we used quantitative risk assessment tools within the HyRAM+ toolkit. These overall reflect a bow tie type of risk assessment and analysis in which component leak frequencies are considered using various fault trees for different leak sizes. And different possible leak outcomes, which include the possibility of no ignition at all, the possibility of a jet fire, and the possibility of an overpressure event are all considered for different leak sizes.
And we use those in conjunction with the estimate on a fatality probability in order to quantitatively calculate the risk for a given system for various leak sizes. And we can do that, we can calculate that overall system risk, at different distances away from the leak point, away from the hydrogen system itself. And at a certain point, that risk will fall below a level of acceptable or comparable risk. In this case, it's comparable to existing risk at gasoline stations, refueling stations, for example.
And then that can give us a risk-based distance. So this is one method that one could use to estimate a type of setback or safety distance is what is the distance at which the risk falls below a certain level. However, because there's so many different assumptions, model choices, and system parameters that go into these risk calculations, these risk-based distances can be highly sensitive to a number of factors. And we kind of illustrate that on the bottom right just by how much some of these purely risk-based distances can change with changing these different inputs to the risk calculations.
Some of these, like pipe diameter or system pressure, are fact functions of the actual design parameters of the system. But others are just unknown or variable parameters that go into the risk calculation itself. So even for a given system, these risk calculations can still be highly variable. That said, we can still use these types of risk calculations to inform what leak size we want to use as the basis.
And the way to do this is to take that sensitivity study that I showed on the previous slide. So essentially varying one parameter at a time, varying it above and below what we would consider a normal or nominal value. And we can see what effect that has on potential risk-based distances and therefore, equivalent fractional hole sizes. In that case, we would take the models that Ethan described, calculate the distance to those harm criteria, the same way we calculate the setback distances, and calculate the equivalent whole size that would give us the same distance as the risk-based distance that I showed on the previous slide.
And that sensitivity study is shown here on the bottom right. Each line is-- and there's a lot of them. There's 26 different lines. Some of them overlap, but each line is a different sensitivity case study in which we varied a particular parameter. And as you can see, most of the lines, including many of the overlapping lines, are clustered here in the low equivalent fractional whole sizes. So the low leak size basis values.
There's only a handful of lines that are really greater than a 5% value. And if we examine those particular cases in more detail, we see that they tend to assume somewhat unrealistic or overly conservative parameters or assumptions in the model choices. This includes assuming that a detonation occurs rather than just a desecration due to the outdoor leak. And it also assumes make some overly conservative assumptions with regard to the harm due to a fire, especially for low radiating hydrogen fires.
And so there's, of course, a couple of different options, but in general, we along with the task group and technical committee felt that the 5% value shown here by the horizontal line was a conservative choice-- was greater than the vast majority of the cases considered, and that the few cases that exceeded this value were overly conservative or unrealistic for one reason or another.
So now that we have an actual fractional leak size to use as a basis for our setback distance calculations, and we have the harm criteria for each of the different exposure groups that Ethan described, we now have everything we need to calculate these consequence-based setback distances and actually obtain values to put into the code. And so the way these setback distances are calculated is that for each exposure group, the distance to the different harm criteria are calculated independently for the different type sizes and system pressures considered.
And so in this case, first we'll calculate the distances for different system diameters, pipe diameters, and system pressures to the ignited concentration of 8% by volume. And we'll get those distances. We'll then calculate, for the same system, we'll calculate the distance to the heat flux criterion for the group 1 exposures. And then we'll calculate the distances to the overpressure criterion and for the group 1 exposures.
So now we have distances for each set of pressures, and what we're going to do is select the longest distance away from the leak point. And that will be the actual value used in the table values for the NFPA 2 code. So all of these different criteria are being considered, and then the longest distance out of each of them is selected for the setback distances. And we do that for each system pressure range as I said and then for a number of possible pipe diameters as well.
In this case, the distances for the group one exposures are, as you could see by those markers, we're driven by the distance to the 8% by volume on ignited concentration. We can do the same thing for the group 2 exposures, in this case, considering the heat flux and overpressure criteria for group 2 exposures that Ethan described previously. And we'll do the same type of calculation.
We look at the same pressure ratings and the same pipe diameters and calculate the distance to these different criteria. And then for each pressure rating, we'll select the longest distance for actual use in the table. In this case, it was the distance to the kilowatts per meter squared heat flux value that drove these distances for the group 2 exposures.
And then lastly, we can do the same thing for the group 3 exposures. This time considering both the visible flame length as well as the heat flux and overpressure criteria that were identified for group 3. And again, in this case, it's the distance to the heat flux criteria that ends up driving the setback distances for the table values.
Ethan: Thanks. So yeah, those are the calculations. And in order to make this digestible in NFPA 2, we tabulated those distances into a couple of tables and those are shown here on this slide. So the right-hand table there, and I don't expect you to read all the details of it, shows the three different exposure groups, what those exposures actually are. And then a single value for each of the three different pressures shown at the top of the table there. And so that's for a typical type size of an inch and a half inner diameter for liquid hydrogen systems.
And then on the left-hand side is a table that's similar. It once again has the three different pressure groupings and a single distance for each of those three different groups. But in this case, what's shown on the left-hand side of that table is four different pipe diameters ranging from half an inch inner diameter out to two inches inner diameters. If you look at the values, and this will become more clear on the next slide, you can see that the pressure ranges don't show a lot of variability in terms of what the distance actually is. But the pipe size can significantly affect the distance. If you want to move on to the next slide, please.
So that's shown more easily in this graphical form here. And so that upper right-hand chart shows what the previous distances were by the blue lines. And you can see that they varied from 50 feet to 75 feet for the individual exposures that are included in group 1 in this case. But then for that characteristic pipe size of an inch and a half and for 60 to 120 PSI system, that new setback distance is 48 feet. So slightly below that 50 foot setback distance for lot lines and ignition sources and a third below that 75 foot setback distance air intakes and wall, openings.
You can see that by reducing the pressure below 60 PSI, you gain a small amount of credit down to 44 feet. But by maintaining that same 60 to 120 PSI pressure and reducing the pipe size down to half an inch, that setback distance now reduces all the way down to 18 feet. Of course, the opposite is also true. If you look at the purple bar, that if you increase the pipe size and/or the pressure that setback distance can increase relative to the typical system.
A similar plot is shown for group 2 exposures on the lower left-hand side. You can see that those exposures were previously 25 and 75 feet. And now relative to the 25 foot setback distance that, setback distance has actually increased a bit out to 36 feet for a typical system. Of course, that does vary once again based on the exact system details shown by the different colored bars.
And then finally, for group 3, you can see that there was previously quite a few of the setbacks were actually down to 5 feet for group 3 exposures and now that's increased pretty significantly out to 31 feet for this typical system. Even for sort of a very small diameter piping system of a half an inch, that distance that was previously 5 feet for some exposures will have increased out to 15 feet.
But if you move on to the next slide, you can see that we also left in language related to credits for fire barrier walls and for insulated portions of the system, which can minimize the effect of those increased setback distances for some of these different exposures.
So we looked to the literature to determine the amount that fire barrier walls can reduce dispersion, heat flux, and overpressure. And there is quite a bit of data out there for gaseous hydrogen systems and very pretty much no data out there for liquid hydrogen systems. But the same features of the fire barrier walls should work to reduce these distances for liquid hydrogen systems because based on the same science.
So there's language in NFPA 2 2023 that does allow the reduction of distances for groups 1 to 2 by 50% by placing a fire barrier wall between the exposure and the bulk liquid hydrogen system. And this does include air intakes. So we looked at some modeling to show that the plume dispersion is reduced pretty significantly by a fire barrier wall being in the way. So we included that credit for fire barrier walls reducing the distance to air intakes in this case.
Fire barrier walls also enable group 3 distances to be reduced all the way down to zero feet. So the fact that they increase isn't a deal breaker for citing these systems. Vacuum-insulated piping, there was previously language in there enabling a reduction by 2/3 for this vacuum-jacketed piping. We thought that remained appropriate for this edition of NFPA 2 because this vacuum-insulated piping reduces the propensity for leaks due to the fact that it's double walls and welded joints.
We did clean up the language to ensure that we were talking about vacuum-jacketed insulation and with no mechanical connections joints or leak sources so that the justification behind that reduction was actually explicitly stated in the language there. And it should be noted that an emergency shutdown system is also required for all public refueling systems to enhance safety of those public facing systems. Next slide, please.
So the real impact of this is we attempted to show in this slide, of course, footprints are a function of a lot of details of the system. But a previous footprint there is shown on the left-hand side, and the large footprint in this case is largely driven by this large 75 foot setback distance between building openings and air intakes and the bulk liquid hydrogen system. And in this case, for a similar system that's 75 foot distance between those air intakes has reduced all the way down to just under 25 feet in the case of this system on the right-hand side using the 2023 edition of an NFPA 2.
The distance a lot lines has perhaps increased slightly depending on the exact system details from 16 and just-- 17 feet up to 25 feet. But you can see that this updated language should enable a reduced footprint for these liquid hydrogen systems using this 2023 edition. Next slide, please.
So we're getting towards wrapping up here but we just like to acknowledge the collaboration that took place to come up with these distances and do all these calculations. This work was part of the NFPA 2 Storage Test Group, which is part of the Hydrogen Technologies Technical Committee. And we want to especially thank the list of folks that are shown there on the right-hand side who participated in many different meetings, helped verify that what we were doing was based on good science and could be reproduced and that these distances are as good as we can come up with the current state of knowledge.
Next slide, please. So just to wrap up here. In the 2023 edition of NFPA 2, there are updated distances for separation distances between bulk liquid hydrogen systems and different exposures. These are simplified, defensible, and well-documented. This documentation enables the assumptions to be changed and incremental improvements to be made over time as the knowledge base increases.
This same framework, now that's documented, could be applied to other setback distances in the future, for example, gaseous setback distances could be revisited. One specific aspect of this work that wasn't included in the gaseous setback distances was the inclusion of unconfined overpressure criteria.
Finally, a couple of things that we think about and are working on is addressing larger systems that currently need science-based codes and standards. Your current limitation within NFPA 2 is about 20 metric tons of liquid hydrogen. For systems above that it requires an independent risk analysis and convincing of the authority having jurisdiction that that safety is included but it would be good to come up with some science-based codes and standards for these larger systems.
And we're also working to address some of the other science that we needed gathering additional data. There is quite limited data for liquid hydrogen systems. One specific issue is mitigations from fire barrier walls specific to liquid hydrogen dispersions in flames is sort of unknown and that's something we plan on studying. And I'll point out that the full report is available there. But I think there's a link in the chat now and we'd love to get feedback from audience members on this analysis. And any other questions you had, we'd be happy to address it. Thanks.
Cassandra: Thank you. Thank you, Brian and Ethan. We're going to work quickly through a few questions. So I'm going to ask you for quick answers so that we can try to get to as many of these great questions as we can. Just a reminder to folks, if you are looking for the Q&A, look for the little chat button and then look to the right. There are three dots, click on that and that will get you to the Q&A and that way we can follow up with you about your question later if we run out of time.
And then for those looking for links, there are quite a few links that have been shared throughout the course of the webinar. Feel free to scroll back in the chat and take a look at those. But other than that, let's just dive straight into questions. First of all, let's talk about whole size. We had a question about what is fractional hole size, perhaps you can give an example about how to think about hole size.
Brian: Yeah, so the fractional hole size is a ratio of fraction of the area of the leak versus the area of the flow diameter of the piping. And so one example there could be similar to if a leak were to be developed through a liquid hydrogen bayonet connector, you could consider the cross-sectional area of the actual flow path of the liquid hydrogen versus the cross-sectional area of a potential leak point.
And we did kind of some basic comparisons of what would it look like for various connector components if the o-ring were to fail for example. This is just one failure scenario, of course, but just to provide an example. And we kind of get a number of various different hole sizes that do tend to be kind of clustered below a 5% fractional area for a liquid hydrogen connector.
Laura: Great, following on along the same lines of talking about different equipment sizes, why are we not looking at piping above two inches? There's a detailed question in the Q&A that I'll encourage you guys to take a look at later but essentially, there are scenarios where pipe diameters above two inches are necessary. And why are we not looking at two inches or above two inches and how should we be thinking about risk and implementing piping above two inches?
Ethan: So it should be said that the hazard distances are fairly linear. So they can be extrapolated slightly outside that two-inch criteria. But this two-inch value was the number that we came up as with the task group with industry input as sort of a characteristic for these liquid hydrogen systems. I guess it should be noted that this is for these smaller systems under 20 tons. I imagine that those larger pipe diameters would only be required for systems with significantly larger volumes. So perhaps as we scale things up, we'll consider systems with larger diameters.
Laura: To that end, does HyRAM+ allow you to take that type of pipe diameter into consideration?
Ethan: Yeah, absolutely it does. It should be said that the simulations are not for a leak from a two inch pipe releasing hydrogen, but it's the 5% area of that 2 inch pipe. That being said, for these larger leaks as well, you're-- for the smaller leaks, there's good evidence that hydrogen will vaporize and start to disperse as a gas before it actually reaches the ground.
For larger leaks that there might be enough coldness in the flow that could actually form pool. And that's something that we're studying experimentally as is the pooling and dispersion from pools behavior. And will be including that in HyRAM as we gather the data.
Laura: Great. Continuing on our conversation about hole size, we had a question about, explain the hole size for liquid hydrogen versus the gaseous hydrogen setback, the basis.
Brian: Yeah, absolutely. So these were different efforts, and the previous effort was done around 10 or 15 years ago for bulk gaseous hydrogen storage. It used a similar type of approach in trying to quantify the risk for a typical or representative type of system. And then ultimately made the decision that a 3 or 10% fractional hole size encompass the vast majority of the potential leak sizes, over 99% of possible leaks, due to the leak frequencies within the risk assessment.
By contrast, in future code revision cycles, the technical committee decided to lower that value to the 1% fractional leak area given the additional conservatism built into the model. It's also worth mentioning, and there's a little bit of an apples and oranges comparison here because the distances for the bulk gaseous setback distances also utilized a safety factor.
So they calculated the distances and then added a safety factor onto the distances once they were calculated using that 1% leak size. Whereas in this case, even though a lot of the same conservative assumptions were made, the committee and the task group decided to just choose a more conservative leak size and then just calculate the distances directly rather than choosing a leak size and a safety factor separately.
Laura: I think we have time for one more quick question. There's a question, is there an assumed system pressure in the new standard when calculating the distances?
Ethan: So the distances are a function of the pressure. So there's three different pressure bins up to 60 PSI, 60 to 120 PSI, and then 120 PSI up to the critical pressure for liquid hydrogen system, because above the critical pressure, you will no longer have a liquid hydrogen system. You'll have a supercritical hydrogen system.
Laura: Great. I thought we might be able to squeeze that in just before the hour. We have hit the end of our time. We have a lot of great questions in the Q&A still. We have everybody's email addresses and we will ask Brian and Ethan to follow up directly with folks later. In the meantime, I'm going to hand it over to Cassie.
Cassandra: Yeah, thank you, Laura. That concludes our H2IQ call for today. Once again, I want to thank all of our presenters. And a recording of today's webinar is going to be available as well as the presentation materials today. Those will be available within the coming weeks. And be sure to subscribe to HFTO news with the link that I post in the chat. But thank you all for attending. We look forward to seeing you at the next H2IQ hour.