Nov. 4, 2025

225 - Battery Energy Storage Systems with Noah Ryder

The player is loading ...

Demand for the energy storage is as high as ever, and is about to triple-quadruple. The development of technology is at unprecedented phase, and even within a single project you may face different cell, battery or container generations. This pace reshapes how we think about battery energy storage safety, from enclosure design to emergency response. We sat down with Noah Ryder from the Fire and Risk Alliance to unpack how BESS has evolved from walk-in containers to dense, modular “refrigerator” units—and how the move to liquid cooling, tighter layouts, and higher amp-hour cells impacts both opportunity and risk.

We explore the real jobs batteries do for the grid: shifting solar and wind, replacing peaker plants, stabilizing frequency, and powering microgrids. Then we zoom into the fast-growing edge case: AI-hungry data centers integrating batteries at the rack level for modularity and speed. That flexibility has a cost. Less free airspace and larger cells mean faster gas accumulation, higher heat flux into insulated enclosures, and a credible explosion hazard from a single failure. We walk through the failure timeline—monitoring anomalies, venting, immediate versus delayed ignition, sustained fire, and potential propagation—and identify practical interventions at each step.

Noah lays out the tradeoffs many teams avoid: accept that a damaged unit is a write-off, or try to save modules at all costs? Should we prefer a known flame over an uncertain blast by using intentional spark ignition? How should NFPA 855’s push toward gas-triggered mechanical ventilation and deflagration venting influence spacing, panel placement, and vent direction? We also dig into enclosure construction—non-combustible insulation, steel skins, coolant flammability—and how better insulation can safely cut spacing by slowing heat penetration and reducing internal temperature rise.

Looking forward, stacking feels inevitable. The smarter approach is to treat batteries not just as a cause but as a fuel, borrowing tested methods from high-rack storage: quantify heat release and radiant exposure, model gas evolution and overpressure, orient vents to manage flame jets, and define acceptable loss before design begins. If you care about real-world energy storage—utility sites, microgrids, or data centers—you’ll leave with a clearer framework to make informed, defensible choices.

If you would like to learn more about Noah and the Fire and Risk Alliance, you can find them online here: https://fireriskalliance.com/

Enjoy the conversation, then subscribe, share this episode with a colleague, and leave a review to help more engineers find the show.

----
The Fire Science Show is produced by the Fire Science Media in collaboration with OFR Consultants. Thank you to the podcast sponsor for their continuous support towards our mission.

00:00 - From Hong Kong To Battery Safety

02:35 - Why Energy Storage Is Booming

06:10 - Power vs Energy Explained

09:06 - What Batteries Do For The Grid

12:10 - Data Centers And On‑Rack Batteries

17:04 - Flexibility Versus External Containers

21:10 - Generations Of BESS Topologies

25:10 - Free Airspace And Cooling Shifts

28:20 - Enclosure Materials And Insulation

31:40 - Spacing, Heat Flux, And Propagation

35:05 - Acceptable Loss And Risk Framing

39:00 - Cell Chemistry, Form Factors, Scaling

42:00 - Failure Sequence: Vent To Fire

46:20 - Igniters, Explosions, And Tradeoffs

50:10 - Ventilation, Panels, And NFPA 855

WEBVTT

00:00:00.080 --> 00:00:02.160
Hello everybody, welcome to the Fire Science Show.

00:00:02.160 --> 00:00:08.880
I've just came back from Hong Kong where I've attended the fourth International Lithium Battery Fire Safety Symposium.

00:00:08.880 --> 00:00:17.440
What an excellent conference that was congratulations to Professor Xinyan Huang and his team at the HK PolyU for organizing this fantastic, fantastic conference.

00:00:17.440 --> 00:00:29.839
I've learned a lot about battery fire safety and also while being there, I've picked some interesting topics or interesting presentations that were shown that I thought could be of high use to the fire science show community.

00:00:29.839 --> 00:00:35.119
Some I've recorded on place, some of I've got the speakers agreed to give me an interview in future.

00:00:35.119 --> 00:00:43.359
But regardless, I'll be releasing them because battery fire safety is one of the most interesting parts of fire safety right now.

00:00:43.359 --> 00:00:49.280
If you like it or not, it's something coming for all of us and we need to be prepared and need to be aware of what's happening.

00:00:49.280 --> 00:00:56.240
While there was a ton of talks about electrolytes, uh separators, uh chemistries, modeling, etc.

00:00:56.240 --> 00:01:02.079
I was looking for talks that are practical and directly useful to our craft.

00:01:02.079 --> 00:01:17.920
And today, first of such talks, my interview with uh Noah Ryder from Fire and Risk Alliance who was talking about battery energy storage systems, BSS, and those are popping like crazy everywhere.

00:01:17.920 --> 00:01:22.319
I had a skewed idea of what a battery energy storage system looks like.

00:01:22.319 --> 00:01:35.280
I had some ideas how to give a safety to the ones that I had in mind, but as the structure evolves, as the idea, the concept of a storage system evolves, uh and the solutions must evolve as well.

00:01:35.280 --> 00:01:50.560
In this talk, I've tried to pick Noah about how those storage systems are constructed, what drives their design, what drives the way, what shapes them, how people decide how to construct them and how to join them into bigger, larger energy storage parks.

00:01:50.560 --> 00:01:56.719
Because all of those are important considerations when we want to provide the fire safety of such a storage system.

00:01:56.719 --> 00:02:02.319
An important question that a fire engineer has to ask what is considered a safe storage system?

00:02:02.319 --> 00:02:15.120
At what point we believe we have provided safety for those devices because this will also also determine how we will construct the safety systems within the cells, batteries, modules, or entire devices.

00:02:15.120 --> 00:02:19.680
So well, anyway, I've talked enough and let's give the microphone to Noah.

00:02:19.680 --> 00:02:24.000
This is recorded in live at the Hong Kong Polytechnic University campus.

00:02:24.000 --> 00:02:27.680
It was a great discussion, and I hope it's a nice listen for you.

00:02:27.680 --> 00:02:30.000
Let's be an intro and jump into the episode.

00:02:30.000 --> 00:02:36.000
Welcome to the Fire Science Show.

00:02:36.000 --> 00:02:39.520
My name is Wojciech Wegrzynski and I will be your host.

00:02:39.520 --> 00:03:22.719
As the UK leading independent fire consultancy, OFR's globally established team have developed a reputation for preeminent fire engineering expertise with colleagues working across the world to help protect people, property and the planet.

00:03:22.719 --> 00:03:39.039
Established in the UK in twelve sixteen as a startup business by two highly experienced fire engineering consultants, the business continues to grow at a phenomenal rate with offices across the country in eight locations, from Edinburgh to Bath and plans for future expansions.

00:03:39.039 --> 00:03:47.840
If you're keen to find out more or join OFR consultants during this exciting period of growth, visit their website at OFR Consultants.com.

00:03:47.840 --> 00:03:49.919
And now back to the episode.

00:03:49.919 --> 00:03:51.039
Hello everybody.

00:03:51.039 --> 00:03:55.039
I am joined today by Noah Ryder from Fire and Risk Alliance.

00:03:55.039 --> 00:03:56.960
Hey Noah, good to have you in the podcast.

00:03:56.960 --> 00:03:58.080
Good to be here.

00:03:58.080 --> 00:03:58.960
Thanks for having me.

00:03:58.960 --> 00:04:01.439
Yeah, thanks for agreeing to do it the live.

00:04:01.439 --> 00:04:07.840
Yay, we're live in Hong Kong at the beautiful um Fourth International Symposium on Battery Fire S afety.

00:04:07.840 --> 00:04:09.039
Are you enjoying it so far?

00:04:09.280 --> 00:04:09.919
It's been great.

00:04:09.919 --> 00:04:10.560
It's been great.

00:04:10.560 --> 00:04:14.080
City is lovely and uh a lot of good technical presentations.

00:04:14.080 --> 00:04:16.399
So uh enjoying my first trip here, actually.

00:04:16.639 --> 00:04:26.720
You know, I really love how when I go for tunneling conferences to Germany because the Swiss German guys are the ones who are building tunnels and have good technology, you learn from them.

00:04:26.720 --> 00:04:30.720
And I can say the same thing about uh going for a battery conference to China.

00:04:31.120 --> 00:04:46.079
There are definitely quite a lot to learn, and uh, I think you know, just from what we've seen in terms of the sheer number of papers here and high quality ones, it's uh a lot obviously a lot of the manufacturing is happening here, so it makes sense that there's quite a bit of uh good information coming out.

00:04:46.399 --> 00:04:57.279
Manufacturing investment, you know, they it feels really strategic in here, and I think they are doing it on a purpose, and I think they understand the purpose very well.

00:04:57.279 --> 00:05:03.360
And I mean, it's also highly we're gonna talk about battery energy storage systems, so but this is so relevant.

00:05:03.360 --> 00:05:12.160
Like I see the developments happening in here because they understand the sheer amount of battery energy storage that we would probably like.

00:05:12.160 --> 00:05:13.040
What do you think?

00:05:13.279 --> 00:05:13.920
Uh 100%.

00:05:13.920 --> 00:05:26.639
I think the you know, as you kind of pointed out, there's been some very, I'd say, high-level strategic thinking and planning in terms of the battery development and just the industries that the batteries can serve.

00:05:26.639 --> 00:05:28.399
So obviously there's the EV side of it.

00:05:28.399 --> 00:05:32.319
I'm more focused, or today we're going to talk more about the energy storage side of it.

00:05:32.319 --> 00:05:40.480
But you know, the amount of effort that's going into improving technologies, manufacturing, safety across the board, I think is pretty impressive.

00:05:40.480 --> 00:05:55.839
I think for them, it's such a critical part of their infrastructure and you know products that they're selling worldwide that they have an inherent interest in safety in a way for the industry in a way that other maybe standalone companies or countries may not have.

00:05:55.839 --> 00:05:58.000
So it's it's it's it's neat to see.

00:05:58.000 --> 00:05:58.240
Yeah.

00:05:58.560 --> 00:06:03.360
If we talk about battery energy storage systems, uh, how many batteries do we need?

00:06:03.360 --> 00:06:05.680
Like what scale are we even talking about?

00:06:05.680 --> 00:06:10.480
Because I I see numbers in megawatts, gigawatts, uh, gigawatt hours.

00:06:10.480 --> 00:06:13.439
It's like a very abstract unit to me.

00:06:13.439 --> 00:06:14.480
How big is that?

00:06:14.720 --> 00:06:16.480
Yeah, no, I mean it's it's a great question.

00:06:16.480 --> 00:06:26.639
So I mean, I think in terms of the the of how many we need, I I think the short answer is if we could snap our fingers and have as much as we need today, we would.

00:06:26.639 --> 00:06:39.839
Um, but when we when we look at that in terms of actual quantities, you know, there's if you start at the base level, just to give some idea, there's we're producing probably in the order of about 10 billion new battery cells each year.

00:06:39.839 --> 00:06:44.720
Not all of those obviously are going towards energy storage, but an increasing number is.

00:06:44.720 --> 00:06:54.000
And so right now within the US, we're sitting at around probably close to about 80, maybe 90 gigawatt of installed capacity.

00:06:54.000 --> 00:06:57.759
China is uh around 230 or so.

00:06:57.759 --> 00:07:05.439
And over the next three to five years, they're expecting that to probably triple the kind of the baseline case.

00:07:05.439 --> 00:07:13.279
And if we kind of push expectations, you may get to you know four or five times uh more installed capacity than we have today.

00:07:13.279 --> 00:07:20.000
So you might be looking at a terawatt hour of uh installed capacity in China by you know 2027.

00:07:20.399 --> 00:07:30.399
So when we define this in gigawatts or or watt hours, the gigawatts reflect to how much you can instantaneously take out of it and what hours means how much it contains?

00:07:30.800 --> 00:07:30.959
Yeah.

00:07:30.959 --> 00:07:42.560
So I mean generally the way you'll see it is you'll hear like a if you talk about an energy storage system, they'll say, oh, this is a hundred megawatt site, or it'll be uh uh 400 megawatt hours.

00:07:42.560 --> 00:07:51.279
And so really the key is is like what is your duration of operation and then what you can simultaneously discharge to the system.

00:07:51.279 --> 00:07:56.079
So some cases they're looking at like a four-hour system, in some cases it's a one-hour system.

00:07:56.079 --> 00:08:04.800
It really just depends on kind of the use case that's there, but that's why uh almost always uh a specific energy storage site will have two numbers associated with it.

00:08:04.800 --> 00:08:09.920
It'll be you know a 100 megawatt, 400 megawatt hour or so for if they're you know, those kind of things.

00:08:10.160 --> 00:08:10.319
Okay.

00:08:10.319 --> 00:08:16.319
Um a power bank is like 50, 60 watt hours, a car is like kilowatt hours.

00:08:16.639 --> 00:08:21.600
Yeah, I mean a car battery, like for a big car, you know, battery pack is a hundred kilowatt hour pack.

00:08:21.600 --> 00:08:22.480
Okay, that's a big car.

00:08:22.480 --> 00:08:23.360
Yeah.

00:08:23.439 --> 00:08:26.319
Um that's the one that gives you 600 kilometers or 400 miles.

00:08:26.959 --> 00:08:30.079
This is like top of the line, go as far as you can, kind of thing.

00:08:30.079 --> 00:08:32.000
And then by scale, right?

00:08:32.000 --> 00:08:37.759
I mean, not that many people have them anymore, but you know, your standard incandescent light bulb is 100 watts.

00:08:37.919 --> 00:08:38.159
Yeah.

00:08:38.399 --> 00:08:44.320
So, you know, when you start looking at scales, um, these are massive systems that we're talking about.

00:08:44.320 --> 00:08:48.639
Obviously, they have to be because they're supporting entire grid infrastructures and homes.

00:08:48.639 --> 00:09:00.159
So most of the ones that we talk about today in terms of size are sufficient to power tens of thousands of homes, sometimes up to you know, a hundred thousand homes, just depending on on where it's at.

00:09:00.480 --> 00:09:02.080
What was the purpose of those?

00:09:02.080 --> 00:09:03.600
What the role do they play?

00:09:03.600 --> 00:09:08.240
Because we we didn't have them 10, 20 years ago and we were just fine.

00:09:08.240 --> 00:09:11.360
Like I know we store the energy in mountains.

00:09:11.360 --> 00:09:16.000
Like you have build a water tank on top of a mountain, you pump water up, then it goes down.

00:09:16.000 --> 00:09:22.080
I I understood that it's it's to conserve some energy when it's cheaper and then spend it when there's a peak.

00:09:22.080 --> 00:09:24.320
What what role do those devices play now?

00:09:24.639 --> 00:09:33.120
So, in some regards, they play a similar role as kind of your hydraulic uh type systems that were storing in based on water or flywheel type systems and more mechanical basis systems.

00:09:33.279 --> 00:09:33.759
Oh, I love those.

00:09:33.759 --> 00:09:34.159
I love those.

00:09:34.320 --> 00:09:37.840
Yeah, so I mean the the concept of energy storage is not new.

00:09:37.840 --> 00:09:38.480
Yeah, yeah.

00:09:38.480 --> 00:09:49.840
What has changed over the past 20 or 30 years, I think, one is is our is our needs, our power needs have skyrocketed and the grid infrastructure has generally not kept up.

00:09:49.840 --> 00:09:54.000
So we haven't been building uh, you know, quite as many power plants and everything else.

00:09:54.000 --> 00:10:00.720
But what has happened is we've built more uh wind and solar power, which is great.

00:10:00.720 --> 00:10:08.879
And the the biggest or one of the bigger challenges I'd say is that if there was excess power that was generated, it just literally went to ground.

00:10:08.879 --> 00:10:14.159
And so energy storage in particular batteries kind of fills that gap.

00:10:14.159 --> 00:10:16.159
And so it can take all that extra.

00:10:16.159 --> 00:10:19.759
It helps you uh not need to spin up like a peaker plant.

00:10:19.759 --> 00:10:22.799
Um, in some cases, you may not have to build the plant at all.

00:10:22.799 --> 00:10:34.720
So it really kind of fits this gap in needs where if there's excess power generation, or well, and that's typically at night or you know, other low use times, then you can store all that for later.

00:10:34.720 --> 00:10:37.360
And so that's really where it's kind of filled in.

00:10:37.360 --> 00:10:44.960
Um, the other big use case that we're seeing now is for entities that want to be more energy independent.

00:10:44.960 --> 00:10:46.639
They're doing their own microgrids.

00:10:46.639 --> 00:11:02.720
And so they they may have a tie and get power from from the made infrastructure, but they have the ability to basically island themselves and they can operate for anywhere from several hours to, in some cases, several days without any outside uh uh uh energy needs.

00:11:02.720 --> 00:11:05.600
So that's kind of a pretty wide range, but yeah.

00:11:05.919 --> 00:11:09.440
In your from your talk, I captured that a lot of this development.

00:11:09.440 --> 00:11:19.519
Obviously, the renewables, like you mentioned just before, is is one of the drivers, but also the data centers and AI, which are extreme uh power consumers.

00:11:19.519 --> 00:11:23.600
What role does play in in those resiliency uh or yeah?

00:11:23.679 --> 00:11:37.519
So I mean data centers are an interesting one because we see, you know, I I think I I saw a recent uh uh number just the other day that the kind of the big guys, if you will, that they spent 400 billion dollars on AI infrastructure in 2025 alone.

00:11:37.519 --> 00:11:40.240
Like a hardware, yeah, yeah, just everything.

00:11:40.240 --> 00:11:45.840
And it's just so I mean I think a lot of it's been the most of it's or a good chunk has been on the hardware because it's it's really expensive.

00:11:46.240 --> 00:11:50.159
I know I bought one GPU for computation, it's like just one.

00:11:50.399 --> 00:11:51.440
Yes, yeah, exactly.

00:11:51.440 --> 00:11:53.759
Yeah, sometimes you know it's like uh it's tough.

00:11:53.840 --> 00:11:57.279
But um I preferred the times when Bitcoin was driving it, you know.

00:11:57.279 --> 00:11:59.279
They those guys needed less.

00:11:59.440 --> 00:12:06.639
Yeah, yeah, and it it was bad enough then, but no, it's definitely it's kind of uh order of magnitude and um they're power hogs.

00:12:06.639 --> 00:12:07.039
Yeah.

00:12:07.039 --> 00:12:10.720
So we kind of see two things happening in the data center space.

00:12:10.720 --> 00:12:25.679
I'd say one is is kind of still kind of the the you know standalone energy storage systems, kind of your standard boxes that are outside, or replacements for your lead-acid batteries that are kind of in your mechanical rooms or wherever, your your UPS applications.

00:12:25.679 --> 00:12:33.759
The other it that's happening that is, I think, frankly, more interesting from a fire safety standpoint is almost everybody is integrating them directly into racks.

00:12:33.759 --> 00:12:40.240
Um so it's you may have a a rack that's half filled with GPUs and then half filled with batteries.

00:12:40.480 --> 00:12:40.799
Okay.

00:12:41.200 --> 00:12:45.919
And they're scaling those out as the as the data center is being filled in.

00:12:45.919 --> 00:12:57.600
So the white space might take a year to fill in, and the stuff that uh gets built first may use one battery design, and by the time they get to the end, it may be a different one.

00:12:57.600 --> 00:12:58.639
So uh yeah.

00:12:58.960 --> 00:13:04.559
I mean, uh that goes a little bit uh away from the subject of this interview, which I hope to be the battery energy storage.

00:13:04.559 --> 00:13:07.039
But it is interesting.

00:13:07.039 --> 00:13:09.200
It's very interesting uh what drives this need.

00:13:09.200 --> 00:13:12.879
I I mean I am of a belief, um we we are talking for the first time.

00:13:12.879 --> 00:13:24.879
I'm of a belief that the fire engineers need to very well understand the reasons why people do things because that allows them to do fire engineering efficiently, and otherwise we just ban stuff to people and they hate us.

00:13:24.879 --> 00:13:32.639
So I I would love to understand why such uh dispersion of the net why put it on the rack?

00:13:32.639 --> 00:13:36.960
What's the benefit of putting it on a rack versus having a container outside of your building?

00:13:36.960 --> 00:13:44.639
Because for me as a fire engineer, hey, your computer is like pretty expensive, like, and this is like pretty fiery.

00:13:44.639 --> 00:13:46.320
Uh how about the move it outside?

00:13:46.799 --> 00:13:49.120
So I think there's the pros and cons discussion.

00:13:49.120 --> 00:13:54.159
Yeah, I think the other aspect is that there's the knowledge and what information they have.

00:13:54.159 --> 00:14:02.639
So I think one of the challenges and one of the things we've seen is is sometimes there's really not a whole lot of thought given to, oh, we're just gonna do this because why not?

00:14:02.639 --> 00:14:03.120
Okay.

00:14:03.120 --> 00:14:11.759
And and so the potential that it may increase the risk to their facility is kind of has been glossed over.

00:14:11.759 --> 00:14:12.240
Okay.

00:14:12.240 --> 00:14:16.000
Um, and and this is going a little bit far afield, but not entirely.

00:14:16.000 --> 00:14:29.679
I uh so if you look at the history of what kind of the cost of the IT equipment has been as you put it into these facilities, if you look 10 years ago at a standard kind of search type of of equipment, it was nominally $10,000 per square foot.

00:14:29.679 --> 00:14:30.080
Okay.

00:14:30.080 --> 00:14:32.720
Assuming you had an eight to ten foot tall rack.

00:14:32.720 --> 00:14:33.200
Okay.

00:14:33.200 --> 00:14:37.600
Now it's three hundred to five hundred thousand dollars per square foot.

00:14:37.600 --> 00:14:53.039
So we we've seen two orders of magnitude increase in the cost of the equipment, but the risk in the thought process in terms of how we approach it from fire safety and what the kind of the concerns are hasn't caught up yet.

00:14:53.039 --> 00:14:57.279
So I I think as far as an industry goes, they're they're starting to get up to speed on that.

00:14:57.279 --> 00:14:58.159
I think that's there.

00:14:58.159 --> 00:15:00.240
And I think that will actually help drive it.

00:15:00.240 --> 00:15:03.519
And there may be some push to do uh some more outside.

00:15:03.519 --> 00:15:11.600
But to answer your the the the original question, part of it is if I put a box outside, I think nobody would would disagree that it's safer.

00:15:11.600 --> 00:15:16.159
However, it means that I'm stuck with that box and whatever it can do for the next 20 to 30 years.

00:15:16.159 --> 00:15:16.559
Okay.

00:15:16.559 --> 00:15:18.879
And I have to pay that infrastructure cost up front.

00:15:18.879 --> 00:15:20.559
I gotta do it, I gotta have it all installed.

00:15:20.559 --> 00:15:26.080
If I can build it out within my white space, I can build out something today for my use today.

00:15:26.080 --> 00:15:30.240
And if my needs change six months from now, I can build that out to address that.

00:15:30.240 --> 00:15:32.320
And I can continue to do that.

00:15:32.320 --> 00:15:40.639
And so, you know, if a project takes three years to complete, what was installed day one may look very different than what was installed day three year year three.

00:15:40.639 --> 00:15:47.600
So it gives them a lot of flexibility uh to adapt to you know, ultimately what their very modular design.

00:15:47.600 --> 00:15:50.080
Yeah, it's it's I I have a fondness for Legos.

00:15:50.080 --> 00:15:55.120
Yeah, and so I always like to think of things in terms of like, hey, this is this is modular, it's blocky.

00:15:55.120 --> 00:15:57.120
Can I put it together and take apart easily?

00:15:57.120 --> 00:15:59.759
And I think that's kind of how they're approaching it.

00:15:59.759 --> 00:16:01.759
Um it's a big benefit too.

00:16:02.000 --> 00:16:08.720
I I thought there will be more direct answer, like, oh yeah, the cable cost is like this, or something, but it's it seems multifactual.

00:16:09.120 --> 00:16:22.240
I I think it's yeah, I think it really comes down to as opposed to a utility type architecture where they're looking at a 30-year time scale for this piece of equipment, they may install it, run it for two or three years, and rip it out and replace it with something new.

00:16:22.399 --> 00:16:22.639
Yeah.

00:16:22.799 --> 00:16:30.799
Uh so they're I don't want to say they're not cost driven, but cost is the secondary or tertiary level you know thing.

00:16:31.039 --> 00:16:54.240
I I mean, I'm sometimes uh not very optimistic, and what they see here is the steam engine case, you know, when the the first industrialization came up, you had the massive steam engine in the middle of your plant, you built everything around the steam engine, then those stuff those things started burning and you start losing a lot of money, and here comes factory mutual and sprinklers and everything, you know.

00:16:54.240 --> 00:16:56.720
So I wonder if this is the same pathway.

00:16:56.720 --> 00:16:57.440
I hope not.

00:16:57.440 --> 00:17:00.639
Uh okay, let's move to the battery energy storage systems.

00:17:00.639 --> 00:17:03.759
In your talk, you gave an overview of topologies.

00:17:03.759 --> 00:17:09.680
I must say, for me, if I think about battery energy storage system, I see a container to which I walk in.

00:17:09.680 --> 00:17:13.680
There are like racks of devices on the walls.

00:17:13.680 --> 00:17:16.480
From your talk, I already know I'm I'm antique.

00:17:16.480 --> 00:17:24.960
Uh so uh let's perhaps walk the listeners through the existing uh technologies and generations of technologies in energy storage systems, how they evolved.

00:17:25.200 --> 00:17:26.480
Yeah, so that's a great question.

00:17:26.480 --> 00:17:30.160
Um so I think if you if you go back, you know, say a a decade.

00:17:30.640 --> 00:17:31.839
Oh, that's ancient times.

00:17:32.160 --> 00:17:37.839
It is, it is, but 10, you know, 10 years was you know, kind of the the state of the art was what you just talked about.

00:17:37.839 --> 00:17:43.519
So it was kind of a a walking container, maybe a building, right?

00:17:43.519 --> 00:17:50.319
So I mean, obviously, most folks that have some knowledge on on energy storage, they're aware of the the surprise Arizona, the McMicken incident.

00:17:50.319 --> 00:17:51.200
That was a building.

00:17:51.200 --> 00:17:54.079
Um, and we can talk about that later a little bit if need be.

00:17:54.079 --> 00:18:05.519
But but in short, there were structures that you could walk into that, as you said, they had batteries lining the walls and for accessibility, maintenance, HVAC purposes, everything, you know, it had some open space and so forth.

00:18:05.519 --> 00:18:19.440
As things started to develop and real estate became more valuable, and in particular, as it transitioned from, say, like utilities building these to more development, you know, type of uh operation.

00:18:19.440 --> 00:18:32.880
So, you know, private equity-backed uh organizations or or whomever that are basically building utility scale, they they kind of looked at it and said, well, why am I wasting you know, effectively half of my space, my surface area?

00:18:33.279 --> 00:18:35.200
So it really matters from a cost perspective.

00:18:35.440 --> 00:18:37.039
Uh tremendously, yeah.

00:18:37.039 --> 00:18:39.920
I mean, if you have unlimited cheap land, it it doesn't really matter.

00:18:40.160 --> 00:18:48.240
I mean, I again it's my distorted image of a data of a data center of something being built in like desert Arizona, like you know, with nothing around.

00:18:48.720 --> 00:18:55.920
So there certainly is some of that, but I think you know, the the locations are also have to be paired with substations and things that are going to support the grid.

00:18:55.920 --> 00:18:59.440
And they increase in density build around them, it's a small city now.

00:18:59.440 --> 00:19:00.000
Yep, yep.

00:19:00.000 --> 00:19:16.480
So so the the you know, your your density per square foot matters a lot, which is why, you know, during this conference, if you see, you know, some people have talked about it a little bit, but you'd see like battery density and what does that translate into in terms of energy per square meter or things like that.

00:19:16.480 --> 00:19:41.440
So what happened was is that you know, I think partly because on on market needs, partly due to manufacturing and transportation, and everything kind of came together and people started moving away from the walk-in style enclosures to self-contained, kind of more reach-in, where it's it may still have been an ISO container, like generally a 20-foot ISO, but they loaded it up with racks, and you could open the doors and reach in, but you can't be occupied.

00:19:41.759 --> 00:19:44.240
So you open it from outside, and there's immediately your batteries.

00:19:44.400 --> 00:19:44.960
Yep, exactly.

00:19:44.960 --> 00:19:50.960
It basically you could think of it just like you know, they almost look just like server racks, and you just have a fully popular thing.

00:19:50.960 --> 00:19:51.759
Yeah, exactly.

00:19:51.759 --> 00:19:57.440
And so the only way to access it is by opening the doors and and pulling the modules in and out.

00:19:57.440 --> 00:20:04.000
So that that started really happening probably six, seven years ago, and is pretty much the default today.

00:20:04.000 --> 00:20:15.599
What we've seen over the past several years is a little bit of a migration uh with more companies moving towards instead of your 20-foot or whatever shipping container style into more modular.

00:20:15.599 --> 00:20:18.079
So I kind of call them the refrigerators.

00:20:18.079 --> 00:20:28.000
And so they're all self-enclosed, but I can take three, four, five, ten, whatever it is, refrigerators, sandwiching them together, and that becomes my unit.

00:20:28.000 --> 00:20:32.400
Um so that is where a lot of entities are moving today.

00:20:32.400 --> 00:20:37.680
If you see some of the new products that are coming out, you know, over the past year or so and and stuff that's been announced.

00:20:37.680 --> 00:20:37.920
Yeah.

00:20:38.319 --> 00:20:43.519
Is this uh refrigerator unit uh a wholly self-contained battery energy storage system?

00:20:43.519 --> 00:20:46.240
Like does it have management, cooling, everything?

00:20:46.480 --> 00:20:47.119
They do, yeah.

00:20:47.119 --> 00:20:54.400
So generally speaking, they they may be linked with the other units or the refrigerators next to them, but they are fully self-contained.

00:20:54.400 --> 00:20:56.720
Um, they may share like an inverter.

00:20:56.720 --> 00:21:03.279
Um, so depending on who makes it, there may be some shared resources that are like a master unit and and and storage units.

00:21:03.279 --> 00:21:04.160
Correct, yeah.

00:21:04.160 --> 00:21:08.400
And uh so it really just depends on how the OEM is configured it.

00:21:08.400 --> 00:21:19.359
But generally speaking, they have at the kind of refrigerator level, they've got the full insight from the BMS controls, they can shut it down, they can do whatever else, independent of the other ones there.

00:21:19.599 --> 00:21:22.559
Uh how did uh free airspace inside evolve?

00:21:22.559 --> 00:21:25.519
I mean, I obviously decreased, but to what extent?

00:21:25.920 --> 00:21:31.920
Well, obviously, so I mean if you had the walk-in enclosure, you you may have had 50, 60, 70% free airspace.

00:21:31.920 --> 00:21:37.359
The systems that we're looking at today are nominally 20% free airspace.

00:21:37.359 --> 00:21:46.240
And so it's uh dropped drastically in terms of uh how much is there, and and that obviously impacts the safety system and anything else.

00:21:46.480 --> 00:21:50.000
I mean to start with cooling, did cooling change significantly also in those?

00:21:50.000 --> 00:21:51.119
So yeah.

00:21:51.279 --> 00:21:55.359
So I mean, basically, there's you can either do air cooling or liquid cooling.

00:21:55.359 --> 00:22:01.039
We've seen an increasing number of folks actually move away from air-cooled systems to liquid cooled systems.

00:22:01.039 --> 00:22:10.160
And so, you know, depending on what approach you've taken uh from a cooling perspective, then obviously the free airspace will impact that as well.

00:22:10.160 --> 00:22:12.480
That's almost entirely on the operational side.

00:22:12.480 --> 00:22:22.480
And one of the key things is that even if you were using an air-cooled system, it's technically supposed to be independent of any mechanical ventilation system you may have for safety.

00:22:22.480 --> 00:22:32.559
But you know, it it's I'd say largely dominated by liquid-cooled systems these days, and they've kind of moved away from the HVAC chiller and kind of approach.

00:22:33.039 --> 00:22:40.240
And the casings themselves, the technologies that we work with, uh what's the material of all of the casing?

00:22:40.240 --> 00:22:42.079
Is it just like a sheet of metal?

00:22:42.079 --> 00:22:43.519
Is it like insulated?

00:22:43.519 --> 00:22:49.039
I mean, it has to be weather controlled and weatherproof, but probably has to have some specific characteristics.

00:22:49.359 --> 00:22:50.880
Yeah, no, that's a good question.

00:22:50.880 --> 00:22:55.440
So most of them are you know steel structure externally.

00:22:55.440 --> 00:22:59.680
They at this point in time almost all have some level of insulation.

00:22:59.680 --> 00:23:06.640
So not driven through safety, uh, but through operations you know, want to keep the temperature and the environment pretty consistent.

00:23:06.640 --> 00:23:14.000
So most of them will have a couple inches of of some type of insulation, like a rock wall type insulation or something else that's there.

00:23:14.000 --> 00:23:18.240
Uh, there was a period where folks had some like polyurethane insulation and things like that.

00:23:18.240 --> 00:23:20.640
Um, fortunately, we've seen a lot of that go away.

00:23:20.880 --> 00:23:25.599
Uh you need to add like a cavity to it, just maybe cover with ACP.

00:23:25.599 --> 00:23:27.039
You're gonna have a very beautiful system.

00:23:27.039 --> 00:23:27.839
Yeah, yeah.

00:23:27.920 --> 00:23:31.680
No, it's something that just, you know, if you just want to make the incident as big as possible, go for it.

00:23:31.680 --> 00:23:32.880
A CLT roof.

00:23:32.960 --> 00:23:33.519
Uh that would be awesome.

00:23:33.759 --> 00:23:34.319
It'd be great.

00:23:34.480 --> 00:23:34.960
Um yeah.

00:23:34.960 --> 00:23:38.960
I mean, we could have a lot of fun with it designing something that would uh you know really burn well.

00:23:38.960 --> 00:23:44.720
But no, no, fortunately, I think they they have moved towards kind of the non-combustible uh insulation.

00:23:44.720 --> 00:23:48.880
And and that generally is covering you know all the walls and in the roof deck for the most part.

00:23:48.880 --> 00:23:56.000
And but that does have a really important impact on the safety calculations uh that they come downstream as well.

00:23:56.000 --> 00:23:58.319
Um, obviously, it's kind of like your oven.

00:23:58.319 --> 00:24:03.680
If I've you know if it does a good job of of insulating and it's gonna keep the heat out if we've got insulating, yeah.

00:24:03.920 --> 00:24:06.079
Well, that's the real reason why I'm asking.

00:24:06.079 --> 00:24:08.160
Also, you know, uh the air cooling.

00:24:08.160 --> 00:24:12.640
I I in my image, like the the device has like some sort of fans on it.

00:24:12.640 --> 00:24:23.039
So I assume there must be to vent out some air from inside, which obviously when you have uh an event inside of the of the thing will be the place where it vents out, you know, the flames, right?

00:24:23.039 --> 00:24:24.559
Uh and and uh the casing.

00:24:24.559 --> 00:24:29.119
If you will consider spread from unit to unit, that's the first thing that's protects the unit.

00:24:29.119 --> 00:24:38.319
If it's just a thin aluminum foil, it's gonna give you a completely different performance than it if it's like a five centimeter sandwich panel with and non-combustible insulation in it.

00:24:38.319 --> 00:24:47.200
So so though those are tremendously different uh characteristics, and again, I have a feeling that those would be driven by other needs rather than fire itself.

00:24:47.200 --> 00:24:51.680
Is the coolant in in liquid systems combustible or a lot of times it is.

00:24:51.759 --> 00:25:01.200
So a lot of times it'll be you know, another layer of uh yeah, you know, wet or glycol mix that usually has some flammability to it, so that can add to it for sure.

00:25:01.200 --> 00:25:11.200
I think you know the the insulation aspect actually what we're seeing is is a little interesting is that safety is starting to slide into it a bit because they need it primarily for operations.

00:25:11.200 --> 00:25:19.359
But when you do an analysis and you say, okay, we're gonna assume this unit is on fire, regardless of cause, but we're gonna assume it's on fire.

00:25:19.359 --> 00:25:23.680
And and and the big question is is it going to propagate to another unit?

00:25:23.680 --> 00:25:27.920
The incident heat flux to a target unit is actually pretty easy to establish, right?

00:25:27.920 --> 00:25:30.799
There's been plenty of tests, there's models, there's all kinds of stuff.

00:25:30.799 --> 00:25:32.480
And so we we can do that pretty well.

00:25:32.480 --> 00:25:35.680
The big question is what happens once it strikes that surface?

00:25:35.680 --> 00:25:38.880
And so uh we're under a wind detector.

00:25:38.880 --> 00:25:39.200
Yeah.

00:25:39.200 --> 00:25:42.960
Oh Lord, we give it a second to die down here, but it came out of nowhere.

00:25:43.200 --> 00:25:43.599
Yeah.

00:25:43.839 --> 00:25:56.160
So uh before we got attacked by the wind, what what's happening from a safety standpoint is when we're trying to evaluate whether propagation from one unit to the next is going to happen, the insulation plays a critical role.

00:25:56.160 --> 00:25:56.640
Yeah.

00:25:56.640 --> 00:26:02.000
So early systems, we looked at, and in many cases didn't have a whole lot of insulation.

00:26:02.000 --> 00:26:08.160
And so, as you can imagine, steel plate, basically thermally thin, that heat just rips right through it.

00:26:08.160 --> 00:26:11.200
And and so the inside of the container heated up pretty quickly.

00:26:11.200 --> 00:26:15.039
And so propagation was was realistic.

00:26:15.039 --> 00:26:16.400
That's not a good outcome.

00:26:16.400 --> 00:26:19.920
At the same time, they wanted to maintain the spacing between these units.

00:26:19.920 --> 00:26:24.559
So if it was six feet or 10 feet, they really wanted it from an energy density standpoint.

00:26:24.559 --> 00:26:29.519
If you told them they had to be spaced 25 feet apart, that you know economically makes a project not viable.

00:26:29.519 --> 00:26:34.000
You just changed the entire design to fit more inside and now you're spacing them further.

00:26:34.160 --> 00:26:34.400
Correct.

00:26:34.799 --> 00:26:35.039
Right.

00:26:35.039 --> 00:26:36.880
So so you you're defeating the purpose.

00:26:36.880 --> 00:26:55.279
So what's happened is that instead of moving out 25 feet, they both from an operational standpoint, but then also from a fire safety standpoint, the insulation is playing more of a role because if I do a better job of insulating it, you can oh yeah, I can now put it at six feet or ten feet safely when I couldn't before.

00:26:55.279 --> 00:26:58.319
So there's more analysis being done on that.

00:26:58.319 --> 00:27:08.240
And I think we're seeing you know more OEMs and every everybody kind of look at that as a as a contributing factor, which is which is kind of cool because you know there's interesting work to do there.

00:27:08.480 --> 00:27:08.799
Yeah.

00:27:08.799 --> 00:27:13.920
I'll put forward a challenging thought, and do you feel free to disagree with it?

00:27:13.920 --> 00:27:31.519
Was if I was thinking about uh a safety of a whole like park of those of those systems, and I had a fire within one of those, I mean, if any kind of fire happened to one of those, I would consider this this refrigerator or this container lost, and that's it.

00:27:31.519 --> 00:27:37.200
I I probably would be okay with losing one, I would be very not okay with losing uh the whole field.

00:27:37.200 --> 00:27:40.160
Maybe I would be okay with losing a few.

00:27:40.160 --> 00:27:46.319
Is this kind of risk-driven thinking a part of the design process for this?

00:27:46.319 --> 00:27:50.960
Or the efforts is to save like all the cells within the one of them?

00:27:51.279 --> 00:27:52.160
Great question.

00:27:52.160 --> 00:27:59.519
And this is really interesting because we've seen I I I think the the thought process varies depending on who you talk to.

00:27:59.680 --> 00:28:00.000
Okay.

00:28:00.319 --> 00:28:13.200
So we've seen some efforts by OEMs and others, and and I think even if it's at the uh the conference here, you know, there's a lot of focus on how do we stop thermal runaway propagation within a module or something to try and even the cell.

00:28:13.200 --> 00:28:13.599
Even the cell.

00:28:13.599 --> 00:28:20.799
Uh I always kind of am like, look, you know, the the cell's a gimme, but that that's you know, that's a different topic.

00:28:20.799 --> 00:28:26.400
But really the propagation within a module or within the unit itself or externally to other units.

00:28:26.400 --> 00:28:41.440
The reality from what we've observed is that, you know, whether it's from heat, whether it's from the smoke, whether it's from the the vent gases that come over firefighter operations, firefighter operations, generally speaking, that unit is a gonna be a write-off.

00:28:41.440 --> 00:28:47.519
Like the insurance company that is insuring it is not going to pretty much allow it to stay in operation.

00:28:47.519 --> 00:28:57.440
So even if in small events, even where there's only been generally like a single module involved, they generally come in, they take that unit away, whatever the size of the enclosure is.

00:28:57.440 --> 00:29:07.920
And this may be one of the benefits of having kind of the refrigerator size versus the 20-footers, is that gets taken away and they replace it for warranty purposes, liability, whatever, there's a whole host of stuff.

00:29:07.920 --> 00:29:18.160
So there is, I don't want to say there's no value in trying to mitigate it, but sometimes I think the thought process is that, oh, this is the panacea.

00:29:18.160 --> 00:29:20.000
We, if we can do this, we've solved the problem.

00:29:20.000 --> 00:29:26.319
This guy, like, well, you didn't fundamentally change the end outcome, which is I'm taking that container away.

00:29:26.319 --> 00:29:29.839
You may have changed PR, you may have changed fire size, so forth.

00:29:29.839 --> 00:29:37.039
So um, depending on kind of who's approaching it um and what their thought process is, they may have different answers to that.

00:29:37.039 --> 00:29:43.200
But generally speaking, I like to say that it's not my job to say what's acceptable.

00:29:43.200 --> 00:29:46.079
My job is to tell you what's going to happen.

00:29:46.079 --> 00:29:49.359
It's your facility, it's your insurer.

00:29:49.359 --> 00:29:52.240
You have to decide whether that's okay or not.

00:29:52.240 --> 00:29:56.559
Most people are gonna say one container is fine, some damage to the ones around it.

00:29:56.559 --> 00:30:00.640
The field absolutely is a loss, uh, or is a, you know, cannot be a Total loss.

00:30:00.640 --> 00:30:08.319
And we've seen pretty good kind of agreement around that, both from the insurance companies as well as you know, developers and everybody else.

00:30:08.640 --> 00:30:17.839
Yeah, I guess it also will uh change with the change in energy density, it will change with the, let's say, uh technical sophistication of the device.

00:30:17.839 --> 00:30:28.400
If you're using like uh reused car batteries for some reason, yeah, versus new uh super high-tech tells that that give you a completely different performance that's gonna probably matter.

00:30:28.400 --> 00:30:39.279
And also if we move to the centralized, as in the battery storage being on top of a rack, then perhaps it's like this the intervention should be at a smaller scale.

00:30:39.599 --> 00:30:41.519
Yeah, so the use case is critical.

00:30:41.519 --> 00:30:41.839
Yeah.

00:30:41.839 --> 00:30:42.319
For sure.

00:30:42.319 --> 00:30:56.319
And I think that's you know, the from the talk you've seen, but you know, I always like to focus on analyzing the risk so we understand, and the risk is really going to be impacted by where this thing is installed.

00:30:56.319 --> 00:31:11.680
And so, you know, you put something out in the middle of a field remote from everything, and you can have a large event, nobody really cares, but you put that in a more urban environment, that's the same fire size and everything, and across some a daycare or whatever, that that profile changes.

00:31:11.680 --> 00:31:14.160
And then if you move it into a data center, it's even worse.

00:31:14.799 --> 00:31:15.440
That's good.

00:31:15.440 --> 00:31:24.480
So, in your talk, uh let's maybe move to what's happening when it goes wrong, or maybe no, maybe uh one more thing about the technology.

00:31:24.480 --> 00:31:25.039
Yeah.

00:31:25.039 --> 00:31:34.720
So the cells that are used in those storage systems, are they like the same as we see in uh e-cigarettes and electric vehicles?

00:31:34.720 --> 00:31:36.319
Are those different cells?

00:31:36.799 --> 00:31:41.440
So they often will have uh a similar chemistry.

00:31:41.440 --> 00:31:47.920
And so, you know, but I it my analogy around that is uh I uh cakes, right?

00:31:47.920 --> 00:31:55.119
Um, and and that is that you know, everybody's got their own recipe for say a chocolate cake, but they still call it a chocolate cake.

00:31:55.119 --> 00:32:08.559
And so battery cells are somewhat similar in that you know, most manufacturers will have an LFP cell or an NMC cell or whatever, but um, and they're all siloed into those kind of chemistries, but they they they will be different.

00:32:08.559 --> 00:32:18.400
Um, so when we look at e-cigarettes or we look at e-bikes and everything else, they may have you know a similar chemistry nomenclature, but the recipes may be different.

00:32:18.400 --> 00:32:21.119
Uh the specific size obviously is going to be different.

00:32:21.119 --> 00:32:27.119
If it's got to fit in an e-cigarette, it's got to be small versus you know, you could support a larger cell in an EV.

00:32:27.119 --> 00:32:29.200
And then finally, the form factor itself.

00:32:29.200 --> 00:32:33.920
Um so you can have the button batteries or you know, your standard cylindrical or prismatics.

00:32:33.920 --> 00:32:41.680
So you will see differences across the board from the more consumer-focused type of products versus energy storage.

00:32:41.920 --> 00:32:46.240
So I guess for energy storage, again, is optimization for the space and and capacity.

00:32:46.240 --> 00:32:47.599
So probably larger, right?

00:32:47.599 --> 00:32:48.400
Better fitting.

00:32:48.640 --> 00:32:49.599
Yeah, larger.

00:32:49.599 --> 00:32:52.640
And so what we've seen is it's kind of crazy.

00:32:52.640 --> 00:32:59.440
You know, four or five years ago, a you know, nominally a 100 amp hour cell was kind of typical for an energy storage system.

00:32:59.440 --> 00:33:05.279
A couple years after that, it was uh, you know, 200, and then last year it was 300.

00:33:05.279 --> 00:33:12.480
We're now seeing 500 amp hour cells in products, and uh a number of companies have just released products that are a thousand amp hours.

00:33:12.480 --> 00:33:13.920
A thousand looks exciting.

00:33:13.920 --> 00:33:14.400
Yes.

00:33:14.400 --> 00:33:18.400
So we've we've got a thousand amp hour products that are that are coming out for sure.

00:33:18.400 --> 00:33:20.880
So that that that's gonna be a game changer a little bit.

00:33:21.039 --> 00:33:21.519
Wow, okay.

00:33:21.759 --> 00:33:27.359
But from a perspective, chemistry, it's like the same like out cathode just how do you pack it, right?

00:33:27.359 --> 00:33:30.559
Yeah, the basic architecture is gonna be similar.

00:33:30.559 --> 00:33:33.599
Um, you know, at a conceptual level, it's all the same.

00:33:33.599 --> 00:33:41.039
Yeah, it's just really the details that are gonna change, and you know, you'll see because everybody has patents, so everybody has to do something different.

00:33:41.039 --> 00:33:41.680
Oh, yeah.

00:33:41.680 --> 00:33:46.079
So you can't have identical things no matter what, just just because of the patents.

00:33:46.559 --> 00:33:48.480
Well, that that that's an interesting driver, yeah.

00:33:48.480 --> 00:33:48.799
Yeah.

00:33:48.799 --> 00:33:53.200
Uh anyway, uh let's move into the um uh the cascading failures.

00:33:53.200 --> 00:34:10.000
Uh, you had this uh this slide on your talk where you were talking about the things that happen, the events that happen, cell-driven events, uh where you mentioned uh cell monitoring, I assume that's where you detect venting, then immediate or delayed ignition, and then perhaps unit propagation.

00:34:10.000 --> 00:34:14.159
They were also indicating that we have opportunities at each of these steps along the way.

00:34:14.159 --> 00:34:18.800
So maybe you can give me an overview of this timeline and then we can talk about the interventions.

00:34:19.039 --> 00:34:19.440
Yeah.

00:34:19.440 --> 00:34:29.519
So if we start out assuming that the initiation of an incident starts with the cell, which isn't always the case, and we could talk about that maybe a little bit later.

00:34:29.519 --> 00:34:35.840
But um if if we look at it from that standpoint, then we have kind of a standard event sequence that you kind of ran through.

00:34:35.840 --> 00:34:41.840
And if everything is operating normally, we're monitoring the cell, the state of health of the system overall.

00:34:41.840 --> 00:34:48.000
If we start to see anything that's abnormal, then ideally we can correct it as soon as possible, right?

00:34:48.000 --> 00:35:00.159
So that's why having a VMS system is really critical, getting as much information as you can down to the cell level, because you may be able to uh basically isolate it, you may be able to discharge.

00:35:00.159 --> 00:35:04.159
Uh, so there may be actions that you can take to do, you know, kind of cut it off.

00:35:04.159 --> 00:35:14.320
Assuming you can't stop the incident then and the cell continues towards failure, then generally the next step is going to be that the cell is going to vent.

00:35:14.320 --> 00:35:15.679
What do you mean by that?

00:35:15.679 --> 00:35:17.280
Well, so exactly that.

00:35:17.280 --> 00:35:17.760
Yeah.

00:35:17.760 --> 00:35:22.079

225 - Battery Energy Storage Systems with Noah Ryder

Listen On

Recent Episodes

Support On