Lithium-Ion Battery Q&A Series: Thermal Runaway

Join ORR Protection experts Lee Kaiser and Aaron Wille as they discuss Lithium-Ion battery safety and fire suppression systems for battery energy storage systems, like those found in data centers.

1. Is a Lithium-Ion battery fire a self-sustaining fire and can thermal runaway be avoided by simply breaking the fire triangle?
2. Where can one find the appropriate sprinkler density for different batteries or technologies?
3. What are some examples of listed devices or approved methods to detect and minimize thermal runaway?

Transcript

With a lithium-ion type fire, is that a self-sustaining fire? Is that something where you can eliminate or stop thermal runaway or stop the fire just by breaking the fire triangle?

What we're seeing in fire incidents is they’re long duration incidents. That should kind of give the answer to your question, that even if there's extinguishment, the hazard doesn't go away, and the hazard can come back. We know that from fire testing. Most of the fire extinguishing and fire suppression systems that we have available to address these systems today, have one major challenge in removing heat from the batteries. The battery gets hot and stays hot and the offending battery, or batteries, as the fire spreads, they get hot and stay hot. The ability of the suppression system to cool is critical. In gaseous systems, we know we can put out flames, we can stop flaming, but we can't cool the batteries. Where we're seeing more success is water-based systems, straight-up sprinklers can help control the spread of the fire and potentially save the room — that's what the testing shows, but it's not addressing the fire and it's not extinguished at the fire.

There’s information about water mist systems and how they're able to both extinguish and have some measure of cooling to reduce the severity of the incident. Encapsulated agents like F500 and their application, show some success and extinguishment and cooling. Those are the leading things when the batteries catch on fire. Where the gaseous agents come back into play and we think there's some promise, is before thermal runaway has occurred when we just have failure of a cell and we're able to detect that through off gas detection, the battery management system of the batteries that know that something went wrong with a battery. We can apply those gaseous systems early and potentially avoid the whole flaming situation. There are some systems that research has been done with that and I think there's a lot of promise there, but we need to know a bit more.

You know, as far as the EPO, like in our traditional electronic settings where you shut the power off to the offending equipment, in this case it's the battery itself, and every rack should have a battery management system installed in it, right? That battery management system has sensors there that look for the state of charge of the batteries, it looks at the voltage that's coming out, looks at the amperage of the batteries, and then there's temperature sensors that are wired back to the battery management system for that rack, and they're watching the overall temperature of each module. When it sees an abnormality, it's able to shut down that rack. It doesn't remove all the charge or anything from it, but it will stop whatever's happening as far as charging or discharging in the battery racks. There’s a lot of safety built into that, but it doesn't completely remove the hazard because if something has happened, let's say there's a cell that's had abuse and it's failed and it's going towards thermal runaway, let's hope that by stopping the electrical movement in or out of the batteries, the situation can be de de-escalated.

Frankly, that's what a lot of the testing shows that's why we are excited about off-gas detection. The Li-on Tamer product is one that a lot of people know. The Li-on Tamer product can help stop that and hopefully just kind of stop the condition of the batteries where they're at and not allow it to get worse and develop into a fire, small fires developing into a big fire, and all the bad things that happen.

One last question, going back to the UL report. Maybe it's a question and a comment as the person asking the question, but I've noticed that on the UL 9540A reports, there may be differences in the recommendations for fire protection or sprinkler density from manufacturer to manufacturer and even within manufacturers from battery model to battery model. Where would they go typically to find required sprinkler density or recommended sprinkler density for different batteries and technologies? Are there other valuable data points from that report that an architect may want to reference, or an engineer may want to reference?

That question indicates that it's easier than it really is. 9540A test results may have that depending on who wrote the report, but I don't think a lot of them have that. Where I think you need to go, or someone just beginning a design for a system needs to go, is they need to obtain the 9540A test results and then give that to a qualified fire protection engineer. We partner with a couple of companies that help us do this process. We build a computer fire model of the batteries using what test results we have, and then we model the severity of the fire, the severity of the off-gas, and the flammable gas coming off of that. We do that for a couple of reasons: One is to determine the fire protection fire suppression approach, the sprinkler density, which should be a minimum of 0.3 GPM per square foot because that's what the code says (it could go up from there, but should be a minimum of 0.3) and then also determine the ventilation rate for when we get into talking about explosion prevention systems where we have exhaust fans to suck out those gases so we don't achieve an explosive atmosphere and go boom, like we know has happened in surprise Arizona when the firemen were hurt back in 2018. We think that the results come out of that fire modeling event. You first model the batteries just as they burn, and then you model the batteries with what happens when you apply a suppression system. They can do that through the software.

That’s how you start to pick what should be the water application rate. We're actively doing that on projects to make that recommendation, but it's not very prescriptive at all. It's all a performance-based design methodology and you've got to get people involved early in the project. If you're making this decision and all the walls and the structure is already up, you're probably starting late. You need to be engaging professionals that can help you early in the project.

Let’s get into the detection of lithium-ion fires or detection technology. We have a question here and it says, “what are some examples of listed devices or approved methods to detect and minimize thermal runaway?”

We're looking for an off-gas detector like Li-on tamer. Essentially what it's doing is it's looking for the electrolytes, when the electrolytes heat up, they turn into steam, which creates an off gas, and the Li-on Tamer product will detect that. Once it gets past that stage and we start getting into a thermal runaway we're going to start seeing signs of smoke. At that point, we would want to use a smoke detection system, much like extras, which is going to pick up the first signs of combustion. The earlier we can get into detecting the off gas and the smoke, the better chances we have to suppress what is happening, turn off a battery maintenance system to take the power away from that charging circuit, and try to cool down those cells so we don't have a thermal runaway condition.