The Limitations of UL 9540A for Battery Energy Storage Thermal Runaway & Fire Testing (Part 1)

This article is intended for readers with existing knowledge of thermal runaway of stationary battery energy storage systems and the applicable Codes and Standards. You can read a digital version of the UL 9540A standard for free on the UL website.

Let's discuss UL 9540A: 2019 (4th Edition), Standard for Safety for Test Method for Evaluating Thermal Runaway Fire Propagation in Battery Energy Storage Systems. There's a lot of confusion in the industry when it comes to thermal runaway and fire testing, particularly on how module and unit level testing is conducted as well as how useful the test results are depending on which hazards you are trying to evaluate. Before we dive into the details, let me begin by saying that writing a test method to cover so many different product types and installation scenarios is extremely difficult. There is a delicate balance between having the right level of detail without being too prescriptive. This test standard is very new and, as with most standards, it has its flaws that will likely be addressed and corrected in later revisions. In fact, I have reviewed and commented on the language for the next revision of UL 9540A which is due to take effect in August 2025. You can read the proposed changes on the UL website by clicking on "CSDS Proposals" on the UL 9540A landing page linked above.

Since this is a LinkedIn article and not a dissertation, I'm only going to focus on the top three issues, in my biased personal opinion, and try to refrain from getting into too much detail

The top 3 issues with UL 9540A: 2019 thermal runaway test standard for BESS:

1. Module level testing requiring cell-to-cell propagation encourages less safe module designs
2. Unit level testing not requiring large-scale fire conditions doesn't truly test the battery enclosure
3. No clear definition for propagation

Now let's being by getting into the details of module level testing.

Issue 1: Module level testing requiring cell-to-cell propagation encourages less safe module designs

UL 9540A (4th Edition) module level testing requires one, or more, cells of a battery module be forced into thermal runaway and that thermal runaway propagates to at least one other cell besides the initiating cell(s). The temperatures at which the initiating cell vents and then goes into thermal runaway are recorded along with some other data but there aren't any requirements for what is considered acceptable test results. There is no limit to how many cells may go into thermal runaway or whether a fire ignites. There is no pass/fail criteria, there is only compliance with the test procedure. The defined "performance criteria" for the module level test requires that unit level testing be done if:

1. the cell vent gases from the cell-level test are flammable, or
2. thermal runaway during module level testing is not contained by the module design

These criteria aren't very useful for lithium-ion batteries since the flammability of the off gases is already known from the cell level test results and it's unclear what is meant by thermal runaway being contained by the module design. If you go on to read the informational Annex A2.3.3, it states that the "effects of thermal runaway (fire and explosion)" should be contained by the module design in order to forego doing a unit level test. For most lithium-ion batteries, the cell vent gasses are flammable so a unit level test will always be required regardless of the results of the module level test.

You may be wondering, then what's the point of doing a module level test?

There are two main reasons for doing a module level test:

1. To satisfy UL 1973 certification requirements, and
2. To compare the safety performance of different module designs

UL 1973 certification requires that the module test results for single-cell initiation of thermal runaway does not result in a fire. Since UL 9540A is the only test standard globally (including IEC standards) for how to do thermal runaway testing of stationary batteries, the results of the UL 9540A module level test is what is used to satisfy the UL 1973 requirement.

Now this is where our issue lies... if I have a module that is designed to limit thermal runaway between cells and heating a single cell does not result in an adjacent cell going into thermal runaway then, per the UL 9540A module test procedure, I'm required to rerun the test with more initiating cells. I may have to rerun the test several more times, heating more and more cells just to get one to propagate and comply with the test procedure. The results of these subsequent tests won't be used for anything else, are more likely to a cause fire or explosion, and won't satisfy the UL 1973 requirement. This also complicates being able to use the module level report to compare different module designs since the tests could have been run with differing number of initiating cells and the position of those cells.

Furthermore, there is no limit to how many iterations of the test have to be done if there continues to be no propagation. It is up to the discretion of the NRTL writing the report, who by the way is getting paid more for every test that is run. With one test costing upwards of $40,000, having to run multiple tests becomes a very high expense for no apparent benefit. It would cost less and save time to design the module with a weak point where the insulation barrier between the cells is removed. Since the test module must be representative of the final product intended for production, the customer would receive a module designed with less safety features. To be clear, I am not in support of that approach. This goes against the original intention of UL 9540A which is a "standard for safety" and ensuring battery modules are designed to be as safe as possible. At a time when battery safety is being highly scrutinized by fire departments and in the news, we, as an industry, should be encouraging safer battery designs and not incentivizing module manufacturers to omit safety features just to comply with a test procedure.

Until the UL 9540A module level test procedure is revised, my guidance for testing modules that have safety features designed to limit cell-to-cell thermal runaway, is to plan to run at least two module tests. The first one being a single-cell initiation that will meet the UL 1973 requirement and a second test with multiple initiating cells. But how do you know how many cells you'll need to heat? CFD modeling and preliminary testing can help guide you, but you won't know until you do the tests. And there lies the problem.

This article is Part 1 of 3, I will address issues 2 and 3 in subsequent articles to be published the next time I decide to spend my Saturday night writing about battery safety instead of socializing. Follow me to stay informed when I post the next article and other educational content.