You haven't really answered this yet. The fact that you don't think it can happen is different than the question "what's the result if it does happen?". Customers can do amazingly stupid things from the viewpoint of the equipment designers. Connecting a battery backwards, even if you need a soldering iron to do it, is so common IRL that I wouldn't even say it's stupid, unless they did it on purpose.
Also consider an overload that isn't a short circuit, if you haven't yet. Like a bad battery that draws extra current but isn't 0Ω.
It sounds like you have multiple failures with similar evidence, I assume from different customers. So, I think it's safe to bet that there is an underlying design issue. Maybe you should try to think of ANY way you could cause this sort of failure, without too much regard to your assumptions about how the product is supposed to be used. I would also seriously consider something like a FMEA (Failure Modes and Effects) study, especially the "effects" part. I think hiring the right product safety consultant (for failure analysis, not just approvals) might be money well spent. Someone like
Exponent* perhaps.
* Not really a recommendation, just an example. I used them a few decades ago and they were OK. YMMV.