It should also be pointed out that those tanks and fuel systems in automotive vehicles have been continuously redesigned for nearly for 75 years to not explode during an accident.
The other thing to remember is that the physics of rail car accidents are much different than the physics of automotive crashes. An automobile typically sheds its kinetic energy in a few moments after an incident where a derailing incident can last up to a minute before the train comes to a stop.
Also: In many tunnels, CNG and others like it common to campers *are banned* because their explosive force is amplified in a tube. It's assumed most automotive crashes won't cause any explosion, whereas a car careening at 60 MPH into a stationary propane tank will.
Well, true, but separate gensets on 18-wheelers haven't been around all that long, and, as far as I know were allowed without any sort of regulatory review when we at EPA began toencourage truckers to use them.
Anyway, auxiliary power units are usually powered by diesel, not propane. (And there are CNG and Propane powered commercial vehicles. I guess they can't use tunnels.)
As far as the 238 rules go, had the regulator been doing it right (and they may well have done so, as I've never seen the rulemaking), They would have conducted a technical analysis that would have considered the physics of rail car accidents, the risk of rail car accidents, the flammability of the fuel, the way the genset and fuel system were mounted, etc. From that they would figure out the number of deaths and injuries avoided and estimated the property damage avoided by banning gensets. Because accidents also disrupt the rail lime and impose more costs, they'd have to estimate any additional costs resulting from accidents with genset-equipped cars.
From all this, the agency could decide whether it was worthwhile to ban gensets or that the extra risks are so small that it wouldn't make enough of a difference to bother everybody. If they were going to exempt private cars, they should estimate PV mileage and risks of accidents, there, too. Of course, PV car miles are probably a small fraction to total passenger rail car-miles, which probably makes the total risk lower, but I'm speculating. If I were doing a rulemaking, I'd have to make an estimate based on hard evidence.
After the agency makes a decision, it gets published as a proposal with all the technical analysis. Then the general public (or more usually, the interested stakeholders) get a chance to comment on the proposal, and the agency has to respond to the comments. Once all that's done, the rule can go final. Of course, that's when disappointed stakeholders will sue the agency to get what they want. I should have gone to law school, after all.
According to policy, EPA requires all of the technical analyses they conduct to have outside peer review before they're incorporated into the rulemaking documents. This is not always followed, as the policy has a few ways to weasel around its intent, and some managers believe that the public comment period is suitable peer review. I'm not sure what the DOT policy on peer review of technical documents.
The bottom line is that while it seems reasonable that gensets on passenger rail cars entails increased risks, I haven't seen any analyses on how much they increase the risks. I would be interested in seeing the original rulemakings for these 238 rules and see how the agency justified them. I sort of shudder at having to dig through the Web to find them, though. I was recently looking for a couple of EPA rulemakings I was involved with 10-15 years ago, and, boy, is that website designed to make it hard to find things!