"Find one that's restoring a steam locomotive." Ah, but that's EXACTLY why I expect a post-apoc society to use steam. Find a museum that's restoring a really early diesel locomotive. There aren't a lot of those. Why? Because, due to lower-precision machining, a steam locomotive shop can make its own spare parts, while a diesel shop is largely limited to installing and servicing spare parts from a factory. Since the earliest diesels are long out of regular service, there are often no aftermarket spare parts suppliers remaining...

In short, steam locomotives do require a significant industrial base to make and use, but diesels require even more.

If you were trying to rebuild an industrialized society, and had the general science and basic engineering, I still think that recapitulating the entire steam engine tech tree, knowing you would abandon it almost as soon as you worked it out, would mean you'd be better off skipping steam. Diesel engines like this dieselnet.com/tech/diesel_history.php were more efficient than steam engines and less complex (in my opinion). The same applies to electric motors vs. the complex engineering of mechanical wheel systems. Diesel-electric is complex, but you don't have to reinvent the wheel.

The issue, as I and others have pointed out, is the limitations of their industrial capacity to make and supply higher-precision higher-strength parts needed to keep diesels running. I'm not assuming they've forgotten how to build diesels; I'm saying that, knowing how to build diesels, under conditions of a weaker industrial base, they WILL revert to steam -- an example of "appropriate technology". And who says it will die out as fast the second time around? It seems an accident of history that "advanced steam" tech wasn't developed that would've kept it competitive for longer...

By "really early," I meant older than those. The MTM's old diesels are postwar EMDs. That's the oldest you'll still find in use with any frequency on non-preserved shortline and industrial railroads, so there are still companies making aftermarket spares for them. I'm talking about diesels so old there are none left working outside museums. Think "Winton 201-A" rather than "EMD 567C".

@Tristan Klassen The 567 (1938) replaced the 201a (1934) within a few years because the 201a wasn't reliable. Your engineers can skip the questionable tech and copy the best of history. But the fact that most of the Cleveland Diesel division's engines went into things other than trains (70% submarines) only proves the point that the Diesel can be one tech applied to endless vehicles, while steam needs to go big or go home. en.wikipedia.org/wiki/Winton_Motor_Carriage_Company

I suspect we're interpreting "post-apocalyptic" in different ways which make a big difference.

"They use the small working Diesels for visitor train excursions because they can't get any of their steam engines working, and keeping up the Diesels is cheap and easy." Yeah, that's common tourist railway practice -- under modern economic conditions. Labor is now expensive and materials relatively cheap.

True, but it also says that things have recovered enough that rail service is both practical and worth the time and effort. That kind of broad recovery suggested to me they were rebuilding their entire infrastructure. In a wood-free, dry desert with supplies of oil, Steam is a hard sell, and for people recovering from an apocalypse, trying to recapitulate the society of the ancients might be even more important than practicality (which I still think Diesel would be).

Rebuilding Diesel engines would be hard. If the Diesel engines were heavier, less reliable, less efficient, and less safe than those we were talking about, they'd be comparable to the heavy, inefficient, maintenance-intense and lethally explosive steam engines.