Your explanation uses the term "crush tolerance." This is not uniform throughout the metal shell, as implied, because different areas of the shell are naturally under more stress than others, no matter what the pressure state, just by virtue of the tank geometry, seam construction geometry, and welding. Therefore, stresses imposed by external pressure will yield the metal at the higher stress concentration areas before those naturally under lower stress. Similarly, the collapse of a steam bubble does not occur all at once throughout the interior volume of the shell because the outer wall is cooler than the central volume, particularly since it is raining. The steam bubble would collapse in over several seconds or minutes from the perimeter towards the center of the volume. Therefore, we have ever-decreasing pressure over the course of seconds acting over a metal surface with a variable "crush tolerance" (i.e. yield point). Under this scenario, the shell would deform in different places over many seconds, but not instantaneously to all shell surfaces at once. There is not an energy sink great enough evident in this video to collapse the steam bubble instantaneously, which is what would be needed to achieve the video's result. Therefore, this tank car did not implode from a temperature transient. For the same reasons, a steam accident does not implode with this rapidity.
Also the unvented hose connection, the sealed tank, and the exterior wetness of the shell do not support the simulation of a steaming or steam cleaning accident scenario, as previously explained. The details of circumstance and caption of the video assimilate a de-pressurization event, and have no evidence supporting an internal temperature transition was in progress.
