Tomo,
A whole series of interesting tradeoffs were available by the mid thirties, to reduce the need for gasoline in surface propulsion. Just as high performance gasoline engines require high mechanical compression, and therefore the knock resistance of relatively highly refined gasoline, pure Diesel engines are fussy about the fuel oil supplied. Here the ignition characteristics, or cetane number of the fuel, as well as its lubricity (in the injection system) are critical to performance and reasonable longevity. In both cases, the percentage of feedstock winding up as usable fuel is relatively small.
A range of modifications to the basic Diesel cycle would permit use of much less critical fuel and substantially lighter engine construction at the cost of some reduction in thermal efficiency. These range from the the early Hesselmann engine (a direct injection multi-fuel spark ignited engine) thru a range of precombustion Diesels, employing L'Orange and, later, Ricardo Comet chambers, fitted with glo plug or sparkplug igniters. The injectors would be much more robust and less critical than those necessary for pure Diesels and Cetane rating control would be unnecessary.
If water supply could be relied on, reciprocating steam (including as much condensing as practical) would be my choice despite its relatively low efficiency, since powdered coal (delivered by stoker screw to a high pressure blower and the burner in the flashtube boiler) would eliminate any petroleum requirement. I understand that there was a major project in the mid fifties, involving Convair in the development of a steam system for tank propulsion.
Dynasoar