Most “v” block engines even if they are liquid-cooled direct more fuel to the rear cylinder(s) to cool the head, either with larger carburetor jets in older engines or with software in newer ones. If you think about it all ICE are cooled by fuel; if you fed them with boiled fuel they would overheat.

Expansion due to heat, the GP mentions a temperature of 300c.

Edit to add: Different materials expand by different amounts due to heat, and different parts of the aircraft are reaching different temperatures.

If you know all the bits are going to expand x amount in normal flight, then you work back to what size they are when cold, if the shrinkage/expansion is too great you possibly could go with a different material (with various tradeoffs) or you just accept the shrinkage as they did here.

It also helps that the SR-71 fuel was almost impossible to ignite. It couldn’t be spark ignited, instead they used triethylborane, an pyrophoric liquid that burns in contact with air. It’s also used to ignite rocket engines as its reactive enough to burn on contact with liquid oxygen.

> Leaking fuel at takeoff sounds extremely dangerous!

The fuel that was used had to withstand fairly extreme conditions before reaching the engine to be burned, which also made it pretty hard to ignite under ordinary conditions. https://en.wikipedia.org/wiki/JP-7

It's lighter and you're inherently accommodating for thermal expansion that would happen anyway and possibly cause issues.

> I wonder how do this kind of crazy engineering ideas happen

Three words: the Cold War. See also: Project Sunshine, the Oak Ridge Experiments, SDI, etc. Once you start writing blank checks to the military-industrial complex the crazy ideas start pouring forth.

If you know everything is going to expand significantly due to heating, I’m sure making it fuel tight at room temperature would involve significant engineering trade offs.