Main problem is the rock melting temperature ~ 1500 to 2000 °C. Even the best performing nuclear fuels can only reach 1600°C in the fuel (TRISO Particle or FCM fuel) for limited periods of time, meaning 1000°C of heat delivered due to major limitations in the steels and heat transfer. They realized this early on in the Subterrene project at Los Alamos, and transitioned to electrically heated electrodes like graphite and tungsten. They even did field tests near Bandelier National Monument, digging a large diameter door sized hole using 100s of small diameter holes to form the perimeter. Even then, there are material degradation problems especially in the presence of air and water, that make rock melting pretty challenging. More recently, an MIT spinoff (Quaize or Quarkz or something like that) is using microwave emitters to vaporize the rock - which reduces material and mechanical challenges. But it's not ideal for large diameter holes - mostly for geothermal, fossil, or utility boring.
It's not clear to me that melting is less energy intensive than digging (and all it's related machinery and material movement).
I think it's pretty clear from an energy perspective that melting/vaporizing or even pulverizing the rock is waaaaay more energy intensive as the force to shear rock is orders of magnitude less than melting it.
One way to think of this is to just imagine the chemical bonds holding the rock together. Tricone drills and TBMs both just sheer the rock, the latter on a massive scale, which is only has to expend that localized bond energy, whereas melting all of it is massively inefficient even with good heat recovery.
One thing I've wondered is if a focused energy beam could just locally break the bonds in a way that'd produce pieces optimized (size, shape, etc.) for transport back through the tunnel. In this way the minimum of energy needs to be expended by reducing the linear distance of cutting.
However, energy beams still have to heat up the rock a lot to vaporize it, so this would still locally be less efficient than cold sheering that current methods use, so when you think about it, TBMs, etc. are actually doing pretty good.
Could you use a nuclear element to directly create pressure waves against the material? Something like: nuclear fuel, graphene drum head, water injectors to drive the drum expansion.
I think this problem may be one if the classic solid state is not actually better than the old-school mechanical solution.
As an experiment, it would be fun to architect 2 designs. One that has a fancy microwave or thermal rock melting setup, one that has a traditional rock drilling approach, both powered by a nuclear power source. Add up the cost of ownership for both, and I'd bet the mechanical solution comes out on top.
With current technology. But the point of this is to get a starting point for solid state and improve afterwards. Very much like how jet engines came to be better than propellers.
It's not clear to me that melting is less energy intensive than digging (and all it's related machinery and material movement).