Even very earth tied machines suffer similar issues. I worked on what was known as a "hot leveller" computer, a PDP-11/73 at the steelworks I was employed at. I

Even very earth tied machines suffer similar issues. I worked on what was known as a "hot leveller" computer, a PDP-11/73 at the steelworks I was employed at. It had something like 9 rolls (maybe 200mm in diameter) that would be applied to a very hot steel plate (maybe 10mm to 150mm thick) after it had been rolled from a maybe 300mm thick slab. The levellers job was to smooth out any waves that might have acquired during the rolling process - almost like a clothes iron. The gap between the rolls needed to be adjusted by hydraulically positioning backup rolls that are even able to bend those work rolls across their width (maybe 3000mm). As you are always intending apply a huge amount of force anyways, to achieve the desired results, it was a mix of metallurgical driven algorithms and hard limits to doing the "setup". While there was always an operator that had to accept the setup before the run, there was always the risk of hitting the machines surfaces too hard, and straining components, and maybe causing a prolonged as expensive outage. Obviously the biggest risks were when there were changes or even experiments by both engineers and metallurgists. It was fun times as a quite junior engineer, and think there were a few times when over zealous setups resulted in some big noises. But I don't think I broke anything fortunately.