My philosophy is to come up with a reference 'clean' design that will succeed. I estimate the cost and state the probability of success is 100%. When I was with the old company, people would then scream and demand changes. For every political change, I estimated how much more it would cost, how much more time, and the change in probability of success. Most hated that! But I will always remember the fun times when they let me run with it, and bring the project in. My breakdowns were associated with massive interference. :)
I did a lot of work on a repository, mainly in the area of seismic stability. For that, we needed a good concept of the stresses, and what an ice age could do. I discovered that our basement rock was at a state of maximum stress, it was being crunched by the settling on a cold mantle. If there were low horizontal stresses, then there was a reason for it (lots of faults). No faults on deep rock meant extreme stresses.
So ideally you wanted a large chunk of rock that was stress-relieved along distant boundaries. Unfortunately, you never got that.
-to be continued, maybe