Part of the problem is that even the state of stress at installation isn't known very precisely. I basically did some pre work on a similar idea. Working through if you could track the displacement of sleepers could you figure it out. Problem is the extension in nominal range is linear and there isn't a simple way to work out where zero is.