LEP Vertical Tunnel Movements: Lessons for Future Colliders

The data from 10 years of vertical surveys verify for all of LEP the previous observation, localized to region P1, that LEP floor movements are predominantly deterministic. This rules out the ATL model as being correct for this tunnel. If generalized, for yearly movements a random ATL model underestimates the possible maximum long-term motions. In contrast, extrapolation of the LEP vertical data to the short-term (hours and days) time-scale shows that the random approach predicts larger short-term movements than the deterministic model. This means that simulations using the ATL hypothesis are overly pessimistic with regard to the frequency of operational re-alignments required. Depending on the constants chosen in the models these differences can be large, of the order of a magnitude and more.This paper deals solely with the direct measured months-to-years tunnel motions in rock, and the extrapolation of such ground motions to hourly or daily time-spans. It does not address the important question of the contribution of hourly-scale movements of the accelerator components, which could have a random part, to the combined motion. Nor does it address the question of movements of accelerator tunnels like HERA or TRISTAN which are built in water and debris, and not in solid rock.