String 1, with one twin aperture quadrupole and three twin aperture 10-m dipoles (MB1, MB2 and MB3) powered in series and operating at 1.9 K, has recently been dismantled after four years of operation interrupted by technical stops and shutdowns for upgrading or exchanging equipment. Following the validation of the main LHC systems (cryogenics, magnet protection, vacuum, powering and energy extraction) the experimental programme was oriented towards the optimisation of the design and the observation of artificially induced fatigue effects. The design study for String 2 has been completed. This facility, which will be commissioned in December 2000, is composed of two LHC half-cells each consisting of one twin aperture quadrupole and three 15-m twin aperture dipoles. A cryogenic distribution line housing the supply and recovery headers runs parallel to the string of magnets. An electrical feedbox is used to power, with high temperature superconductor current leads, the circuits as in the regular part of an LHC arc. This paper reviews the experiments carried-out with String 1 and summarises the results obtained after more than 12800 hours of operation below 1.9 K and 172 quenches. It also describes the layout and the components of String 2 and explains the objectives pursued by its designers.
STRING OPERATIONThe facility was operated between December 1994 [1] and December 1998. During this time, the magnets experienced 172 quenches. 144 were provoked by firing quench heaters. 70 quenches were above nominal current (12.4 kA). The magnets remained excited at nominal current for 314 hours [2]. Each cool-down lasted 3 to 5 days depending on the limitations imposed on temperature gradients across individual magnets. Following a quench, automatic procedures took 6 to 12 hours to cool-down the magnets from approximately 30 K [3,4]. The temperature of the magnets was controlled by a Joule-Thomson valve with very stringent operational constraints (0.025 K control band) imposed by the superconducting magnets characteristics, the capacity of the cryogenic system, the variability of heat loads and the accuracy of instrumentation (± 0.01 K).Model-based predictive control (MBPC) algorithms were investigated in order to obtain a narrower control band compared to standard PID control loops [5,6]. Preliminary results were encouraging but the temperature operational range was limited because only linear approximations of the process were used. Future developments are the implementation of non-linear models into the MBPC controller. Before being able to power the magnets an in-situ calibration of the temperature sensors was necessary. The observed reproducibility between sensors was better than 0.01 K and no degradation was measurable during the 4 years of operation. The String was controlled and monitored [7] from a dedicated control room but could be controlled from any remote terminal with appropriate privileges. Over 600 process variables were archived during the lifetime of the String every second and, transients on volt...