Predictive temperature control of the LHC superconducting magnets

This paper describes the application of Model Based Predictive Control (MBPC) to the temperature regulation of the superconducting magnets for the future Large Hadron Collider (LHC), the next particle accelerator being built at the European Center for Particle Physics (CERN). This controller has the capability of having different sampling times for the prediction and the controlled variable. Through real experiments on the plant, significant improvements have been demonstrated using a standard MBPC algorithm in comparison with the former PID implementation. A multirate concept is included in the controller to overcome the problem of using large horizons and its consequences: large computation and ill-condition matrixes.