Simulation Study on Application of Nonlinear Model Predictive Control to the Superfluid Helium Cryogenic Circuit ?

In the 27 km circumference Large Hadron Collider, the temperature of more than 1600 main superconducting magnets is stabilized below 2 K by the Superfluid Helium Cryogenic Circuit. The key component of the circuit’s Standard Cell is an over 100 m long bayonet heat exchanger with two phase flow of superfluid helium, He II, that is integrated into the magnets submerged in a static bath of He II. The magnets operate under constraints, at variable conditions and their temperature dynamics is highly nonlinear, exhibiting variable dead times of response. We present a simulation study on the application of Nonlinear Model Predictive Control for the temperature stabilization. The controller is based on a simplified, first principles model of the circuit and C/GMRES online optimization algorithm. The good performance with small computational cost of the C/GMRES and real-time feasibility of NMPC are highlighted.