A Novel Tuning Scheme for Fast Norm Optimal Iterative Learning Controller

While operation of radio frequency (RF) cavity in a particle accelerator, there is need to stabilize electric field (E) within RF cavity by control system. In pulse mode of RF cavity operation, a particle beam injected into it perturbs amplitude and phase of E, commonly termed as beam loading disturbance problem. For this disturbance compensation, Iterative Learning Controller (ILC) algorithm using proportional integral derivative (PID) type formulation have been reported for Tesla cavity. However, better ILCs based on optimization of errors and control inputs as used in Norm Optimal ILC (NOILC) have not been explored much for beam loading disturbance compensation. This paper elaborates on RF cavity model, Fast NOILC (FNOILC) algorithm and its application to beam loading disturbance problem. Modeling and simulations are explained to highlight strengths of FNOILC. More importantly for best performance, FNOILC needs tuning of weight parameters. A method to tune FNOILC is a novel contribution of this work.