Speaker
Description
For advanced high-Q SRF linacs like LCLS-II, precise cavity resonance control is crucial for ensuring stable operations. Inadequate control can lead to a significant increase in RF power demands, escalating both operational and capital expenses due to the need for additional RF power sources. To address this challenge, we have developed an innovative data-driven model predictive controller that incorporates a highly efficient surrogate model. This model is designed to manage the complex dynamics of cavities affected by microphonics and nonlinear Lorentz forces. Its efficacy has been thoroughly validated with real SRF cavities at SLAC. The MPC is implemented in a soft processor and is currently being integrated into the resonance control of an LCLS-II LLRF-like system. This foundational work paves the way for extending the model to broader motion control applications where extremely low-tolerance vibration control is essential. In this paper, we will showcase the model and share the latest test results
