KIND: A Kalman-Inspired Adaptive Estimator for SRF Cavity Detuning

TL;DR

This paper presents KIND, a novel data-driven estimator combining DMD and Transformer models for accurate, uncertainty-aware detuning estimation in superconducting RF cavities, enhancing resonance control.

Contribution

KIND introduces a hybrid Kalman-inspired neural approach that fuses DMD and Transformer models, enabling improved detuning estimation and anomaly detection in SRF cavities.

Findings

KIND outperforms classical Kalman filtering in detuning estimation accuracy.

KIND provides learned uncertainty signals for regime change detection.

Operational data demonstrates KIND's potential for future control applications.

Abstract

Superconducting radio frequency cavities with a high quality factor enable energy-efficient accelerator operation but are very sensitive to mechanical disturbances that detune their resonance. Accurate detuning estimation is therefore essential for efficient resonance control and stable beam conditions. This paper introduces Kalman-Inspired Neural Decomposition (KIND), a data-driven estimator that fuses a Dynamic Mode Decomposition model for stationary modal behavior with a Transformer-based predictor for transient dynamics. KIND further outputs learned uncertainty signals that indicate regime changes, enabling anomaly detection. Using operational cavity data, we compare KIND with a classical Kalman filtering baseline and discuss its potential as a foundation for future uncertainty-aware, forecast-based control.