Speaker
Description
Ion profile monitors (IPMs) are important non-destructive diagnostics for transverse profile measurement, but space-charge forces and other effects can distort the ion distribution, biasing inferred beam size and emittance. We are developing a simulation- and ML–based framework to improve the accuracy and reliability of IPM measurements in the Alternating Gradient Synchrotron (AGS) at Brookhaven National Lab. Using datasets generated with the Warp particle-in-cell code, we train supervised models that infer the true transverse beam size from short scans of collector voltage and simple profile features, without explicit knowledge of other bunch parameters. To improve profile robustness, we also investigate deep-learning methods for signal denoising and quality control, using autoencoder and transformer models to learn compact profile representations. These tools identify noisy or atypical measurements and reduce errors in downstream fitting. Together, they form an end-to-end AGS IPM analysis chain that removes noise, flags anomalies, and applies ML corrections for physical distortions, enabling robust emittance monitoring for current operations and future machines.
| I have read and accept the Privacy Policy Statement | Yes |
|---|---|
| Paper submission software | LATeX |