A Machine Learning Approach for the Tune Estimation in the LHC

The betatron tune in the Large Hadron Collider (LHC) is measured using a Base-Band Tune (BBQ) system. processing of these BBQ signals often perturbed by 50 Hz noise harmonics present beam. This causes measurement algorithm, currently based on peak detection, to provide incorrect estimates during acceleration cycle with values that oscillate between neighbouring harmonics. LHC feedback (QFB) cannot be used its full extent conditions as it relies stable and reliable estimates. In this work, we propose new estimation algorithms, designed mitigate problem through different techniques. As ground-truth real does not exist, developed surrogate model, which allowed us perform comparative analysis simple weighted moving average, Gaussian Processes deep learning simulated dataset train models was also improved variant Generative Adversarial Networks (GANs) called SimGAN. addition, demonstrate how methods respect algorithm.