Physics models, particularly for online operations, such as for our MAD-X or Bmad models, depend on a good understanding of the magnet characteristics. While we often measure the magnets or some subset of the magnets, those measurements are only meant to verify that the magnets meet specifications before being installed. We often have magnets that are not precisely understood. As a result, we end up adjusting the coefficients in our models to match beam-based measurements with little or no theoretical basis. In this work, we present a new method for deriving these coefficients using orbit response matrix (ORM) methods. This new approach utilizes a neural network (NN) surrogate model to establish the mapping between ORM measurements and quadrupole kicks. The NN model is trained to identify quadrupole kick as a source of error by observing the difference between measured ORM and model ORM with no quadrupole kick. With actual kick values from the NN model and power supply current values from the control system, we can calculate the magnet transfer function coefficients using a polynomial fit. We will present results from preliminary beam studies in the AGS Booster.

The Alternating Gradient Synchrotron (AGS) is a particle accelerator at Brookhaven National Laboratory (BNL) that accelerates protons and heavy ions using the strong focusing principle. In this work, we perform simulation studies on the AGS ring of a machine error detection method by comparing simulated and measured orbit response matrices (ORMs). We also present preliminary results of building two machine learning (ML) surrogate models of the AGS system. The first ML model is a surrogate model for the ORM, which describes mapping between orbit distortions and corrector settings. Building a self-adaptive model of ORM eliminates the need to re-measure ORM using the traditional time-consuming procedure. The second ML model is an error identification model, which maps the correlation between measurement errors (differences between measurement and model) and sources of such errors. The most relevant error sources for the error model are determined by performing sensitivity studies of the ORM.