This talk covers our work on errant beam prognostics at the Spallation Neutron Source (SNS), focusing on the end-to-end process from data collection to the development and deployment of predictive models in specific. A short overview of AIML work done for accelerators and current trends will be presented. We will walk through key steps involved in creating robust Machine Learning (ML) models, including model training, validation, and deployment in an operational setting. In addition to presenting our technical approach, we will share valuable lessons learned, emphasizing the importance of infrastructure to support the continuous adaptation of models to evolving data and system behaviors. This talk will provide insights into the challenges and solutions involved in applying ML to real-world operational environments, with a particular focus on managing data drift and changes in accelerator setup while ensuring model resilience over time.

AA new harp has been installed in the Ring To Target Beam line (RTBT) section of Spallation Neutron Source. The harp is made of two planes with 32 titanium 50 micron wide wires each plane. The narrow, low-Z wires versus the 100-micron tungsten wires of the original harp, are to minimize the beam scattering. This harp will be both a backup and a complement to the existing harp further downstream. The newly created data-acquisition system is also suitable to replace the existing’s harp data-acquisition system, now over 20 years old. We show the use of a cRIO platform as a cost-effective way to process many channels and sample the beam profile at the full 60 Hz beam repetition rate. We also describe the performance of the titanium wires. A passive analog board is used to lengthen the signals to allow sampling at <= 10kS/s/ch. The data is acquired by the FPGA, passed on to the real-time OS, LabVIEW RT, and through the SNS EPICS Channel Access server presented to the control room.