LHC BLM SYSTEM AUTOMATIONS FOR FAULT DETECTION, DIAGNOSIS, AND ISSUE TRACKING

One of the principal roles of CERN’s Beam Loss Monitoring (BLM) system pertains to the protection of LHC’s superconducting magnets against quench-inducing beam losses. Thus, the continuous surveillance of the BLM system’s performance is essential for the high reliability and availability of the LHC. This paper focuses on the architecture of a novel data pipeline with implementations on monitoring the communication status between the 864 acquisition tunnel modules that digitize the analog loss measurements, and the 432 processing surface modules that determine and act upon the criticality of the beam losses. The discussed pipeline replaces an older batch Extraction-Transformation-Loading (ETL) process, which published daily BLM status reports, in favor of a streaming ETL process. The new pipeline expands beyond the daily publication of static status reports by exploiting real-time data analysis and processing enabling the live assessment of the system’s status via online fault detection and web-based dashboards. Future development on the implemented pipeline envisions online machine learning and automated Jira issue generation features permitting fault prognosis and issue tracking.