WEPD
-
WEPD Posters
24 Sep 2025, 16:30 -
18:00
WEPD001
Laser Megajoule facility status report
use link to access more material from this paper's primary code
The Laser MegaJoule, a 176-beam laser facility developed by CEA, is located near Bordeaux. It is part of the French Simulation Program, which combines improvement of theoretical models used in various domains of physics and high performance numerical simulation. It is designed to deliver about 1.4 MJ of energy on targets, for high energy density physics experiments, including fusion experiments. The LMJ technological choices were validated on the LIL, a scale-1 prototype composed of 1 bundle of 4-beams. The first bundle of 8-beams was commissioned in October 2014 with the implementation of the first experiment on the LMJ facility. The operational capabilities are increasing gradually every year until the full completion by 2026. By the end of 2025, 22 bundles of 8-beams will be assembled (full scale) and 19 bundles are expected to be fully operational. As the assembly of the laser bundles is coming to an end and before to be in full operation, we propose to make a status report on the LMJ/PETAL installation. We will present the major software developments done for theses 2 past years, the latest experimental results and the new challenges to keep this facility at its best operating level. Key words: Laser facility, LMJ, PETAL, Control System Glossary: LMJ: Laser MegaJoule CEA: Commissariat à l’Energie Atomique et aux Energies Alternatives LIL : Ligne d’Intégration Laser
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG001
About: Received: 05 Sep 2025 — Revised: 21 Sep 2025 — Accepted: 23 Oct 2025 — Issue date: 25 Nov 2025
WEPD002
ATLAS DEMO Inheritance commissioning and performance testing using SCADA and PLC based automatized procedures
1039
By the end of 2024, to cope with needs of future ATLAS tracking detector (ITk), surface integration and testing facility at CERN was fully equipped with a new large power and low temperature cooling system nick-named “DEMO Inheritance” based on the 2PACL technique. This paper will discuss implemented control system solutions for both the cooling plant and the distribution system installed in the detector proximity at the assembly clean room. The PLC and SCADA software has been fully deployed, following CERN UNICOS framework and allowed for successful and rapid initial performance test with 50kW dummy load. In this paper we will describe in details SCADA and PLC based procedures used for automatic system performance tests with no operator attendance as the first step in future commissioning of the final detector cooling systems of ATLAS and CMS detectors at the High Luminosity era of the Large Hadron Collider at CERN. These procedures running unattended allow for significant commissioning time reduction which is a key requirement for very challenging schedule.
Paper: WEPD002
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD002
About: Received: 04 Sep 2025 — Revised: 22 Sep 2025 — Accepted: 23 Oct 2025 — Issue date: 25 Nov 2025
WEPD003
Extension of the SPIRAL2 PLC-based control system for the integration of DESIR and NEWGAIN
1043
The SPIRAL2 heavy ion accelerator is currently undergoing several extension projects under development. In Phase 1+, the DESIR experimental hall will receive very low energy radioactive ion beams from S3 or SPIRAL1. In Phase 1++, a new cryogenic ion source and a new injector will be implemented to broaden the range of heavy ions currently accelerated. This contribution presents the integration of these new facilities into the SPIRAL2 PLC-based control system. The integration builds on recent technological upgrades while maintaining system consistency and ensuring compliance with the specific constraints related to radiation protection and operational safety. Some existing PLCs have been extended, and new ones have been added to automate processes related to beam operation (machine protection system, equipment insertion, vacuum, RF, cryogenics), building infrastructure (nuclear ventilation, refrigeration, alarm management), and nuclear safety (access control units, radiation monitoring). This integrated approach ensures a coherent, maintainable control system that supports safe and efficient operations for the DESIR and Newgain projects.
Paper: WEPD003
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD003
About: Received: 05 Sep 2025 — Revised: 23 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD004
First phase of control system for compact Muon Linac at J-PARC
use link to access more material from this paper's primary code
A muon linear accelerator (Muon Linac) for the muon g-2/EDM experiment is currently under construction at Japan Proton Accelerator Research Complex (J-PARC). The objective of this project is to accelerate thermal muons (25 meV at 300 K) to 212 MeV, marking the world’s first implementation of muon acceleration. Development of the control system for the Muon Linac began in 2024, with the implementation of the Ultra-Slow Muon (USM) section -- the initial acceleration stage up to 5.6 keV -- nearly completed in April 2025. The system adopts the standard EPICS framework and features a compact architecture consisting of (a) a QNAP NAS for disk storage and LDAP-based user authentication, (b) two operator terminals, and (c) two commercial micro servers serving as the EPICS IOC and the archiver server, respectively. Core functionalities of the control system are scheduled for verification during May and June, followed by beam commissioning of the USM section in December 2025. This paper reports on the status of the control system development for the USM section, as part of the first phase of the Muon Linac project. Toward the full commissioning of the entire Muon Linac in 2028, the prospects for extending the present control system to the main Linac components are discussed.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG003
About: Received: 23 Sep 2025 — Revised: 26 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD005
Development of an EPICS-based control platform for electron beam commissioning at ELI-NP
1048
The ELI-NP (Extreme Light Infrastructure - Nuclear Physics) gamma beam system employs a 234-742 MeV linear accelerator to generate gamma rays via laser-electron interactions. Accelerator operation requires coordinated control of multiple subsystems (RF, vacuum, magnets, diagnostics). During pre-commissioning, we developed an EPICS-based platform to enable equipment debugging and beam tuning simulations before control system completion. The platform integrates both soft IOCs (for device emulation and algorithm validation) and device IOCs (for hardware control). This dual-IOC architecture allows: 1) High-level control software testing without physical hardware, and 2) Physics application interfaces for advanced functionality. The system successfully simulated control operations and supported tuning algorithm development. Testing demonstrated the platform's effectiveness for subsystem debugging and realistic beam tuning simulations. By bridging the gap between hardware availability and control system readiness, the platform is expected to accelerate the commissioning process and enhance the operational efficiency of the ELI-NP gamma beam system in the future.
Paper: WEPD005
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD005
About: Received: 06 Sep 2025 — Revised: 22 Sep 2025 — Accepted: 30 Oct 2025 — Issue date: 25 Nov 2025
WEPD006
The EuAPS betatron radiation source control system
use link to access more material from this paper's primary code
EuAPS (EuPRAXIA Advanced Photon Source) is a project carried out in the EuPRAXIA context and financed by the Italian Ministry in the Recovery Europe plan framework. A new advanced betatron radiation source, obtained by exploiting plasma LWFA, is currently being realized at the Laboratori Nazionali di Frascati of INFN in Italy and will be operated as user facility. Several elements of EuAPS are remotely controlled, such as laser diagnostic devices, motors and vacuum system components. In order to efficiently run the facility, the realization of a robust and performing control system is crucial. The EuAPS control system is based on EPICS (Experimental Physics and Industrial Control System) open-source software framework. Functional safety systems such as Machine Protection System (MPS) and Personnel Safety System (PSS) in accordance with IEC-61508 standards are also integrated for interlocks control, anomalies monitoring and for protecting the personnel from hazardous areas. In this contribution, details on how the EuAPS control system has been projected and realized will be provided.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG004
About: Received: 09 Sep 2025 — Revised: 25 Sep 2025 — Accepted: 23 Oct 2025 — Issue date: 25 Nov 2025
WEPD007
Commissioning and operation of vacuum control system for SPES project
1052
The SPES (Selective Production of Exotic Species) project aims to create a facility based on particle accelerators to produce radioactive ion beams. The second phase of the project foreseen the transport of the non-reaccelerated radioactive ion beam from the TIS (Target Ion Source) to the low energy experimental hall. This part of SPES beam lines involves the major requirements in terms of complexity for the vacuum control system, as interface with GRS (Gas Recovery System), GSS (Global Safety System) and MPS (Machine Protection System), management of centralize pumping system for the exhaust, and different configurations of each section. The VCS (Vacuum Control System) of the TIS and of the following beam line sections are based on modular control units which are highly configurable to be used in different installation and with different equipment. Each unit is constituted by a SIEMENS S7-1500 PLC and 10” touch panel for the local/remote configuration and operation by expert operator, while high level control system is done in EPICS (Experimental Physics and Industrial Control System) and CSS (Control System Studio). This paper describes the commissioning phase of the VCS for the SPES facility, and the operation of the about 20 systems running at LNL (Laboratori Nazionali di Legnaro)*.
Paper: WEPD007
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD007
About: Received: 05 Sep 2025 — Revised: 25 Sep 2025 — Accepted: 28 Oct 2025 — Issue date: 25 Nov 2025
WEPD008
SKA control system in 2025
use link to access more material from this paper's primary code
It is 2025 and the SKA Telescope Control System has come a long way since the start of construction. The outline of the software architecture and some key technology decisions (including the choice of Tango) were made early. To keep the geographically distributed teams engaged, and avoid creating silos and fragmentation, development of virtually all the software components started in parallel; often while the detailed designs for the custom hardware was still evolving and before the COTS equipment was selected. The deployment strategy was adjusted to align with the industry trends. From designing a software system for hardware that does not exist we arrived at the point where we can prove that the software can actually work with the hardware. However, the software design and implementation meeting reality uncovered some issues, forcing us to make changes (ska-tango-base) and learn hard lessons (naive implementation of event callbacks). Are we ready to deliver a large distributed control system? We realize that scalability will be a challenge. This paper provides an honest overview of what works and what did not work so well, and how we address issues.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG002
About: Received: 10 Sep 2025 — Revised: 26 Oct 2025 — Accepted: 27 Oct 2025 — Issue date: 25 Nov 2025
WEPD009
LCLS-II cavity heater controls: design, operation, and accuracy
use link to access more material from this paper's primary code
The SLAC National Accelerator Laboratory's upgrade to the LCLS-II, featuring a 4 GeV superconducting linear accelerator with 37 cryomodules and two helium refrigeration systems supporting 4 kW at 2.0 K, represents a significant advancement in accelerator technology. Central to this upgrade is a 2K system with five stages of centrifugal cold compressors, operating across a pressure range from 26 mbar suction to 1.2 bara discharge*. These dynamic centrifugal compressors have a limited operational envelope hence maintaining stable pressure and flow is critical for its operation. This paper describes how SLAC achieved stable LINAC pressures in each of the 37 Cryomodule using electrical heaters compensating actively to the changes in RF power to maintain constant flow through the system. Additionally, this paper details the power accuracy of these heaters, which can be useful not only for control, but also when measuring cavity efficiency.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG005
About: Received: 21 Sep 2025 — Revised: 26 Sep 2025 — Accepted: 14 Oct 2025 — Issue date: 25 Nov 2025
WEPD010
SOLEIL II: enhancing data management and computing for tomorrow’s science
1056
Operational since 2008, SOLEIL [1] offers users access to a wide array of experimental techniques through its 29 beamlines, covering a broad energy spectrum from THz to hard X-rays. In response to evolving scientific and societal needs, SOLEIL is undergoing a major upgrade through the SOLEIL II project. This transformative initiative includes the development of a new Diffraction Limited Storage Ring (DLSR) [2], designed to dramatically increase brilliance, coherence, and flux. The upgrade also encompasses the modernization of beamlines to support state-of-the-art experimental techniques, along with a comprehensive digital transformation centered on data and user-oriented workflows. This poster presents the current status of the digital transformation efforts within the SOLEIL II framework. It outlines the project's overall progress, with a particular focus on advancements in computing and data management. A central element of this transformation is the implementation of a unified Data Platform. Key developments include the deployment of a data catalog, upgrades to the IT infrastructure, user interface (UI) research, and the integration of robotics. The platform leverages shared infrastructure and software patterns to support both beamline and accelerator teams. Additionally, ongoing evaluations of data streaming technologies—such as ASAPO, LIMA2, and Dranspose—aim to enhance real-time data acquisition and processing capabilities.
Paper: WEPD010
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD010
About: Received: 06 Sep 2025 — Revised: 27 Oct 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD012
Waveform monitoring system
1061
The Advanced Photon Source (APS) recently completed a significant upgrade to its storage ring, replacing all existing components with new ones. One of the newly introduced systems is the waveform monitoring system. This system is a 2U rackmount chassis with 8 ADC inputs, MRF event link and 16 channels of TTL I/O. This is an FPGA-based 8-channel 4 GS/s digitizer that monitors the decoherence and injection high voltage pulser waveforms. The main function of this system is to qualify each pulse and make a decision on whether to proceed with injection or not. This paper presents in detail the development and features of this system.
Paper: WEPD012
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD012
About: Received: 08 Sep 2025 — Revised: 21 Sep 2025 — Accepted: 04 Nov 2025 — Issue date: 25 Nov 2025
WEPD013
Control and assembly of complex bend magnet for proposed NSLS-II upgrade
use link to access more material from this paper's primary code
The Complex Bend (CB) is a novel lattice concept proposed for the NSLS-II upgrade, utilizing permanent magnets instead of traditional electromagnets. This innovative design aims to reduce horizontal emittance from 700 pm to 40 pm and increase beam energy from 3 GeV to 4 GeV, significantly enhancing beam brightness. However, as a new lattice architecture, the CB introduces substantial technical challenges in design, assembly, and verification, particularly in meeting stringent magnetic field requirements. Unlike electromagnets, permanent magnets cannot be adjusted after assembly, making precise design and fabrication critical. These challenges are further compounded by the nonlinear behavior of magnetic fields with respect to magnet position and geometry. To address these issues, we propose integrating advanced FPGA-based hardware with EPICS-based software into a comprehensive control and tuning system. Real-time sensor data, including position, pressure, magnetic field strength, and temperature will be continuously collected and analysed. In addition, AI/ML algorithms will support optimizing magnet positioning and alignment to meet the required field specifications for each CB unit. This presentation will cover the CB mechanical assembly system, electrical hardware design, low-level control software design, and high-level tuning software implementation.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG007
About: Received: 24 Sep 2025 — Revised: 30 Oct 2025 — Accepted: 30 Oct 2025 — Issue date: 25 Nov 2025
WEPD014
Design of an upgraded analog signal digitizer to replace the MADC system at RHIC
1064
A new general-purpose analog signal digitizer has been designed and prototyped to serve as an upgrade to the legacy Multiplexed Analog to Digital Converter (MADC) system currently in use around the RHIC accelerator and injector complex at BNL. The new system is a standalone rackmount chassis with an embedded System on a Chip (SoC). This is a departure from the traditional VME form factor used by most legacy controls equipment within the Collider Accelerator Department. New features include completely independent channels, real time digital signal processing, large sample buffers, built-in timing links, and high bandwidth network connectivity. Support is included for the legacy timing links as well as future compatibility with the EIC Timing Data Link. The core features, system architecture, and scheme for integration with the controls system network is presented.
Paper: WEPD014
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD014
About: Received: 30 Aug 2025 — Revised: 25 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD015
Modernizing FPGA development using the DESY FPGA firmware framework
use link to access more material from this paper's primary code
Brookhaven National Laboratory (BNL) is currently developing new hardware description language (HDL) code and embedded software for the Electron-Ion Collider (EIC) control system. Part of this effort is modernizing the development process itself, leveraging methodologies and tools that were initially targeted at the software world. These methods include effective source control and project management, modularization and rapid deployment of updated code, automated testing, and in many cases automated code generation. HDL designers additionally face unique challenges compared to software designers, particularly with vendor locking and dependency on particular tools and IP. The FPGA Firmware Framework (FWK), developed by DESY, is a set of tools that helps to both apply these modern methods and to overcome some of those unique challenges. This paper will cover the workflow, successes, and challenges faced when using the FWK. In particular, we will focus on the experience using this workflow to develop a customizable delay generator IP targeting a Zynq FPGA.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG009
About: Received: 06 Sep 2025 — Revised: 26 Sep 2025 — Accepted: 23 Oct 2025 — Issue date: 25 Nov 2025
WEPD017
WREN: A versatile White Rabbit Event Node for CERN’s timing system renovation
1067
WREN is a versatile White Rabbit (WR) node developed for CERN's event-based timing system renovation. Thousands of WRENs are expected to be deployed across the entire CERN accelerator complex from 2027 onwards. Equipped with dedicated hardware and gateware, WREN integrates synchronisation in both TAI (International Atomic Time) and RF (accelerator Radio Frequency) timing. It can function as a TAI event transmitter and receiver, a Beam Synchronous (RF) transmitter and receiver, and is also capable of FPGA-based time-to-digital conversion and fine-delay generation. WREN is highly adaptable for various timing and trigger distribution systems. It is available in multiple form factors, including PCIe, VME, PXIe, and uTCA. All boards are based on the Zynq UltraScale+ System-on-Chip (SoC), designed using the open-source KiCad tool and licensed under the CERN Open Hardware License (OHL). The gateware and software are also open source. This paper presents the WREN hardware modules, the gateware architecture, and potential customisations for applications beyond CERN. It also shares insights from the initial pilot deployments at CERN.
Paper: WEPD017
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD017
About: Received: 05 Sep 2025 — Revised: 24 Sep 2025 — Accepted: 30 Oct 2025 — Issue date: 25 Nov 2025
WEPD018
Open source event timing system at ALS-U
1073
The Advanced Light Source Upgrade (ALS-U) is a major upgrade project for the existing light source at Lawrence Berkeley National Laboratory. One of the key challenges of the upgrade is that the new accelerator sections cannot operate at the existing RF frequency. The injector will operate at the current RF frequency (f1 = 499.64MHz) and the new Accumulator Ring and Storage Ring will operate at a rationally related frequency (f2 = p/q * f1 = 500.39MHz). As a result, the Timing Event Generator (EVG) must be synchronous to both RF frequencies and generate separate event streams for each frequency domain. To support beam transfer from the injector (f1) to the Accumulator Ring (f2), the EVG must detect the coincidence of the two frequencies and synchronize counters to the coincidence rate.. Due to extensive need for timing channels, Event Fanout (EVF) and Event Receiver (EVR) chassis were also designed as part of this effort. This paper describes the main concepts of the Dual EVG, EVF and EVR projects, open source gateware, embedded software and EPICS IOC implementations, as well as the tests and deployment at the current ALS.
Paper: WEPD018
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD018
About: Received: 18 Sep 2025 — Revised: 21 Sep 2025 — Accepted: 30 Oct 2025 — Issue date: 25 Nov 2025
Fast archiving for BPM data at ALS-U
The Advanced Light Source Upgrade (ALS-U) is a major upgrade project for the existing light source at the Lawrence Berkeley National Laboratory. There is a growing interest in the community to employ ML/AI methods to use predictive analysis, optimization and error fault analysis. In order to enable those methods, a rich dataset must be available and integrated into the control system. This project aims at collecting, storing and providing methods to retrieve, initially, BPM data from the ALS-U Storage Ring, by using an additional, passive node connected to the Fast Orbit Feedback network. The data rate from just the BPMs alone would be on the order of 1Gbps (updating at a rate of 10kHz) with the requirement that the system must be able to store the data for 1 week, in a continuous manner, totaling dozens of TB of data. This paper describes a conceptual design and prototype details of the Fast Archiver. The authors are confident that the archiver can be easily extended to archiving other useful metrics of the ALS-U, like power supply setpoints and monitoring data.
WEPD020
Modernizing software and hardware for LANSCE EVR with FPGA and Real Time Linux
1079
This paper describes the approach to modernizing Los Alamos Neutron Science Center’s (LANSCE) Event Receiver (EVR) by replacing the Micro Research Finland (MRF) EVR with Xilinx UltraScale+ Multi-Processor System on a Chip (MPSoC). The Xilinx UltraScale+ MPSoC architecture has been chosen for this project due to its use by other teams across LANSCE and around the industry. The EVR modernization project will utilize open-source FPGA design, mrf-openevr, along with in-house implementations for interfacing. The EVR will: produce timing patterns from Event Generators (EVG) via an event link within existing time constraints, manage new and reoccurring entries into the Per Cycle Data Buffer (PCDB), and provide diagnostic tools in an easy-to-use Real Time Linux interface. The EVR modernization project is in the evaluation stage where minimum viable product criteria is being evaluated on development boards. LA-UR-25-24009
Paper: WEPD020
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD020
About: Received: 10 Sep 2025 — Revised: 16 Sep 2025 — Accepted: 22 Oct 2025 — Issue date: 25 Nov 2025
WEPD021
Object-oriented industrial I/O for EPICS on NI cRIO: reusable LabVIEW-FPGA bitfiles via the NI C API
use link to access more material from this paper's primary code
Abstract The Los Alamos Neutron Science Center (LANSCE ) has completed a significant modernization effort, migrating from the legacy RICE control system to an entirely EPICS-based infrastructure. A key enabler of this transition has been the development and deployment of modular, object-oriented Industrial I/O (IIO) architectures on National Instruments (NI) cRIO platforms. The Industrial I/O framework provides a reusable and scalable system for controlling and monitoring sensors and instruments. It is built around precompiled FPGA bitfiles accessed through NI’s C application programming interface. Where necessary, LabVIEW real-time code integrates seamlessly with EPICS IOCs. This architecture enables clear separation between control logic and hardware interfaces, supports future maintenance with minimal overhead, and accommodates both modern Linux RT cRIO and legacy VxWorks systems. The result is a flexible and resilient method for managing and improving complex control architectures across LANSCE. This contribution outlines how IIO enables hardware reuse by treating NI cards as modular components with shared logic, abstracting low-level FPGA interaction, and standardizing configurations through parameterized bitfiles and EPICs startup files. The poster and discussion focus on how this approach supports object-like behavior to improve maintainability, scalability, and cross-platform deployments of EPICS-compatible systems. LA-UR-25-24051
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG008
About: Received: 09 Sep 2025 — Revised: 22 Sep 2025 — Accepted: 24 Oct 2025 — Issue date: 25 Nov 2025
WEPD022
Proton pulse charge calculation algorithm in Beam Power Limiting System at the Spallation Neutron Source
use link to access more material from this paper's primary code
A proton pulse charge calculation algorithm in the Beam Power Limiting System (BPLS) at the Spallation Neutron Source (SNS) was developed and implemented in an FPGA. The algorithm calculates one-minute running average of the pulse charges and issues a fault to the Personal Protection System (PPS) and the Machine Protection System (MPS) when a limit is reached. A bit-accurate model of the algorithm was first developed and tested in Matlab® and then implemented and simulated in VHDL using Vivado® design environment. Finally, the algorithm was verified on a µTCA-based hardware platform.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG010
About: Received: 04 Sep 2025 — Revised: 25 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD023
Development status of FPGA-based FOFB system for PLS-II
1083
The third-generation synchrotron accelerator Pohang Light Source-II (PLS-II) at Pohang Accelerator Laboratory uses a Fast Orbit Feedback (FOFB) system to maintain beam orbit stability in the storage ring. The FOFB system operates in real time to suppress orbit perturbations in both horizontal and vertical directions. Currently, the system uses VME-based Single Board Computers (SBCs) and Reflective Memory (RFM) technology, achieving a feedback repetition rate of about 1kHz. However, the aging hardware is causing difficulties in maintenance and performance upgrades. To solve this issue, a new FOFB system based on Zynq UltraScale+ FPGA high-speed digital processing technology is under development, aiming to increase the feedback rate to 10 kHz. The new design distributes twelve independent FOFB controllers throughout the storage ring to minimize latency from Beam Position Monitor (BPM) Fast Acquisition (FA) data reception to the output of control signals to the magnet power supplies. The system is being developed to work stably with the existing Fast Magnet Power Supplies at 1 kHz and also to support future high-performance supplies capable of operating at higher rates. The FPGA-based FOFB system is currently under development, with a goal of achieving a control bandwidth greater than 100 Hz and significantly improving maintainability and scalability. This paper introduces the design concept and the current development status of the new system.
Paper: WEPD023
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD023
About: Received: 04 Sep 2025 — Revised: 24 Sep 2025 — Accepted: 27 Oct 2025 — Issue date: 25 Nov 2025
WEPD024
An FPGA-based autoencoder model for real-time RF signal denoising for industrial accelerators
use link to access more material from this paper's primary code
A challenge that industrial particle accelerators face is the high amounts of noise in sensor readings. This noise obscures essential beam diagnostic and operational data, limiting the amount of information that is relayed to machine operators and beam instrumentation engineers. Machine learning-based techniques have shown great promise in isolating noise patterns while preserving high-fidelity signals, enabling more accurate diagnostics and performance tuning. Our work focuses on the implementation of a real-time FPGA-based noise reduction autoencoder, tested on a Xilinx ZCU104 evaluation kit with the intention of being deployed on industrial particle accelerators in the near future.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG006
About: Received: 05 Sep 2025 — Revised: 18 Sep 2025 — Accepted: 21 Oct 2025 — Issue date: 25 Nov 2025
WEPD026
Modular scientific SCADA suite with Sardana and Taurus – latest developments
use link to access more material from this paper's primary code
Sardana* and Taurus\*\* are community-driven, open-source SCADA solutions that have been used for over a decade in scientific facilities, including synchrotrons (ALBA, DESY, MAX IV, SOLARIS) and laser laboratories (MBI-Berlin). Taurus is a Python framework for building both graphical and command-line user interfaces that support multiple control systems or data sources. Sardana, is an experiment orchestration tool that provides a high-level hardware abstraction and a sequence engine. It follows a client-server architecture built on top of the TANGO control system\*\*\*. In the last two years, significant developments have been made in both projects. Sardana focused on enhancing continuous scans, introducing multiple synchronization descriptions to support passive elements (e.g. shutters) and detectors reporting at different rates. The configuration tool has also been extended, following the roadmap defined by the community\*\*\*\*. Taurus has seen substantial performance gains, particularly in GUI startup times, as part of an optimization effort that started nearly three years ago. Latest improvements take profit of new TANGO event subscription asynchronous modes\*\*\*\*\*. Continuous codebase modernization is underway, and support for Qt6 is planned for the July 2025 release. This presentation will overview these recent advancements in both Sardana and Taurus and outline their current development roadmap.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG017
About: Received: 05 Sep 2025 — Revised: 24 Sep 2025 — Accepted: 04 Nov 2025 — Issue date: 25 Nov 2025
WEPD027
EPICS IOC Extension Points: Old, Recent, and Proposed
1087
The EPICS Input/Output Controller (IOC) has always been extensible, enabling applications to add functionality without modifying the core software. Since the early EPICS 3.14 releases in 2002 the Core Developers have introduced ten new extension mechanisms that IOC application developers can use in individual IOCs or shared support modules. This paper reviews the plugin interfaces available in EPICS 7.0.9 and suggests a couple of areas where new extension points could be added in the future.
Paper: WEPD027
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD027
About: Received: 06 Sep 2025 — Revised: 22 Sep 2025 — Accepted: 16 Oct 2025 — Issue date: 25 Nov 2025
WEPD028
Upgrading the ATLAS tune archiving system
use link to access more material from this paper's primary code
The Argonne Tandem Linear Accelerating System (ATLAS) is a U.S. DOE national user facility that delivers stable and radioactive ion beams from hydrogen to uranium for low-energy nuclear physics research [1]. Operators routinely expedite setup by restoring previously optimized machine parameters sets (“tunes”). The legacy tune archiving system, implemented in Corel Paradox (1999), has become a maintenance and operational bottleneck due to recurrent table corruption, single-user access, limited integration, and proprietary language. We present the ATLAS Time Machine (ATM), a modern replacement comprised of PySide6 for the UI, FastAPI for backend services, MariaDB for experiment metadata, and InfluxDB v2 for time-series device data. ATM supports multi-user access, direct integration with the ATLAS control system, and automated beamline-aware data collection based on a dynamically generated beam path view. Initial results from beta operations indicate improved reliability, streamline operator workflows, and more convenient operator access. We conclude with lessons learned and a roadmap toward full production deployment.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG015
About: Received: 22 Sep 2025 — Revised: 23 Sep 2025 — Accepted: 30 Oct 2025 — Issue date: 25 Nov 2025
WEPD029
Automated sample identification and registration system for the MOGNO beamline at SIRIUS
1091
Mogno* is a micro- and nano-tomography beamline at the Brazilian Synchrotron Light Source, SIRIUS. It performs fast tomographies with tender (22 keV, 39 keV) and hard (67 keV) X-rays at resolutions down to 500 nm, supporting classic, 4D (time-resolved), zoom (continuous magnification) and high-throughput experiments. Two stations are available: a nano-station for external users and a micro-station in scientific commissioning, each equipped with an automatic sample-exchange system using robotic arms and magazines holding 21 and 88 samples, respectively. To enhance automation and user experience under fast-measurement, high-throughput conditions, we developed a sample-registration and cataloging system. Samples are registered via a Data Matrix at a dedicated station using a PyQt desktop interface, which sends requests to a RESTful FastAPI service backed by PostgreSQL. During operation, an orchestrator routine coordinates the magazine, robot arm and code reader to identify each sample holder, fetch its name via the SOA services, and display status on the main beamline control UI. This architecture streamlines workflows, reduces manual errors and enables traceable, high-volume sample handling.
Paper: WEPD029
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD029
About: Received: 12 Sep 2025 — Revised: 23 Sep 2025 — Accepted: 03 Nov 2025 — Issue date: 25 Nov 2025
WEPD030
Performance and reliability improvement of a Python-based EPICS IOC by switching to pyDevSup
use link to access more material from this paper's primary code
The power supplies used for FOFB correctors at SIRIUS expose only electrical current values, making it necessary to perform conversions to and from beam kick values. To take advantage of the canonical Python implementation of this conversion, a separate IOC was developed using pyEPICS and PCASPy. This technology stack imposed some limitations, making it necessary to limit the update rate, and, even then, requiring one independent instance of the IOC per ring sector (20 in total) to avoid PV timeouts and disconnects; disconnection events when one of the power supplies was down also had cascading issues with reconnection and memory corruption. This motivated us to pursue more modern alternatives for integrating Python code into an IOC, specifically one that could take advantage of the Channel Access (CA) integration already present in EPICS databases, avoiding any of the bridges between CA and Python. We evaluated the pythonSoftIOC project and the pyDevice and pyDevSup support modules, which we present in this work. We settled on pyDevSup due to the development experience it provided. This work also presents benchmarks comparing the performance gains with the new IOC and aims to explore the architecture differences that enabled them.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG014
About: Received: 08 Sep 2025 — Revised: 23 Sep 2025 — Accepted: 30 Oct 2025 — Issue date: 25 Nov 2025
WEPD031
epics-in-docker: a small framework for building slim IOC and EPICS tooling container images
1096
The SIRIUS accelerators have used containers for IOCs for years, but build definitions and launch scripts were often duplicated, and image sizes could be over 3GB. On the other hand, the SIRIUS beamlines, until recently, used IOCs installed in a shared NFS, which complicated application management, especially across different OS versions. To address these issues, we have developed a framework for building slim IOC container images (e.g. ADAravis takes 300MB) using a curated set of dependencies (and their versions) and simple and short build definitions. We avoid duplicating shared information by using git submodules, which aids in versioning the base images used. The resulting container images include a standard set of installed packages and scripts, making them ready for deployment in a wide range of container orchestration setups. The shared interface provided by the EPICS build system allows us to also create images with EPICS tools, including CA and PVA gateways and epics-base utilities. For beamlines, it was necessary to adapt the IOC orchestration to also support containerized applications, keeping the same user interface for managing IOCs for beamline and support staff. This article aims to explain the epics-in-docker architecture, the user experience, and how SIRIUS manages containers. It also aims to explore the different tradeoffs made in epics-in-docker and other frameworks, such as our choice to not support different versions of dependencies.
Paper: WEPD031
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD031
About: Received: 06 Sep 2025 — Revised: 25 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD033
The Tango AlarmHandler: advancements in core functionality and tools
1102
The AlarmHandler system is a key component for ensuring operational safety and efficiency in complex control systems. Key updates include improved support for array data types within the alarm evaluation logic, enabling more sophisticated and flexible condition definitions directly involving array or matrix data. Furthermore new tools, designed to extend the AlarmHandler's reach and usability, are now available. A dedicated notification service allows for configurable, multi-channel alerting, such as email and messaging platforms, facilitating timely operator awareness and response. Complementing this, new management utilities have been created to streamline the configuration, deployment, and maintenance of alarm definitions across distributed systems, significantly simplifying administrative tasks. This contribution details the architectural changes, implementation specifics, and the benefits these advancements bring in terms of system robustness, operator efficiency, and overall monitoring capability.
Paper: WEPD033
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD033
About: Received: 05 Sep 2025 — Revised: 18 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD034
HL-LHC Inner Triplet String controls and software architecture
use link to access more material from this paper's primary code
The High Luminosity-Large Hadron Collider (HL-LHC) project at CERN aims to increase the integrated luminosity of the Large Hadron Collider (LHC). As an important milestone of the HL-LHC project, the scope of the Inner Triplet (IT) String test facility is to represent the various operation modes and the controls environment to study and validate the collective behaviour of the different systems. As for the HL-LHC, the IT String operation requires a wide-ranging set of control systems and software for magnet powering, magnet protection, cryogenics, insulation vacuum, and the full remote alignment. An overview of the control systems and their interfaces is presented with a particular focus on the software layers essential for the powering and magnet protection tests during the IT String validation program. Ensuring integration of the new HL-LHC device types and their operational readiness requires close collaboration between development teams, equipment owners and the IT String operation team which is validated by dedicated Dry Run tests. These tests aiming to validate the functionalities of new device types within the control and software applications are described in detail, with the goal of achieving a smooth transition to the magnet powering phase. The IT String facility presents a unique opportunity to validate all control and software layers ahead of the HL-LHC hardware commissioning (HWC) within the LHC complex and their operation in the High Luminosity era.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG018
About: Received: 08 Sep 2025 — Revised: 15 Sep 2025 — Accepted: 21 Oct 2025 — Issue date: 25 Nov 2025
WEPD035
An overview of the FGC4 – CERN’s new power converter controller
1107
The CERN’s Electrical Power Converters group manages over 5000 power converters, 4000 of which are controlled, monitored, and diagnosed by a few generations of the Function Generator/Controller (FGC) devices. However, the current generation (FGC3) is now facing performance limitations and component obsolescence. To address this and accommodate future installations at CERN and other laboratories, a fourth generation of FGC is under development. Built with cutting-edge technology and modern standards, FGC4 features a Linux-based System-on-Chip (SoC), delivering an order-of-magnitude improvement in regulation rate, extensive configuration options, and significantly enhanced diagnostics. While designed to fit CERN’s accelerator control system, reusability beyond CERN has been a core design principle from the outset, enabling compatibility with EPICS and TANGO frameworks. This paper provides an overview of the FGC4 project, with a primary focus on its software architecture and highly modular design, which facilitates extensibility and ensures a future-proof solution. Additionally, it discusses the hardware architecture, including a CERN-developed System-on-Module hosting a Xilinx SoC.
Paper: WEPD035
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD035
About: Received: 05 Sep 2025 — Revised: 18 Sep 2025 — Accepted: 21 Oct 2025 — Issue date: 25 Nov 2025
WEPD036
Validating VSlib, the voltage source control library used in the fourth generation of function/generator controllers at CERN
1112
The fourth generation of power converter control at CERN, known as Function Generator/Controller 4 (FGC4), is currently under development. The chosen hardware is based on a quad-core A53 ARM-architecture CPU within an AMD Zynq UltraScale+ MPSoC System-on-Chip (SoC), where one core is dedicated to voltage source control. This architecture necessitates a new approach to voltage control software, previously implemented on a Digital Signal Processor (DSP) card and an FPGA, where high reliability and performance are crucial. To achieve these goals in a highly integrated environment, a new library called VSlib (Voltage Source library) has been developed. VSlib serves as a toolkit, providing all necessary building blocks for user-defined voltage regulation algorithms, such as filters, controllers, and lookup tables. Additionally, it supports digital and analogue signal logging, inter-core data exchange, and communication with other FGC4s using an in-house Ethernet-based protocol. The development process was test-driven, focusing on performance, determinism, and reliability. The library adheres to best industrial practices, including version control, static analysis, and automated testing. Tests were conducted against power convert models running on a Speedgoat Hardware-in-the-Loop system.
Paper: WEPD036
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD036
About: Received: 06 Sep 2025 — Revised: 24 Sep 2025 — Accepted: 04 Nov 2025 — Issue date: 25 Nov 2025
WEPD037
Implementation and scalability analysis of TSPP for Vacuum Framework
use link to access more material from this paper's primary code
SCADA (Supervisory Control and Data Acquisition) systems traditionally acquire data from PLCs through polling. The Time Stamped Push Protocol (TSPP), on the other hand, enables a PLC to timestamp and push data to the SCADA at its own discretion. The Vacuum Control Systems for CERN accelerators are primarily built on a dedicated Vacuum Framework, which relies on polling and is therefore subject to its limitations. Implementing TSPP would thus be an important improvement. TSPP needs software on the PLC – a Data Manager - to determine what data to push, when to push it, and how to package it into the correct format. Due to its particular data model, implementing TSPP for the Vacuum Framework required the development of a dedicated Data Manager. Additionally, while most current systems with TSPP have a single PLC per SCADA instance, Vacuum Framework applications often involve hundreds. Given that no data was available on the impact that large numbers of PLCs pushing data to a SCADA system might have, extensive testing was required. In particular, the relationship between server load and the effective rate of received values was studied to assess performance at scale. This paper details the implementation of TSPP for the Vacuum Framework, its Data Manager design, and the testing carried out to validate the protocol and assess its performance limits in order to ensure a smooth deployment.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG019
About: Received: 05 Sep 2025 — Revised: 25 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD038
Enabling high-performance PLC communication through open standards: OPC UA PubSub
1117
The growing diversity of PLC models and brands in industrial controls systems is increasing the complexity of the communications at the supervisory (SCADA) and control layers. At CERN, the in-house framework UNICOS-CPC manages this diversity through the integration of proprietary and open communication protocols, as well as bespoke implementations at the SCADA level to manage the incoming process data received by different communication drivers, including Modbus, S7, and S7Plus. To reduce this complexity, this paper proposes unifying PLC-SCADA communications across all platforms using OPC UA PubSub, a lightweight and highly performant publisher-subscriber protocol specified in the IEC 62541 industrial standard. This approach simplifies integration with new vendors and technologies, while enabling direct communication between PLCs. It positions OPC UA once more as a homogenizing middleware layer on top of heterogeneous hardware, which has proven to be a reliable and scalable solution for other use cases at CERN, such as power supplies, powered crates and custom electronics. The paper outlines the design, prototyping and testing phases involved in integrating the OPC UA PubSub protocol into industrial applications. It also presents the challenges encountered in the integration process, and concludes with the promising results achieved in both PLC-PLC and PLC-SCADA communication setups.
Paper: WEPD038
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD038
About: Received: 04 Sep 2025 — Revised: 25 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD039
CANModule: a lightweight, vendor-neutral CAN bus abstraction library for simplified integration and diagnostics
1122
This paper presents CANModule, an open‑source, cross‑platform library that provides a unified abstraction layer for vendor‑specific Controller Area Network (CAN) bus implementations. It supports ethernet‑CAN gateways from Analytica and Linux’s SocketCAN out of the box, and offers an open architecture for adding further gateways. Requiring only standard C++17, CANModule is lightweight and framework‑independent, unlike Qt CAN support, which introduces extra dependencies. The library standardises CAN communication via a generic API in C++ and Python, reducing the effort of integrating multiple vendor APIs. Built‑in diagnostic tools mirror SocketCAN’s canutils but work transparently with many vendors and OS, easing development for heterogeneous CAN environments. CANModule is integrated into CERN’s Quasar Framework, enabling numerous OPC UA servers that control large‑scale experiments and infrastructure. It has proven reliable in settings such as ATLAS detector control and power‑supply control. Beyond CERN, the library suits industrial applications—including automotive and robotics—by providing a scalable, extensible foundation for CAN‑based systems and abstracting vendor‑specific complexities. CANModule streamlines CAN bus integration, providing a flexible, dependable, and efficient foundation for both research and industrial use cases.
Paper: WEPD039
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD039
About: Received: 04 Sep 2025 — Revised: 25 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
Study of PLC hardware integration within the CERN controls environment
In view of the consolidation of the LHC Beam Dumping System (LBDS) control planned during CERN Long Shutdown 3 (LS3), a study was conducted to evaluate industrial Ethernet-based protocols to simplify the current multi-fieldbus architecture connecting PLCs and distributed I/O. Initial assessments used the 32-bit PROFINET driver as an alternative communication interface between the Front-End Software Architecture (FESA) and PLCs, reducing system complexity and enhancing flexibility to adapt to new data acquisition needs without recompilation. While early results were promising, limitations arose from the 32-bit kernel’s integration into CERN’s middleware. The release of a new SIEMENS® PROFINET driver based entirely on libraries now enables 64-bit integration on Debian 12 systems. This development offers new possibilities for interfacing the Slow-Control and Surveillance System (SCSS) of the LBDS, using PROFINET ring communication. This paper evaluates the performance and deployment scenarios of industrial protocols (Open User Communication, SIEMENS® SOFTNET, and OPC UA) within CERN’s control infrastructure. It compares them with the existing system design and highlights the initial operational results of the PROFINET driver integration within SCSS on the Beam Transfer kicker test bench.
WEPD041
Control system design for the new SMH16 pulsed current generator in the CERN PS extraction system
1127
The existing direct-drive septum magnet (PE.SMH16), in operation since 1994, is reaching the end of its lifetime under increased extraction frequency and will be replaced by a new eddy-current septum magnet, requiring a redesigned 30 kA pulsed generator. To meet the demanding flat-top stability requirement of ±0.05% over an 11 µs window and ±0.05% pulse-to-pulse repeatability over a year, a dedicated regulation control system was developed and validated. The objective of this work is to demonstrate that the control system achieves the required closed-loop performance, delivering repeatable magnet current waveforms under representative test bench conditions. The control system employs real-time regulation of flat-top amplitude and flatness, supported by a thermally stabilized enclosure to mitigate acquisition drift. Measurements confirm that the closed-loop system consistently maintained flatness requirement over an 8 µs window, with long-term repeatability and all observed deviations, including spurious glitches up to 400 ppm, remaining within specification. The restriction of 8 µs stems from test-bench limitations rather than control capability, which is discussed in detail. These results show that the regulation system fulfils its role up to specification; final confirmation of absolute flat-top accuracy will come from further qualification of acquisition chain elements in a dedicated test campaign, followed by beam-based validation in operation.
Paper: WEPD041
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD041
About: Received: 05 Sep 2025 — Revised: 22 Sep 2025 — Accepted: 30 Oct 2025 — Issue date: 25 Nov 2025
WEPD042
Performance characterisation of real-time software in C++. A real-life example of digital camera-based acquisition systems at CERN
1132
The performance of real-time software is critical in accelerator control systems, where precision and reliability are essential. This paper presents a method for performance characterisation of real-time software developed in C++, using a digital camera-based acquisition system at CERN as a case study. Key performance metrics, including execution time, latency, memory footprint and network use are analysed to evaluate the system's ability to meet strict real-time and scalability constraints. In addition, the profiling of task execution serves as a means of ensuring that the software continues to adhere to agreed specified behaviour, particularly as the system evolves and undergoes changes. The impact of software architecture, multithreading strategies, and hardware optimization on overall system performance is also discussed. Finally, the tools used to extract key performance metrics are presented, emphasising their generic nature and potential use for other real-time software developed in CERN’s accelerator sector.
Paper: WEPD042
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD042
About: Received: 06 Sep 2025 — Revised: 25 Sep 2025 — Accepted: 04 Nov 2025 — Issue date: 25 Nov 2025
WEPD043
PLC Integrator: A modern tool for PLC-EPICS integration at ESS
1137
PLC Integrator is a newly developed tool to integrate PLC based control systems with the EPICS (Experimental Physics and Industrial Control System) framework, replacing the legacy PLC Factory at the European Spallation Source (ESS). PLC Factory depended on the ESS Controls Configuration Database (CCDB) service and REST APIs to generate PLC code and interface via Modbus TCP/IP, but its outdated Python base and the scheduled decommissioning of CCDB made its continued support unsustainable. PLC Integrator reproduces all of PLC Factory's functionality while removing obsolete dependencies, and brings native support for modern protocols such as OPC UA and Beckhoff ADS, which are critical for upcoming integration efforts like Target Remote Handling systems and Neutron Scattering Systems. Designed for sustainability, extensibility and ease of maintenance, the tool enables automation engineers to rapidly implement new features as requirements evolve. This paper describes the design motivations, implementation strategies and integration outcomes of PLC Integrator, demonstrating how it modernizes control workflows at ESS, reduces technical debt, and enhances EPICS interoperability.
Paper: WEPD043
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD043
About: Received: 05 Sep 2025 — Revised: 25 Sep 2025 — Accepted: 05 Nov 2025 — Issue date: 25 Nov 2025
WEPD045
Physics application software for FRIB: from commissioning to operational excellence
1141
The physics application software is a critical part of the FRIB accelerator’s control and beam tuning infrastructure. Development of high-level applications (HLAs) and online modeling tools began well before initial beam commissioning to support early machine setup, diagnostics, and operational readiness. As the accelerator transitioned to routine beam delivery, a broader suite of applications was developed and iteratively refined through close collaboration between physicists, control system engineers, and software developers. These applications leverage model-based control techniques, online beam dynamics simulations, and automated optimization algorithms to enhance tuning efficiency and improve beam delivery reliability across a wide range of operating conditions. Robust data management has also become essential, enabling the capture, organization, and rapid access to operational and diagnostic data critical for real-time decision-making and post-run analysis. This paper presents an overview of the current software ecosystem, highlights key applications for lattice characterization, beam tuning, and outlines future directions, including expanded automation, tighter integration with machine learning frameworks, and improvements in scalability and maintainability.
Paper: WEPD045
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD045
About: Received: 06 Sep 2025 — Revised: 21 Sep 2025 — Accepted: 27 Oct 2025 — Issue date: 25 Nov 2025
Using C++ templates for correct and efficient hardware access
Programming language compilers utilize sophisticated optimization techniques to translate high-level abstractions into performant machine code. These transformations, such as instruction scheduling and data pre-loading, are deemed correct if they preserve the program's observable behavior. However, such optimizations often fail to maintain correctness when interacting with hardware peripherals due to side effects and timing constraints. This paper presents a C++ template, meta-programming approach to generate hardware access routines that are both correct and exhibit optimal machine code generation.
A serverless control system
Serverless refers to a set of principles and practices that offload the complexities of provisioning, managing and scaling infrastructure to a cloud computing provider. At Fermilab, the Controls department has been investigating how to bring the promise of Serverless on-premise by providing similar cloud-computing infrastructure to software development teams. In this paper we discuss how Fermilab is utilizing Fission, a product that brings Functions-as-a-Service to Kubernetes as a platform for hosting core “business logic” of its control system in a developer-friendly and scalable environment.
WEPD048
Use cases for consistent robust processing of data models
use link to access more material from this paper's primary code
Many control algorithms or optimisation procedures profit from a consistent set of data which is available with a high frequency: e.g. machine learning or automated commissioning. Modern distributed control systems allow combining and presenting data based on data models, which are then transported consistently over the network: e.g. EPICS7 introduced these data models as normative types or their combination. In this paper we present use cases that profit from combining data sub-models to a consistent higher order data model. These are today typically implemented in some programming language. The authors present use cases that can profit from a consistent robust combination of data sub-models of many devices to a higher order model. Finally common patterns are presented which could be reasonable to implement independently.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG012
About: Received: 06 Sep 2025 — Revised: 25 Sep 2025 — Accepted: 28 Oct 2025 — Issue date: 25 Nov 2025
Flexible containerised deployment of EPICS IOCs via CI/CD
Managing a large number of EPICS Input/Output Controllers (IOCs) in a control system presents significant challenges in configuration, deployment, and maintenance. This poster introduces a streamlined deployment pipeline that efficiently manages IOC lifecycles within containerized environments, leveraging CI/CD practices for robust configuration management and automated updates. A key feature of this system is its fine grained deployment control: it supports both the rollout of revised IOC images across all instances and the selective deployment of updated database (db) files to individual IOC instances without impacting others in the same Docker stack. This flexibility enables rapid, low risk updates and simplifies the orchestration of complex EPICS-based infrastructures, ensuring scalability, maintainability, and operational reliability.
WEPD050
Status of development and application of the Pyapas at HEPS
use link to access more material from this paper's primary code
To meet the stringent requirements of beam commissioning at the High Energy Photon Source (HEPS), China’s first fourth-generation high-energy synchrotron light source, a new high-level application (HLA) framework named Pyapas was developed entirely in Python. Designed for flexibility and maintainability, Pyapas serves as the foundation for all HLAs at HEPS, supporting tasks such as orbit correction, optics measurement, and machine modeling. Since early 2023, Pyapas-based HLAs have been successfully applied during the commissioning of the Linac, booster, and storage ring, contributing to key milestones including first light in October 2024. This paper summarizes the major developments and applications of HLAs at HEPS and outlines the direction of future work.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG016
About: Received: 25 Sep 2025 — Revised: 26 Sep 2025 — Accepted: 22 Oct 2025 — Issue date: 25 Nov 2025
WEPD051
Libera instruments integration with control systems
1146
Libera instruments have been used with various control systems for several years. In line with the latest security and functionality upgrades, Libera control system interfaces have also been upgraded. EPICS interface in Libera instruments has been upgraded to support the latest EPICS BASE version 7.0.9, which enables users to use the PVA protocol and retrieve more signal data in a single call. Group PVs, allowing atomic access to all signal components, were also added. Furthermore, similar parameters, such as sensors, can be grouped on the Libera side already and provided to PVA clients in a single PV. The TANGO interface has been upgraded to version 9.5. It supports flexible configuration where DeviceClasses can be configured for each board type individually. The interface has also been extended with TANGO alarm and logging functionalities. Both interfaces, EPICS and TANGO, can run on the Libera instrument or they can now be compiled and run from an external server station. This network architecture enables easier maintenance and upgrades. This paper details all recent updates and improvements to the Libera control system interfaces and presents possible use cases.
Paper: WEPD051
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD051
About: Received: 03 Sep 2025 — Revised: 22 Sep 2025 — Accepted: 22 Oct 2025 — Issue date: 25 Nov 2025
WEPD052
Design study for integrating EPICS-based control systems with medical treatment apparatus
1148
This contribution proposes a design study focused on the future integration between an EPICS-based accelerator control system and a medical treatment environment within the ANTHEM project. As the accelerator subsystems (source, RFQ, MEBT, etc.) evolve toward operational readiness, a conceptual architecture is needed to bridge high-level beamline control with treatment room systems. The goal is to anticipate integration challenges by identifying critical requirements, data flows, and safety constraints typical of medical environments. This study explores design approaches to minimize bottlenecks at the final integration phase, particularly where patient positioning, safety interlocks, and real-time synchronization with beam delivery intersect. The study will not implement a prototype, but rather provide a high-level blueprint and system requirements, laying the groundwork for future harmonization between scientific instrumentation and clinical operation contexts.
Paper: WEPD052
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD052
About: Received: 03 Sep 2025 — Revised: 22 Sep 2025 — Accepted: 14 Oct 2025 — Issue date: 25 Nov 2025
WEPD053
Swarm and bayesian optimization strategies for the PIAVE-ALPI accelerators at LNL
1152
The ALPI linear accelerator at the Legnaro National Laboratories serves as the final superconducting stage in a complex chain designed to accelerate heavy ions—from carbon to uranium—for nuclear and applied physics experiments. It also plays a key role in the SPES project, aimed at re-accelerating exotic radioactive ion beams. Within the TANDEM-PIAVE-ALPI (TAP) complex, the PIAVE injector provides superconductive acceleration of very low velocity ions before they enter ALPI. Managing the interface between these two systems poses significant operational challenges: manual tuning is often required, resulting in lengthy setup procedures and reduced transmission efficiency. Beam instabilities further complicate operations, requiring frequent manual re-adjustments. To address these limitations, advanced optimization strategies based on swarm intelligence and Bayesian algorithms have been applied. These methods enable coordinated control of multiple subsystems, including beam optics, RF settings, and ion source parameters, offering a more autonomous and adaptive tuning process. Experimental results demonstrating the effectiveness of this approach will be presented.
Paper: WEPD053
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD053
About: Received: 01 Sep 2025 — Revised: 09 Sep 2025 — Accepted: 23 Oct 2025 — Issue date: 25 Nov 2025
WEPD055
Modernizing legacy Python applications at LCLS
1158
Between the start of LCLS in 2009 and Python 2’s end of life status in 2020, many control system tools and user interfaces were created using home-built Python 2 environments. With the end of official support for Python 2 comes a host of maintenance issues for all the legacy applications that survive to this day. This poster will contain techniques for modernizing these applications, common pain points in Python 2 to 3 conversions, and advice for testing and redeployment with recent examples from LCLS.
Paper: WEPD055
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD055
About: Received: 09 Sep 2025 — Revised: 21 Sep 2025 — Accepted: 28 Oct 2025 — Issue date: 25 Nov 2025
WEPD056
PvPlot: A live software oscilloscope library for accelerator control systems
1161
When operating a complex accelerator facility such as the Los Alamos Neutron Science Center (LANSCE), it is often necessary to observe live signal waveforms for various purposes. Traditionally, this has been done using dedicated physical oscilloscopes, whether permanently installed alongside equipment or temporarily deployed on rolling carts. At times, the screens of these oscilloscopes were even transmitted by video link over coaxial cable to secondary television monitors, which was a remarkable convenience at the time, but is considered cumbersome and limited today. With modern control system software and network infrastructure, the inconvenience of physical co-location and dedicated long-distance cabling with dedicated secondary equipment can be eliminated in favor of a flexible and dynamic distributed software approach which reduces complexity while adding significant capability. Here we present a solution using the Experimental Physics and Industrial Control System (EPICS) and Python as part of a comprehensive control system UI library that allows connection to arbitrary signal sources, simultaneous viewing from multiple remote networked locations, and instant reconfiguration or selection of alternate signal sources. Library architecture and various other available UI tools are also discussed.
Paper: WEPD056
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD056
About: Received: 11 Sep 2025 — Revised: 25 Sep 2025 — Accepted: 30 Oct 2025 — Issue date: 25 Nov 2025
WEPD057
Overview and current status of the SKA-Low Monitoring, Control and Calibration Subsystem (MCCS)
1164
SKA-Low is the low frequency radio telescope currently under construction in Western Australia. At its final extent, it will consist of 512 stations up to 74 km apart, each containing 256 antennas which can be used in different combinations to digitally “point” the telescope. The Monitoring, Control and Calibration Subsystem (MCCS) is responsible for performing calibration and providing local monitoring and control of all of the LFAA (Low Frequency Aperture Array) hardware components. This includes managing the allocation of resources for an observation, and the aggregation of health status. SKAO has adopted the Tango control system framework, and the MCCS software comprises upwards of 18 different Tango devices, some of which are replicated dozens of times for a single station. This complexity poses a significant challenge to computing resources and reliability when considering how to scale up the system, first to the 16 stations to be constructed and integrated by January 2026, then 68 stations at the end of 2026, and targeting 307 stations by mid-2028. This paper will describe the MCCS architecture, report on our latest performance profiling, and discuss how we are preparing for the AA 2 construction milestone which will need to support 68 stations.
Paper: WEPD057
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD057
About: Received: 04 Sep 2025 — Revised: 22 Sep 2025 — Accepted: 22 Oct 2025 — Issue date: 25 Nov 2025
WEPD058
Operational Sequencer for ESS facility
1169
The European Spallation Source (ESS) is set to become the world's most powerful neutron source, enabling groundbreaking research across a wide range of scientific disciplines. A key tool used in its operation is the ESS Sequencer — a software tool designed to automate commonly executed high-level Main Control Room procedures required for the facility’s functionality. By using predefined sequences, it improves repeatability and reliability of the processes by minimizing the potential for human error. We will discuss the technical challenges addressed by the ESS Sequencer, including task execution types, architectural design, and system scalability. Additionally, we will highlight recent upgrades and future developments aimed at further enhancing the framework’s capabilities. The implementation of the ESS Sequencer marks a major milestone for the ESS Integrated Control System Software group and is expected to significantly contribute to ESS operations and reduction of time required to deliver neutrons for experiments.
Paper: WEPD058
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD058
About: Received: 06 Sep 2025 — Revised: 27 Oct 2025 — Accepted: 31 Oct 2025 — Issue date: 25 Nov 2025
WEPD059
AI-powered scientific chatbot for accelerator operations
1172
We present the design of a retrieval-augmented generation (RAG) based scientific chatbot, tailored for control room operators at particle accelerators and laser facilities. The chatbot integrates with institutional knowledge bases, including operational manuals, control system documentation, incident logs, and structured machine data, to provide real-time, context-aware responses to operator queries. This tool is designed to support critical operational workflows such as troubleshooting, shift handovers, beamline setup, and safety procedures. By leveraging secure deployment options (e.g. on-premise or cloud environments), it ensures compliance with data governance and cybersecurity policies typical in large-scale research infrastructures. The system reduces cognitive load, improves onboarding of new staff, and enhances efficiency by enabling intuitive natural language access to complex technical knowledge, automatic logs and reports creation, and many other optimizations of daily responsibilities. We will discuss the system architecture, data integration challenges, evaluation with pilot users, and the broader potential of AI assistants in control room environments.
Paper: WEPD059
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD059
About: Received: 05 Sep 2025 — Revised: 30 Oct 2025 — Accepted: 31 Oct 2025 — Issue date: 25 Nov 2025
WEPD060
AI-driven device driver generator
1176
We present a web-based application that significantly simplifies and accelerates the development of Tango Controls device servers by integrating large language models (LLMs) into the code generation process. The tool allows users to define device attributes, commands, and properties through an intuitive graphical interface, and optionally upload device documentation in PDF format. Using retrieval-augmented generation, the system extracts relevant content from the documentation and generates Python code for Tango device servers, tailored to the specific device functionality. The backend leverages FastAPI and LangChain to interface with various LLMs such as GPT, Claude, and Gemini. Tests on devices like power supplies and teslameters show that the generated code often requires limited manual adjustments. While the application improves development efficiency and accuracy, it also highlights certain limitations, including occasional command mismatches and the need for better retrieval strategies. Future enhancements include automated test code generation, improved document parsing, support for additional programming languages, and integration of open-source models for broader applicability.
Paper: WEPD060
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD060
About: Received: 05 Sep 2025 — Revised: 29 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD061
Simplifying cryogenic process control at ESS LINAC through automation: development and integration of an automatic control sequence
1180
This paper presents the Automatic Control Sequence (ACS) developed and implemented to simplify the control of cryogenic processes in the linear accelerator (LINAC) at the European Spallation Source (ESS), which includes 27 cryomodules and 43 valveboxes. The main objectives of the project were to reduce the risk of human error and minimize manual operations — while maintaining full decision-making authority in the hands of the operator. The sequence is designed through interdisciplinary collaboration, with Excel serving as a central platform for information exchange. A custom Python script is then used to generate PLC code in SCL programming language based on the defined logic. The final sequence is deployed on a master PLC and 43 dedicated PLCs, fully integrated with the EPICS control system and interconnected via Profinet for optimized system synchronization. A user-friendly operational interface was developed using CS-Studio, serving as both a monitoring and control layer. It provides visibility across all levels of the control system — from individual devices, through local PLC sequences, up to inter-system synchronization. This paper provides an overview of the development of the Automatic Control Sequence and discusses key lessons learned and future improvements in cryogenic control at ESS.
Paper: WEPD061
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD061
About: Received: 06 Sep 2025 — Revised: 26 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD062
Field deployment and iterative enhancement of the dish structure qualification (DiSQ) software for SKA-Mid
1185
As part of the construction of the SKA-Mid telescope in South Africa’s Karoo desert, each of the 133 new mid-frequency radio dish structures, supporting a 15m diameter dish, must undergo a thorough qualification process before they are integrated into the array. To support this work, the SKAO Wombat team has developed the Dish Structure Qualification (DiSQ) software: a tailored suite of tools designed to interact with the dish structure’s PLC-based control system via an OPC-UA interface. DiSQ comprises a user-focused engineering GUI, a synchronous Python API for automated testing with bespoke scripts, and a high-performance data logger that captures engineering parameters in HDF5 format. Since 2024, DiSQ has been successfully deployed during testing and commissioning activities by Dish Structure engineers, operating in the field on the first delivered dish structures. Its modular design enabled rapid adaptation to differences between simulation environments and real hardware, with updates informed by continuous feedback from SKAO and SARAO Dish Structure engineers. This paper presents the current status of DiSQ, highlights lessons learned from deployment, and details enhancements made to improve usability, resilience, and compatibility with evolving control interfaces. DiSQ’s evolution exemplifies the value of iterative development and close collaboration between software development teams and end-users in delivering robust tools for complex scientific engineering tasks.
Paper: WEPD062
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD062
About: Received: 04 Sep 2025 — Revised: 24 Sep 2025 — Accepted: 28 Oct 2025 — Issue date: 25 Nov 2025
WEPD063
A high-precision motion profile data stream pipeline for LCLS-II fast wire scanner
use link to access more material from this paper's primary code
The LCLS-II is the first X-ray Free Electron Laser (XFEL) to utilize continuous-wave superconducting accelerator technology (CW-SCRF), capable of delivering X-ray pulses at repetition rates up to 1 MHz. The LCLS-II fast wire scanner motion control system, based on the Aerotech Ensemble controller, is designed to measure the beam profile across both high and low repetition rates. To effectively and timely analyse the motion trajectory of the fast wire scanner, we have developed a data stream pipeline that transmits high-precision profile data from the Ensemble controller to the LCLS-II server. This system integrates the motion profile into the EPICS control system, displaying the scan profile in real time via a PyDM GUI. This paper outlines the design of the data transmission pipeline and the software development process.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMG011
About: Received: 05 Sep 2025 — Revised: 21 Sep 2025 — Accepted: 21 Oct 2025 — Issue date: 25 Nov 2025
WEPD064
Logging infrastructure for EPICS-based control systems using Loki and Promtail
1190
To support multiple scientific facilities at SLAC, a modern logging system capable of high message throughput and easy record filtering is required. A new logging system deployed at SLAC meets these requirements by blending an existing EPICS logging server with open-source technologies for log storage and retrieval. To save log data, Loki was chosen as the data storage technology. Promtail was selected to push existing log files generated by various EPICS clients into Loki. To encourage increased accelerator operator engagement, multiple interfaces are provided for interacting with logs. Grafana dashboards offer a user-friendly way to build displays with minimal code, while a custom PyQt-based user interface displays the results of direct queries to Loki in a table-based display. A wrapper script around LogCLI called qlog provides a command-line interface for interacting with the logs. The rationale behind these decisions and their integration into our controls infrastructure will be considered.
Paper: WEPD064
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD064
About: Received: 11 Sep 2025 — Revised: 22 Sep 2025 — Accepted: 03 Nov 2025 — Issue date: 25 Nov 2025
WEPD065
pvAccess and virtualisation
1192
At the ISIS accelerators we are migrating to an EPICS control system using the pvAccess protocol, with most of our IOCs and equivalents running in containers. By default, the pvAccess protocol relies on UDP broadcasts for discovery of PVs. Issues that may arise due to this in a containerised environment are discussed, and solutions both general and specific to Docker Swarm are presented. We discuss an open-source UDP broadcast relay tool - called SnowSignal - developed for use internally within our Docker Swarm environments, and the configuration of our PVA Gateways.
Paper: WEPD065
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD065
About: Received: 05 Sep 2025 — Revised: 06 Oct 2025 — Accepted: 04 Nov 2025 — Issue date: 25 Nov 2025
WEPD066
Advanced p4p usage at the ISIS Neutron and Muon Source
1198
The p4p library is a Python wrapper for the C++ pvxs library allowing Python developers to access client functionality to put, get, and monitor pvAccess PVs. Server functionality allows the creation of PVs and implements the structure of the most commonly used Normative Types (e.g. NTScalar) and their fields (e.g. alarm, control, etc.). To facilitate the transition to EPICS underway at the ISIS Neutron and Muon Source accelerators, an implementation of the logic of the Normative Type fields and a subset of other IOC functionality such as CALC records has been developed. We present our uses of this work and highlight parts which may be applicable to other facilities interested in using Python.
Paper: WEPD066
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD066
About: Received: 05 Sep 2025 — Revised: 23 Sep 2025 — Accepted: 31 Oct 2025 — Issue date: 25 Nov 2025
WEPD069
Leveraging local LLM for enhanced log analysis: integrating Ollama into electronic logging systems
use link to access more material from this paper's primary code
Modern systems for electronic log keeping and trouble log management generate vast datasets of issues, events, solutions, and discussions. However, extracting actionable insights from this information remains a challenge without advanced analysis tools. This paper introduces an enhancement to two log programs—Electronic Log Keeping (elog) and Trouble Logging (TroubleLog) - used in the RHIC control system at Brookhaven National Laboratory. The enhancement integrates Ollama, an open-source, locally deployed large language model (LLM), to facilitate intelligent log analysis. We present a framework that combines MySQL database queries with Retrieval-Augmented Generation (RAG), enabling users to generate period-based summaries (e.g., daily, weekly) and retrieve topic-specific information- such as issues and solutions - through natural language queries. By indexing log data using vector embeddings and interfacing with Ollama’s API, the system provides accurate, conversational responses while ensuring data privacy by avoiding external data sharing. The paper details implementation aspects, including SQL query optimization and prompt engineering, and evaluates performance using real-world log data sets. Results demonstrate improved usability and significant reductions in manual analysis time. This work the potential of local LLMs in domain-specific log management, offering a scalable and privacy-preserving solution for accelerator control systems.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMR011
About: Received: 04 Sep 2025 — Revised: 24 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD070
Image processing with ML for automated tuning of the NASA Space Radiation Laboratory beam line
use link to access more material from this paper's primary code
Research conducted at the NASA Space Radiation Laboratory (NSRL) seeks to increase the safety of space exploration. The NSRL uses beams of heavy ions extracted from Brookhaven's Booster synchrotron to simulate the high-energy cosmic rays found in space. To accomplish this, the source machines provide many potential beam species, ranging in atomic number (Z) from 1, hydrogen/protons, to 83, bismuth and we have gone as high as Uranium. To test large-area samples, beams can be shaped to the user's specifications from a small-format 1-cm radius circular beam up to 20-cm by 20-cm uniform-area rectangular beams. This requires a complex transfer line of 24 magnets, including 9 quadrupole and 2 octupole magnets. Given the wide range of beam rigidity and size possibilities, operators tune the optics by hand while observing the beam profile on a phosphor screen imager. Successful tests have been conducted using a machine learning (ML) workflow for tuning. We capture the beam image, then process and parameterize the beam to assess centroid, shape, tilt, edge thickness, and uniform area size. These parameters are fed to the Badger software stack to avoid re-inventing a UI, using an Xopt-based Bayesian optimization algorithm for iterative tuning. The requirement to start from an image, which can be very noisy, and quantify it, makes the workflow more complex than the standard ML cookie-cutter approach of reading from traditional beam instrumentation fed to an algorithm.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMR009
About: Received: 05 Sep 2025 — Revised: 24 Sep 2025 — Accepted: 27 Oct 2025 — Issue date: 25 Nov 2025
WEPD072
Plans and strategy for edge AI/ML at the Electron-Ion Collider at Brookhaven National Laboratory
use link to access more material from this paper's primary code
Scheduled to begin Operations in 2035, the Electron-Ion Collider (EIC) is being built at Brookhaven National Laboratory (BNL) and will be the only operating particle collider in the United States. It may also be the only large collider built in the world in the next 20-30 years, during the “Age of Artificial Intelligence (AI)”. Recognizing the potential for AI and machine learning (ML) to enhance operations and create more research opportunities, the EIC is being envisioned and designed as a large-scale AI-ready state-of-the-art facility. Specifically, it will support three core areas of AI/ML capabilities, referred to as Edge, End-to-End, and Bottom-Up. Edge capabilities are intended to address what are expected to be some of the most demanding AI/ML applications in the world in terms of timescales by anticipating the infrastructure, hardware, and local compute resources needed for success. At the same time, considerable care must be taken to ensure that these capabilities are manifested in an efficient, safe, and secure Controls ecosystem and Operations environment. We report on our plans and strategy for high-performance edge AI/ML at the EIC.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMR003
About: Received: 06 Sep 2025 — Revised: 26 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD073
Using computer vision for online calibration of beam instruments at CERN
1202
Accurate calibration of beam instrumentation is critical for the optimal operation of particle accelerators. This work presents a case study of a beam imaging system at CERN’s Antiproton Decelerator (AD) target, composed of a light-emitting screen interacting with the beam and an observation camera. During operational use, the system required frequent online recalibrations to address temperature-induced image drifts. To resolve this issue, a fully automated procedure was developed that periodically acquired images and applied multiple computer vision techniques. These techniques included custom curve-fitting methods applied to pre-processed regions of interest and SIFT-based (Scale-Invariant Feature Transform) feature detection to track and correct positional shifts. By automatically performing recalibrations at regular intervals, the approach has significantly enhanced consistency and reliability, enabling continuous and precise beam monitoring in varying environmental conditions. This stabilization technique has subsequently contributed to the optimization of antiproton production at the AD facility. The paper first introduces the challenges associated with calibrating the beam imaging instrumentation of the AD target. It then presents the chosen image analysis techniques, followed by a discussion of the results and measurement errors of the tested methods. Finally, an outlook on potential future improvements is provided.
Paper: WEPD073
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD073
About: Received: 05 Sep 2025 — Revised: 13 Sep 2025 — Accepted: 17 Oct 2025 — Issue date: 25 Nov 2025
WEPD074
Towards safe and robust neural network controllers at CERN: a review of methods and challenges
1207
Advances in optimization and machine learning algorithms have shown great potential when applied to control systems of many industries, such as automotive, avionics and aerospace. At CERN, we also find many initiatives applied to our particle accelerators and industrial facilities. In recent years, neural networks are increasingly being explored as components or even full replacements for model-based control systems, which rely on handcrafted rules or hard optimization schemes. In contrast, neural networks promise near-optimal performance while being trainable purely on existing data. However, for critical control systems it is of great importance that any control-policy or dynamics-model conforms to predictable behavior and adheres to strict requirements. While model-based control realizes this via construction, neural networks models are known to exhibit unpredictable behavior, such as adversarial examples. Due to this, the use of formal methods for guaranteeing properties on neural networks has been widely explored in the literature. In this paper, we present an overview of the safety, robustness and stability challenges posed by neural network-based control systems at CERN. We examine how these challenges can be specified as formal properties and discuss state-of-the-art techniques for verifying and mitigating them.
Paper: WEPD074
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD074
About: Received: 05 Sep 2025 — Revised: 23 Oct 2025 — Accepted: 03 Nov 2025 — Issue date: 25 Nov 2025
WEPD075
Bayesian active learning for converging posteriors in latent variable inference for control systems
1214
Inferring latent variables, such as Courant-Snyder parameters in particle accelerators, is challenging due to noisy, partial observations that often produce multi-modal posterior distributions, despite the true latent variable being unique. We present a Bayesian Active Learning (BAL) framework to enhance latent variable inference in simulation-equipped control systems. BAL actively selects control settings (e.g., quadrupole magnet configurations) to maximize information gain, efficiently refining multi-modal posteriors into unimodal ones for improved inference accuracy. Using an ensemble of physics-informed beam envelope simulations in PyTorch, our approach approximates posterior sampling and mutual information to guide data acquisition. This interpretable framework holds broad potential for improving latent variable inference in control systems.
Paper: WEPD075
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD075
About: Received: 13 Sep 2025 — Revised: 21 Sep 2025 — Accepted: 04 Nov 2025 — Issue date: 25 Nov 2025
WEPD076
Geoff developments in 2025
1219
The complexity of the CERN and GSI/FAIR accelerator facilities requires a high degree of automation to maximize beam time and performance for physics experiments. GeOFF, the Generic Optimization Framework & Frontend, is an open-source tool developed within the EURO-LABS project by CERN and GSI to streamline access to classical and AI-based optimization methods. It provides standardized interfaces for optimization problems and utility functions to speed up implementation. Plugins are independent packages with their own dependencies, allowing scaling from simple prototypes to complex state machines that communicate with devices in different timing domains. This contribution presents GeOFF’s design, features, and current applications. At GSI, multi-objective Bayesian optimization was applied to SIS18 multi-turn injection, building a Pareto front from experimental data. At CERN, GeOFF and ML/AI contributed to a record ion beam intensity for the LHC in 2024 through LEIR and SPS optimization. In addition, GeOFF underwent major updates in 2025, aligning it with the latest developments in Python-based numerical and machine-learning software.
Paper: WEPD076
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD076
About: Received: 08 Sep 2025 — Revised: 08 Oct 2025 — Accepted: 31 Oct 2025 — Issue date: 25 Nov 2025
WEPD077
Long short-term memory of recurrent neural network to the analysis of temperature rise on the production target in Hadron Experimental Facility of J-PARC
1223
Hadron Experimental Facility (HEF) is designed to handle intense slow-extraction proton beam from 30-GeV Main Ring (MR) of Japan Proton Accelerator Research Complex (J-PARC). The production target in HEF is operated under severe conditions in which its temperatures periodically vary from 30 to over 300 °C by heat deposit of irradiated beams. In a long term, a careful evaluation of damage to the target and its lifetime is quite important. If the target temperatures are accurately predicted from the existing data, including beam intensity, duration of beam extraction and beam position, it could be possible to verify the cumulative damage by comparing the predicted temperature rise with the measured one. The predicted temperature rise was calculated from the existing data using linear regression with a machine learning library, scikit-learn. However, the predictions of temperature rise by linear regression were not fully satisfactory due to changes of beam optical and accelerator conditions. Therefore, an enhanced prediction method for the temperature rise on the proton target has been developed using Long Short-Term Memory (LSTM), a type of Recurrent Neural Network (RNN) architecture. A systematic study was also conducted to investigate the effects of hyperparameters, including a sequence length and hidden layer. The present paper reports the status of the prediction system of the temperature rise on the production target using the LSTM analysis in detail.
Paper: WEPD077
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD077
About: Received: 04 Sep 2025 — Revised: 26 Sep 2025 — Accepted: 14 Oct 2025 — Issue date: 25 Nov 2025
WEPD078
Machine learning for ISIS Controls
1227
Machine learning methods have been demonstrated significant promise when applied in the context of accelerator controls. This work highlights the ongoing efforts using ML methods to improve and optimise the operation of the accelerator. We present current applications in anomaly detection, optimisation and digital twinning. As well as control applications the machine learning operations (MLOps) work required for rapid prototyping and reliable and repeatable deployment is presented. Finally, we discuss the current and future impact of ML onthe ISIS Control system.
Paper: WEPD078
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD078
About: Received: 06 Sep 2025 — Revised: 29 Sep 2025 — Accepted: 31 Oct 2025 — Issue date: 25 Nov 2025
WEPD079
Design of an intelligent inspection system for particle accelerator facilities
use link to access more material from this paper's primary code
Aiming at the limitations of the traditional manual inspection method for particle accelerator facilities, an intelligent inspection system for particle accelerator facilities based on multi-modal sensors and artificial intelligence technology is provided. Multi-modal sensors can collect various types of data from the accelerator facilities, such as temperature, audio, images, and water leakage information, which provides comprehensive information for a thorough understanding of the equipment status. Artificial intelligence technology is capable of conducting in-depth analysis of massive amounts of data, uncovering fault patterns and rules. Through technologies such as fault modeling and data analysis, it can achieve early warning and diagnosis of faults. The intelligent inspection system for particle accelerator facilities effectively addresses the limitations of traditional inspection methods. It enables real-time monitoring of the accelerator facilities and rapid alarming of abnormalities, significantly improving work efficiency and the operational safety of the equipment. This system provides strong support for the stable operation of particle accelerator facilities.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMR007
About: Received: 15 Sep 2025 — Revised: 21 Sep 2025 — Accepted: 27 Oct 2025 — Issue date: 25 Nov 2025
Reinforcement learning approaches for parameter tuning in particle accelerators
Recent developments at the INFN laboratories in Legnaro have demonstrated the effectiveness of Bayesian optimization in automating the tuning process of particle accelerators, yielding substantial improvements in beam quality, significantly reducing setup times, and shortening recovery times following interruptions. Despite these advances, the high-dimensional parameter space defined by numerous sensors and actuators continues to pose challenges for fast and reliable convergence to optimal configurations. This paper proposes a machine learning-based framework that combines surrogate modeling of the accelerator with reinforcement learning strategies for closed-loop optimization, with the goal of further accelerating commissioning procedures and enhancing beam performance.
WEPD082
Toward particle accelerator machine state embeddings as a modality for large language models
1233
Understanding and diagnosing the state of a particle accelerator requires navigating high-dimensional control system data, often involving hundreds of interdependent parameters. We propose a novel multimodal embedding framework that jointly learns representations of machine states from both numerical control system readouts and natural language descriptions. This enables the translation of complex machine conditions into human-readable summaries while maintaining fidelity to the underlying physical system. The obtained embeddings are subsequently adapted to an open-weights large language model via cross-attention conditioning. We demonstrate a first implementation trained on European XFEL machine state data. This work covers the embedding model architecture, training methodology, and presents initial examples demonstrating the model's capabilities in action. Due to the general concept of machine state, the model can be easily adapted to other facilities and control system environments.
Paper: WEPD082
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD082
About: Received: 05 Sep 2025 — Revised: 22 Sep 2025 — Accepted: 03 Nov 2025 — Issue date: 25 Nov 2025
WEPD083
Overview and status of the Machine Learning Data Platform project
1239
The Machine Learning Data Platform (MLDP) is a product providing full-stack support for data science, Artificial Intelligence, and Machine Learning (AI/ML) applications at particle accelerator and large experimental physics facilities. It supports AI/ML applications from front-end, high-speed acquisition of heterogeneous, time-series data, through data archiving and management, to back-end analysis. The MLDP represents a “data-science ready” platform for the diagnosis, modeling, control, and optimization of these facilities. It offers a consistent, data-centric interface to archived data, standardizing implementation and deployment of AI/ML algorithms for different facility configurations, or between facilities. The MLDP is also deployable for experimental data collection, archiving, and analysis. It can acquire and archive heterogeneous data from experimental diagnostics (e.g., images, arrays, etc.) along with control system configurations, and any metadata required for provenance. Thus, the MLDP can manage experimental data through its entire lifecycle, from acquisition and archiving, through analysis and investigation, to release and final publication. The MLDP is a public-domain product and available to the community. The archive management system is fully independent with an installation and deployment utility (see https://github.com/osprey-dcs/data-platform). We present a brief MLDP project overview then detail the status and notable achievements.
Paper: WEPD083
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD083
About: Received: 05 Sep 2025 — Revised: 22 Sep 2025 — Accepted: 04 Nov 2025 — Issue date: 25 Nov 2025
Integrated denoising for improved stabilization of RF cavities
use link to access more material from this paper's primary code
Typical operational environments for industrial particle accelerators are less controlled than those of research accelerators. This leads to increased levels of noise in electronic systems, including radio frequency (RF) systems, which make control and optimization more difficult. This is compounded by the fact that industrial accelerators are mass-produced with less attention paid to performance optimization. However, growing demand for accelerator-based cancer treatments, imaging, and sterilization in medical and agricultural settings requires improved signal processing to take full advantage of available hardware and increase the margin of deployment for industrial systems. In order to improve the utility of RF accelerators for industrial applications we have developed methods for removing noise from RF signals and characterized these methods in a variety of contexts. Here we expand on this work by integrating denoising with pulse-to-pulse stabilization algorithms. In this poster we provide an overview of our noise reduction results and the performance of pulse-to-pulse feedback with integrated ML based denoising.
Reinforcement learning for automation of accelerator tuning
use link to access more material from this paper's primary code
For more than half a decade, RadiaSoft has developed machine learning (ML) solutions to problems of immediate, practical interest in particle accelerator operations. These solutions include machine vision through convolutional neural networks for automating neutron scattering experiments and several classes of autoencoder networks for de-noising signals from beam position monitors and low-level RF systems in the interest of improving and automating controls. As active deployments of our ML products have taken shape, one area which has become increasingly promising for future development is the use of agentic ML through reinforcement learning (RL). Leveraging our substantial suite of ML tools as a foundation, we have now begun to develop an RL framework for achieving higher degrees of automation for accelerator operations. Here we discuss our RL approaches for two areas of ongoing interest at RadiaSoft: total automation of sample alignment at neutron and x-ray beamlines, and automated targeting and dose delivery optimization for FLASH radiotherapy. We will provide an overview of both the ML and RL methods employed, as well as some of our early results and intended next steps.
WEPD086
Xopt and Badger: a machine learning ecosystem for real-time accelerator control and optimization
use link to access more material from this paper's primary code
Machine learning (ML)-based black-box optimization algorithms have demonstrated significant improvements in accelerator optimization speed, often by orders of magnitude. However, deploying these algorithms in real-time facility control remains challenging due to the specialized expertise and infrastructure required. To bridge this gap, we introduce the Xopt ecosystem, a versatile suite of tools designed to make advanced ML-based optimization accessible to the broader accelerator community. This ecosystem includes Xopt, a modular Python framework that facilitates the integration of ML-based optimization algorithms with arbitrary control problems, and Badger, a graphical user interface built on top of Xopt, which enables seamless deployment of ML algorithms in real-time control systems. The Xopt ecosystem has been successfully applied towards solving challenging real-time control problems at leading international accelerator facilities, including SLAC, LBNL, Argonne, Fermilab, BNL, DESY, and ESRF, demonstrating its effectiveness in real-world optimization tasks. In this presentation, we provide an overview of Xopt’s capabilities and illustrate its impact through case studies from SLAC accelerator facilities including LCLS, LCLS-II, and FACET-II.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WESV001
About: Received: 18 Sep 2025 — Revised: 23 Sep 2025 — Accepted: 21 Oct 2025 — Issue date: 25 Nov 2025
WEPD087
Machine learning–based longitudinal phase space control for X-ray free-electron laser
use link to access more material from this paper's primary code
Precise control of the longitudinal phase space (LPS) in X-ray free-electron laser (XFEL) is critical for optimizing beam qualities and X-ray pulses properties required by the experimental stations. We present results of using machine learning techniques for LPS shaping and control with Bayesian optimization.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMR005
About: Received: 10 Oct 2025 — Revised: 21 Oct 2025 — Accepted: 31 Oct 2025 — Issue date: 25 Nov 2025
WEPD088
Exploring AI-based models in accelerators: a case study of the SOLARIS synchrotron
use link to access more material from this paper's primary code
The National Synchrotron Radiation Center SOLARIS, third generation light source, is the only synchrotron located in Central-Eastern Europe, in Poland. The SOLARIS Center, with seven fully operational beamlines, serves as a hub for research across a diverse range of disciplines. The most important aspect of such research infrastructure is to provide stable working conditions for the users, operators and the conducted projects. Due to its unique properties, problem complexities, and challenges that require advanced approaches, the problem of anomaly detection and automatic analysis of signals for the beam stability assessment is still a huge challenge that has not been fully developed. To address this problem, different AI-based projects are under discussion and development, i.e. automatic analysis of diagnostic signals on the example of transverse beam profiles or beam position FFT windows classification. The best proposed solution, based on the InceptionV3 architecture, can assess beam quality automatically, based solely on the image itself with 94.1% accuracy and 96.6% precision. Discussion on the current developments and deployments in SOLARIS on that field, both for the accelerator and beamlines, will be covered.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMR001
About: Received: 05 Sep 2025 — Revised: 21 Sep 2025 — Accepted: 22 Oct 2025 — Issue date: 25 Nov 2025
WEPD089
AI and ML integration for beamline optimization and virtual assistance at the SOLARIS synchrotron
use link to access more material from this paper's primary code
The future of synchrotron beamline operations is poised for a transformative leap with advancements in artificial intelligence (AI) and machine learning (ML). While SOLARIS National Synchrotron Radiation Centre* has yet to integrate these technologies, their potential to revolutionize experiments, data analysis, and user interactions is immense. AI-driven automation promises real-time assistance in optimizing beamline experiments, minimizing manual intervention while enhancing precision. Machine learning algorithms will unlock deeper insights from complex datasets, facilitating faster, more accurate interpretations. Additionally, intelligent virtual agents could redefine how researchers interact with beamline controls, offering predictive guidance and adaptive optimization. As SOLARIS expands its capabilities, embracing AI and ML will position it at the forefront of scientific innovation, ensuring seamless, efficient, and accessible synchrotron research for future generations.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMR002
About: Received: 06 Sep 2025 — Revised: 22 Sep 2025 — Accepted: 27 Oct 2025 — Issue date: 25 Nov 2025
WEPD092
BOLT: beamline operations and learning testbed for EPICS and Bluesky integration
use link to access more material from this paper's primary code
The Beamline Operations and Learning Testbed (BOLT) is a portable, cost-effective platform developed at the Advanced Light Source, inspired by a similar device developed at Diamond Light Source, to test experimental control systems without disrupting user operations. BOLT simulates a beamline endstation with minimal hardware (two motors, one detector) while implementing the complete EPICS/ Bluesky software stack. This testbed serves two purposes: firstly, developers can prototype new control software, user interfaces, and infrastructure layouts, e.g. virtualizing beamline computers, and secondly, beamline scientists and users can safely explore new developments and provide feedback before production deployment. This approach is intended to improve usability and accelerate adoption. To help lower the barrier of entry for facilities interested in adoption, we will open-source CAD models, wiring diagrams, and bill of materials. In terms of experimental technique, BOLT implements photogrammetry (3D reconstruction from images), conceptually similar to tomography but using reflected rather than transmitted signals. We demonstrate the system running on a terminal-based EPICS/ Bluesky integration, a browser-based interface developed in-house and presented in a separate contribution, as well as the traditional LabVIEW-based ALS controls system. By creating a safe learning environment, BOLT aims to accelerate the adoption of open-source controls tools across the synchrotron community.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMR014
About: Received: 06 Sep 2025 — Revised: 09 Oct 2025 — Accepted: 14 Oct 2025 — Issue date: 25 Nov 2025
WEPD095
Particle accelerator simulation using GPUs in Accelerator Toolbox
1245
In storage ring light sources, two important performance parameters are injection efficiency and lifetime. Accelerator particle tracking in digital twins and simulators is an approach to efficiently derive these quantities and guide the optimization of the physical accelerator machine. These calculations are resource-intensive processes that often require high-performance computing. GPU usage can be an effective strategy to achieve good performance at a reasonable cost. This paper presents the current state of GPU particle tracking code implemented in Accelerator Toolbox (AT) using OpenCL or CUDA. The GPU kernel code is dynamically generated according to the lattice definition, allowing for optimal performance and selection of the symplectic integrator. We present accuracy and performance comparisons obtained with GPUs versus those obtained with AT CPU code. Trade-offs necessary to achieve optimal results are also discussed. Additionally, we discuss GPU kernel synchronization techniques needed by collective effect elements, as well as the object-oriented code structure to facilitate the integration of other GPU APIs.
Paper: WEPD095
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD095
About: Received: 21 Aug 2025 — Revised: 08 Sep 2025 — Accepted: 30 Oct 2025 — Issue date: 25 Nov 2025
WEPD096
APOLLO: a facility-scale differentiable virtual accelerator at Fermilab FAST/IOTA
1249
As the design complexity of modern accelerators grows, there is more interest in using advanced simulations that have fast execution time or yield additional insights like gradients. The FAST/IOTA facility has been working on implementing and experimentally validating an end-to-end digital twin that is both fast and gradient-aware, allowing for rapid prototyping of new software and experiments with minimal beam time costs. Our framework integrates physics and ML codes for linac and ring simulation through a set of generic interfaces between surrogate and physics-based sections. To reproduce device inputs and outputs, system state is exposed as a deterministic event loop in a specialized discrete event simulator architecture. Because Fermilab is undergoing control system transition, several APIs were implemented as final user interfaces - a fully asynchronous EPICS soft IOC, a gRPC-based Data Pool Manager (DPM), and legacy ACNET protocols. We discuss implementation details as well as challenges handling live data assimilation and future plans to extend modelling to main complex proton accelerators like PIPII and Booster.
Paper: WEPD096
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD096
About: Received: 21 Sep 2025 — Revised: 26 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD097
Digital twin framework for PIP-II linac: AI-driven multi-scale modeling from ion source to 800 MeV
1253
The PIP-II superconducting linac at Fermilab is designed to deliver multi-megawatt proton beams for neutrino physics and other high-intensity applications. To expedite commissioning and enhance operational reliability, we have developed an EPICS-based data flow framework that seamlessly integrates digital twins (DT) with physical twins (PT). These digital twins comprise high-fidelity beam dynamics models or data-driven surrogate models connected to their physical counterparts through real-time diagnostics and advanced machine-learning algorithms. Central to this framework is Linac_Gen, an accelerated simulation tool that incorporates convolutional neural networks, random forests, and genetic algorithms to provide up to a tenfold speedup in optimizing the accelerator geometry model. An EPICS translator layer ensures interoperability by efficiently mapping lattice parameters across diverse simulation platforms. Our EPICS-based framework supports multiple operational modes—monitoring, passive learning, closed-loop control, and online learning—covering the entire machine lifecycle. By leveraging HPC resources and multi-objective optimization techniques, the digital twin enables adaptive trajectory correction, real-time fault detection, and predictive modeling of beam stability. This comprehensive approach paves the way for robust, high-intensity operation and data-driven accelerator R&D at Fermilab.
Paper: WEPD097
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD097
About: Received: 26 Sep 2025 — Revised: 05 Oct 2025 — Accepted: 31 Oct 2025 — Issue date: 25 Nov 2025
WEPD098
Fermilab's control system development with digital twin
use link to access more material from this paper's primary code
Control Systems development is often the last thing considered when designing and building new equipment, e.g. a new detector or superconducting RF LINAC; however when the new equipment is installed, it is the first thing desired to be operational for testing. Due to frequent delays in building new equipment and project deadlines, control system development and testing is often curtailed. A way to alleviate this problem is to simulate the control system, though this will be challenging for complex systems. The Fermilab PIP-II (proton improvement plan - II) project is being constructed at Fermilab to deliver $800\,MeV$ protons of $>1\,MW$ beam power to replace the present LINAC for the remainder of the existing accelerator complex. The new LINAC consists of a warm front end (WFE), 23 superconducting RF cryomodules (of 5 types), and a beam transfer line (BTL) to the existing complex. The accelerator physics group has a parallel project to create a digital twin (DT) of the PIP-II accelerator. We have coupled the EPICS controls to this DT and are developing both the DT and EPICS software in parallel. This will allow us to develop the EPICS software framework, the HMIs, sequences, high level physics applications, and other services for use in a fully functional control system. This presentation will detail the work that we have performed to date and show demonstrations of controlling and monitoring the status of the accelerator, as well as future plans for this work.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WESV003
About: Received: 06 Sep 2025 — Revised: 27 Oct 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD099
Real time computations of cryogenic He properties
1257
The Fermilab PIP-II (proton improvement plan - II) project is being constructed at Fermilab to deliver $800\,MeV$ protons of $>1\,MW$ beam power to replace the present LINAC and provide protons to the remainder of the existing accelerator complex. The new LINAC consists of a warm front end, 23 superconducting RF cryomodules, and a beam transfer line to the existing complex. The cryomodules (CMs) are to be tested at Fermilab's CryoModule Test Facility (CMTF). An important measurement in cryogenic testing is the heat load of each CM. Traditionally, at Fermilab, these measurements were made collecting archived data offline and analyzing it. The new control system for PIP-II is being developed with the EPICS (Experimental Physics and Industrial Control System) framework, which allows us to compute the heat load in real time using the HePak library. We are exploring other $He$ properties, such as flow, where flow meters are not available, which can also be calculated in real time and fed back to the cryogenics engineers. This paper details the real time heat load calculation and $He$ flow software developed for CM testing at CMTF, as well as the first results from the prototype HB650 CM. Future plans for 2-phase $LHe$ flow will also be outlined.
Paper: WEPD099
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD099
About: Received: 06 Sep 2025 — Revised: 08 Oct 2025 — Accepted: 04 Nov 2025 — Issue date: 25 Nov 2025
WEPD100
Integrating CODAC in ITER Plant Simulator
use link to access more material from this paper's primary code
The use of Digital Control Systems (DCS) with process simulators for engineering purpose, control system validation, virtual commissioning, or operator training is increasingly demanded in large and increasingly complex industrial projects. Coupling a DCS with a process simulator requires to support specific functionalities: ability to operate on a simulated time basis and to save/restore states to load different scenarios starting point, or to jump back in time, which is traditionally achieved by emulating or simulating the DCS. The ITER Control, Data Access and Communication (CODAC) system uses EPICS at its core, which is not designed to operate with such constraints. In the frame of the ITER Plant Simulator project, we leveraged advanced Linux features (libfaketime, namespaces, and CRIU) combined with a custom interface between CODAC and the simulator to meet these requirements. This approach allows integration of a wide range of CODAC tools (HMI, Archive, Alarms, Logbook, Operations Sequencer), synchronized with the simulator, with a lightweight and efficient solution.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMR013
About: Received: 06 Sep 2025 — Revised: 26 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
Development of virtual beamline technology for advanced light sources : simulation and application of key components
This study develops a virtual beamline technology for advanced light sources, with a focus on simulating the fundamental operational functions of critical devices including motors, double-crystal monochromators (DCM), fluorescent screens (FS), X-ray detectors, and X-ray beam position monitors (XBPM). By establishing parametric models, the simulation of device actions is achieved. It supports users in setting the displacement of motors, adjusting the Bragg angle of the DCM, and configuring the parameters of the slit aperture, and generates the corresponding state feedback signals of the devices.An interactive visualization interface is designed. Based on the state feedback signals of the devices, it generates the spot images on the fluorescent screen and synchronously displays the position trajectory of the beam measured by the XBPM, providing a visual reference for the beam tuning process. Through preliminary beam tuning simulations, the platform enables standardized operational workflows (e.g., energy selection) and optimizes parameter configuration sequences, effectively reducing trial-and-error adjustments during physical commissioning. The lightweight simulation framework proposed in this work offers a scalable and practical reference for advancing virtual commissioning technologies in synchrotron radiation facilities and other large-scale scientific installations.
Graphical interfaces and integration tools for particle accelerator digital twins
Particle accelerator modeling tools are largely based on open-source codes that utilize specialized command line interfaces. This makes them well suited for high performance computing workflows, however difficult for proficient usage. Moreover, the associated customized computing environments makes models difficult to share. RadiaSoft developed a browser-based toolkit enabling accelerator simulation interfacing and constructing end-to-end simulations. In a parallel partnership with SLAC, development of infrastructure for deploying our models leveraging online models integrated with LUME. Additionally, we have developed infrastructure for connecting our GUIs directly to beamlines through a control-system layer, bluesky. Our goal is to provide integration services for deploying models on the machine utilizing either a physics back-end or a machine learning back-end. Here we will introduce our toolkits, provide example demonstrations, and describe our plans for machine learning integration.
WEPD103
Accelerator digital twin development through simulation modeling and MLOps using the LUME ecosystem at SLAC
use link to access more material from this paper's primary code
Accelerator digital twins can enable real-time optimization and predictive control helping streamline complex facility operations and reduce setup time. Machine Learning(ML) models can enhance digital twin capabilities by leveraging prior experiments, known parameters, and real-time measurements. These require robust infrastructure and open-source software tools for accurate beam modeling and system integration. SLAC has developed the Lightsource Unified Modeling Environment (LUME) to enable large-scale modeling of X-ray free electron laser performance. The team is developing an ecosystem of tools towards start-to-end simulations for a much broader set of accelerators beyond light sources. LUME components include python wrappers for physics simulation which also work with a snapshot of the current accelerator reading through Kafka. LUME-services offers core infrastructure for model workflows including contextualized file service, model and results databases, and scheduling via Prefect for automated runs. LUME-Model holds the data structures used in the LUME modeling toolset, encapsulating physics simulations and ML models. LUME-EPICS is a dedicated API for serving LUME model variables with EPICS. MLflow is being integrated to manage the full machine learning lifecycle. This infrastructure is currently being developed and deployed at SLAC, alongside collaborators from other laboratories (especially LBNL). Here we provide an overview of the LUME ecosystem and example use cases.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMR012
About: Received: 05 Sep 2025 — Revised: 21 Sep 2025 — Accepted: 27 Oct 2025 — Issue date: 25 Nov 2025
WEPD104
MeerKAT antenna positioner emulator test bench project
1260
The MeerKAT Antenna Positioner Emulator (APE) project is being built to imitate the functionality of the antenna positioner servo system (AP). The AP is the mechanical component mainly responsible for the mounting and pointing of the antenna structure on the MeerKAT radio telescope. The project aim is fault finding, obsolescence mitigation, software debugging, and technical training. APE will use existing MeerKAT AP building blocks with a new mechanical design. The project will be designed for remote access, this involves network configurations. Internal interfaces include monitoring with electrical and internal infrastructure. The external interfaces include connections to Control And Monitoring software, time and frequency reference (TFR), development servers, and on-site hardware. The project is also focused on electronics design review, cable manufacturing, power supply considerations, mechanical aspects such as load tests and weight distribution. Project management activities include discussions on asset transfers, procurement, foundation design, and remote access considerations. Systems engineering involves documentations, discussions around user interface, and coordination with various stakeholders.
Paper: WEPD104
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD104
About: Received: 06 Sep 2025 — Revised: 26 Sep 2025 — Accepted: 30 Oct 2025 — Issue date: 25 Nov 2025
WEPD106
Enhancing SIRIUS fast orbit feedback actuators using IIR filters
1263
Studies with the SIRIUS Fast Orbit Feedback (FOFB) system revealed that its power supplies were operating near saturation, which limited the achievable FOFB controller gain. Further analysis identified avoidable noise sources in Beam Position Monitors (BPMs) and adjacent magnet power supplies as the main causes of this behavior. This paper presents a model-based approach employing Infinite Impulse Response (IIR) filters to attenuate the effects of such noise sources on the control effort as well as equalization of actuators' dynamic responses without any hardware modifications. The proposed method effectively reduces actuator effort and provides response equalization adding negligible phase shift in the band of interest, resulting in higher FOFB loop gain and improved rejection of orbit disturbances up to 1 kHz.
Paper: WEPD106
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD106
About: Received: 05 Sep 2025 — Revised: 18 Sep 2025 — Accepted: 24 Oct 2025 — Issue date: 25 Nov 2025
WEPD107
Designing the High-Dynamic Double Crystal Monochromators (HD-DCM-Lite) control system for fast energy scans and beam sub-nanometer stability at SIRIUS
use link to access more material from this paper's primary code
Two new High-Dynamic Double Crystal Monochromators (HD-DCM-Lite) have been successfully deployed on the SAPUCAIA (SAXS) and QUATI (quick-EXAFS) beamlines at SIRIUS. Building on previous work, that introduced the dynamic modeling and initial stabilization control strategies*, this paper details the mechatronic architecture, commutation schemes, and control strategies that enabled these systems to meet stringent operational requirements during online beamline validation. The contributions of this work can be summarized in three key areas: (i) the development of commutation rules that enable closed-loop motion control of the 3-phase brushless rotary stages; (ii) the control approach for coordinated motion of two goniometers to achieve single-degree-of-freedom movement; and (iii) the design of controllers for the high-bandwidth Short-Stroke system. At SAPUCAIA, the HD-DCM-Lite achieved sub-5 nrad RMS parallelism stability in the pitch direction, essential for ultra-low-noise scattering experiments. At QUATI, the system was able to reach high-speed energy scanning while maintaining the beam in fixed-exit condition, crucial for quality assurance in time-resolved spectroscopy. These results highlight the HD-DCM-Lite as a state-of-the-art mechatronic platform. Experimental data on crystal and beam stabilities during online operation confirm the effectiveness of the designed control system and commutation strategies.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMR017
About: Received: 06 Sep 2025 — Revised: 24 Oct 2025 — Accepted: 03 Nov 2025 — Issue date: 25 Nov 2025
WEPD109
Study design of a model-based controller for time varying delay compensation in a cryogenics process
1266
For the High Luminosity LHC (HL-LHC) project at CERN, a dedicated cryogenic test facility—HL-LHC IT String—is being commissioned. This cryogenic system including a helium refrigerator, a 100 m-long cryogenic distribution line, and a low-pressure pumping system with a cold compressor (CC). A key challenge is the management of time-varying delays in the system's thermal and pressure responses, especially in the presence of dynamic components, which requires control due to its axial-centrifugal configuration and thermal constraints. This paper presents a study on the design of a model-based controller employing a Smith Predictor architecture to compensate for time-varying delays in the cryogenic process. The controller is specifically tailored to the cold compressor’s operating sequence, which includes conditioning, regulation, and controlled stop phases—all of which impose strict operational and safety requirements due to the thermal and mechanical sensitivity of the system. Simulation results using simulation software are presented alongside experimental data from the commissioning of the IT String. The proposed control strategy enables stable operation of Line B during transient phases, such as magnet pre-loading and cooldown, and ensures the preservation of system integrity by limiting thermal gradients and compressor acceleration. The study gives some perspective in advance control model-based delay compensation in complex cryogenic infrastructures like those at CERN.
Paper: WEPD109
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEPD109
About: Received: 29 Aug 2025 — Revised: 21 Sep 2025 — Accepted: 29 Oct 2025 — Issue date: 25 Nov 2025
WEPD110
Scheduler for cooling and ventilation plants: feedback on easy and low cost method for energy savings
use link to access more material from this paper's primary code
In industrial engineering, scheduling is a well-established strategy for optimizing resource use and minimizing operational costs. At CERN's Engineering department, the Cooling and Ventilation group has implemented an automatic scheduling solution to reduce electricity consumption by selectively shutting down plants during nights and weekends, when their operation is not required. Given that CV systems account for a significant share of CERN's total electricity use, even simple scheduling strategies can yield substantial energy savings - up to 75% in some cases. This paper presents the motivation, methodology, and preliminary results of scheduler deployments across multiple CV plants between 2023 and 2025, including recent pilots at Point 5 of the Large Hadron Collider (LHC). Two types of scheduler conditions were implemented: calendar-based (e.g., operating only during working hours) and temperature-based (e.g., starting only when zone temperature thresholds are exceeded). Operational safety was carefully assessed - a CO₂ measurement campaign was conducted at Point 5 to confirm compliance with environmental and safety requirements. Preliminary results from several sites show significant reduction in consumption without compromising performance. This low-cost approach demonstrates how simple digital solutions can lead to impactful energy savings in large-scale technical infrastructures.
DOI: reference for this paper: 10.18429/JACoW-ICALEPCS2025-WEMR015
About: Received: 06 Sep 2025 — Revised: 20 Sep 2025 — Accepted: 24 Oct 2025 — Issue date: 25 Nov 2025