MC6: Beam Instrumentation, Controls, Feedback and Operational Aspects
T33 Online Modeling and Software Tools
Paper Title Page
TUZPLM1 Adding Data Science and More Intelligence to Our Accelerator Toolbox 1191
 
  • S. Biedron
    University of New Mexico, Albuquerque, USA
  • S. Biedron
    Element Aero, Chicago, USA
 
  Requirements for recent accelerators are becoming more and more stringent and sophisticated machine tuning is necessary. A large amount of data is acquired from accelerator components as an assistant of machine tuning. It is hard for operators to utilize all the accelerator data for machine tuning. Therefore, machine learning, data mining and big data handling are recently applied to accelerators. For instance, Bayesian optimization is used for maximizing a target performance, a clustering algorithm is used for anomaly detection, and hidden correlation finding is utilized for discovering new aspects of a machine. This talk reviews recent progress of machine learning applications and big data handling in accelerators.  
slides icon Slides TUZPLM1 [11.978 MB]  
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2019-TUZPLM1  
About • paper received ※ 20 May 2019       paper accepted ※ 16 June 2019       issue date ※ 21 June 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
TUZZPLM1 Operational Results of LHC Collimator Alignment Using Machine Learning 1208
SUSPFO053   use link to see paper's listing under its alternate paper code  
 
  • G. Azzopardi, A. Muscat, G. Valentino
    University of Malta, Information and Communication Technology, Msida, Malta
  • S. Redaelli, B. Salvachua
    CERN, Geneva, Switzerland
 
  A complex collimation system is installed in the Large Hadron Collider to protect sensitive equipment from unavoidable beam losses. The collimators are positioned close to the beam in the form of a hierarchy, which is guaranteed by precisely aligning each collimator with a precision of a few tens of micrometers. During past years, collimator alignments were performed semi-automatically*, such that collimation experts had to be present to oversee and control the alignment. In 2018, machine learning was introduced to develop a new fully-automatic alignment tool, which was used for collimator alignments throughout the year. This paper discusses how machine learning was used to automate the alignment, whilst focusing on the operational results obtained when testing the new software in the LHC. Automatically aligning the collimators decreased the alignment time at injection by a factor of three whilst maintaining the accuracy of the results.
*G.Valentino et al., "Semi-automatic beam-based LHC collimator alignment", PRSTAB, no.5, 2012.
 
slides icon Slides TUZZPLM1 [6.060 MB]  
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2019-TUZZPLM1  
About • paper received ※ 10 May 2019       paper accepted ※ 21 May 2019       issue date ※ 21 June 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
TUZZPLM2 Status of Automated Optimization Procedures at the European XFEL Accelerator 1212
 
  • S. Tomin
    EuXFEL, Schenefeld, Germany
  • L. Fröhlich, M. Scholz
    DESY, Hamburg, Germany
 
  The European XFEL is in the operational stage since fall 2017. Since then, tuning of the FEL performance (e.g. of the photon pulse energy) has become increasingly important. Due to a large number of parameters to which FEL facilities are highly sensitive and their complex correlations, controlling and optimizing them in a speedy manner is becoming a very important and challenging task. Several automated optimization procedures were developed to optimize the FEL beam quality. In this work, we present the status and the results of these activities, as well as the optimization statistics.  
slides icon Slides TUZZPLM2 [5.882 MB]  
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2019-TUZZPLM2  
About • paper received ※ 13 May 2019       paper accepted ※ 21 May 2019       issue date ※ 21 June 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
THPRB001 Applications of Online Optimization Algorithms for Injection at the Australian Synchrotron 3795
 
  • R. Auchettl, R.T. Dowd
    AS - ANSTO, Clayton, Australia
 
  At the Australian Synchrotron, accelerator tuning predominantly occurs via manual optimization or traditional optimization techniques such as the Linear Optics from Closed Orbits (LOCO) algorithm. While we have had distinct success with the implementation of LOCO* and manual tuning, these strategies are not without their downsides. Some situations (such as the optimization of synchrotron beam dynamics) produce a design space too large and multifaceted for manual tuning while implementing LOCO can be computationally expensive. Also, without sufficient diagnostic systems, both LOCO and manual tuning do not necessarily guarantee that the optimal solution will be found. Motivated by the successful implementation of online optimization algorithms at SPEAR3**, this paper outlines the application of online optimization algorithms to improve the performance of the Australian Synchrotron injection system. We apply the efficient Robust Conjugate Direction Search (RCDS) Algorithm to reduce beam loss along the Booster-to-Storage ring (BTS) Transfer line and Storage Ring and compare against the LOCO method.
* R. Dowd et al. (2011), Phys. Rev. ST: AB, 14, 012804.
** X. Huang et al. (2013), Nucl. Instr. Meth. A., vol. 726, pp. 77-83.
 
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2019-THPRB001  
About • paper received ※ 08 April 2019       paper accepted ※ 21 May 2019       issue date ※ 21 June 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
THPRB012 Toward Autonomous Phasing of ISAC Heavy Ion LINACs 3827
 
  • O. Shelbaya, R.A. Baartman, O.K. Kester, S. Kiy
    TRIUMF, Vancouver, Canada
 
  Ongoing development work at TRIUMF aims to implement a model-based tuning approach for accelerators, with the goal of automation of tuning tasks and minimizing tuning times. As a part of this, work is underway toward the development of an analytical model of the linacs using the methodology of Hamiltonian based beam envelope dynamics. The TRIUMF High-Level Applications (HLA) project has been developing software that allows direct interfacing with the control system. The envelope code TRANSOPTR is now being extended to simulate the ISAC-II Superconducting Linac. Within the emerging HLA framework, this will allow for automated phasing and tuning of the linac. The steps of the model development will be presented in this contribution.  
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2019-THPRB012  
About • paper received ※ 10 May 2019       paper accepted ※ 22 May 2019       issue date ※ 21 June 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
THPRB028 Redesign of the JavaFX Charts Library in View of Real-Time Visualisation of Scientific Data 3868
 
  • R.J. Steinhagen, H. Bräuning, A. Krimm, T. Milosic
    GSI, Darmstadt, Germany
 
  The accurate graphical representation of accelerator- or beam-based parameters is crucial for commissioning and operation in any modern accelerator. Charts are one of the most visible but at the same time often underappreciated accelerator control system components even though these are crucial for easing and improving a quick intuitive understanding of complex or large quantities of data, which in turn is used to efficiently control, troubleshoot or improve the accelerator performance. While the Java SDK and other third-party libraries provide some charting components, we found that these lack either functionality, performance, or are based on outdated complex APIs. Based on earlier GSI and CERN designs and careful analysis of missing functionalities, performance bottlenecks, and long-term maintenance risks for the necessary workarounds, we decided that it was worth to re-engineer a new scientific charting library that preserves the functionality of established other libraries while addressing the performance bottlenecks and APIs issues. The new library offers a wide variety of plot types common in the scientific community, a flexible plugin system to extend the functionality towards chart interactors as well as online parameter measurements commonly found in oscilloscopes. Tailored towards high performance, it achieves real-time update rates up to 25 Hz for data sets with a few 10k up to 5 million data points. The new API shields the complexity from and eases the library’s use by normal users, while still being modular and having explicitly open interfaces that allow more-inclined developers to modify, add or extend missing functionalities. This contribution provides a performance and functionality comparison with other existing Java-based charting libraries.  
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2019-THPRB028  
About • paper received ※ 15 May 2019       paper accepted ※ 18 May 2019       issue date ※ 21 June 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
THPRB038 ALARM SYSTEM OF IRFEL AT NSRL 3896
 
  • X. Chen, C. Li, G. Liu, Z.X. Shao, Y. Song, J.G. Wang, K. Xuan
    USTC/NSRL, Hefei, Anhui, People’s Republic of China
 
  An InfraRed Free Electron Laser Light (IRFEL) is under commissioning at National Synchrotron Radiation Laboratory (NSRL). The control system of IRFEL is a distributed system based on Experimental Physics and Industrial Control System (EPICS). The alarm system is an essential part of the control system. It is developed based on the software Phoebus. The module named "Alarms" in Phoebus can store states and configuration information of the Process Variable (PV) in the Kafka topics. To meet our requirements, 3 kinds of alarm message distribution applications are developed, i.e. Web-Based GUI, WeChat and SMS. This paper will introduce the alarm system architecture and the implementations of the applications for alarm message distribution.  
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2019-THPRB038  
About • paper received ※ 17 April 2019       paper accepted ※ 21 May 2019       issue date ※ 21 June 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
THPRB053 Upgrade of the Historical Data Query and Analysis System for HLS-II 3928
 
  • Z.Y. Xie, C. Li, G. Liu, Z.X. Shao, Y. Song, J.G. Wang, K. Xuan
    USTC/NSRL, Hefei, Anhui, People’s Republic of China
 
  The current historical data query and analysis system for the Hefei Light Source II (HLS-II) was developed with Apache Struts2. However, Apache Struts2 need to be fixed from time to time to avoid being attacked. Therefore, a new system based on Spring Boot and Vue.js is developed. Meanwhile, the performance of the system is optimized, and the radiation monitor module is added. This paper will detail the system architecture and software implementation.  
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2019-THPRB053  
About • paper received ※ 24 April 2019       paper accepted ※ 23 May 2019       issue date ※ 21 June 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
THPRB077 Optics Corrections Using Machine Learning in the LHC 3990
 
  • E. Fol, J.M. Coello de Portugal, R. Tomás
    CERN, Meyrin, Switzerland
  • G. Franchetti
    GSI, Darmstadt, Germany
 
  Optics corrections in the LHC are based on a response matrix approach between available correctors and observables. Supervised learning has been applied to quadrupole error prediction at the LHC giving promising results in simulations and surpassing the performance of the traditional approach. A comparison of different algorithms is given and it is followed by the presentation of further possible concepts to obtain optics corrections using machine learning.  
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2019-THPRB077  
About • paper received ※ 14 May 2019       paper accepted ※ 21 May 2019       issue date ※ 21 June 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
THPRB080 Automatisation of the SPS ElectroStatic Septa Alignment 4001
 
  • S. Hirlaender
    ATI, Vienna, Austria
  • M.A. Fraser, B. Goddard, V. Kain, J. Prieto, L.S. Stoel, F.M. Velotti
    CERN, Geneva, Switzerland
  • M. Szakaly
    Wigner Research Centre for Physics, Institute for Particle and Nuclear Physics, Budapest, Hungary
 
  An electrostatic septum composed of 5 separate tanks is used to slow-extract the 400 GeV proton beam resonantly on the third integer resonance from the CERN SPS. The septa are all mounted on a single support structure that can move the ensemble coherently and, in addition, the internal anode and cathode of each tank can be moved independently. The septum is aligned to the beam by measuring and minimising the induced beam loss signals in the extraction region following an alignment procedure that is usually carried out manually at the beginning of each year. The large number of positional degrees of freedom complicates the procedure and until recently each tank was aligned one after the other semi-manually, typically requiring 8 hours. It is not uncommon that the septum has to be re-aligned later in the run taking time away from physics programme. To tackle this issue, a simplified beam dynamics and scattering simulation routine was developed to permit error studies with a large number of seeds to be carried out in a reasonable computation time. In this contribution, the simulation model will be described before the results of its exploitation to understand the efficacy of alignment procedures based on different optimization algorithms are discussed and compared to the present operational procedure. The effort culminated with the implementation of an automated alignment procedure based on a Powell optimisation algorithm that reduced the time needed to align the septum by over an order of magnitude.  
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2019-THPRB080  
About • paper received ※ 14 May 2019       paper accepted ※ 22 May 2019       issue date ※ 21 June 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
THPRB091 APPLICATION PROGRAMS FOR TPS BEAM TRIP ANALYSIS 4032
 
  • B.Y. Chen, T.W. Hsu, B.Y. Huang, C.S. Huang, C.H. Kuo, T.Y. Lee, W.Y. Lin
    NSRRC, Hsinchu, Taiwan
 
  For the Taiwan Photon Source (TPS), the orbit inter-lock system is one of the most important machine pro-tection systems. It is the fastest and the most preferred system to detect abnormalities to prevent possible dam-ages caused by magnet power supply failures or subsys-tems failures. In order to monitor electron orbit changes during a beam trip, we developed the ’orbit monitoring and recording tool’, the ’TBT BPM analysis tool’ and the ’magnet power supply recording and analysis tool’ to assist us in the failure analysis as will be discussed in this paper.  
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2019-THPRB091  
About • paper received ※ 15 May 2019       paper accepted ※ 22 May 2019       issue date ※ 21 June 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
THPRB095 A Simulation Framework for Photon-Particle Interactions for Laserwires and Further Applications 4045
SUSPFO112   use link to see paper's listing under its alternate paper code  
 
  • S. E. Alden, S.M. Gibson, L.J. Nevay
    JAI, Egham, Surrey, United Kingdom
 
  A model has been developed for simulating photon-particle interactions with Beam Delivery Simulation (BDSIM). BDSIM is a high energy physics program that utilises the Geant4, CLHEP, and ROOT libraries to seamlessly track particles through an accelerator. The photon-particle interactions introduce the capability for modelling a range of applications in accelerator physics. One such application is a laserwire which is a minimally invasive diagnostic technique to measure beam profiles and emittance. In this paper we describe the recent implementation of inverse Compton scattering and electron stripping of Hydrogen ions. This is demonstrated on an example beamline.  
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2019-THPRB095  
About • paper received ※ 15 May 2019       paper accepted ※ 22 May 2019       issue date ※ 21 June 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
THPRB099 Applications of Dimension-Reduction to Various Accelerator Physics Problems 4060
 
  • W.F. Bergan, I.V. Bazarov, C.J.R. Duncan, D. L. Rubin
    Cornell University (CLASSE), Cornell Laboratory for Accelerator-Based Sciences and Education, Ithaca, New York, USA
 
  Funding: DE-SC 0013571 DGE-1650441 OIA-1549132
Particle accelerators contain hundreds of magnets, making dimension-reduction techniques attractive when attempting to tune them. We apply this procedure to two different problems: correcting the orbit in the Cornell synchrotron and maximizing the dynamic aperture in the Cornell Electron Storage Ring (CESR). Cornell’s rapid cycling synchrotron accepts a 200 MeV beam from the linac and accelerates it to 6 GeV for injection into the CESR. ‘Kicker coils’ (dipole correctors) are used to correct for residual fields which would otherwise cause beam loss at the low energies. In such cases, it is usually advisable to measure and correct the orbit. However, one cannot measure the orbit without first getting the beam to circulate a few hundred times, by which point the low-energy orbit would already be mostly corrected. In order to speed up the process of empirical orbit tuning, we form knobs which have the largest effect on the global orbit error, so that the dimensionality of the space which must be searched may be greatly reduced. A small dynamic aperture in CESR will have adverse effects on beam lifetime and injection efficiency, and so ought to be maximized by tuning sextupoles. However, it is often unclear which sextupoles one ought to tune to alleviate the problem. Moreover, once the chromaticity is properly adjusted, it should not be changed. Since we expect resonance driving terms (RDTs) to have a large impact on the dynamic aperture, we develop sextupole knobs which change the RDTs as much as possible while leaving the chromaticity fixed.
 
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2019-THPRB099  
About • paper received ※ 14 May 2019       paper accepted ※ 23 May 2019       issue date ※ 21 June 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
THPRB100 A Generic Software Platform for Rapid Prototyping of Online Control Algorithms 4063
SUSPFO123   use link to see paper's listing under its alternate paper code  
 
  • C.J.R. Duncan, M.B. Andorf, I.V. Bazarov, I.V. Bazarov, C.M. Gulliford, V. Khachatryan, J.M. Maxson, D. L. Rubin
    Cornell University (CLASSE), Cornell Laboratory for Accelerator-Based Sciences and Education, Ithaca, New York, USA
  • I.V. Bazarov
    Cornell University, Ithaca, New York, USA
 
  Funding: US Department of Energy DE-SC 0013571
Algorithmic control of accelerators is an active area of research that promises significant improvements in machine performance. To facilitate rapid algorithm prototyping, we have developed a generic interface between accelerator controls, beam physics modelling software and modern scripting languages. The work-flow of a project using this interface begins with testing algorithms of choice offline in simulation. After off-line testing, the same code can be deployed on real machines via the Experimental Physics and Industrial Control System (EPICS) API. We include noise in our simulations in order to mimic realistic accelerator behaviour and to evaluate robustness of algorithms to experimental uncertainties and long-term drifts. The results of test cases of using this framework are presented, including emittance tuning of the Cornell Electron Storage Ring (CESR), correction of diurnal drift in CESR steering and orbit correction on CESR and the Cornell-BNL ERL Test Accelerator (CBETA).
 
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2019-THPRB100  
About • paper received ※ 14 May 2019       paper accepted ※ 23 May 2019       issue date ※ 21 June 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)