This material is based upon work supported by the U.S. Department of Energy Office of Science under Cooperative Agreement DE-SC0000661, the State of Michigan and Michigan

State University. Michigan State University designs and establishes FRIB as a DOE Office of Science National User Facility in support of the mission of the Office of Nuclear Physics.

Physics High-level Applications and Toolkitfor Accelerator System

An overview of FRIB high-level physics applications development

Tong ZhangControls Physicist

E-mail: zhangt@frib.msu.eduPresented at EPICS Collaboration Meeting, 11-15 June 2018, APS, ANL

Acknowledgements

Dylan Maxwell, Martin Konrad, Daron Chabot, Eric Berryman,Steven Beher, Diego Omi�o, Guobao Shen, Michael Davidsaver,Steven Lidia

Kei Fukushima, Harry He, Tomofumi Maruta, Takashi Yoshimoto,Jonathan Wong, Alexander Plastun, Qiang Zhao, Yue Hao, Steve Lund,Peter Ostroumov, Masanori Ikegami

1 / 14

Outline

1 Designed ArchitectureIntroductionDevice Abstraction

2 Key FeaturesVirtual AcceleratorOnline Model

3 Conclusions

2 / 14

Outline

1 Designed ArchitectureIntroductionDevice Abstraction

2 Key Features

3 Conclusions

3 / 14

Introductions to High-level Physics Applications

Accelerator System

particle source, beam transport, end stations, ...

devices: optics, diagnostics, ...

distributed controls units: EPICS input & output controllers (IOCs)

High-level Physics Applications

Final goal: operating accelerator facility

Purpose: have robust and functional beam tuning algorithms

Solution: so�ware environment for high-level physics controls

High-level Physics Applications = Physics Algorithms + Controls So�ware

3 / 14

Introductions to High-level Physics Applications

Accelerator System

particle source, beam transport, end stations, ...

devices: optics, diagnostics, ...

distributed controls units: EPICS input & output controllers (IOCs)

High-level Physics Applications

Final goal: operating accelerator facility

Purpose: have robust and functional beam tuning algorithms

Solution: so�ware environment for high-level physics controls

High-level Physics Applications = Physics Algorithms + Controls So�ware

3 / 14

Introductions to High-level Physics Applications

Accelerator System

particle source, beam transport, end stations, ...

devices: optics, diagnostics, ...

distributed controls units: EPICS input & output controllers (IOCs)

High-level Physics Applications

Final goal: operating accelerator facility

Purpose: have robust and functional beam tuning algorithms

Solution: so�ware environment for high-level physics controls

High-level Physics Applications = Physics Algorithms + Controls So�ware

3 / 14

Introductions to High-level Physics Applications

Accelerator System

particle source, beam transport, end stations, ...

devices: optics, diagnostics, ...

distributed controls units: EPICS input & output controllers (IOCs)

High-level Physics Applications

Final goal: operating accelerator facility

Purpose: have robust and functional beam tuning algorithms

Solution: so�ware environment for high-level physics controls

High-level Physics Applications = Physics Algorithms + Controls So�ware

3 / 14

Introductions to High-level Physics Applications

Accelerator System

particle source, beam transport, end stations, ...

devices: optics, diagnostics, ...

distributed controls units: EPICS input & output controllers (IOCs)

High-level Physics Applications

Final goal: operating accelerator facility

Purpose: have robust and functional beam tuning algorithms

Solution: so�ware environment for high-level physics controls

High-level Physics Applications = Physics Algorithms + Controls So�ware

3 / 14

Introductions to High-level Physics Applications

Accelerator System

particle source, beam transport, end stations, ...

devices: optics, diagnostics, ...

distributed controls units: EPICS input & output controllers (IOCs)

High-level Physics Applications

Final goal: operating accelerator facility

Purpose: have robust and functional beam tuning algorithms

Solution: so�ware environment for high-level physics controls

High-level Physics Applications = Physics Algorithms + Controls So�ware

3 / 14

Introductions to High-level Physics Applications

Accelerator System

particle source, beam transport, end stations, ...

devices: optics, diagnostics, ...

distributed controls units: EPICS input & output controllers (IOCs)

High-level Physics Applications

Final goal: operating accelerator facility

Purpose: have robust and functional beam tuning algorithms

Solution: so�ware environment for high-level physics controls

High-level Physics Applications = Physics Algorithms + Controls So�ware

3 / 14

Software Solution upon Python Programming Language

Fundamental Requirements

�ick prototyping: dynamic programming language

Functional: plenty of third-party packages

Agile development: develop → build → test → deploy

4 / 14

Software Solution upon Python Programming Language

P H A N T A S YPhysics High-level Applications aNd Toolkit for Accelerator SYstem

Features Highlight

Device configuration management

Device abstraction

Online modeling

Python interactive scripting environment for high-level controls

Virtual accelerator based on EPICS control environment

Web service integration (channelfinder, UNICORN, scanserver)

Deployment

Target OS: Debian 8 (Jessie)

Main packages: python-phantasy, phantasy-machines

Physics model engines: python-flame, python-impact

4 / 14

Physics Applications Architecture

Toolkit

CLI commands, data management, convenient scripts, ...

Applications

Virtual accelerators, orbit correction, parameters scan/optimization, ...

5 / 14

Modeling Architecture: Overview

6 / 14

Modeling Architecture: Device Abstraction

Information abstraction and aggregation:

7 / 14

Modeling Architecture: Device Abstraction

Information abstraction and aggregation:

7 / 14

Outline

1 Designed Architecture

2 Key FeaturesVirtual AcceleratorOnline Model

3 Conclusions

8 / 14

Virtual Accelerator

8 / 14

Create EPICS controls environment for development, physics behaviorsimulated by model engine (flame, impact).

Virtual Accelerator

8 / 14

Create EPICS controls environment for development, physics behaviorsimulated by model engine (flame, impact).

Interactive Python scripting Environment

Create a full-featured high-level abstracted so�ware environment,Accelerator Physicists focus on solving physics problems.

9 / 14

UNICORN: Unit Convertion Web application

REST APIs to evoke scaling laws: Python-client or web page

Represent devices with an informative way

Manage scaling rules in a friendly wayDebian package:

Web application: unicorn-webappPython interface: python-unicorn, python3-unicorn

10 / 14

UNICORN: Unit Convertion Web application

10 / 14

UNICORN: Unit Convertion Web application

10 / 14

UNICORN: Unit Convertion Web application

10 / 14

Deployment (I)

FRIB controls networkgit → stash → jenkins → puppet → target workstations

Local developmentVirtualBox Appliance

Cloud developmentDocker container based web computing platform (configurable-proxy,docker images)

11 / 14

Deployment (II)

12 / 14

Deployment (II)

12 / 14

Outline

1 Designed Architecture

2 Key Features

3 Conclusions

13 / 14

Conclusions

Established Python-based so�ware infrastructure for high-levelphysics controls

The solution for systematic high-level device abstraction

Dedicated web application and Python interface for unitsinterpretation

Continuous integration and delivery at FRIB

Future Plans

Operation: develop mature physics algorithms into so�-IOCs

Python ecosystem: data management

13 / 14

Thank you for your a�ention!