Code for simulating the production FACET accelerator with the FACET2E Bmad model. There are two primary layers of infrastructure here:
- BmadLiveModel: monitors the production controls system and updating a local instance of PyTao to simulate the FACET-II linac
- FACET live model PVA server: publishes model data to PVs
A python class for monitoring the live accelerator controls system via EPICS Channel Access and AIDA-PVA as needed, and updating a local instance of PyTao to match extant machine settings. The live model is also responsible for assessing the current klystron complement and estimating the live momentum profile of the beam, as well as suggested lattice magnet settings BLEM.
- Link to full design overview: https://confluence.slac.stanford.edu/display/FACET/Bmad+live+modeling
- API documentation: https://f2-live-model.readthedocs.io/en/latest/
- Built using PyTao: https://bmad-sim.github.io/pytao/
- FACET design lattice: https://github.com/slaclab/facet2-lattice
The simplest method of using this class is to only serve the design model. This is basically like using PyTao with extra steps.
>>> from F2_live_model.bmad import BmadLiveModel
>>> f2m = BmadLiveModel(design_only=True)There are two methods of actually monitoring the production accelerator, the first of which is to update data only when asked. This is called "instanced" mode.
>>> from F2_live_model.bmad import BmadLiveModel
>>> f2m = BmadLiveModel(instanced=True)
>>> # your code goes here
>>> f2m.refresh_all()
>>> # even more code!The second method is to stream data in real-time, which is more resource intensive. This can be handled either by using a context manager (preferred) or with the provided .start and .stop functions.
>>> from F2_live_model.bmad import BmadLiveModel
>>> with BmadLiveModel() as f2m:
>>> # your code here, updates happen in the background>>> from F2_live_model.bmad import BmadLiveModel
>>> f2m = BmadLiveModel()
>>> f2m.start()
>>> # your code here, updates happen in the background
>>> f2m.stop()The BmadLiveModel object stores some live & design model data locally as python data structures.
>>> f2m.design.twiss.beta_x
<np.ndarray>>>> f2m.live.p0c
<np.ndarray>>>> f2m.get_rmat('QE10525')
<6x6 np.ndarray>>>> f2m.live.quads['QE10525']
<structs._Quad>The local instance of PyTao is also publicly accessible.
>>> f2m.tao.cmd('set ele QE10525 b1_gradient -1.5')
<results snipped>Full interface details can be found in the online documentation.
The FACET model server is responsible for running a live-streaming BmadLiveModel, and publishing data from it to PVs via PVA. The provided RMAT and twiss tables are compatible with the python meme service.
- Documentation for the meme service is here: https://github.com/slaclab/meme
The live model server also published LEM data, which is not ingested by meme, but is used by the LEM server app (coming soon ...).
This server publishes the following PVs:
BMAD:SYS0:1:FACET2E:DESIGN:TWISS BMAD:SYS0:1:FACET2E:DESIGN:RMAT BMAD:SYS0:1:FACET2E:DESIGN:URMAT BMAD:SYS0:1:FACET2E:LIVE:TWISS BMAD:SYS0:1:FACET2E:LIVE:RMAT BMAD:SYS0:1:FACET2E:LIVE:URMAT BMAD:SYS0:1:FACET2E:LEM:DATA BMAD:SYS0:1:FACET2E:LEM:L0_AMPL BMAD:SYS0:1:FACET2E:LEM:L1_AMPL BMAD:SYS0:1:FACET2E:LEM:L2_AMPL BMAD:SYS0:1:FACET2E:LEM:L3_AMPL BMAD:SYS0:1:FACET2E:LEM:L0_CHIRP BMAD:SYS0:1:FACET2E:LEM:L1_CHIRP BMAD:SYS0:1:FACET2E:LEM:L2_CHIRP BMAD:SYS0:1:FACET2E:LEM:L3_CHIRP BMAD:SYS0:1:FACET2E:LEM:L0_FUDGE BMAD:SYS0:1:FACET2E:LEM:L1_FUDGE BMAD:SYS0:1:FACET2E:LEM:L2_FUDGE BMAD:SYS0:1:FACET2E:LEM:L3_FUDGE