Mark settings as good and retrieve them

Workshop conducted on 15 July 2021

Participants

• Andreas Schaller
• Christoph Hessler
• David Ondreka
• Frank Herfurth
• Hanno Hüther
• Jens Stadlmann
• Jutta Fitzek
• Martin Stein
• Miriam Klich
• Oksana Geithner
• Stephan Reimann
• Wolfgang Geithner

Host: Christian Hillbricht

Possible save point contents

The term 'save point' is used here for what is most often called a 'save'. This doesn't necessarily refers to a certain point in the trim history, it could also be a (partial) copy of the current settings plus other information stored elsewhere. How this is solved technically will be decided later.

Name	Category	Retrieved automatically	Facility scope	Why it should be saved?	Remarks
Energy	Beam Setting	Always	Typically global	Matching with target experiment	May change somewhere in the beamline. Typically read at last ring extraction
Element	Beam Setting	Always	Typically global	Matching with target experiment	May change somewhere in the beamline. Typically read at last ring extraction
Charge state	Beam Setting	Always	Typically global	Matching with target experiment	May change somewhere in the beamline. Typically read at last ring extraction
Isotope	Beam Setting	Always	Typically global	Matching with target experiment	May change somewhere in the beamline. Typically read at last ring extraction
Hierarchy version	Data supply metadata			Matching with target experiment	Available?
Ion source	Beamline	Always	Global	Matching with target experiment	Could be the ion source terminal or the actual ion source e.g. VARIS or both
Beam line	Beamline	Always	Global	Matching with target experiment	Beamline could be e.g. particle transfer transfer lines vias
Beam destination	Beamline	Always	Global	Matching with target experiment
Time range / point in time	Metadata	Always	Global	Matching with target experiment	This could be e.g. Last Trim time Save time(Last) scheduled time range (end time not available when creating save point) Validity period
Intensity	Measured value	No permanent measurement guaranteed	Various locations	The higher the better / Matching with desired intensity	Possible measurement locations: Experiment Last ring (flattop)
Transmission	Measured value	No permanent measurement guaranteed	Various locations	The higher the better
Partial transmissions					needs clarification
Author	Metadata	no	Global	Knowledge who created the safe/optimization Some users are being associated with creating good saves.	This could be e.g. User who created the save User who performed the optimization
Keywords	Metadata	no	Global	Improve findability
Predefined properties	Metadata	no	Global	Improve findability	Weakened beam Beam position
Usage number	Statistics	yes	Global	The higher the better	How often was the save point used? How often was the associated context scheduled?
Experiment and shift information (OLOG)	Metadata	yes	Global	Improve findability	Shift Spokesperson Experiment number
Stability					How to quantify stability? How to detect a wobbling beam? Alternativally juse a 'predefined property' or even just a reference marker.
Radiation protection	Measured value		Various locations		Certain location or along the beamline?
Low magnet stimulation		yes		The lower the better	Derive set of magnets from chain or user-defined set?
Settings close to theory settings	Beam Setting	yes		"Settings closer to theory work better most of the time"	Certain setting(s) or all settings?
Focus				Reuse of equal or similar focus	Relevant only at experiment? How to save? Alternativally juse a 'predefined property' or even just a reference marker.
Overall rating	Metadata	no		Summarizes subjective overall impression of beam	Added by C. Hillbricht Possible way to save could be e.g: Star rating
Associated context(s)		yes		Some information is context-related not save point-related e.g. when storing the settings-only it’s not clear when these context containing the settings was scheduled	Added by C. Hillbricht
Comment	Metadata	no		Improve findability	Added by C. Hillbricht Mentioned in workshop but not added to board
Console information	Metadata	yes		Improve findability	Added by C. Hillbricht This could be the tcl name or a console group name like 'SIS console'

Save point details

• Measured values should be collected automatically when available. If no measured values are available when creating the save point (e.g. detector not positioned), provide a way to add this information manually.
• Created safe points should be available over a long period of time. As the LSA database requires a lot of storage which might lead to deletion of some contexts and their settings, ideas have been collected how availability over a long period of time could be achieved.
• Store save points independently from LSA contexts
• Only store parameters displayed in ParamModi standard view
• Store LSA settings in archive system, expected amount of storage is negliable in archiving system dimensions
• Save storage by cleaning up settings history e.g. removing unmarked settings regularly

• It has been discussed briefly if using keywords can fully replace predefined properties. Keywords could evolve into a self-organized system of ordering. They don’t provide a key-value mechanism. An alternative way would be to provide the opportunity to add keys/properties dynamically, or display some predefined keys on demand. Too many visible keys/properties could lead to overload and might not always apply to all safe points.
• Should we save non-multiplexed settings such as slits? It must be considered that foil positions may change over time. Current approach is to use the comment field.

Limit save point settings scope

Sometimes users optimize only a certain category of settings along the beamline. The other settings are not necessarily better than theory settings. Operators who use these save points should not be tempted to consider the whole save point as optimized - especially when the user who optimized the settings is a well known expert. This could be the categories in ParamModi such as injection or.

In case of Benno this would be magnet strengths such as HKICK/VKICK of steerers and bending magnets, KL of quadropols and sextupol magnets and a few other settings for documentation purposes. These parameters pretty much align with the extraction line category in ParamModi. Details provided by C. Hessler.

Setting a reference and going back to it

• Operators would like to save a reference to the current settings for the following reasons:
  • In order to return to marked settings later, that have been perceived as good
  • Marking the so called 'Schichtreferenz'
  • Switching between different operation scenarios for a single chain e.g. later returning back to normal intensity before weakening a beam.
• There could a predefined set of typical markers and a simple way to add custom markers.
• Setting a reference could be performed as a marker to the certain trim (instead of actually creating a save point).
• In addition to markers going back to settings active at a certain point in time is required. Operators might remembering a certain point in time that can be associated with a good or at least working setting.

Retrieving (good) settings

• Users need to know for beam the settings should match. This can be retrieved from the beamtime schedule or experimentors wishes. In the context of creating a new chain, this information should already be present.
• Proper performance is a requirement. Current search is way too slow (which one?). Exact performance requirements missing
• One idea was to provide a base view of most often needed filters and an extended view with all filters (D. Ondreka).
• How to deal with missing information such as missing intensity? Include/exclude in search results? Make certain information mandatory?
• How to know when to stop searching? Set filters may exclude too many or too less safe points. One idea could be to give hints which set filters exclude many search results (C: Hillbricht).
• When the target beam has a different beamline than in the beamline of the preferred safe point, only matching settings must be loaded into the target chain. Note: This should already work today (but not always).

Safe point based setting comparison and analysis tool

This is not exactly about retrieving good settings. Based on a tool that can be used to find existing settings, a search result, (consisting of two or more save points) that meets certain criteria could also be used to compare analyse settings. Comparing certain settings contained in these save points could e.g. highlight larger deviations, scattering or outliers. Details provided by F. Herfurth.

Workshop board

View workshop board

Board contents exported on 29th July 2021. Board can still be edited.

-- ChristianHillbricht - 29 Jun 2021