DAQ Use Cases (for discussion purposes)

• Use Case 1-1: Analyze and display the beam energy statistics over a given (and rather long) time range [e.g. 15 hours of 800 bunches / 3mA]
  • Time interval for averaging should be configurable [e.g. 10 minutes, 2 minutes, etc]
  • Are there advantages of knowing how the DAQ is stored (in files for small time frames) that can improve efficiency?
  • How should the retrieved data be stored to potentially be used again without duplicating large DAQ accesses?
  • Provide statistics for …
    • All bunches for all pulses
    • Specific bunches per pulse, e.g. first 10, middle 10, last 10
    • Provide a switch to discard pulses terminated early
• Use Case 1-2: Analyze BLM trips over a given (and rather long) time range [e.g. 15 hours of 800 bunches / 3mA]
  • # of trips per BLM over time
  • Integrated alarm or instantaneous?
  • Where in the pulse train did the trip occur?
  • Plot histogram of # of beam trips vs beam energy for each BLM
• Use Case 1-3: Same as Use Case 1 but fetch average orbit (vector of beam positions) over time

• Use Case 1-4: Same as Use Case 2 but correlate BLM trips vs. orbit
  • How best to correlate a vector with trips?

• Use Case 1-5: Sources of Energy Distribution – Use results of Use Case 1 and search for sources of large energy variations

• Use Case 2-1: Identify pulses (or time periods) that match a set of conditions:
  • Pulses that were terminated by a BLM trip && Charge > 2.8nC && Orbit was outside a given window
• Use Case 2-2: Extract additional data (e.g. LLRF envelopes) from the pulses identified above

* Use Case 3-1: “When was the last time we had stable operation with > N bunches and for longer than M minutes? Pull out the DAQ data, but remove any pulses where the bunch train was terminated early

* Use Case 3-2: “The rf phase is jumping occasionally. When it jumps, what other machine parameters are jumping? When else in the past 12hrs did the phase jump, and did the same things jump then too?”

* Use Case 3-3: The machine isn't working now, but I think I have the same machine setup as yesterday when it was working just fine. What's different and when did it change? Have things drifted or did something jump? If it jumped, did someone change something?” [JC comment: we assume the 'good' machine configuration we want to compare with would come from some average over N pulses and that any short pulses would be removed before averaging to avoid skewing the data.]

* Use Case 3-4: “The machine is working but unstably. What's jittering and what are the most likely causes?” JC comment: we assume two broad parts to this…

1. DAQ tools would identify what's jittering and analyze the correlations
2. Send this information to some online physics model or analysis module that will then try to invert the effects to point back at possible causes. To avoid over-complicating the use-case, assume the analysis module or physics model is done by someone else.