BOOSTR: A Dataset for Accelerator Control Systems
BBOOSTR: A Dataset for Accelerator Control Systems
Diana Kafkes and Jason St. John Fermi National Accelerator Laboratory, Batavia, IL 60510, USA * [email protected] ABSTRACT
BOOSTR (Booster Operation Optimization Sequential Time-Series for Reinforcement) was created to provide cycle-by-cycletime series of readings and settings from instruments and controllable devices of the Booster, Fermilab’s Rapid-CyclingSynchrotron (RCS) operating at 15 Hz.
BOOSTR provides a time series from 55 device readings and settings which pertainmost directly to the high-precision regulation of the Booster’s Gradient Magnet Power Supply (GMPS). To our knowledge, thisis the first known dataset of accelerator device parameters made publicly available. We are releasing it in the hopes that it canbe used to demonstrate aspects of artificial intelligence for advanced control systems, such as reinforcement learning andautonomous anomaly detection.
Background & Summary
Tuning and controlling particle accelerators is challenging and time consuming. Even marginal improvements to acceleratoroperation can translate very efficiently into improved scientific yield for an experimental particle physics program. The datareleased here was collected in the hopes of achieving improvement in precision for the Fermilab Booster Gradient MagnetPower Supply (GMPS) regulatory system, which is detailed below.The Fermilab Booster receives a 400 MeV proton beam from the Linear Accelerator and accelerates it to 8 GeV throughsynchronously raising accelerator cavity radiofrequency and instigating a controlled magnetic field to steer the beam withcombined-function bending and focusing electromagnets, known as gradient magnets. These magnets are powered by theGMPS, which operates on a 15 Hz cycle between a minimum current (at injection) and a maximum current (at beam extraction).The GMPS is realized as four power supplies, evenly distributed around the Booster, and each powers one fourth of the gradientmagnets. The role of the GMPS regulator is to calculate and apply small compensating offsets in the GMPS driving signalin order to improve the agreement of the resulting observed minimum and maximum currents with their set points. Withoutregulation, the fitted minimum of the magnetic field may vary from the set point by as much as a few percent.At beam injection, a perturbation of only a percent is enough to significantly decrease beam transfer efficiency and therebyreduce the beam flux ultimately available to the high-energy particle physics experiments run at the lab. Disturbances to themagnet current can occur due to ambient temperature changes, other nearby high-power pulsed radio-frequency systems, andelectrical ground movement induced by power supplies of other particle accelerators at the complex. The current GMPSregulation involves a PID (Proportional-Integral-Derivative) control scheme (see Figure 1 for schematic). The regulatorcalculates estimates for the minimum and maximum currents of the offset-sinusoidal magnetic field from the previous 15 Hzcycle. These values are then used to adjust the power supply program and decrease systemic error in the next cycle’s current,such that it more closely matches the set point. Presently, the PID system achieves regulation errors corresponding to roughly0.1% of the set value.Although some 200,000 entries populate the device database of Fermilab’s accelerator control system , the 55 device valuetime series presented here in BOOSTR were collected in accordance with suggestion by Fermilab accelerator subject matterexperts (SMEs). These values exhibit correlations with GMPS power supply perturbations. The full data were collected duringtwo separate periods: from June 3, 2019 to July 11, 2019 — when the accelerator was shut down for regular maintenance —and from December 3, 2019 to April 13, 2020 — when the accelerator was shut down in response to the Covid-19 Pandemic.Data from a single day of BOOSTR was previously described in a Datasheet . A proof-of-concept paper (submittedto Physical Review Accelerators and Beams ) used this subset of
BOOSTR and demonstrated the viability of training areinforcement learning agent to control GMPS regulation better than the existing PID system. Relative to the originalDatasheet , this manuscript is expanded with more SME input, describes more than 100 times more data, and includesdocumentation of validation not presented in the original Datasheet. a r X i v : . [ phy s i c s . acc - ph ] J a n ethods Collection Process
A data collection node was created and set to request data at 15 Hz from the Data Pool Manager of the Accelerator ControlNetwork (ACNET) . The created scheme involved front-end nodes, each managing their respective devices, replying withtimestamped values at the stated rate barring differences of clock speed, input-output (I/O) lag time variations due to networktraffic fluctuations, and higher-priority interruptions from competing processes on the front-end node. These inconsistencieswere later addressed through a time-alignment process discussed in the Data Processing Section. The collection node stored thedata in a circular buffer approximately 10 days deep.A Python script managed by a nightly cron job polled the data collection node for the most recent midnight-to-midnight 24hours of timestamped data for each of the 55 time series identified by SMEs. A second cron-managed script did the same forrelevant accelerator control events issued in the same period. These event data correspond to important cycles achieved throughcontrolling the devices at the accelerator. Event data were requested by a separate data collection node.Each day’s data harvest was originally stored in HDF5 (Hierarchical Data Format Version 5) files. Any data instancesmissing from the parquet files released here were not included in the original data buffers from which this dataset was drawn. Data Processing
Each instance was created through a concatenation of each device’s timestamp data table within every HDF5 file and thensaved in parquet format. A similar procedure was undertaken for one of the accelerator control event signals polled,
Event0C ,as its broadcast is synchronized with the GMPS magnetic field minimum.
Event0C was collected to correct a problem inthe observed sampling frequency: there was an issue of the sampling of each device being nominally at 15 Hz, but in realitysynchrony was demonstrably imperfect, and the time intervals between successive timestamps display varying lags.Since
Event0c serves as the baseline or heartbeat of the Booster at approximately 15 Hz and is synchronized with thesmoothly varying electrical current GMPS regulates, we used
Event0c to time-align our data. The alignment approximatesthe data available to the GMPS regulator operating in real time. We used the GMPS-synchronized
Event0C ’s timestamp asthe moment to begin forward inference, taking the value for each device time series which had the most recent correspondingtimestamp. In practice, this required timestamp-sorted series for each device and finding the most recent device value, relative to
Event0c timestamp, in a lookback window equal to the maximum interval between device timestamps (necessarily excludingthe five month gap between our two data collection periods). This time-alignment step was run over the whole dataset inmultiple parallel processes using Apache Spark.Notably, the data recorded for
Event0c was missing the period from July 1 to 11, 2019. Therefore aligning on thisvariable discarded some of the data collected during our first period of collection.
Data Records
The data release is stored on Zenodo . Each instance is a zip compressed parquet of one of the 55 aligned time series withcolumns corresponding to the aligned time stamp, original time stamp, difference between time stamps, and the reading/settingvalue . The original timestamp and time difference is included to demonstrate the mechanics of our alignment process andenable a check for reproducibility. All timestamps are in Greenwich Mean Time (UTC).Our data release contains device data from each of the four gradient magnet power supplies, the GMPS PID regulator, andthe Main Injector, where the beam is directed after acceleration via the Booster. Minimum and maximum current informationreadings and settings, the feedback and transductor gain settings, and the feed-forward start trigger are collected as part ofthe current PID regulation scheme. The “inhibit” value controls whether the GMPS regulator accepts settings changes forparameters other than the minimum and maximum current, such as the gain settings (any positive value will prevent changes).Additionally, ˙ (cid:126) B , the rate of change of the magnetic field, is recorded as a proxy for the magnetic field we are interested inregulating. Timing information derived from ˙ (cid:126) B = 0 synchronizes the current PID regulator system.We acknowledge that the ACNET parameter names are by no means standardized across different particle accelerators andthat they will appear especially abstruse for those well-versed in control systems who are new to working with accelerators.In Table 1, we detail explanations of each of the parameters read from devices (devices whose first letter is followed by :)and indicate whether the device setting was included in the dataset (devices whose first letter is followed by _ and appear inSetting column), since describing these corresponding pairs would be redundant. In Figures 2 and 3, we visualize metadatatrends for each “nonconstant" parameter in each data collection period (see Table 2 for a list of values we considered to bevirtually unchanging within the two periods) and also provide the mean and standard deviation of device readings across the twocollection periods in Table 1. Furthermore, Table 1 includes dates missing in the data recorded for each reading. As a reminderthe data were collected during two separate periods: from June 3, 2019 to June 30, 2019 (July is missing due to time-alignment ith Event0C ) and from December 3, 2019 to April 13, 2020. Finally, in Figure 4 we demonstrate the centrality of eachrecorded parameter with a heatmap of histogram values.Additionally, we provide the PID regulator status values
B|GMPSSC (ACNET status parameters include |), whose 16 bitscontain various motherboard states. Here we are concerned with bit 3, which indicates whether or not the GMPS regulator wason (1), and bit 7, which indicates whether the Booster is in its normal alternating current (AC) mode (1) or “coasting beam”direct current (DC) mode at constant beam energy (0). Unlike the rest of the devices, this status value is presently recordedat only 1 Hz because it was not included in our initial data node request and was relegated to an archived data node at lesserfrequency. While the same time-aligning described above was applied to align
B|GMPSSC , due to the slow sampling rate, wecaution the user to refer closely to the original timestamp such that they might make decisions about whether to use data whenGMPS was off and to inform them of potential problems when interpolating in a region immediately before or after a statuschange. See Figure 5 for more details on
B|GMPSSC values.
Technical Validation
In order to verify the quality of this dataset, we pored over the electronic logbook (Elog) that Fermilab Booster techniciansand operators use to record changes to device settings as well as observations while in operation. We used these Elog entriesto authenticate our data’s viability across timescales. First, we used the Elog to corroborate expert acknowledgement of themajor spikes observed in Figures 2 and 3. These outlier changes, typically seen in the value’s mean and standard deviation,represented major changes made on that specific day, including when the Booster was switched from alternating current (AC)to direct current (DC) mode (see Figure 5) as well as when the GMPS regulator was turned off altogether. These reconciliationsare presented in Table 3.Furthermore, we pinpoint changes in the AC vs. DC settings according to the Elog for June 24, 2019 and March 11, 2020in Figure 6. Here applying a bitmask reveals that a B|GMPSSC value of 159 indicates AC mode/GMPS on, while 31 indicatesDC mode/GMPS on, and 407 indicates AC mode/GMPS off. In this figure, the plotted timestamps were offset to Central Time(UTC-5) in order to align with times given in the Elog, which were not recorded in UTC. On June 24, the trace of
B|GMPSSC clearly shows GMPS regulation briefly switching off before commencing DC studies from 8:00 AM - 6:00 PM with a value of31, then being turned back to 159. On March 11,
B|GMPSSC is at 159 before 6:00 AM, off at 407 from 6:00 - 9:50 AM, andthen is set to AC mode from 9:50 AM - 12:45 PM, to DC mode from 12:45 - 3:49 PM, and back to AC mode for the rest ofthe day, as per the Elog . The close correspondence of these changes in our data to the recorded actions and observations ofBooster personnel boost our confidence in the quality and relevance of the collected dataset.Additionally, we plot settings changes on March 10 and 11 documented in the Elog in Figure 7. The blip in B:ACMNPG from 6.5 to 13.5 is visible as is the slight decrease in
B_VIMIN around 4:00 PM CST, which were mentioned in Table 3.
Usage Notes
BOOSTR could be used to train various control networks for accelerator regulation, to construct “digital twins” of the FermilabBooster regulator’s control environment, or to develop anomaly detection/categorization capabilities. Please note: there are nolegal or ethical ramifications of using this data as it was collected from a machine, and not collected from or representative ofpeople.In the future, the dataset could feasibly be expanded to include more of the 200,000 available ACNET system parametersand therefore be used to control, mimic, or monitor further aspects of the particle accelerator complex. One could argue thatthis initial dataset might become the foundation upon which substantial fine-tuning of particle accelerators could depend.
Code availability
The preprocessing code is available here. When using
BOOSTR data, the authors recommend ordering by time immediately, asthe parquet files do not store the data entries sequentially . Acknowledgements
This dataset was created as part of the “Accelerator Control with Artificial Intelligence” Project conducted under the auspicesof the Fermilab Laboratory Directed Research and Development Program (Project ID
FNAL-LDRD-2019-027 ). The manuscripthas been authored by Fermi Research Alliance, LLC under Contract No. DE-AC02-07CH11359 with the U.S. Departmentof Energy, Office of Science, Office of High Energy Physics and is registered at Fermilab as Technical Report Number
FERMILAB-PUB-21-005-SCD .We are extremely grateful for Brian Schupbach and Kent Triplett for lending their Booster technical expertise, withoutwhich we could not have validated our dataset. Additionally, we would like to acknowledge Burt Holzman for guidance on etting set up in the cloud, and the help of the Databricks federal support staff, in particular Sheila Stewart. Furthermore,we would like to recognize Jovan Mitrevski and Aleksandra Ciprijanovic for useful discussions and a careful reading of thismanuscript.
Author contributions statement
J.S. created the data collection script, and set up and maintained the cron jobs to record the data in HDF5 files. D.K. migratedthe data from on-premise storage to the cloud, wrote the preprocessing and time-alignment code, and validated the data. Bothauthors reviewed this manuscript.
Competing interests
The authors declare no competing interests.
References Cahill, K. & et. al. The fermilab accelerator control system.
ICFA Beam Dyn. Newslett. , 106–124 (2008). Kafkes, D. & St. John, J. BOOSTR: A Dataset for Accelerator Control Systems (Full Release), 10.5281/zenodo.4382663(2021). Kafkes, D. & St. John, J. BOOSTR: A Dataset for Accelerator Control Systems (Partial Release), 10.5281/zenodo.4088982(2020). John, J. S. et al.
Real-time artificial intelligence for accelerator control: A study at the fermilab booster (2020). 2011.07371. Figures & Tables
Reference system : B coil, transductor, d B /d t coil, zero-crossing GMPS control rack
Programmable logic target settings
Power supplies 1-4 measurements (& errors)control signalsseriesconnect
Accelerator Control Network sampled I min , I max Figure 1.
Overview of current GMPS control system . Presently, a human operator specifies a target program for B:VIMIN and
B:VIMAX , the GMPS compensated minimum and maximum currents respectively, via the Fermilab Accelerator ControlNetwork that is transmitted to the GMPS control board. able 1.
Description of BOOSTR dataset parameters. Here “GMPS" denotes the Gradient Magnet Power Supplies (1-4), “MI"means main injector, “MDAT" refers to Fermilab’s Machine Data communications protocol. Device parameters that begin with B relate to the accelerator Booster, whereas device parameters that begin with I relate to the Main Injector. Parameter meanand standard deviation have been truncated to two decimal points. Parameter Details (Units) Setting Mean (Std) Missing Dates
B:ACMNIG
Min AC integral feedback gain
B_ACMNIG
B:ACMNPG
Min AC proportional feedback gain
B_ACMNPG
B:ACMXIG
Max AC integral feedback gain
B_ACMXIG
B:ACMXPG
Max AC proportional feedback gain
B_ACMXPG
B:DCIG
DC integral feedback gain
B_DCIG
B:DCPG
DC proportional feedback gain
B_DCPG
B:GMPS1V
GMPS1 output voltage (V) 81.82 (89.56) 2019: 6/14, 12/12, 12/30-31and 2020: 1/01-4/12
B:GMPS2V
GMPS2 output voltage (V) 85.29 (96.00) 2019: 6/14, 12/12, 12/30-31and 2020: 1/01-4/12
B:GMPS3V
GMPS3 output voltage (V) 63.67 (61.19) 2019: 6/14, 12/12, 12/30-31and 2020: 1/01-4/12
B:GMPS4V
GMPS4 output voltage (V) 61.08 (65.19) 2019: 6/14, 12/12
B:GMPSBT ∂ B / ∂ t offset trigger (s) B_GMPSBT
B|GMPSSC
Binary status control of GMPS N/A (N/A) 2019: 6/07, 6/12, 6/14-15and 2020: 1/18, 3/08, 3/15
B:GMPSFF
Feedforward start trigger (s)
B_GMPSFF
B:IMAXXG
Max transductor gain (A/V)
B_IMAXXG -117.12 (1.80) 2019: 6/14, 12/12
B:IMAXXO
Max transductor offset (A)
B_IMAXXO
B:IMINXG
Min transductor gain (A/V)
B_IMINXG -11.73 (0.23) 2019: 6/14, 12/12
B:IMINXO
Min transductor offset (A)
B_IMINXO
B:IMINER
Discrepancy from setting at min (0.1 A) 1.93 (3.86) 2019: 6/14, 12/12
B:IMINST ∂ B / ∂ t sample off B_IMINST
B:IPHSTC
Phase regulator time constant
B_IPHSTC
B:LINFRQ
60 Hz line frequency offset (mHz) -0.44 (16.31) 2019: 6/03-7/11, 12/1212/30-31 and 2020: 1/01-06
B:NGMPS
Number of GMPS suppliers 4.00 (0) 2019: 6/14, 12/12
B:PS1VGM
GMPS1 V- to ground (V) -2.30 (23.56) 2019: 6/14, 12/12, 12/30-31and 2020: 1/01-4/12
B:PS2VGM
GMPS2 V- to ground (V) -21.29 (27.52) 2019: 6/14, 12/12, 12/30-31and 2020: 1/01-4/12
B:PS3VGM
GMPS3 V- to ground (V) -15.13 (14.11) 2019: 6/14, 12/12, 12/30-31and 2020: 1/01-4/12
B:PS4VGM
GMPS4 V- to ground (V) -26.27 (17.22) 2019: 6/14, 12/12, 12/30-31and 2020: 1/01-4/12
B:PS1VGP
GMPS1 V+ to ground (V) 52.00 (34.82) 2019: 6/14, 12/12, 12/30-31and 2020: 1/01-4/12
B:PS2VGP
GMPS2 V+ to ground (V) 26.53 (30.53) 2019: 6/14, 12/12, 12/30-31and 2020: 1/01-4/12
B:PS3VGP
GMPS3 V+ to ground (V) 20.17 (13.74) 2019: 6/14, 12/12, 12/30-31and 2020: 1/01-4/12
B:PS4VGP
GMPS4 V+ to ground (V) 9.14 (12.75) 2019: 6/14, 12/12, 12/30-31and 2020: 1/01-4/12
B:VIMAX
Compensated max GMPS current (A)
B_VIMAX
B:VIMIN
Compensated min GMPS current (A)
B_VIMIN
B:VINHBT
Inhibit value
B_VINHBT
B:VIPHAS
GMPS ramp phase wrt line voltage (rad)
B_VIPHAS
I:IB
MI lower bend current (A) 2275.71 (2571.03) 2019: 6/03-19, 12/12
I:MDAT40
MDAT measured MI current (A) 2400.57 (2576.37) 2019: 6/03-7/11, 12/12
I:MXIB
Main dipole bend current (A) 2279.68 (2564.53) 2019: 6/03-11, 6/13-19, 12/12 able 2.
Summary of nearly constant variables in both periods of data collection. Here “nearly constant" denotes variableshaving a standard deviation less than 10 − across both periods.Device Setting Constant Value B:DCIG B_DCIG B:GMPSBT B_GMPSBT
B:IMINST B_IMINST B:IMINXO B:IMINXO Table 3.
Summary of Booster-related electronic log (Elog) entries corresponding to spikes in Figures 2 and 3. OriginalCentral Time times are given with values in parenthesis designating UTC. Here “RF" denotes radiofrequency.Date Elog Entry Summary6/8/2019 GMPS in AC mode until 8:00 AM (13:00) on 6/08, then switched off until 6/176/10/2019 GMPS was locked/tagged out for outage, West Booster gallery RF off from8:30-10:00 (13:30-3:00) for work6/17/2019 GMPS turned back on and put in AC mode6/22/2019 High energy physics beam turned off at 8:00 PM (1:00 +1) (GMPS remained in AC mode)6/24/2019 DC studies from 8:00 AM - 6:00 PM (13:00 - 23:00), back to AC mode6/26/2019 Alternated between AC and DC mode, GMPS off for 30 min around 5:30 PM (22:30)6/27/2019 GMPS in AC mode all day, but removing certain study events caused bias to creep up,eventually tripping the RF6/28/2019 Alternated between AC and DC mode12/8/2019 B_VIMIN adjusted, GMPS in AC mode12/12/2019 AC mode, operators reset virtual machine environment locally12/28 - 12/30/2019 No beam from injector, GMPS in AC mode12/31/2019 Booster injection back, GMPS in AC mode1/1/2020 GMPS off for 15 min around 9:30 AM (14:30), in AC mode for rest of day2/4/2020 RF sparking in gallery due to reverting of RF capture settings, GMPS in AC mode2/5/2020 GMPS off from 6:00 AM - 3:30 PM (11:00 - 20:30), then in AC mode for rest of day2/6/2020 Lowered beam intensity to users, but GMPS was in AC mode all day3/5/2020 Beam tails were large, so turned
B_VIMIN down3/10/2020 GMPS in AC mode all day,
B:ACMNPG changed from 6.5 to 13.5,
B_VIMIN decreasedfrom 103.440 to 103.4203/11/2020 GMPS in AC mode from 12:00 AM - 6:00 AM (5:00 - 11:00), off from 6:00 - 10:00 AM(11:00 - 15:00), then alternated between AC and DC mode,
B_VIMIN adjusted from103.418 to 103.3863/13/2020 GMPS off from 9:30 AM - 11:00 AM (14:30-16:00), back on and put in AC mode3/20/2020 Booster turned off on account of Covid-19 pandemic at 12:00 PM (17:00)
Values D a t e s C o ll e c t e d Variable Metadata
B:IMAXXGB_IMAXXG
Values D a t e s C o ll e c t e d Variable Metadata
B:IMINXGB_IMINXG
Values D a t e s C o ll e c t e d Variable Metadata
B:ACMNIGB:ACMXIGB:VINHBTB:VIPHASB_ACMNIGB_ACMXIGB_VINHBTB_VIPHAS
Values D a t e s C o ll e c t e d Variable Metadata
B:ACMNPGB:ACMXPGB:GMPSFFB:IMAXXOB:IMINERB:NGMPSB_ACMNPGB_ACMXPGB_GMPSFFB_IMAXXO
100 50 0
Values D a t e s C o ll e c t e d Variable Metadata
B:DCPGB:IPHSTCB:LINFRQB:PS2VGMB:PS3VGMB:PS4VGMB_DCPGB_IPHSTC 100 50 0 50 100 150
Values D a t e s C o ll e c t e d Variable Metadata
B:PS1VGMB:PS1VGPB:PS2VGPB:PS3VGPB:PS4VGPB:VIMINB_VIMIN0 100 200 300
Values D a t e s C o ll e c t e d Variable Metadata
B:GMPS1VB:GMPS2VB:GMPS3VB:GMPS4V
Values D a t e s C o ll e c t e d Variable Metadata
B:VIMAXB_VIMAX
Values D a t e s C o ll e c t e d Variable Metadata
I:IBI:MDAT40I:MXIB
Figure 2.
Metadata variable trends for Period 1: June 3, 2019 to June 30, 2019. The graphs show the mean for each variableon the given date and shades in the standard deviation of that variable on that date.
20 118 116 114
Values D a t e s C o ll e c t e d Variable Metadata
B:IMAXXGB_IMAXXG
Values D a t e s C o ll e c t e d Variable Metadata
B:IMINXGB_IMINXG
Values D a t e s C o ll e c t e d Variable Metadata
B:ACMNIGB:ACMXIGB:VINHBTB:VIPHASB_ACMNIGB_ACMXIGB_VINHBTB_VIPHAS
Values D a t e s C o ll e c t e d Variable Metadata
B:ACMNPGB:ACMXPGB:GMPSFFB:IMAXXOB:IMINERB:NGMPSB_ACMNPGB_ACMXPGB_GMPSFFB_IMAXXO
100 50 0
Values D a t e s C o ll e c t e d Variable Metadata
B:DCPGB:IPHSTCB:LINFRQB:PS2VGMB:PS3VGMB:PS4VGMB_DCPGB_IPHSTC 100 0 100
Values D a t e s C o ll e c t e d Variable Metadata
B:PS1VGMB:PS1VGPB:PS2VGPB:PS3VGPB:PS4VGPB:VIMINB_VIMIN
100 200 300
Values D a t e s C o ll e c t e d Variable Metadata
B:GMPS1VB:GMPS2VB:GMPS3VB:GMPS4V
Values D a t e s C o ll e c t e d Variable Metadata
B:VIMAXB_VIMAX
Values D a t e s C o ll e c t e d Variable Metadata
I:IBI:MDAT40I:MXIB
Figure 3.
Metadata variable trends for Period 2: December 2, 2019 to April 13, 2020. The graphs show the mean for eachvariable on the given date and shades in the standard deviation of that variable on that date.
Bin
B:ACMNIGB:ACMNPGB:ACMXIGB:ACMXPGB:DCIGB:DCPGB:GMPS1VB:GMPS2VB:GMPS3VB:GMPS4VB:GMPSBTB:GMPSFFB:IMAXXGB:IMAXXOB:IMINERB:IMINSTB:IMINXGB:IMINXOB:IPHSTCB:LINFRQB:NGMPSB:PS1VGMB:PS1VGPB:PS2VGMB:PS2VGPB:PS3VGMB:PS3VGPB:PS4VGMB:PS4VGPB:VIMAXB:VIMINB:VINHBTB:VIPHASB_ACMNIGB_ACMNPGB_ACMXIGB_ACMXPGB_DCIGB_DCPGB_GMPSBTB_GMPSFFB_IMAXXGB_IMAXXOB_IMINSTB_IMINXGB_IMINXOB_IPHSTCB_VIMAXB_VIMINB_VINHBTB_VIPHASI:IBI:MDAT40I:MXIB D e v i c e Counts Figure 4.
Heatmap of histogram distributions for each reading and setting variable with equal sampling. This is only meant tocharacterize the centrality of each recorded value. See Fig. 2 and 3 for actual metadata value ranges.
Dates Collected V a l u e s Period 1: AC vs. DC Setting
B|GMPSSCAC, GMPS ONDC, GMPS ONAC, GMPS OFF
Dates Collected V a l u e s Period 2: AC vs. DC Setting
B|GMPSSCAC, GMPS ONDC, GMPS ONAC, GMPS OFFDC, GMPS OFF
Figure 5.
Daily values of
B|GMPSSC (should be interpreted as taking the mode for each day) whose bits encode relevantBooster statuses.
Time V a l u e s June 24, 2019
B|GMPSSC 0:00 4:59 9:58 14:58 19:58
Time
March 11, 2020
B|GMPSSC
Figure 6.
Values of status
B|GMPSSC corresponding to Table 3 entries for June 24, 2019 and March 11, 2020 (timestampswere put in Central Time to align with Elog). Recall: a value of 159 indicates AC study/GMPS on, 31 indicates DCstudy/GMPS on, and 407 indicates AC study/GMPS off. These traces display this value at a much greater granularity thanFigure 5. time V a l u e s B:ACMNPG on 3/10
B:ACMNPG time V a l u e s B_VIMIN on 3/11
B_VIMIN
Figure 7.
Switches corresponding to Table 3 entries for March 10, 2020 and March 11, 2020:
B:ACMNPG changed from 6.5to 13.5 and
B_VIMIN decreased from 103.418 to 103.386 (timestamps were put in Central Time to align with Elog). Thesudden large increase in
B_VIMIN from 12:45 - 3:49 PM CST to a value off the plotted region corresponds to the DC modeobserved in Figure 6.from 12:45 - 3:49 PM CST to a value off the plotted region corresponds to the DC modeobserved in Figure 6.