ATLAS

Expert System Updates

Ignacio Asensi
University of Valencia

9/26/19 Ignacio Asensi 1

 Introduction
Representation of the
ATLAS detector

l ATLAS is a general-purpose particle physics experiment at the LHC

l Its major components are
n Magnet, Muon, Inner detector, Calorimeters
n Many others like computing, Control and Safety systems
l Confident knowledge on many systems of the detector is critical for
Control room maintenance, upgrade operations control and monitoring
9/26/19 Ignacio Asensi 2
 What is ATLAS?

A physicists thinks this is ATLAS A engineer thinks this is ATLAS

A Safety person thinks this is ATLAS General public

9/26/19 Ignacio Asensi 3
 What is ATLAS for us?
ATLAS is…

§ Database
§ User Interface
§ Questions
§ Answers
§ Decisions

Tasks!
https://its.cern.ch/jira/projects/ATLASTCES/issues Database

Number of issues done What happens if EXD1 trips??

and pending vs time What happens if the TRT cooling loop 3 stops?
What happens if fire detection system looses power?
What happens if….

Questions

9/26/19 Ignacio Asensi 4

 Simulating ATLAS behavior
Using graphical interface
§ Individual systems can be found
by locations, types or groups info
switch
§ Systems can be switch off and state
alarms be triggered
§ Systems are represented as
boxes with up to 3 icons
(switch, state, info)
§ When there is an interaction,
the inference engine
determines the consequences
and displays the new scenario

9/26/19 Ignacio Asensi 5

 Descriptions
Electricity Cryogenics Gas Magnet Racks Water Safety
Electricity Electricity Gas Summary Cryogenics SH1 Racks SR1 Water Distribution Access System
Electricity 18kV Cryogenics Distribution Gas Racks Cryogenics USA15 Racks SDX1 Water SF1 Elevators
Diesel Electricity Cryogenics SH1 Gas CSC Cryogenics UX15 Racks USA15 Water SPS Smoke Centrals
UPS SDX1 Cryogenics USA15 Gas MDT Magnet Cryogenics Racks US15 Water SUX1 Flammable Gas Centrals
UPS SX1 Cryogenics Racks USA15 Gas RPC Magnet Vacuum Racks UX15 Water SH1 Sniffer Racks
UPS USA15 LAR Heater Racks Gas TGC Magnet Electricity Access System Water SU1 Firemen Boxes
UPS US15 Cryogenics UX15 Gas TRT Smoke Centrals Water USA15 DSS Racks
Racks SR1 Cryogenics ANRS Gas TFC Flammable Gas Centrals Water US15 DSU
Racks SDX1 Cryogenics Argon Sniffer Racks Backup Chiller Minimax
Racks USA15 Cryogenics Argon PLC13 Firemen Boxes Cooling
Racks US15 Cryogenics Argon PLC14 Lighting Water Distribution
Racks UX15 Cryogenics Argon PLC15 DSS Racks CV Room
Lighting Cryogenics Argon PLC16 Gas Racks Cooling US15 DSS
Magnet Electricity Magnet Cryogenics CV Room Cooling UX15 DSS Racks
Cooling US15 Cooling TRT DSU
Muon Cooling
Muon Cooling Loops Loops DSS Alarms
Cryogenics Racks USA15 Detector Cooling Inhibit Request
Evaporative
LAR Heater Racks Cooling
Minimax Thermosyphon
IBL Cooling
9/26/19 Ignacio Asensi 6
 Sub detector descriptions - CSC

9/26/19 Ignacio Asensi 7

 Sub detector descriptions - TGC

9/26/19 Ignacio Asensi 8

 Simulation of power cut

• Understanding the consequences

of a power cut in a main line
• Minimize impact in subsystems

9/26/19 Ignacio Asensi 9

 Impact on IBL

9/26/19 Ignacio Asensi 10

 Impact on ATLAS

9/26/19 Ignacio Asensi 11

 Expert System portability
Database Server Web application

Objects Logic Communication

with server

Relationships Database class Diagrams

helpers

Model Controller View

• Maintenance of • Respond to user • User interaction

simulation state input • Scenario rendering
• Store and retrieve • Interpret user
data request with
data from model
General
ATLAS specific

9/26/19 Ignacio Asensi 12

 Summary and status

§ The ATLAS Expert System by ATLAS Technical

coordination is a diagnostic tool for the maintenance of
the experiment.
§ Description of critical systems like electricity, gas,
detectors and others is reaching the desired
granularity.
§ The Expert System has been proven useful evaluating
the impact of interventions.
§ Simulations have been compared with actual
intervention outcomes during LS2.
§ The portability to other detectors is possible.
§ Future plans include the ability to search the causes of
current state in the simulations

9/26/19 Ignacio Asensi 13

 Back up

9/26/19 Ignacio Asensi 14

 ATLAS systems in Expert System

9/26/19 Ignacio Asensi 15

 Performance
200 ChangeSwitch (97 ± 30 ms)

GetCurrentState (122 ± 35 ms)

180
None (77 ± 30 ms)
160
Ping (79 ± 30 ms)
140 GetTriggeredAlarms (84 ± 33 ms)

120
100
80
60
40
20
0
50 100 150 200 250 300
Time [ms]
Server:
Shows good performance
<latexit sha1_base64="43SypcAUNIUa3N0xVwMlJR8GkhE=">AAAB9XicbVDLTgIxFL2DL8QX6tJNIzHBDen4iC5cEN24xEQeCYykUzrQ0M5M2o6GTPgPNy40xq3/4s6/scAsFDzJTU7PuTe99/ix4Npg/O3klpZXVtfy64WNza3tneLuXkNHiaKsTiMRqZZPNBM8ZHXDjWCtWDEifcGa/vBm4jcfmdI8Cu/NKGaeJP2QB5wSY6WHuIwr+Pz46hRjqVG3WLLPKdAicTNSggy1bvGr04toIlloqCBat10cGy8lynAq2LjQSTSLCR2SPmtbGhLJtJdOtx6jI6v0UBApW6FBU/X3REqk1iPp205JzEDPexPxP6+dmODSS3kYJ4aFdPZRkAhkIjSJAPW4YtSIkSWEKm53RXRAFKHGBlWwIbjzJy+SxknFxRX37qxUvc7iyMMBHEIZXLiAKtxCDepAQcEzvMKb8+S8OO/Ox6w152Qz+/AHzucPL6WQWQ==</latexit>
p(0.05) < 300ms

Most Probable Value (MPV) for “get current state”,

function that loads the simulation is ~100 ms

9/26/19 Ignacio Asensi 16