Transcript LHC Controls Project - lhc-div-mms.web.cern.ch - /lhc-div-mms

LHC production data expected for the
LHC Control System
Ronny BILLEN
SL/MR/DBS
for the LHC Controls Project
Data for LHC-CP - slide 1
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
Outline
1. LHC Controls Project
2. Data management for LEP
3. Databases for LEP Controls
Layout, Controls, Measurement, Logging, Alarms
4. How will LHC be different than LEP
5. Conclusions and Recommendations
Part 2: LHC as an INB installation
Data for LHC-CP - slide 2
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
LHC Controls Project

Mandate, scope & objectives:
 Design & implementation of future LHC control system
 Overall architecture of HW & SW
 Manage production of application SW

Intermediate objective: control system for Sector Test
 it should not be different than the final system

Project Leader: Robin Lauckner
+ core team, link-people in equipment groups and projects


LHC-CP reports to SLTC and member of Controls Board
web site:
 http://lhc-cp.web.cern.ch/lhc-cp/
 EDMS site for documents (LHC Machine  LHC Controls)
Data for LHC-CP - slide 3
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
Let’s look back: Data Management for LEP
 Identified as a major issue before LEP construction
 In 1982 Oracle RDBMS was purchased for this reason
 Important data management projects:
 project planning
 +40000 activities; worst-case scenarios
 cables installation
 thousands of kilometres; selecting best access path
 equipment specification
 different constructors, teams of experts, project phases
 Machine experts specified and built their own data
management applications.
 Standardised DBMS allows sharing of data
Data for LHC-CP - slide 4
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
Databases for LEP Controls
off-line
analysis
layout
data
simulations
LAYOUT
operator
control
logging
data
LOGGING
layout
data
CONTROL
MEASURE
static
alarms
data
read
measurements
set
LEP
Data for LHC-CP - slide 5
dynamic
alarms
data
alarms
screen
ALARMS
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
Layout database: the machine blue-print




Off-line Database
Relatively static data, mostly numeric
dimensions & relative positions of all components
Information must be in place before real installation
 allow computer simulations




yearly update after technical stop
Versioning for comparison between years
Main users: accelerator physicists
Data subset copied to Controls Database
Data for LHC-CP - slide 6
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
Databases for LEP Controls
off-line
analysis
layout
data
simulations
LAYOUT
operator
control
logging
data
LOGGING
layout
data
CONTROL
MEASURE
static
alarms
data
read
measurements
set
LEP
Data for LHC-CP - slide 7
dynamic
alarms
data
alarms
screen
ALARMS
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
The LEP Controls

Controlling thousands of parameters






open/close vacuum valves
switch on/off cooling circuits
set magnet currents
tweak high-voltage equipment
…
All controls information stored in Controls Database
 names, actions, actual settings, references

History of actions and settings over time stored as well
 allowed heuristic approach of fine-tuning LEP

Software 100% data driven
 mainly Pro*C with graphical interfaces running on UNIX
Data for LHC-CP - slide 8
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
LEP Databases for Controls
off-line
analysis
layout
data
simulations
LAYOUT
operator
control
logging
data
LOGGING
layout
data
CONTROL
MEASURE
static
alarms
data
read
measurements
set
LEP
Data for LHC-CP - slide 9
dynamic
alarms
data
alarms
screen
ALARMS
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
Measurement Database


Equipment control requires adequate feedback
There is no uniform measurement feedback
 different measurement rates
 different data structures
 different communication methods


Same information requested by many users or systems
simultaneously causing real bottlenecks
How did we solve this bottleneck?
 Disable direct access to equipment for these measurements
 put the real-time measurement data in the Oracle database

The data is generally overwritten (on-line use)
 archiving is possible (using Oracle triggers)
Data for LHC-CP - slide 10
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
LEP Databases for Controls
off-line
analysis
layout
data
simulations
LAYOUT
operator
control
logging
data
LOGGING
layout
data
CONTROL
MEASURE
static
alarms
data
read
measurements
set
LEP
Data for LHC-CP - slide 11
dynamic
alarms
data
alarms
screen
ALARMS
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
Data Logging - Graphical visualisation
Data for LHC-CP - slide 12
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
LEP Databases for Controls
off-line
analysis
layout
data
simulations
LAYOUT
operator
control
logging
data
LOGGING
layout
data
CONTROL
MEASURE
static
alarms
data
read
measurements
set
LEP
Data for LHC-CP - slide 13
dynamic
alarms
data
alarms
screen
ALARMS
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
Alarm messages via database
Data for LHC-CP - slide 14
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
How different will LHC be?




Of course, LHC will be more complex and difficult with
newer technologies and ideas...
…but we do have the experience with LEP
…and we will surely build on this experience
From (my) data management point of view, LHC will not
be very different, except for:
 Configuration data: the structure will be in the database as well
(meta-data for the control system itself)
 Post Mortem data: combine and correlate huge amounts of data
from different systems, synchronised with millisecond precision
Data for LHC-CP - slide 15
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
What does controls need from production?

From Layout database
 naming for addressing equipment
 geometric, magnetic data on all installed elements
 optics data

From Survey Database
 alignment errors, corrections

From Magnetic Measurement database
 calibration curves, hysteresis, temperature effect, ...

From Powering database
 powering circuits, magnet correspondence, actions

From other Equipment databases (vacuum, cryo)
 naming, actions, settings, measurements
 logging and post-mortem analysis
Data for LHC-CP - slide 16
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
Conclusions and Recommendations







Similar to the control of LEP, data for LHC controls will
be managed with a set of centralised databases
In order to feed these LHC Controls databases, data
from various production databases will be needed
Oracle is the chosen RDBMS
We will push Oracle to the limits, real time capabilities
LHC could be controlled through the database as much
as possible: The database is the machine (i.e. my
vision)
LHC-CP will need to look into your data, so please keep
your data in a reasonable structure
If your data is already in Oracle, we can work it out...
Data for LHC-CP - slide 17
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
So much for control of the LHC...
… now for some real problems!
Traceability for LHC in the
INB framework
Data for LHC-CP - slide 18
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
LHC Quality Assurance Plan





Implement a Total Quality Management System based on
defect prevention & continuous process improvement
Quality Assurance Plan defines quality activities &
responsibilities, procedures, standards
Quality is essential to make LHC a great success
All LHC parts that will be manufactured and installed must
be identified according to the QAP rules, ensuring
traceability through its entire lifecycle
However, traceability in the INB framework is slightly
different...
Data for LHC-CP - slide 19
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
LEP Traceability









LEP is being dismantled completely
We have to follow the rules for dismantling an
“Installation Nucléaire de Base (INB)” for French
authorities
Track all equipment movements coming out of LEP
Know exact location at any time
Conventional and Radioactive material (zoning!)
Use only Authorised disposal channels
Temporary storage for reused equipment
Radioactive waste remains at CERN for the moment
Subject to INB inspections
Data for LHC-CP - slide 20
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
LEP Dismantling Traceability database




30000 tons of material to be evacuated
18-month period, carefully planned
Full-proof INB traceability
Inventory & barcode stickers
 Dedicated traceability Database
 UML modelling; Java, C++ & PL/SQL
implementation
 barcode readers with remote
database connectivity
 result-oriented contract for
tracking personnel
Data for LHC-CP - slide 21
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
SPS Traceability



SPS is an INB since July 2000 (surprise!)
Temporary system put in place for equipment that
comes out of SPS (a lot!)
Working Group is active on the future system
 no complete inventory
 there is too much
 component going in  component coming out
 components can be conditioned
 identification when it comes out
 implies less work
 naming can change
 labels can deteriorate
Data for LHC-CP - slide 22
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
LHC Traceability

For LEP traceability, we started 15 years too late

For SPS traceability, we started 25 years too late

… For LHC traceability, we could start on time

Again, we should learn from our experience (read mistakes)
 How
could the data management look like?
Data for LHC-CP - slide 23
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
How could LHC traceability look like?
Distributed database
data administrator
LHC INB Trace data
LHC Installation & Layout data
LHC Components Production data
Data for LHC-CP - slide 24
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS
LHC Traceability




For LEP traceability, we started 15 years too late
For SPS traceability, we started 25 years too late
… For LHC traceability, we could start on time
Again, we should learn from our experience (read mistakes)
 How


could the data management look like?
There is no hurry yet but it should be considered
If the data is available in Oracle, we can work it out...
Data for LHC-CP - slide 25
LHC Production database Workshop - 1 March 2001
R. Billen, SL/MR/DBS