Choose timezone
Your profile timezone:
MP3 meeting
Participants
Adriaan Rijllart
Alick Macpherson
Andrea Musso
Andrzej SIEMKO
Antonio Perin
Arjan Verweij
Bernhard Auchmann
Christian Giloux
Christine Darve
Ezio Todesco
Frederic Bernard
Giorgio D'Angelo
Glyn Kirby
Hugues Thiesen
Jean-Philippe Tock
Jim Strait
Karl Hubert Mess
Markus Zerlauth
Michele Modena
Mike Koratzinos
Odd Oyvind Andreassen
Rob van Weelderen
Robert Henry Flora
Roberto Saban
Rüdiger Schmidt
Sandor Feher
Walter Venturini Delsolaro
MP3 meeting 23 September 2009
The new QPS is included. Data from the new QPS are doubled (4Hz and 0.1 Hz). Data flow will be higher where the dead band is the LSB, and may lead to slow downloads.
The QPS team need to define the parameters for the trend plots
Mike Koratzinos asked whether there is any averaging in the data
Frederic answered no, the difference is only in the filtering
ALick asked about security policies
The system is not yet RBAC, however there are 5 privilege levels managed through the nice login.
Masking is only allowed to interlock experts
The safety policy for rebooting of servers through DIAMON is managed by the DIAMON team
Ruediger opined that the masking of QPS OK is probably the most critical.
Markus recalled that masking of the cryo interlocks still leaves in place the hardware protections
An important point concerned the MP3 experts when they want to approve a test in presence of a formally failed sequence.
In this case the PIC expert should be contacted BEFORE signing the test, to reduce the work needed to change the various data bases
Sandor pointed out that to the expert who does the analysis it might not be clear that the sequence has failed
Odd picked up the point, saying that he is going to get the information from LSA and display it in the PM analysis tool
Alick proposed to repeat the talk at the LHC section meeting
Ruediger said it will be as well accessible from the HCC pages
Alick asked the experts to always put a comment when failing a test
The latter has had 5 quenches
Walter commented that this is what we expect, and that for the RCD it is not worring as the circuit contains 77 magnets. Some MQTL are known to be weak.
Karl Hubert Mess argued that we don't know if the quenches are in the magnets or in the bus bar, he thinks the latter is more likely
Andrzej explained the reason why he had asked Andrea to do the investigation: it was to get an idea of how many circuits were retraining. He proposed to establish a rule of maximum 2 quenches for this commissioning run. The motivation is to avoid unnecessary risks once the minimum level for 3.5 TeV operation is achieved.
This proposal was accepted
It was also agreed to open an Non Conformity in MTF for each circuit which has more than 2 quenches
Andrea said that a button to do this quickly will be added to the analysis tool
Bernhard proposed to inform the MP3 members of the new rule with an email
Action-> Walter
Ruediger said that a list of NC will be available on the HCC web pages
The modification needs to be integrated into the document
Giorgio proposed to write a new document dedicated to the 120A of the triplet.
Markus said that we should try to keep the circuit types to a minimum
Jean Philippe asked who can write the document
The autors of the existing document will meet and agree on the way to follow
Sandor went on to talk of the 600A of the triplet, which share the same type of leads and control system
Here the situation seems not to require a change in the procedure.
It has been recently agreed to downgrade I_PNO to 400A, which is considered sufficient for the next LHC run
Christine recalled that some of the trips at 550A were caused by the new "square root" interlock which triggers at 550A with no margin.
She and Sandor are confident that the regulation will work well at 400A
Ruediger said that a NC documenting the downgraded current should be issued
Sandor took the opportunity of his talk to submit to the panel three issues related to the commissioning of the triplet. The first issue concerns a proposal for a different approach to Hipot, the second one the need to monitor and interlock the presence of the fuse, the third the installation of additional current sensors to monitor the individual magnet currents in the tripled with reconfigured.
Ruediger commented that for this a procedure need to be circulated and approved
On the second issue, Andrzej commented that the monitoring is in place and will be presented by Reiner, while the interlocking had not been discussed with the specialists and looked problematic
Markus confirmed, but he added that if a signal of "fuse present" is made available, it could be linked to the QPS OK.
Concerning the lead regulation Andrzej asked the cryogenics team to reconsider the alternative solutions. (See also the minutes of MP3 meeting 8 July 2009, AOB)
Karl Hubert Mess said that the next shutdown, in one year time, will probably be long, so starting now it should be possible to solve the problem
Andrzej asked to have a presentation with a proposed solution at one of the next MP3 meetings
The conclusions are that the propagation times depend on
1) The ratios of quenching magnet volume to the remaining.
2) The position of the quenching magnet with respect to the hydraulic plugs, buffer volumes and relief valves, and
3) The setting pressure of the relief valve.
In particular the lower the latter, the faster the propagation
Arjan asked what the time constant of the current discharge was
Rob answered it was similar to that of the RB circuit.
(from one of Rob's slides one can see the time constant was about 100s)
Arjan asked if it was possible to have 8 K in an interconnect after 15 seconds
ROb's answer was affirmative, (also this is supported by one of the plots shown )
Walter observed that from the evidence shown, early opening of the relief valves is a mistake, as it accelerates the propagation, and so the interconnects quench at higher currents
Arjan asked whether somebody was going to work to adapt and extend the model to the present LHC configuration
Andrzej said that it is foreseen
Walter remarked that in the short term and for practical purposes the qualitative understanding which we have must be used to identify the dipoles that have the fastest propagation of quench to the neighbours; so that it can be measured during PGC.
He is saluted by the applause of the colleagues.
Special topic "New developments/features of PVSS SCADA systems used during HCC (Frederic Bernard)
Frederic went through the new features of the PVSS SCADA systems (see slides).The new QPS is included. Data from the new QPS are doubled (4Hz and 0.1 Hz). Data flow will be higher where the dead band is the LSB, and may lead to slow downloads.
The QPS team need to define the parameters for the trend plots
Mike Koratzinos asked whether there is any averaging in the data
Frederic answered no, the difference is only in the filtering
ALick asked about security policies
The system is not yet RBAC, however there are 5 privilege levels managed through the nice login.
Masking is only allowed to interlock experts
The safety policy for rebooting of servers through DIAMON is managed by the DIAMON team
Ruediger opined that the masking of QPS OK is probably the most critical.
Markus recalled that masking of the cryo interlocks still leaves in place the hardware protections
FAQ in hardware commissioning (Markus Zerlauth)
The presentation addressed frequent causes of test/sequence failures during harware commissioning.An important point concerned the MP3 experts when they want to approve a test in presence of a formally failed sequence.
In this case the PIC expert should be contacted BEFORE signing the test, to reduce the work needed to change the various data bases
Sandor pointed out that to the expert who does the analysis it might not be clear that the sequence has failed
Odd picked up the point, saying that he is going to get the information from LSA and display it in the PM analysis tool
Alick proposed to repeat the talk at the LHC section meeting
Ruediger said it will be as well accessible from the HCC pages
Alick asked the experts to always put a comment when failing a test
Observed retraining in 600A correctors (Andrea Musso)
Andrea found that only a couple of circuits are retraining (see slides), an RCD and an RQTL circuitThe latter has had 5 quenches
Walter commented that this is what we expect, and that for the RCD it is not worring as the circuit contains 77 magnets. Some MQTL are known to be weak.
Karl Hubert Mess argued that we don't know if the quenches are in the magnets or in the bus bar, he thinks the latter is more likely
Andrzej explained the reason why he had asked Andrea to do the investigation: it was to get an idea of how many circuits were retraining. He proposed to establish a rule of maximum 2 quenches for this commissioning run. The motivation is to avoid unnecessary risks once the minimum level for 3.5 TeV operation is achieved.
This proposal was accepted
It was also agreed to open an Non Conformity in MTF for each circuit which has more than 2 quenches
Andrea said that a button to do this quickly will be added to the analysis tool
Bernhard proposed to inform the MP3 members of the new rule with an email
Action-> Walter
Ruediger said that a list of NC will be available on the HCC web pages
New procedure for the 120A correctors in the triplet (Sandor Feher)
Sandor illustrated the agreed procedure which includes the tuning of the lead control loopThe modification needs to be integrated into the document
Giorgio proposed to write a new document dedicated to the 120A of the triplet.
Markus said that we should try to keep the circuit types to a minimum
Jean Philippe asked who can write the document
The autors of the existing document will meet and agree on the way to follow
Sandor went on to talk of the 600A of the triplet, which share the same type of leads and control system
Here the situation seems not to require a change in the procedure.
It has been recently agreed to downgrade I_PNO to 400A, which is considered sufficient for the next LHC run
Christine recalled that some of the trips at 550A were caused by the new "square root" interlock which triggers at 550A with no margin.
She and Sandor are confident that the regulation will work well at 400A
Ruediger said that a NC documenting the downgraded current should be issued
Sandor took the opportunity of his talk to submit to the panel three issues related to the commissioning of the triplet. The first issue concerns a proposal for a different approach to Hipot, the second one the need to monitor and interlock the presence of the fuse, the third the installation of additional current sensors to monitor the individual magnet currents in the tripled with reconfigured.
Ruediger commented that for this a procedure need to be circulated and approved
On the second issue, Andrzej commented that the monitoring is in place and will be presented by Reiner, while the interlocking had not been discussed with the specialists and looked problematic
Markus confirmed, but he added that if a signal of "fuse present" is made available, it could be linked to the QPS OK.
Concerning the lead regulation Andrzej asked the cryogenics team to reconsider the alternative solutions. (See also the minutes of MP3 meeting 8 July 2009, AOB)
Karl Hubert Mess said that the next shutdown, in one year time, will probably be long, so starting now it should be possible to solve the problem
Andrzej asked to have a presentation with a proposed solution at one of the next MP3 meetings
Quench propagation in string 1 (Rob Van Weelderen)
Rob presented results of thermofluidodynamics experiments in string 1 where studies of quench propagation had been done as a function of the setting pressure of the relief valvesThe conclusions are that the propagation times depend on
1) The ratios of quenching magnet volume to the remaining.
2) The position of the quenching magnet with respect to the hydraulic plugs, buffer volumes and relief valves, and
3) The setting pressure of the relief valve.
In particular the lower the latter, the faster the propagation
Arjan asked what the time constant of the current discharge was
Rob answered it was similar to that of the RB circuit.
(from one of Rob's slides one can see the time constant was about 100s)
Arjan asked if it was possible to have 8 K in an interconnect after 15 seconds
ROb's answer was affirmative, (also this is supported by one of the plots shown )
Walter observed that from the evidence shown, early opening of the relief valves is a mistake, as it accelerates the propagation, and so the interconnects quench at higher currents
Arjan asked whether somebody was going to work to adapt and extend the model to the present LHC configuration
Andrzej said that it is foreseen
Walter remarked that in the short term and for practical purposes the qualitative understanding which we have must be used to identify the dipoles that have the fastest propagation of quench to the neighbours; so that it can be measured during PGC.
AOB
Karl Hubert thanks everybody for the long collaboration, and bids farewell.He is saluted by the applause of the colleagues.
There are minutes attached to this event.
Show them.
The agenda of this meeting is empty