Transcript Event of 14 th May 2011

Update on the analysis of the event
of 14th May 2011
Emmanuele Ravaioli
Thanks to M. Zerlauth, G. Cumer, H. Thiesen, Z. Charifoulline, A. Verweij
TE-MPE-TM
18-08-2011
Event of 14th May 2011
Sequence of events
Impact on the QPS
Emmanuele Ravaioli
Short and mid-term
actions
TE-MPE-TM
18-08-2011
Long-term actions
2
Event of 14th May 2011
Impact on the
QPS
Sequence of events
1. LHC at injection current, no ramp, pilot beam in LHC, power
converters supply 760 A in steady state.
2. AUG triggering due to a defective AUG switch.
3. Loss of the 18 kV network around BA7, SMI2 and LHC point 2.
4. RB power converter trip in Sector 12 and Sector 23: PCs followed a
wrong reference voltage. But no PIC signal was generated.
5. Voltage across the power converters of both sectors decreased up to
~ -152 V, the filter at the PC output charged up.
6. Earth fault detected in the power converters (21 ms after the 18 kV
loss): a PIC signal was generated.
7. Only at this moment were the thyristors of the power
converter’s crowbar fired.
8. Current exchange between capacitive and inductive branches of the
PC filter, which lead to a substantial change of the magnet current
(±0.3 A) and to large oscillations at the output of the power
converters.
9. The voltage waves propagated through the dipole chains and caused
the triggering of the QPS and the subsequent firing of the quench
heaters of 9 dipoles.
10. The heaters of one further magnet were fired by the nQPS
(symmetric quench detector) due to lost reference, as the other two
magnets present in its crate triggered and fired the quench heaters.
Emmanuele Ravaioli
TE-MPE-TM
Short and midterm actions
Long-term
actions
Chaine AUG vers BA7
Chaine AUG vers UW25
For more info
Major Event Report (EDMS document: 1144995)
Rapport préliminaire (EDMS document : 1147728)
LMC Meeting on 18 May 2011 - TI2 AUG on Saturday 14
May, G. Cumer
Review AUG TI2, G. Cumer, 21 July 2011(EDMS document:
1154682)
18-08-2011
3
Event of 14th May 2011
Sequence of
events
Short and midterm actions
Impact on the QPS
• The dipoles whose QPS triggered were the
ones observed in the past to have apertures
with unbalanced impedances, and thus be
oversensitive to any voltage wave propagating
through the chain.
• The incident occurred at injection current and
during a plateau (760 A, 0 A/s). At higher
current the consequences would have been
heavier and lead to a larger number of
quenched dipoles.
• Two possible worst cases
• 2 kA, 10 A/s: The unbalanced dipoles
oscillate more when they are ramping
around 2 kA.
• 6 kA: The quenches are more dangerous
for the dipoles.
Emmanuele Ravaioli
TE-MPE-TM
Long-term
actions
For more info
LMC Meeting on 18 May 2011 - HTS quenches and
consequences of AUG, R. Schmidt
TE-MPE-TM on 28 July 2011 - Event of 14th May 2011 what we understood and what we did not, E. Ravaioli
14.05.2011 – Analysis of heater firings in RB.A12 and
RB.A23 following spurious AUG in TI2 (EDMS document:
1157870)
18-08-2011
4
Event of 14th May 2011
Sequence of
events
Impact on the
QPS
Short and mid-term actions
Long-term
actions
1. Incident report
2. Adding the signal from the AUG chain EUB1/2I into
the sum signal sent to the powering interlock
controllers around IR2 (EN-EL during next technical
stop, end of August)
(Approval of LMC needed)
3. Improved model of the power converter, same results
• Current source → Voltage source
• When the 18 kV loss occurred, PCs followed a
wrong reference voltage (
)
• Thyristors of the PC switched-off at the moment
of the 18 kV loss (gate current to 0)
• Thyristors of the crowbar fired only after 22 ms,
when an earth fault is detected
4. Establishing whether similar events (no current in the
PC branches but no PIC signal sent) may occur in the
future, and assessing the consequences (voltage at the
output of the PC, firing of the quench heaters)
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
5
Event of 14th May 2011
Sequence of
events
Impact on the
QPS
Short and midterm actions
Long-term actions
1. Modification of the AUG chains of TI2 (and
later on TI8) is currently under study by ENEL, to realign current inconsistencies between
the powering layout, sector doors and the
AUG chains of TI2 and TI8. This
modification will allow disconnecting again
the upstream parts of TI2/TI8 from the PIC
interlock logic, as AUG activations in this
zone will not affect LHC powering anymore.
(not possible before LS1)
2. Study of alternative and more radical
solutions to the problems caused by the
voltage waves travelling through the dipole
chain (example: Knud’s idea: Changing the
earthing point of the RB circuit)
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
6
Event of 14th May 2011
Sequence of events
Impact on the QPS
Short and mid-term
actions
Long-term actions
Chaine AUG vers BA7
Chaine AUG vers UW25
For more info
LMC Meeting on 18 May 2011 - HTS quenches
and consequences of AUG, R. Schmidt
For more info
Major Event Report (EDMS document:
1144995)
Rapport préliminaire (EDMS document :
1147728)
TE-MPE-TM on 28 July 2011 - Event of 14th
May 2011 - what we understood and what we
did not, E. Ravaioli
14.05.2011 – Analysis of heater firings in
RB.A12 and RB.A23 following spurious AUG
in TI2 (EDMS document: 1157870)
LMC Meeting on 18 May 2011 - TI2 AUG on
Saturday 14 May, G. Cumer
Review AUG TI2, G. Cumer, 21 July
2011(EDMS document: 1154682)
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
7
Event of 14th May 2011
Overview of implementation of AUG chains in TI2
Courtesy of G. Cumer
Chaine AUG vers BA7
Chaine AUG vers UW25
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
8
Event of 14th May 2011
Explanation of the event
1. The circuit is operating at constant current during injection @ 760 A .
2. A FPA is triggered. Normally the thyrstors of the crowbar should be fired and conduct most of the current,
but they fail to conduct. The only path through which the current of the circuit can flow is the capacitive
branch of the PC filter, which is charged. The voltage across the PC reaches high negative values (~ -150 V)
while the current in the circuit decreases with high ramp-rate (~ -15 A/s). There is no voltage ringing coming
from the PC since the inductive branch of the filter is blocked.
3. After 22 ms, the thyristors are fired. The current starts to flow through the crowbar. A current flows back and
forth between the inductors and the capacitors of the filter, which is discharging. A wave is generated and
travels along the dipole chain.
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
9
Event of 14th May 2011
Dipoles that triggered the QPS – Comparison with a past event – S12
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
10
Event of 14th May 2011
Dipoles that triggered the QPS – Comparison with a past event – S23
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
11
Event of 14th May 2011
Incident Report - Abstract
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
12
Event of 14th May 2011
New model of the PC: Voltage at the output of the PC (V_MEAS)
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
13
Event of 14th May 2011
New model of the PC: Current flowing in the circuit (I_A)
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
14
Event of 14th May 2011
New model of the PC: Voltage difference between the apertures of a dipole (QSO)
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
15
Event of 14th May 2011
New model of the PC: Voltage across a dipole (nQPS)
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
16
Event of 14th May 2011
New model of the PC: Main current in the circuit
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
17
Event of 14th May 2011
Limitation des interdépendances – Propositions EN/EL à valider
Courtesy of G. Cumer
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
18
Event of 14th May 2011
QSO signal – Comparison between measurements and simulations
Effect of the change of the earthing point of the circuit
Emmanuele Ravaioli
TE-MPE-TM
18-08-2011
19