Power converters implications for Booster Energy Upgrade
Jean-Paul Burnet, Serge PittetLIU day, 01.12.2010
TE EPC
MPS load characteristic
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 2
TE EPC
Magnets quantity
Total resistance [mOhms]
Total inductance [mH]
Dipoles outer ring 64 172 64
Focusing quadrupoles 128 74 8
Reference magnet 4 10 4
Defocusing quadrupoles 64 51 6
Dipoles inner ring 64 172 64
Total 324 479 146
16 periods, 4 rings
1.4 GeV Booster MPS (1/2)
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 3
TE EPC
BHZ Ring 1&4
Reference magnet
GND
QFO
QDE
BHZ Ring 2&3
MPS3600V4000Apk
2300ARMS
TRIM A1900V50Apk
30ARMS
QFO520V300Apk
200ARMS
QDE520V300Apk
200ARMS
Load: 0.5 Ohms / 0.18 H1.4GeV margins:• MPS voltage: -2%• MPS peak current: -2%• MPS rms current: 3%• TRIM A peak current: 40%• TRIM Q peak current: 0%
1.4 GeV Booster MPS (2/2)
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 4
TE EPC
• From 4000A to 5500A, the current on the Trim A must be increased by a factor 4 to compensate the Outer rings saturation.
%
• The existing MPS can run faster but cannot provide more peak current
2 GeV Booster MPS
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 5
TE EPC
64 BHZ Ring 1&4
Reference magnet
GND
128 QFO Ring 1-4
64 QDE Ring 1-4
64 BHZ Ring 2&3
MPS3600V5300Apk
TRIM A1900V270Apk
QFO520V420Apk
QDE520V420Apk
Upgrade scaling on one cycle• MPS peak current: +30%• MPS rms current: +35%• 18kV apparent power: +77%• 18kV peak power: +45%• TRIM A peak current: +540%• TRIM Q peak current: +40%
18kV network
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 6
TE EPC
0 0.2 0.4 0.6 0.8 1 1.2 1.40
0.5
1
1.5
2
2.5x 10
7
X: 0.8475Y: 2.081e+007
Reactive Power
X: 0.8676Y: 1.582e+007
1.4GeV
2.0GeV
0 0.2 0.4 0.6 0.8 1 1.2 1.4-1
-0.5
0
0.5
1
1.5
2
2.5x 10
7 Active Power
X: 0.6925Y: 1.368e+007
X: 0.742Y: 2.053e+007
1.4GeV
2.0GeV
Fast transients with 2GeV cycle which can not be compensated by Meyrin SVC.
50% increase of active power, 30% increase of reactive power
Upgrading existing supply
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 7
• All power components must be upgraded to the new peak current (Thyristors, transformers, filters,…)
• Four consecutive 2GeV cycles has to be considered as a DC value for power semiconductors.
• Replace the existing TRIM A (+ 1 spare).• Replace the existing QFO and QDE (+1 spare).
Side effects:• Major upgrade of magnets cooling circuit
needed.• Power DC cables section must be increased.• 18kV average power capability to be checked.• 18kV network will probably not be able to
absorb the new peak power pulse. Flickers to be foreseen!
TE EPC
POPS MPSB existing
MPSB 2Gev Hyp
Imax 6kA 4kA 6kA
Vmax 10kV 3.6kV 3.6kV
Irms (dc side) 3.2kA 2.3kA 3.1kA
Plossrms 4MW 2.9MW 5.2MW
Emagpeak 12MJ 1.5MJ 2.5MJ
“POPS” basic principles
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 8
The energy to be transferred to the magnets is stored in capacitors.
• DC/DC converters transfer the power from the storage capacitors to the magnets.• Four flying capacitors banks are not connected directly to the mains. They are charged via the magnets.• Only two AC/DC converters (called chargers) are connected to the mains and supply the losses of the
system.
TE EPC
DC3
DC
DC +
-
DC1
DC
DC
DC5
DC
DC +
-
DC4
DC
DC-
+
DC2
DC
DC-
+
DC6
DC
DC
MAGNETS
+
-
+
-
CF11
CF12
CF1
CF21
CF22
CF2
AC
CC1
DC
MV7308
AC AC
CC2
DC
MV7308
AC
18KV ACScc=600MVA
OF1 OF2
RF1 RF2
TW2
Crwb2
TW1
Lw1
Crwb1
Lw2
MAGNETS
AC/DC converter - AFE
DC/DC converter - charger module
DC/DC converter - flying module
Chargers
Flying capacitors
Active power of the magnets
-60000000
-40000000
-20000000
0
20000000
40000000
60000000
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Time [s]
Po
wer
[W
]
.
Power to the magnets:
+50MW peak
Resistive Losses and charger power
0
2000000
4000000
6000000
8000000
10000000
12000000
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Time [s]
Po
we
r [W
] .
Power from the mains
10MW
POPS based proposal (1/2)Benefits• Overall voltage available increases and would
allow a reduction of the RMS current using a faster ramping.
• The capacitor bank totally absorbs the peak power on the 18kV network. Meyrin SVC would then become optional.
• Spare sharing between MPS A and B and eventually with POPS.
• Only a few new cables needed between the reference magnet (BCER) and the MPS.
• New B-field regulation to minimize eddy currents and saturation effects impact at higher current and acceleration rate.
Drawbacks• Cost estimation 14MCHF using POPS module.• RF acceleration has to be increased.
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 9
TE EPC
Reference magnet
Ring 2&3
GND
MPS B2500V5500A
QFO600V450A
QDE600V450A
GND
MPS A2500V5500A
64 BHZ Rings 1&4
128 QFO Rings 1,2,3&4
64 QDE Rings 1,2,3&4
64 BHZ Rings 2&3
Reference magnet
Ring 1&4
Ring 1 B-filedmeasurement
Ring 2 B-filedmeasurement
POPS based proposal (2/2)
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 10
TE EPC
Upgrade scaling on one cycle• MPS peak current: +30%• MPS rms current: +3%• 18kV apparent power: -26%• 18kV peak power: -78%
DC3
DC
DC +
-
MAGNETS
V_F3
UdcF3
AC
CC1
AC
MV7308
DC
DIP+QF+REF
DC3
DC
DC +
-
V_F3
UdcF3
AC
CC1
AC
MV7308
DC
DIP+QD+REF
DC3
DC
DC +
-
UdcF3
AC
CC1
AC
MV7308
DC
Layout of new building
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 11
TE EPC
New
MPSB
(2 GeV)
(26m x 16m)
TT
6
Cap bank #1 Cap bank #2 Cap bank #3
Control roomLocal for Cooling& Air Conditioning
AC/DC3 spare (1.2m x 10m) DC/DC3 floating (1.2m x 10m)
AC/DC1 (1.2m x 10m) AC/DC2 (1.2m x 10m)
DC/DC1 (1.2m x 10m) DC/DC2 (1.2m x 10m)
Out
put c
ubic
le
QF QD QS16
m
26m
4m4m
• 420m2 required, only 190m2 available in bdg. 271 ground floor.• Only 200kg/m2 allowed in bdg. 271 first floor.• No space left in bdg. 361.
With this new building:• We can install and commission during Booster operation.• We have a backup power supply during the first years.• Easy connection to existing cables and cooling services.
Effect of idle cycles
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 12
TE EPC
RMS current can also be reduced keeping RF acceleration used today (3.6kV on the MPS) and providing 2GeV cycles to LHC only:
Only 10% RMS current increase if ZERO cycles are inserted. ZERO 12 and 18 would allow a slow setting up of the transfer line:
No increase of the RMS current (compared to all cycles at 1.4GeV) if Isolde has no physics during LHC proton filling:
PSB LHC25A LHC25B ISO LHC25A LHC25B ISO LHC25A LHC25B ISO LHC25A LHC25B ISO EASTB ISO ISO TOF TOF ISO
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Energy 2GeV 2GeV 1.4GeV 2GeV 2GeV 1.4GeV 2GeV 2GeV 1.4GeV 2GeV 2GeV 1.4GeV 1.4GeV 1.4GeV 1.4GeV 1.4GeV 1.4GeV 1.4GeV
RMS partial 2822 ARMS 2822 ARMS 2822 ARMS 2822 ARMS 2267 ARMS 2267 ARMS
RMS total 2650 ARMS
PSB LHC25A LHC25B ZERO LHC25A LHC25B ZERO LHC25A LHC25B ZERO LHC25A LHC25B ZERO EASTB ISO ISO TOF TOF ZERO
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Energy 2GeV 2GeV 2GeV 2GeV 2GeV 2GeV 2GeV 2GeV 1.4GeV 1.4GeV 1.4GeV 1.4GeV 1.4GeV
RMS partial 2517 ARMS 2517 ARMS 2517 ARMS 2517 ARMS 2267 ARMS 1873 ARMS
RMS total 2380 ARMS
PSB LHC25A LHC25B ZERO LHC25A LHC25B ZERO LHC25A LHC25B ZERO LHC25A LHC25B ZERO EASTB ZERO ZERO TOF TOF ZERO
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Energy 2GeV 2GeV 2GeV 2GeV 2GeV 2GeV 2GeV 2GeV 1.4GeV 1.4GeV 1.4GeV
RMS partial 2517 ARMS 2517 ARMS 2517 ARMS 2517 ARMS 1371 ARMS 1873 ARMS
RMS total 2263 ARMS
Existing supply with idle cycles
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 13
• All power components must be upgraded to the new peak current (Thyristors, transformers, filters,…)
• Four consecutive 2GeV cycles has to be considered as a DC value for power semiconductors.
• Replace the existing TRIM A (+ 1 spare).• Replace the existing QFO and QDE (+1 spare).
Side effects:• Major upgrade of magnets cooling circuit
needed.• Power DC cables section must be increased.• 18kV average power capability to be checked.• 18kV network will probably not be able to
absorb the new peak power pulse. Flickers to be foreseen!
TE EPC
POPS MPSB existing
MPSB 2Gev Hyp
Imax 6kA 4kA 6kA
Vmax 10kV 3.6kV 3.6kV
Irms (dc side) 3.2kA 2.3kA 2.5kA
Plossrms 4MW 2.9MW 3.2MW
Emagpeak 12MJ 1.5MJ 2.5MJ
New magnet requirementsIn order to be powered by the existing power supply 4kA/3.6kV, magnet coil
copper cross section must increase• By 30% to increase the number of turns and provide the additional field
with the MPS rated current.• By 30% to compensate the additional coil length.• By 400% to compensate the increasing inductance (L n2) and keep
enough voltage margin for the desired acceleration (U = RI + L dI/dt).
Resistive losses divided by 4but
Copper weight multiplied by 7
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 14
TE EPC
And what if… (1/2)
… we increase reasonably the amperes-turns and the section on all magnets:• Existing current rating would be sufficient.• Additional voltage needed to compensate the increasing magnets
inductance.• Reactive power still increases and the Meyrin SVC capability would be an
issue within 1.2s cycle.• At the end a POPS based supply would be needed.
…we increase the amperes-turns and/or the section on the bending only:• Existing supply could be used to power the quadrupoles.• A smaller POPS based supply could be used for the bending magnets.
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 15
TE EPC
And what if… (2/2)
…we decrease the magnet gap (and vacuum chamber size) by 30%• Existing power supply can be used. The magnetic cycle would have to be
adapted to compensate for the 30% increase in load inductance.
…we install a new MPS adapted to the existing magnets and they fail after a few years of operation (they were designed for 0.8GeV).
In any cases several MCHF would be saved on the new supply and/or on electricity consumption with new magnets:
15GWh per year at 1.4GeV → 18 MCHF over 25 years10% increase in current → 21% increase in power → 3.8MCHF
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 16
TE EPC
Booster Ring (others)
• Dipoles correctors and multipoles are mainly used at low energy and have enough margin. They will be consolidated during the shutdown 2011-2012.
• The Qstrips are only used at injection. Any upgrade would be part of the linac 4 project and not of the Booster energy upgrade.
• BDLs are used at ejection but have enough margin.
• The DBS are dedicated to destructive beam measurement and will probably not be used at 2GeV.
• The shavers are only used at injection. Any upgrade would be part of the linac 4 project and not of the Booster energy upgrade.
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 17
TE EPC
PSB Ejection
• BE.SMH:2GeV setting not exactely defined yet. The capacitor bank size will have to be adapted to provide the additional energy and a new capacitor charger may have to be foreseen.
• BE1 to 4.DHZ 11L1, BE1 to 4.DHZ 4L1, BE1 to 4.DVT 11L1 and BE1 to 4.DVT 4L1:Even approaching their rated limits, these converters will not need to be upgraded for 2GeV operation.
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 18
TE EPC
BT and BTM transfer
• The transfer line includes some huge bendings (RL load similar to MPS!). An important power transfer is needed for PPM operation.
• The BT and BTM transfer lines will have to provide PPM operation between 1GeV or 1.4GeV Isolde and 2GeV PS.
• The natural discharge of the magnet is estimated by the current reached after 0.9s starting from 2GeV current.
• The capacitor bank size of ALG-1 model will have to be adapted to provide the additional energy.
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 19
TE EPC
BTP transfer
• The BTP transfer is not used by Isolde. PS cycles only will go through.• It is a non-PPM only line and quadrupoles should be redesigned for PPM
operation, allowing optics flexibility on various beams.
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 20
TE EPC
EPC Conclusions
01/12/2010 TE-EPC - Serge Pittet - EDMS 1106116 21
MPS:• The existing supply can not endorse the additional RMS and peak current, even
for one single pulse.• An increase of peak power, using traditional thyristor technology, would have a
significant negative effect on power quality of the Meyrin network 18 kV.• The solution will probably be a design similar to the new POPS for the PS, using
DC capacitors to store the energy for the pulsating load (civil engineering work required, but less demanding on resources, procurement and commissioning).
• A higher voltage ratings on each magnet would allow to limit the increasing RMS current, pushing RF acceleration to their limits. Impact of the upgrade on magnets, DC cables, 18kV average power could then be notably reduced. The alternatives would be to reduce Isolde physics program.
Transfer lines:• Pulse option must be studied with TE-MSC to reduce the overall RMS current
(many magnets could then stay in place).
TE EPC