Reliability-oriented activities around the MAX 700 MHz...
Transcript of Reliability-oriented activities around the MAX 700 MHz...
Reliability-oriented activities around the MAX 700 MHz horizontal cryomodule
Frédéric Bouly (IPN Orsay – CNRS/IN2P3)
2nd Open Collaboration Meeting SLHiPP-2
INFN, Catania, SicilyFriday, May 4th 2012
Superconducting MYRRHA Linac
2
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012
Introduction
3 sections : 48 Spoke Resonators : βg = 0.35 - 352.2 MHz - Eacc(@ βopt) = 6.2 MV/m 34 5-cell elliptical cavities : βg = 0.47 - 704.4 MHz - Eacc(@ βopt) = 8.2 MV/m 60 5-cell elliptical cavities : βg = 0.35 – 704.4 MHz - Eacc(@ βopt) = 11 MV/m
Reliability & Goals
3Introduction
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012
Demonstrate the ADS Concept & Transmutation Coupling : Accelerator
+ spallation source + subcritical reactor
Avoid beam trips longer than 3 seconds to minimise thermal stresses and fatigue on target, reactor & fuel assemblies. Specifications : Less than 10 trips per 3 months operation cycle.
Consequences on Linac Design :
“Overdesign” of the component : Operation margin between operation and technological limitations (from beam optics to R&D)
Redundancy & Fault tolerance :Big phase and energy acceptance of superconducting cavities
Capabilities of fast retuning for “fast fault-recovery scenarios” and failures compensations.
Requires LLRF & control system for cavity tuning system operating in CW.
Fault scenario
4Fault tolerance & reliability study
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012
Typical case : One cavity loose its RF power supply. Region of interest of the MYRRHA superconducting Linac : βg = 0.47
Cavity voltage
Cavity synchronous phase
Losses along the linac
Without linac retuning the beam is lost.
Fault recovery scenario : compensation
5Fault tolerance & reliability study
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012
Compensation Strategy : the failure of a cavity is compensated by the 4 neighbours cavities: 2 cavities placed upstream the failed one 2 cavities placed downstream the failed one
Cavity voltage
Cavity synchronous phase
Phases & accelerating fields have to be modified in less than 3 seconds.
No beam during the procedure.
Feasibility of retuning procedures
6Fault tolerance & reliability study
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012
To study the retuning procedure of superconducting cavities a Matlab SimulinkTM model have been established. 2 feedback loops were modelled :
RF feedback loop Frequency tuning system of the cavity
feedback loop.
Model based on a βg = 0.47 5-cell elliptical cavity prototype (from TRASCO project).
Motor:«slow» action (~ 1 s)Large detuning (~ 250 kHz).
Piezoelectric actuators : «Fast» action (~ 1 ms) Fine detuning (~ 2 kHz).
Feedback loops modelled
7Fault tolerance & reliability study
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012
RF feedback loop : Based on an existing digital LLRF system
Modelled as Delay + ZOH + BP filters + IOT transfer function. PI correctors tuned to minimise Beam loading
Fast tuning system feedback loop
Needed to minimise power consumption & compensate perturbations (Lorentz, microphonics)
Transfer function of piezo action have been measured and modelled.
Optimum frequency detuning :
Study of a retuning scenario (compensation example)
8Fault tolerance & reliability study
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012
Cavity set points have to be changed in order to compensate another failed cavityin the linac
Beam switched off and on duringthe procedure
Synchronous phase has to bechanged
Accelerating field ramped up toenable the tuning system tocompensate Lorentz forces
Resonance frequency of cavityhas to be changed
Retuning scenario
9Fault tolerance & reliability study
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012
RF power (kW)
Power delivered to the beam (kW)
Accelerating field (MV/m)
Synchronous phase (°)
Cavity detuning (Hz)
The cavity can be retuned in200 ms (i.e. one order ofmagnitude lower than 3 s)
When Eacc ramped-up anoverconsumption of RF poweris observed This iscompensated by the piezosfeedback loop in less than~100 ms.
A PI controller is used for thetuning system loop.
3 strong microphonicsperturbations are assumed :
@ 1Hz with ∆fperturbation = +/- 10 Hz@ 71 Hz with ∆fperturbation = +/- 10 Hz@ 120 Hz with ∆fperturbation = +/- 20 Hz
Control of the tuning system
10Fault tolerance & reliability study
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012
Different control strategies studied
Simulations showed that an adaptive andpredictive system enable to minimise thepower consumption compare to a PIcorrection. It also adapt its correction effectwhen an unforeseen perturbation occurs.
Controller
Band-pass Filters
+∆fcav-errorPiezoscontrol action
Identified microphonics perturbations canbe compensated by applying a correction onthe perturbation frequency thanks to band-pass filters.
(Isaias Martin Hoyo, ADEX)
Tuning system requirements
11Fault tolerance & reliability study
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012
The results shows that it is conceivable to retune the linac in less than 3 sec with a fast tuning system.
Another important results : To avoid beam “deceleration” in the unused cavity it is necessary to change its resonance frequency.
It has to be detuned by 13 kHz in less than 3 sec.Minimum detuning speed of 5 kHz per second.
An hardware scheme is currentlyunder development
Both “PID + Bandpass FiltersApproach” and “ADEX + BandpassFilters Approach” will be tested on theCold Tuning System of the 5 cellelliptical cavity prototype.
The horizontal cryomodule prototype
12Reliability test bench
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012
A prototype of cryomodule has been installed in anexperimental pit at IPN Orsay . It holds a 5-cell ellipticalcavity equipped with its blade tuner.
Development achieved by LASA (INFN Milano) for thecavity + cryogenic vessel and by IPNO for the valve box.
Development achieved by LASA (INFN Milano) for thecavity + cryogenic vessel & by IPNO for the valve box.
GOAL :Dispose of a “real scale” experimental facility to Carryout Reliability-oriented experiments with a fully-equipped 700 MHz prototypical cryomodule controlledby feedback systems (RF and Fast Cold Tuning System).
Component robustness retuning procedures
Module description
13Reliability test bench
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012
First experimental results
14Reliability test bench
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012
Stepper motor capabilitiesMax detuning range (@ 4K): ~270 kHz Observed hysteresis 2 kHz
Several tries needed tomeasure the Q0 curves at 2Kbecause of cryogenic leaks.
Q0 obtained in critical couplingMultipacting easily processed Quench value a bit lower thanexpected (field unflatness) Still above MYRRHA requirements
Conclusions & Prospects
15Conclusion
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012
A model of cavity + feedback system – based on measurement – has been developed tostudy fast fault recovery scenarios in the MYRRHA linac.
Results show it is possible to locally retune the linac to compensate a failure, but uponcondition that :
Dispose of a fast and efficient LLRF systemalready developed and tested The tuning system must be fast enough to correct Lorentz detuning and microphonics. It also has to detune the cavity on a large frequency range with a minimum speed of
5kHz/secMotor + Piezos & development of an hardware for adaptive and predictive control.
Robustness of many different components of the module has been tried and improved. Leaks on Helium Tank, modification of circuitry to improve cooldown speed, automation
of all the cryogenic procedures.
Promising results obtained with the test of the cavity in critical coupling (Q0
measurement in agreement with expectations)
Next step is the High Power tests. The conditioning of the power coupler is currently inprogress.
16
THANK YOU !
Frédéric Bouly 2nd Open Collaboration meeting SLHIPP, Catania May 4th 2012