Performance diagnostics for iron post Nb Sn magnets at CERN · Type of performance limits of a...

Performance diagnostics for

Nb3Sn magnets at CERN

Gerard Willering

23 September 2019, Mon-Mo-Or3-011

August 2019: First series MBH 11T

magnet qualified for installation in

the LHC.

vertical pad

iron yoke

iron post

100 mm

aperture

11T dipole MQXF Quadrole

HL-LHC

FRESCA2

14.6 T dipole

Thanks to many persons who have made the magnets and supported the tests.

Special thanks to the direct test team, M. Bajko, H. Arnestad, H. Bajas, M. Duda,

F. Mangiarotti, J. Feuvrier, V. Desbiolles, E. Karentzos, G. Ninet

FRESCA2 14.6 T dipole

attached to insert

2

Typical results as they are presented after a magnet test

0

2

4

6

8

10

12

14

0 1 2 3 4 5 6 7

Cu

rre

nt

[kA

]

Event number

MBHB-002 training

D1U

D1L

D2L

D2U

No quench

Nominal

Ultimate1.9 K 4.5 K 1.9 K

the

rma

lcyc

le

CD 1 CD 1 CD 2 CD 24.5 K7000

8000

9000

10000

11000

12000

13000

14000

0 5 10 15 20 25 30 35

Qu

ench

cu

rren

t (A

)

Quench number

MBHSP101

MBHSP102

MBHSP103

MBHDP101

MBHSP104

MBHSP105 A training plot is the first

important indicator of how the

magnet performs.

However:

6000

7000

8000

9000

10000

11000

12000

13000

0 5 10 15 20

Que

nch

curr

ent (

A)

Quench number

MBH-proto 1.9 K

MBH proto 4.5 K

D2U

CSD

2UCS

D2U

NCS

D2U

NCS

D2U

NCS

D2U

NCS

1 A

/s -

D2U

NCS

D2U

CS

D1L

CS

-spl

ice

cycl

e

10 A

/s -

D2U

NCS

50 A

/s -

D2U

CS

100

A/s

-D

2UCS

150

A/s

-D

1L

150

A/s

-D

2U

There is much more to learn!

Gerard Willering, Performance

diagnostics Nb3Sn magnets

Contents

1. Performance limits and indicators

2. Tool box for performance diagnostics

3. Case studiesCase 1: Nb3Sn magnet limited by mechanics

Case 2: Nb3Sn magnet limited by homogeneous JcB,T,ɛ) conductor degradation

Case 3: Nb3Sn magnet limited by local conductor degradation

3

Note: This presentation focusses on the diagnostics,

attempting to avoid the “magnet” discussion.



Performance limits dominated by mechanical transients

4

Dominated by mechanical transients- Movements between different magnet components (slip-stick)

- (micro) cracking of epoxy

- Debonding of turns with impregnated pole

- Movement between strands

- Training, Detraining, Memory after Thermal Cycle

NbTi magnet investigations are mainly limited to this category.



Type of performance limits of a magnet

5

Conductor dominated- Design short sample limit Iss

- Iss homogeneously affected by strain state of conductor (Jc(B,T,ε))

- Iss homogeneously affected by filament damage (too high stresses at some moment in

conductor lifetime)

- Local damage to conductor (one or a few strands)

- Local damage on filament level (self-field instabilities, low RRR, etc.)

We find that many Nb3Sn accelerator magnets have limits dominated the conductor.

Ultimate performance

for each magnet

Two workshops in the past year underlined the importance of the topic:- October 2018: “Nb3Sn Rutherford cable characterization for accelerator

magnets”, CEA Paris-Saclay, France. https://indico.cern.ch/event/743626/

- April 2019: “Instrumentation and Diagnostics for Superconducting Magnets”,

Berkely, California, USA. https://idsm01.lbl.gov/



https://indico.cern.ch/event/743626/

https://idsm01.lbl.gov/

Instrumentation tool box

At CERN there were not too many changes in the toolbox over the last years:

Basic instrumentation:

Voltage taps

a) In model magnets small segments are measured

b) In series magnets only the full magnet length.

Pickup coils/Quench Antenna, typically in the bore of the magnet

1. Long coils (1.2 m) for “normal” investigation

2. Short coils (4 cm) “trouble shooting”.

6

Changes mainly in Data Aquisition and Diagnostics:

a. V-I measurements

b. Signal analysis

c. Combining diagnostics data



7

Origin

Quench # 3:

Vibrations in

both heads

Fully trained

magnet:

Some vibrations

left in connection

lead end

First powering after

thermal cycle:

Good memory, a few

vibrations still left.

Healthy magnet.

Define origin, amplitude and activity

for each vibration and visualize in

color map. Use earthquake analogy.

Toolbox: Vibrations and Magnet training



Contents




Case 2: Nb3Sn magnet limited by homogeneous Jc(B,T,ɛ) conductor degradation


8Gerard Willering, Performance


9

10

11

12

13

14

15

0 2 4 6

Bo

re F

ield

(T

)

therm

alcycle

FR

ES

CA

2a

FR

ES

CA

2b

FR

ES

CA

2b

0 2 4 6 8 10 12 14 16

Quench Number

1.9K

4.5K therm

alcycle

FRESCA2c

Case study 1: Coil limited by mechanical transients

- Training ends in an “erratic” range of ±100 A

- All training quenches preceded by large precursor

- Reaching higher fraction of short sample limit at

4.5 K compared to 1.9 K, but similar quench

characteristics.

- All quenches in the same coil layer.

In this case, the “circumstantial evidence” hints

in the direction of a mechanical limit, so we take

a look.



Vibration pattern analysis- Very repetitive pattern for vibration, main activity starting

at about 13.6 T

- No training effect from ramp to ramp

- Also 4.5 K quenches fall in the zone with high vibration

activity.

- Main activity next to Quench Antenna 5, end of straight

segment, start of flared end.

Conclusion for this magnet:

- Coil is mechanicall stable up to 13.6 T and shows more

activity in mechanical transients which limits the magnet to

14.6 T.

13.6 T

10

14.6 T

Note that this is a World Record FRESCA2 dipole

magnet with 100 mm bore has potential to go higher.

This diagnostics can help to improve a magnet.

Courtesy CERN, CEA Paris-Saclay.

Case study 1: Coil limited by mechanical transients

13.6

13.8

14

14.2

14.4

14.6

14.8

0 10 20 30 40 50 60

B (

T)

Ramp rate(A/s)

1.9 K

4.5 K

0.78

0.79

0.8

0.81

0.82

0.83

0.84

0 10 20 30 40 50 60

I/Is

s (-

)

Ramp rate(A/s)

1.9 K

4.5 K



Contents








12

7000

8000

9000

10000

11000

12000

13000

14000

0 5 10 15 20 25 30 35

Qu

ench

cu

rren

t (A

)

Quench number

MBHSP101

MBHSP102

MBHSP103

MBHDP101

MBHSP104

MBHSP105

MBHDP102

Case: 11 T double aperture model magnet.

Case 2: Nb3Sn magnet limited by homogeneous (JcB,T,ɛ) conductor degradation

- Training immediately at coil limit.

- All quenches exactly same voltage buildup,

location and pattern.

- No precursor or vibration for any of the quenches.

- Same fraction of Iq/Iss at 4.5 K as at 1.9 K, but far

below expected value.

- “Normal” ramp-rate dependency shape.



13

-5

0

5

10

15

20

8000 9000 10000 11000 12000

Vo

ltag

e (

uV

)

Current (A)

Voltage on midplane segment 109 II-I1

4.5 K

1.9 K

4.5 K vs 1.9 K

V-I measurement of the

quenching segment of

1.3 meter long mid-plane

Results in this case:

- SC-to-normal transition visible.

- Expected shift in from 4.5 K to 1.9 K.

- Stable voltage of 18 µV at 11.3 kA shows no current redistribution effect.

Case 2: Nb3Sn magnet limited by homogeneous (JcB,T,ɛ) conductor degradation

Method:

Measure voltage (high precision, low frequency) over

short coil segments. Use 2 minute current plateaus.

Conclusion

Homogeneous critical current reduction (affecting all strands over width or over > cable twist pitch).

Could be reversible strain effects or irreversible degradation or combination of both.



Contents








logo

area


At normal ramp of 20 A/s

- Absence of vibrations and precursors.

- Repetitive quench levels (no training effect)

- 1.9 K quench current higher than at 4.5 K

- At 4.5 K large improvement in quench current after 1

hour flat-top.

- Each temperature has its own quenches pattern.

- Different fraction of Iq/Iss at 4.5 K compared to 1.9 K,

- “Abnormal” ramp-rate dependency shape.

-> Gives a suspicion of local conductor degradation.

15Gerard Willering, Performance diagnostics Nb3Sn magnets

16


At 4.5 K

At 1.9 K

V-I measurement over the short quenching segment

At 4.5 K

- We stop ramping just before quench.

- Voltage was building up, but decays in short time

- When ramping slowly at 1 A/s later, voltage builds

up again until quench.

- Clear sign of current redistribution effect.

At 1.9 K

- Same current cycle

- Too far from local critical surface to see any

voltage buildup.

Why quenches at 1.9 K at a current lower than 4.5 K

without voltage buildup?

- Spikes in voltages, only seen at 1.9 K, suggest

self-field instabilities that quench the magnet.



At 1.9 K

Complex case combining multiple types of coil limits

Performance Diagnostics Summary

17

Even a small tool box can give a lot of possibilities for diagnostics

Combining results of different tests with correct interpretation is

very powerful for giving feedback to magnet performance.



Thank you

18

Suggestions for talks this conference touching this topic

Tu-Mo-PL2-01, Helen Felice, Advances in Nb3Sn Superconducting Accelerator Magnets

Wed-Mo-PL4-02, Franco Mangiarotti, “Superconducting Magnet Testing: The Art of Giving Feedback on Magnet Design”

Fri-Mo-Or25-07, Maxim Marchevsky, “Analysis of the transient mechanics behind superconducting accelerator magnet training”

Wed-Af-Or14-06, Bernardo Bordini, The effect of transverse loads on Nb3Sn Rutherford cables for accelerator magnets

With many thanks to all groups give us nice magnets to “play” with:

CERN, TE-MSC group

CEA, Paris-Saclay

Ciemat, Madrid



19

Backup slides



20

Typical work flow for investigation of performance limits

Look for circumstantial evidence

Variability of quench current

Temperature dependency (1.9 K and 4.5 K)

Holding current tests

Inversed or normal ramp rate dependency

Look for presence of precursors/vibration at

quench onset

Perform deeper investigation

V-I curve on voltage segments

- Look for SC-NC transition

- Look for decay at flattop

- Look for negative voltages

- Look for high-field instability spikes

Look at possible ….. Etc….

Other effects to verify

- Very fast longitudinal quench propagation

- Very slow training (seen in layer jump issues in 11T short models, explained as combination of some

strands having small margin and more susceptible to mechanical transients).

- Variations in strain gages

- Look at variations in higher order field errors with rotating coils

If any anomaly

shows up:



21

Normal for Nb3Sn:

From 1.9 K to 4.5 K about 8 to

10 % reduction of ISS

What is normal: Temperature dependency (Nb3Sn)

൘𝐼𝑄,1.9𝐾

𝐼𝑆𝑆,1.9𝐾= ൘𝐼𝑄,4.5𝐾

𝐼𝑆𝑆,4.5𝐾



22

Expected in perfect case: IQ = Iss at low ramp rate

Homogeneous degradation:Expect similar curve shape, but at lower fraction.

Non-homogenous degradation/self field instabilitiesExpect abnormal shape, for example highest Iq at intermediate current

levels.

What is normal: Ramp rate dependency



23

-2

0

2

4

6

8

10

12

14

8.0 10.0 12.0 14.0 16.0

E (u

V/m

)

Current (kA)

Measured 116-I7-I8

Measured 109 II-I1

Peak field, expected

Block 3, expected

Midplane, expected

Reduced strand Ic and n - block 3

Reduced strand Ic and n - midplane

𝐸(𝐼, 𝐵) = 𝐸𝑐𝐼

𝑓𝐼𝑐𝐼𝑐(𝐵)

𝑓𝑛𝑛(𝐵)

Expected vs reduced electric field

measurement in a degraded conductor.

Simply adding

- a reduction factor (𝑓𝐼𝑐) for Ic- a reduction factor (𝑓𝑛) for n

What is normal: V-I curves?

V-I curves

Note:

A reduction of Ic by 10 % at fixed field (cable/strand test characterisation )

corresponds to about 3 % of reduction on the load line for a magnet

Note 2: Generally two main causes of Ic variation are distinghuished

- Filament breakage: permanent degradation

- Strain state of the filament: can be reversible

It is difficult to separate the two in magnet measurements.

Ic(B) and n(B) characteristics measured in B163



Performance diagnostics for iron post Nb Sn magnets at CERN · Type of performance limits of a...

Documents

Transcript of Performance diagnostics for iron post Nb Sn magnets at CERN · Type of performance limits of a...