Alignment And Stability Of The TPS Storage Ring Auto-tuning Girder System

Tse-Chuan Tseng NSRRC, Taiwan

IWAA’16 , OCT. 6, 2016

1

Outline

• Girder system design briefing • Girders Installation at TPS tunnel • Survey alignment works • Auto alignment preparation and

implementation • Girder stability during commissioning

period • Conclusions

2

NSRRC Site Aerial View

3

TLS

TPS

Taiwan Photon Source

4

The Dedication Ceremony of Taiwan Photon Source Inauguration

5

The Ceremony Officially Made the Facility Available to Researchers Worldwide With 7 beamlines in top-up mode 300mA

Design goals of the girder system for TPS • Firm support and precise positioning of magnets (30mm RMS),

Vacuum Chamber & BPM

• Higher nature frequency (above 30 Hz is better)

• Alignment accuracy between girders within 0.1mm(relative) Traditional alignment network simulation reveals an accuracy of 0.3mm typically. It

needs to be iterated several times to reduce to 0.1mm

• Precise resolution (mm)

Manual adjustment mechanism is of poor resolution and time consuming

• Toward whole ring automatic alignment (optional)

How to align the girders precisely and quickly with less manpower? Considering the deformation of the floor and limited space in the tunnel also frequent earthquakes in Taiwan .

6

A 6-axes motorized adjusting mechanism was thus proposed!

TPS Girder System Timeline

7

Date Event

2004 June TPS feasibility study started

2005 Dec. A preliminary study prototype bending section with 3 girders established

2009 Oct. A nearly real size testing mock-up system was set up in NSRRC lab .

2010~2013 Parts manufacturing, sub-systems assembling and calibration in a rental factory

2012 May A third mock-up system was set up

2013 Jan. Pedestals were to be set out, anchor bolts drilling and installed

2013 July 2 bending sections installed as an on-site mock-up testing system

2013 Oct. Entire system installation began

2014 Mar. The Last girder was craned

2014 Aug. A whole ring girder automatic alignment was really performed

2014 Aug. Phase I accelerator system test and commissioning started

2014 Oct. A second automatic alignment was performed

2014.12.31 First synchrotron light from TPS

Cam mover type adjustment mechanism

8

Girder of Swiss Light Source

From “Handbook of Accelerator Physics and Engineering” Alexander Wu Chao, P377.

From “Precision Machine Design” Alexander H. Slocum, P377.

Types of kinematic mounting

A 3 grooves type kinematic mounting modification to 6 stands girder design

9

One girder system configuration

More contact points with locking system to raise natural frequency and reduce deflection. All contact points persist rolling contact condition when adjusting to reduce friction and

remain high mobility . Contact stress less than elastic limitation to reduce friction wear and keep high reliability

Cam mover type mechanism modification and adjusting algorithm

Two coordinate systems are to be established at girders and movers separately, from the rigid body assumption, the adjustment of center position of each ball can be calculated.

Due to the kinematic V groove type arrangement, the girder moving range in horizontal direction is only ½ of the mover range in horizontal(X) direction (4.5mm) and √3/2 in longitudinal(Z) direction (7mm)

10

Modification of point contact type cam to line contact to reduce stress

Kinematic mounting situation preserved. The contact position of the ball and the cam

remains the same for adjusting algorithm. the contact situation changes from point contact to

line contact. the stress is reduced drastically to 12.4% and far

below the elastic limitation of the cam. Steps of the stepping motor is 5000 and plus with the gearbox ratio of 160, it comes with a resolution of 800000 per turn and also refers to a step resolution of at least 0.03um in girder coordinate system

11

Heidenhain ECN425

harmonic drive 1: 160

constant torque : 35 kg.m

Heidenhain

ECN425

rotary encoder

Magnets assembling on girder design

12

• 3 magnet installation referencing channels precisely machined within 15mm tolerance • The magnet mounting base also precisely machined within 15mm tolerance＊ • two side channels can keep the magnets on the same height with minimized rotation • the mounting base will be pushed to closely touch the side of center channels with an

inclined clamper • The clamper produces a horizontal pushing force to make sure the magnet base

contact with the certer reference channel • The right screw for the clamper is locked with a constant 1400 kg.cm torque, while

the left screw 1000 kg.cm (with a shim underneath)

Shim

＊ J. C. Jan, et al., "Magnet Design and Control of Field Quality for TPS Booster and Storage

Ring", IPAC’15.

13

TPS storage ring girder system design

One girder section(1/24) with magnets and vacuum system

One girder section

1/6 ring symmetry super-period configuration

3 kinds and 5 types of girder

Leica Nevil220 tilting sensors

Heidenhain Acanto

AT1218 absolute length gauge

No sensor for length measurement of straight section

Control system Architectures

14

• Each local control system controls 3 girders of a bending section.

• A NI PC-based PXI platform • Consists of 6 sub-systems: 1. cam mover control(18) 2. rotational encoder (18) 3. touch sensor reading(16) 4. tilting sensor reading(3) 5. PSD reading(4) 6. locking control(18)

* H.S. Wang et al, “Design and testing of a girder control system at NSRRC”, ICALEPCS(2011)

With dipole mag. Max.:18um Without dipole mag. Max.:45um

Girder flatness measurement after magnets assembled

15

Natural Frequency and damping testing of the mock-up system at the rental plant

Amplification with damper

The 1st NF can be raised from 24Hz to

33 Hz with locking mechanisms

The Peak amplitude is reduced

with dampers

One girder section was installed with dampers

and locking system for testing

16

Survey data accumulation

Survey network implementation

17

TPS tunnel variation during 2012~2013

With the survey data from the sockets at tunnel walls, a small displacement of (-3,-1) and rotation of 0.0021 degree (about 5mm at lattice position) clockwise of the virtual center were derived and a coordinate values adjustment of all components were decided accordingly.

18

Shielding wall (y dir.)

Shielding wall (x dir.)

Pedestal set out, anchor bolts implanting, alignment ,grouting and cam movers installation

19

Started from Jan. 2013

Transportation of girders with magnets to TPS

20

Lifting and installation girders

21 It spends about 40 min to an hour to install a girder

Girder installation deviations

-3.00

-2.00

-1.00

0.00

1.00

2.00

3.00

mm

position

X

-3.00

-2.00

-1.00

0.00

1.00

2.00

3.00

mm

position

Y

-2.00

-1.00

0.00

1.00

2.00

mm

position

Z

before alignment

after alignment

22

Magnet centralizing

• Adjust Position Jig and Circular PSD jig

• Install two position jigs with PSD on girder

• Adjust laser to parallel and have equidistance to girder datum plane

• Replace the Position jig with the quadrupole and sextupole magnets

• Insert Circular PSD jig on the center of quadrupole and sextupole magnets

• The offset of beam position can be detected by PSD

• Insert the steel shims between magnet and girder for error compensation

23

Magnet Centralizing Results

-80.0

-60.0

-40.0

-20.0

0.0

20.0

40.0

60.0

Magnet Error at Vertical Magnets of TPS

-80.0

-60.0

-40.0

-20.0

0.0

20.0

40.0

60.0

Magnet Error At Horizontal Magnets of TPS

RMS 18.53um

RMS 20.42um

24

Most magnets were acceptable but a few were still shimmed after double checked

Girder system completed installation

In May 2013, 2 mockup sections (R19 and R20) were installed at storage ring.

The full ring installation processes began from Oct. 2013 and the last girder was finished installation at March 2014.

25

Hard stop, Locking system and damper assembling

26

Control rack installation and wiring

27

Control system problems encountered

28

•Motor malfunction → check and changed

•Touch sensor or connector damaged → check and changed

•Touch sensor signal direction inversed → check and adjusted

•Laser optical fiber damaged → check and changed

•Laser holder being moved → check and adjusted

•PSD module holder replaced (conflicted with frontend)

The data base for auto-alignment

• Linear encoder distance measurement data between 2 adjacent girders

• Girder reference point distance measurement data with a laser interferometer

• Laser interferometer distance measurement data compensation chart

• Linear encoder distance measurement data compensation chart

• PSD beam profile curve fitting • PSD1 / PSD2 coefficient

calibration • PSD to linear encoder calibration

29

Girder sensor reading variations with respect to temperature change

1℃ in temperature change induces 0.1mm deviation between girders

30

Laser tracker measurement data

Girder adjusting individually with a laser tracker (a few motor failures encontered)

Girder auto-alignment in 1/3 ring each (1mm drop protection)

-0.50-0.40-0.30-0.20-0.100.000.100.200.300.400.500.600.700.80

SR1

_…

SR1

_…

SR2

_…

SR2

_…

SR3

_…

SR3

_…

SR4

_…

SR4

_…

SR5

_…

SR5

_…

SR6

_…

SR6

_…

SR7

_…

SR7

_…

SR8

_…

SR8

_…

SR9

_…

SR9

_…

SR1

0…

SR1

0…

SR1

1…

SR1

1…

SR1

2…

SR1

2…

SR1

3…

SR1

3…

SR1

4…

SR1

4…

SR1

5…

SR1

5…

SR1

6…

SR1

6…

SR1

7…

SR1

7…

SR1

8…

SR1

8…

SR1

9…

SR1

9…

SR2

0…

SR2

0…

SR2

1…

SR2

1…

SR2

2…

SR2

2…

SR2

3…

SR2

3…

SR2

4…

SR2

4…

:mm

Girder position

TPS SR girder position deviations (2014,7)

Z(Longitudinal)

X(Transverse)

Y(Vertical)

Girder auto-alignment in 1/3 ring each (vertical direction only) 31

Laser tracker measurement data

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

SR1

_G

1-1

SR1

_G

2-3

SR2

_G

1-1

SR2

_G

2-3

SR3

_G

1-1

SR3

_G

2-3

SR4

_G

1-1

SR4

_G

2-3

SR5

_G

1-1

SR5

_G

2-3

SR6

_G

1-1

SR6

_G

2-3

SR7

_G

1-1

SR7

_G

2-3

SR8

_G

1-1

SR8

_G

2-3

SR9

_G

1-1

SR9

_G

2-3

SR1

0_

G1

-1

SR1

0_

G2

-3

SR1

1_

G1

-1

SR1

1_

G2

-3

SR1

2_

G1

-1

SR1

2_

G2

-3

SR1

3_

G1

-1

SR1

3_

G2

-3

SR1

4_

G1

-1

SR1

4_

G2

-3

SR1

5_

G1

-1

SR1

5_

G2

-3

SR1

6_

G1

-1

SR1

6_

G2

-3

SR1

7_

G1

-1

SR1

7_

G2

-3

SR1

8_

G1

-1

SR1

8_

G2

-3

SR1

9_

G1

-1

SR1

9_

G2

-3

SR2

0_

G1

-1

SR2

0_

G2

-3

SR2

1_

G1

-1

SR2

1_

G2

-3

SR2

2_

G1

-1

SR2

2_

G2

-3

SR2

3_

G1

-1

SR2

3_

G2

-3

SR2

4_

G1

-1

SR2

4_

G2

-3

mm

TPS SR girder position deviations (2014,8)

Z(Longitudinal)

X(Transverse)

Y(Vertical)

First girder auto-alignment in full ring

32

-0.80

-0.60

-0.40

-0.20

0.00

0.20

0.40

0.60

0.80

SR1

_G

1-1

SR1

_G

2-3

SR2

_G

1-1

SR2

_G

2-3

SR3

_G

1-1

SR3

_G

2-3

SR4

_G

1-1

SR4

_G

2-3

SR5

_G

1-1

SR5

_G

2-3

SR6

_G

1-1

SR6

_G

2-3

SR7

_G

1-1

SR7

_G

2-3

SR8

_G

1-1

SR8

_G

2-3

SR9

_G

1-1

SR9

_G

2-3

SR1

0_

G1

-1

SR1

0_

G2

-3

SR1

1_

G1

-1

SR1

1_

G2

-3

SR1

2_

G1

-1

SR1

2_

G2

-3

SR1

3_

G1

-1

SR1

3_

G2

-3

SR1

4_

G1

-1

SR1

4_

G2

-3

SR1

5_

G1

-1

SR1

5_

G2

-3

SR1

6_

G1

-1

SR1

6_

G2

-3

SR1

7_

G1

-1

SR1

7_

G2

-3

SR1

8_

G1

-1

SR1

8_

G2

-3

SR1

9_

G1

-1

SR1

9_

G2

-3

SR2

0_

G1

-1

SR2

0_

G2

-3

SR2

1_

G1

-1

SR2

1_

G2

-3

SR2

2_

G1

-1

SR2

2_

G2

-3

SR2

3_

G1

-1

SR2

3_

G2

-3

SR2

4_

G1

-1

SR2

4_

G2

-3

Un

it :

mm

TPS SR girder Position Deviations(2014.10.27)

Z(Long.)

X(Hori.)

Y(Vert.)

Second girder auto-alignment in full ring

Auto alignment processes

33

1. Load laser tracker measurement data (XL,YL) 2. Load touch sensor matrix and calibration coeff. , PSD calibration coeff. ,

girder length… initial data 3. Read touch sensor, PSD, Nivell, Encoder initial data 4. Calculate local dz(longitudinal deviation), dx(transverse deviation) Touch sensor matrix and PSD reading to calculate local (dzs , dxs) Local (dzs , dxs) transfer to global (dXs,dYs) (dXs,dYs) of each girder Calculate new position by average with the laser tracker data

• P2(X_new,Y_new) = ½ *[ P1(XL,YL) + ½ *(dXs1,dYs1) +1/2*(P2(XL,YL)- P1(XL,YL))+ P3(XL,YL) – ½ *(dXs2, dYs2) - 1/2*( P3(XL,YL) - P2(XL,YL) )]

• Calculate all P1(X_new, Y_new)~P72(X_new, Y_new) • Compare with (P1_theory~P72_theory) to calculate adjusting values • Adjust girder 1 of each section and record touch sensor 1st readings • Adjust girder 2 of each section and record touch sensor 2nd readings • Adjust girder 3 of each section and record touch sensor 3rd readings • Accord to the 3 touch sensor readings to calculate the real

movement of each girder • Renew each girder postions P(X’,Y’) = P(X_new,Y_new)+movement

P(dX,dY) • Iteration till converge

-0.800

-0.600

-0.400

-0.200

0.000

0.200

0.400

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71

d_move_x d_move_y d_move_z

SR girder auto alignment first adjustment values

Second adjustment values

Mistakes

To be check

Auto alignment calculated values

The full ring adjustment takes about 1900 seconds (32min) in 5 iteration

longitudinal(mm)

transverse(mm)

34

Girder positions comparison between from sensors and laser tracker

35

0 50 100 150-40

-20

0

20

40

電子

束方

向lo

cal

0 50 100 150-30

-20

-10

0

10

20

垂直

電子

束方

向lo

cal

Longitudinal

Transverse

Due to sensor initial conditions are corrupted, two steps are to be adopted 1. Applied optimizing method to calculate 2. Re-calibration

Booster ring alignment

3 teams with 3 laser trackers and 4 weeks to align the booster ring manually

36

TPS COD before correction

37

1.78 mm rms horizontal 1.04 mm rms vertical

Measured COD with all correctors off After LOCO and BBA

Simulated COD with all correctors off from alignment errors, dipole field errors

＊ C. C. Kuo, et al., "Commissioning of the Taiwan Photon Source", IPAC’15 .

The measured data were even better shows good alignment conditions

The deviation between girders from a Richter scale 6.6 earthquake

Girder

Section

Max. Deviation Between

Earthquake

Sensor ID

Max. Deviation

After Earthquake

Sensor ID

R1 0.0033 16 0.0015 3

R2 0.0062 16 0.0035 16

R3 0.0025 16 0.0013 7

R4 0.0025 7 0.0013 11(Y)

R5 0.0017 3 0.0016 5(Y)

R6 0.004 3 0.0028 3

R7 0.0052 8 9.00E-04 5(Y)

R8 0.0027 16 0.0015 16

R9 none none none None

R10 0.0054 12 0.0034 16

R11 0.0043 3 0.0034 3

R12 0.0017 16 0.0011 16

R13 none none none none

R14 0.0083 3 0.0026 3

R15 0.0029 16 0.0018 16

R16 0.0018 12 9.00E-04 12

R17 0.0043 16 0.0015 16

R18 0.0016 5(Y) 0.0015 5(Y)

R19 0.0017 12 0.0011 3

R20 0.0078 7 0.0018 8

R21 0.0023 7 0.0019 11(Y)

R22 0.0048 16 0.0039 16

R23 0.0016 8 1.00E-03 8

R24 0.0035 3 0.002 3

38 However, the intensity in Hsinchu is only 2~3

Laser tracker measurement data(2015-2016)

-1.00

-0.50

0.00

0.50

1.00SR

1_

G1

-1

SR1

_G

3-1

SR2

_G

2-1

SR3

_G

1-1

SR3

_G

3-1

SR4

_G

2-1

SR5

_G

1-1

SR5

_G

3-1

SR6

_G

2-1

SR7

_G

1-1

SR7

_G

3-1

SR8

_G

2-1

SR9

_G

1-1

SR9

_G

3-1

SR1

0_

G2

-1

SR1

1_

G1

-1

SR1

1_

G3

-1

SR1

2_

G2

-1

SR1

3_

G1

-1

SR1

3_

G3

-1

SR1

4_

G2

-1

SR1

5_

G1

-1

SR1

5_

G3

-1

SR1

6_

G2

-1

SR1

7_

G1

-1

SR1

7_

G3

-1

SR1

8_

G2

-1

SR1

9_

G1

-1

SR1

9_

G3

-1

SR2

0_

G2

-1

SR2

1_

G1

-1

SR2

1_

G3

-1

SR2

2_

G2

-1

SR2

3_

G1

-1

SR2

3_

G3

-1

SR2

4_

G2

-1

Un

it :

mm

TPS SR Girder Position Deviations (X Transverse)

ALL2015-06-08_USMN(RMS = 0.22 mm)

SR2015-07-13_USMN(RMS = 0.27 mm)

SR2016-01-04_USMN(RMS = 0.19 mm)

SR2016-07-06_USMN(RMS = 0.25 mm)

-1.00

-0.50

0.00

0.50

1.00

SR1

_G

1-1

SR1

_G

3-1

SR2

_G

2-1

SR3

_G

1-1

SR3

_G

3-1

SR4

_G

2-1

SR5

_G

1-1

SR5

_G

3-1

SR6

_G

2-1

SR7

_G

1-1

SR7

_G

3-1

SR8

_G

2-1

SR9

_G

1-1

SR9

_G

3-1

SR1

0_

G2

-1

SR1

1_

G1

-1

SR1

1_

G3

-1

SR1

2_

G2

-1

SR1

3_

G1

-1

SR1

3_

G3

-1

SR1

4_

G2

-1

SR1

5_

G1

-1

SR1

5_

G3

-1

SR1

6_

G2

-1

SR1

7_

G1

-1

SR1

7_

G3

-1

SR1

8_

G2

-1

SR1

9_

G1

-1

SR1

9_

G3

-1

SR2

0_

G2

-1

SR2

1_

G1

-1

SR2

1_

G3

-1

SR2

2_

G2

-1

SR2

3_

G1

-1

SR2

3_

G3

-1

SR2

4_

G2

-1

Un

it :

mm

TPS SR Girder Position Variations (Y Vertical)

ALL2015-06-08_USMN(RMS = 0.32 mm)

SR2015-07-13_USMN(RMS = 0.33 mm)

SR2016-01-04_USMN(RMS = 0.31 mm)

SR2016-07-06_USMN(RMS = 0.23 mm)

儲存環磁鐵Girder定位與設計差值 Z,X,Y=±0.50 mm -1.00

-0.50

0.00

0.50

1.00

SR1

_G1

-1

SR1

_G3

-1

SR2

_G2

-1

SR3

_G1

-1

SR3

_G3

-1

SR4

_G2

-1

SR5

_G1

-1

SR5

_G3

-1

SR6

_G2

-1

SR7

_G1

-1

SR7

_G3

-1

SR8

_G2

-1

SR9

_G1

-1

SR9

_G3

-1

SR1

0_G

2-1

SR1

1_G

1-1

SR1

1_G

3-1

SR1

2_G

2-1

SR1

3_G

1-1

SR1

3_G

3-1

SR1

4_G

2-1

SR1

5_G

1-1

SR1

5_G

3-1

SR1

6_G

2-1

SR1

7_G

1-1

SR1

7_G

3-1

SR1

8_G

2-1

SR1

9_G

1-1

SR1

9_G

3-1

SR2

0_G

2-1

SR2

1_G

1-1

SR2

1_G

3-1

SR2

2_G

2-1

SR2

3_G

1-1

SR2

3_G

3-1

SR2

4_G

2-1

Un

it :

mm

TPS SR Girder Position Deviations (Z Longtudinal)

ALL2015-06-08_USMN(RMS = 0.26 mm)

SR2015-07-13_USMN(RMS = 0.28 mm)

SR2016-01-04_USMN(RMS = 0.24 mm)

SR2016-07-06_USMN(RMS = 0.27 mm)

Ground Variation Due to Temperature and Tide

40

10M quartz glass bar on 4 girder section

Commissioning team found there is about 90~100um circumference variation correlated to daily environment temperature change and tide

Conclusions

• TPS girder system including storage ring, booster ring and transport line were finished installation August 2014.

• A laser tracker survey data based full ring auto-alignment had been performed and shows good conditions for smoothly commissioning and the stability is still good during 2 years operation.

• The laser tracker survey results show that the full ring accuracy is about ± 0.5mm.

• The sensor’s durability is quite a problem for the auto-alignment system, however, the quartz glass bar test shows a feasible solution toward the sensor based girder auto-alignment.

41

42

Thank you for your attention!