measurements. ENRIS2019, 16-18 June 2019 1 Acknowledgements - - PDF document

▶
measurements
SMART_READER_LITE
LIVE PREVIEW

measurements. ENRIS2019, 16-18 June 2019 1 Acknowledgements - - PDF document

Nov 08, 2022 254 likes •740 views

6/11/2019 Setting up a very low noise continuous vibration monitoring system and evaluating the measurements. ENRIS2019, 16-18 June 2019 1 Acknowledgements Dorien van der AA, project manager 2015-2018 H.L. Offerhaus, chairholder E.H.

slide-1
SLIDE 1

6/11/2019 1

Setting up a very low noise

continuous vibration monitoring

system and evaluating the

measurements.

ENRIS2019, 16-18 June 2019 1

Acknowledgements

ENRIS2019, 16-18 June 2019 2

E.H. Ruiter, maintenance G.J Roelofs, head MESA+ Nanolab H.L. Offerhaus, chairholder Dorien van der AA, project manager 2015-2018

slide-2
SLIDE 2

6/11/2019 2

Motivation

An increasing number of instruments require a “quiet” floor

  • Real time web page (“can I blame it on the floor”).
  • Checking equipment susceptibility for vibrations.
  • Identifying noise sources (construction work, traffic,

machines, people,….

  • Check if building contractors works “quiet”.
  • Check equipment for anomalies (worn bearings).

ENRIS2019, 16-18 June 2019 3 ENRIS2019, 16-18 June 2019 4

Offices Cleanroom, VC-D Labs, VC-D Facilities & Technical Area VC-G Labs

Nanolab Overview

Clean- Room VC-G

slide-3
SLIDE 3

6/11/2019 3

ENRIS2019, 16-18 June 2019 5

Nanolab cross section

Plenum VC-G labs Plenum Cleanroom VC-G labs Cleanroom Basement crawlspace Technical area

ENRIS2019, 16-18 June 2019 6

Nanolab construction, schematic drawing

slide-4
SLIDE 4

6/11/2019 4

Hardware

(commercially available)

ENRIS2019, 16-18 June 2019 7

Measurement setup

ENRIS2019, 16-18 June 2019 8

Sensor CME-6211 3-axial, 60s–50Hz Data-acq. NI USB-6341 (BNC) Power supply + filtering Web server, control, (HP elitebook) Optional: Second sensor for:  Redundancy,  Excluding locally induced vibrations,  Noise source Identification,  Validating the sensors by “coherent transfers”.

http://r-sensors.ru/

slide-5
SLIDE 5

6/11/2019 5

Location sensors

ENRIS2019, 16-18 June 2019 9

Molecular Electronic Transducer

https://en.wikipedia.org/wiki/Molecular_electronic_transducers

ENRIS2019, 16-18 June 2019 10

Anode: 3 I− → I3

− + 2 e−

Cathode: I3

− + 2 e− → 3 I−

slide-6
SLIDE 6

6/11/2019 6

CME-6211 noise spectral density

ENRIS2019, 16-18 June 2019 11

seconds Acceleration Power Spectral Density [10Log(m2/sec4)/Hz]

Lowest earth background noise, NLNM Our sensor Lowest noise geophones

Units can be hard to interpret

Software

(NOT commercially available)

ENRIS2019, 16-18 June 2019 12

  • Measurement, real time
  • Post processing
slide-7
SLIDE 7

6/11/2019 7

What timeframe to measure?

Ability to detect peak disturbances Ability to cover the VC-range (1-80 Hz)

ENRIS2019, 16-18 June 2019 13

1 Hz is minimum

Vibration Criterion (VC) is velocity power spectral density (rms) integrated over one third octave bands from 1 Hz to 80 Hz. The maximum value in the VC-curve defines the VC class. For more information see:

Highest grade floor = VC-G, v < 0.78 μm/s, next: VC-F, 1.6 μm/s, VC-E, 3.2 μm/s, ….., until VC-A.

Shorter is better

Vibration transfers to measure the performance of vibration isolated platforms on site using background noise excitation, F. B. Segerink, J. P. Korterik, and H. L. Offerhaus, REVIEW OF SCIENTIFIC INSTRUMENTS 82, 065111 (2011).

What to show on the real time screen?

Narrowing down the data-amount!

  • Peak disturbances are most important.
  • Each timeframe ( 1 s) is converted to 1 number called

“trillingsgetal”, defined as the peak VC value in all directions relative to the VC-G level.

  • Our reference VC-G level will be depicted by a green line.

ENRIS2019, 16-18 June 2019 14

slide-8
SLIDE 8

6/11/2019 8

ENRIS2019, 16-18 June 2019 15

Measurement web page

http://utwks61920.roaming.utwente.nl:8000/CMEmain02.html

ENRIS2019, 16-18 June 2019 16

High pass filtering, Barkan formula

slide-9
SLIDE 9

6/11/2019 9

ENRIS2019, 16-18 June 2019 17

Post processing software

Data is only valuable when processed!

  • 1. Time slices within one file (~1 hour)
  • 2. VC terts bands vs time over long time (~1 month)

Evaluating measurements

What did we learn so far (4 years)?

ENRIS2019, 16-18 June 2019 18

slide-10
SLIDE 10

6/11/2019 10

Traffic and height differences

ENRIS2019, 16-18 June 2019 19

cantine

Traffic and height differences

ENRIS2019, 16-18 June 2019 20

“trillingsgetal” = 9.2 Conclusion: Height differences add very significant to floor vibrations!

slide-11
SLIDE 11

6/11/2019 11

ENRIS2019, 16-18 June 2019 21

Traffic and height differences

problem solved!

Separated floors

A well known concept, hardly ever quantified…

ENRIS2019, 16-18 June 2019 22

Sensor 1

Isolated floor

Sensor 2

Noisy floor

Coherent transfer measurement setup. Excitation: background noise. Floor separation Coherent transfer

slide-12
SLIDE 12

6/11/2019 12

Post processing software

VC-terts bands over long time

26 days of measurements, VC-terts-band versus time, starting Friday February 15, 2019. Time trace interval length = 32768 points (rms), Hamming window. Many peaks Offset, sensor noise.

North- South East- West Vertical, Z

slide-13
SLIDE 13

6/11/2019 13

Increasing the frequency to the next terts band Lower offset, sensor noise. Earthquake Equador Mw 7.5

slide-14
SLIDE 14

6/11/2019 14

Earthquakes show at very low frequencies No more offset, some background signal comes up. Background increases, origin still unknown. Not wind-speed related.

slide-15
SLIDE 15

6/11/2019 15

Background, typically between 0.1 and 0.3 Hz. Background at maximum.

slide-16
SLIDE 16

6/11/2019 16

Background gradually decreasing Background gradually decreasing

slide-17
SLIDE 17

6/11/2019 17

Both background and earthquake gradually disappear at higher frequencies. Just below 1 Hz there is not much happening.

slide-18
SLIDE 18

6/11/2019 18

Still not much happening. Still very quiet

slide-19
SLIDE 19

6/11/2019 19

Approaching 1 Hz, a daily pattern starts to appear. Weekends and working days show up

slide-20
SLIDE 20

6/11/2019 20

Weekends and working days show mostly in the x-y direction Weekends and working days, z-direction is joining

slide-21
SLIDE 21

6/11/2019 21

Weekends and working days, z-direction dominates Weekends and working days impact at maximum

slide-22
SLIDE 22

6/11/2019 22

Weekends and working days gradually decreasing Weekends and working days decreasing more.

slide-23
SLIDE 23

6/11/2019 23

Weekends and working days gradually decreasing further. The z-direction shows extra signal, related to the air handling system.

slide-24
SLIDE 24

6/11/2019 24

Air conditioning and distribution, dominant in z-direction, increases. Air conditioning and distribution, dominant in z-direction around 10 Hz

slide-25
SLIDE 25

6/11/2019 25

Air conditioning and distribution, rapidly decreasing above 10 Hz Local construction work

slide-26
SLIDE 26

6/11/2019 26

Local construction work No high amplitudes at higher frequencies

slide-27
SLIDE 27

6/11/2019 27

No high amplitudes at higher frequencies No high amplitudes at higher frequencies

slide-28
SLIDE 28

6/11/2019 28

Upper limit detector = 50 Hz Zooming in on the Equador earthquake, Mw 7.5, 9783 km from the Nanolab.

slide-29
SLIDE 29

6/11/2019 29

Zooming in on the earthquake. Adjusting the z-scale

slide-30
SLIDE 30

6/11/2019 30

Zooming in more. And more

slide-31
SLIDE 31

6/11/2019 31

43 minutes Zooming in on a typical working day, Tuesday February 26, 2019.

slide-32
SLIDE 32

6/11/2019 32

Tuesday February 26, 2019, 00:00h – 24:00 h. Well below VC-G Well below VC-G, restricted access until 22:00h. Mainly human induced vibrations Curiosity: Tuesday February 26, 2019, peaks at very low frequency, around 09:00h, Predominantly in the East-West direction, not in the z-direction.

slide-33
SLIDE 33

6/11/2019 33

Zooming in on the peaks at very low frequency. 6 min. From observation it showed that the truck for cantine supplies parks near the sensor for unloading. Lab tilted Lab tilted back

Post processing software

Time slices

slide-34
SLIDE 34

6/11/2019 34

Time signal post processing, 1 hour 9 minutes data file with earthquake event. 60 seconds Pre shocks suddenly appear.

slide-35
SLIDE 35

6/11/2019 35

Further following the time line… The big one! 150 s

slide-36
SLIDE 36

6/11/2019 36

End of post processing software demo “noticeable events”

(last part of the talk)

ENRIS2019, 16-18 June 2019 72

Noise increase triggers maintenance

Sunday June 25, 2017, 10 Hz terts band Monday Tuesday 12.7 Hz terts band Start excess noise Wednesday

slide-37
SLIDE 37

6/11/2019 37

ENRIS2019, 16-18 June 2019 73

Negotiating construction work

VC-G

ENRIS2019, 16-18 June 2019 74

Asphalt milling at ~1 km, not negotiable…

VC-G

slide-38
SLIDE 38

6/11/2019 38

ENRIS2019, 16-18 June 2019 75

Storm batters Nanolab and Carré

S1 S2

ENRIS2019, 16-18 June 2019 76

Storm batters Nanolab and Carré

slide-39
SLIDE 39

6/11/2019 39

ENRIS2019, 16-18 June 2019 77

King’s-Day

400 m

podium FC Twente

ENRIS2019, 16-18 June 2019 78

King’s-Day

VC-G

slide-40
SLIDE 40

6/11/2019 40

ENRIS2019, 16-18 June 2019 79

GOALS in the Golsch Veste

VC-G

Wrapping up: Impact of vibrational disturbances

ENRIS2019, 16-18 June 2019 80

Generic groups I. Damage to equipment and sample

tip crash in UHV vacuum STM (2-3 weeks time loss)

  • II. Damage to sample only

e-beam write errors, TEM imaging, FIB milling, Heidelberg mask writer error

  • III. Failed experiments

most imaging applications (SEM, TEM, AFM), pulsed laser deposition

  • IV. Not susceptable for vibrations

most of the equipment has low floor vibration demands, cleaning, sputtering, etching, spinning,…..

slide-41
SLIDE 41

6/11/2019 41

Thank you for your attention.

ENRIS2019, 16-18 June 2019 81

Extra slides

ENRIS2019, 16-18 June 2019 82

slide-42
SLIDE 42

6/11/2019 42

ENRIS2019, 16-18 June 2019 83

Carré air handling machines

Top floor (5th) Sensor position

ENRIS2019, 16-18 June 2019 84

Carré 5th floor and basement

slide-43
SLIDE 43

6/11/2019 43

ENRIS2019, 16-18 June 2019 85

Carré 5th floor and basement, zoom-in Earthquakes, background.

  • Richter magnitude ML = 10log[A/A0(δ),

A is the maximum excursion of the Wood-Anderson seismograph A0(δ) is an empirical function of a reference excursion depending on distance 1 micron at 100 km distance = mag. zero Accurate between mag 3 and mag 7 (saturates)

  • Moment magnitude scale (Mw)

from 1970’s, based on seismic moment, numbers compare approximately to the Richter scale The press always (still) uses the name “Richter scale”.

  • Energy scales with 3/2 power of shaking amplitude, so 31.6 times for a difference of mag 1

(0.2 mag difference means 2 times the energy)

  • Impact varies with depth and geological structures,

hard to compare with “equivalents TNT”, which usually detonates above ground.

ENRIS2019, 16-18 June 2019 86

slide-44
SLIDE 44

6/11/2019 44

Sensor types, solid state mass

ENRIS2019, 16-18 June 2019 87

CME-6211 photo’s

ENRIS2019, 16-18 June 2019 88

slide-45
SLIDE 45

6/11/2019 45

Current-voltage diagram of MET sensor

ENRIS2019, 16-18 June 2019 89

Behaves like pentode

  • r FET, where the

liquid movement transports charge to the electrodes and thus corresponds with the grid or gate voltage.

ZERO G

ENRIS2019, 16-18 June 2019 90

CME-6211, a great spy tool

slide-46
SLIDE 46

6/11/2019 46

Zooming in on Tuesday February 26, 2019, a typical working day. Tuesday February 26, 2019, a typical working day.

slide-47
SLIDE 47

6/11/2019 47

Hamming window Applying a few FFT windows Hanning window Applying a few FFT windows

slide-48
SLIDE 48

6/11/2019 48

Rectangle window Applying a few FFT windows