Properties of seismic noise at Virgo

Irene FioriINFN Firenze

for the Virgo Collaboration

5th Amaldi Conference on Gravitational Waves Tirrenia (Italy) 7-11 July 2003

Seismic Monitoring Network

Low Frequency Accelerometers Tri-axial, 0.2-200 Hz(Kinemetrics - “Episensors”)

High Frequency Accelerometers Vertical Piezo, 10-10000 Hz(Bruel&Kjaer)

3km

3km

Virgo Central building

MC End b.

West End b.

North End b.

Also:• Microphones• Magnetometers• Pressure probes• Humidity • Temperature

.

..

ADC

ADC

MAIN FRAME

BUILDER

MAIN FRAME

BUILDER

ETHERNET LINK

C1 C2 DISK…

Anti-aliasingfilters

Other chs.

Seism General Properties

Seism amplitudes (m)• 10-5 @ 0.1 Hz • 10-7 @ 1.0 Hz• 10-9 @ 10. Hz

Displacements (m/√Hz)

10-7

f 2(m/√Hz)d ≈

~

Acceleration spectral amplitudesTri-axial accelerometer

• 0-200 Hz

• On concrete floor of Virgo Central B.

x-y-z directions

Approx. flat acc. spectrum in 3D

DAY vs. NIGHT seism accelerations

• Human & Naturalseismic sourcesbelow ≈ 10 Hz

Daily Seismic Variations

NIGHT vs. WINDY NIGHT• Wind & Sea Wavesseismic sourcesbelow ≈ 1 Hz

0.4Hz0.8Hz

Monitor of RMS Seismic Motion

in 3 frequency bands: 0.2-1 Hz ; 1-4 Hz ; 4-10 Hz

• 1 - 4 Hz: day-night-weekend variations à local traffic ? • 0.2 - 1 Hz: slow motion à correlated with wind & sea-waves ?

• 4 - 10 Hz: spikes between 10AM-6PM à human activities on-site.

mic

rons

/s2

RMS = ( |a(f)|2df )½∫ ~

f1

f2

Time (DAYS) since July 3rd 2002 00:00 (local)

On-line seismic monitor

• 1 week data at 10 minutes steps

• Continuously running on 50Hz re-sampled data

http://wwwcascina.virgo.infn.it/commissioning/Monitoring_50Hz

Earthquake shakes

31 Oct. 2002, Magn. 5.9, 540 km SE from Virgo 2 Nov. 2002, Magn. 4.2, 270 km E from Virgo

Two shakes recorded at Virgo site :

3

SEISMIC ARRAY ANALYSIS

5 tri-axial velocimeters(Lennartz 3D/5 sec)

Ø Coherent spectral peak at ≈ 0.3 Hzà signals are band-pass filtered (0.1÷0.5 Hz) à

Up-Down

N-S

E-W

Investigating the 0.2-1 Hz seism: Oceanic MicroSeism ?

SEA 15 km

Oceanic MicroSeism at Virgo

Ø Consistent with OMS typical properties:• Peak freq. is double the sea waves freq. [R.A.Haubric et al.] • Peak freq. changes slightly with weather conditions [T.Braun et al.]

≈ East àWest

• Plane wave-front• peak frequency ≈ 0.3 Hz• speed 750-800 m/s

0 200 400 600 Time (seconds)

LAG

(sec

onds

)

+6

+4

+2

0

-2

-4

-6

Vi1Vi2

Vi4Vi3

Vi5Travel time between each pair of stations :

Investigating the 1-4 Hz seism:Local Traffic ?

ü “bursts” at ≈ 3 Hz

ü approx. 20 s long

ü ≈ 1 / minute during daytime

ü Minimum at 2AM (local)

ü Very reduced during week-ends

Fre

quen

cy [

Hz]

Hypothesis:Trucks exciting oscillations of nearby bridges

Road bridges around VirgoØ 1-3 km far from Virgo arms ends

Bridges Oscillations 1/3

Ø Resonant frequencies: 2, 6, 11,… Hz

Ground vibrations under bridge:

Road bridges around VirgoØ 1-3 km far from Virgo arms ends

Bridges Oscillations 2/3

Correlation measurement(preliminary):

x-correlation

@Virgo

@bridgeFiltered 1÷ 4 Hz

Road bridges around VirgoØ 1-3 km far from Virgo arms ends

Bridges Oscillations 3/3

A(s,ω) = A0 exp(- )ω s2vQ

Ø A(6Hz)/A(2Hz) < 0.1 (@1km)

à Only 2 Hz component survives at 1 kmand seen by Virgo seismometers

Ø A(3km)/A(1km) < 0.1 (@2Hz)

à “bursts” are not coherent among seismic stations at Virgo arms ends

Sedimentary soil at Virgo :

Q = 2-10 , v ≈ 1 km/s

Which seismic sources are relevant for Virgo?

Seismic isolation at Virgo

8m

Seism

GROUND

LASER BEAMMirror

Inverted Pendulum

ØActively controlled todamp horizontal SAresonances below 2. Hz( “Inertial damping” )

• Multiple mechanical pendulum= I.P. + 5 filters + marionetta + mirror

• Inverted Pendulum

Super Attenuator: (SA)

= Ultra-low frequency oscillator

Ground seism attenuation: ≥ 1013 @ 4 Hz

Low pass filter in all d.o.f.

Seismic isolation at Virgo

8m

Seism

GROUND

LASER BEAMMirror

Inverted Pendulum

GOAL: seismic noise below thermal noise above ≈2 Hz

“Seismic wall”

Crossing point

• Seismic noise accounted for all of the DF noise below 1.6 Hz (à mainly Oceanic MicroSeism)

• Entered exciting IP Vertical resonances :à improvements for final Virgo from implementing IP vertical control

2

SA vertical resonances

Seismic noise to Dark FringeCITF commissioning, eng. Run E3

Dark FringeWESNVerticalTOTAL

Two recent results: a good one, … and a not-so-good one

Other seismic sourcesCoherence DF vs. seism (CITF - Run E4)

Frequency (Hz)

Ø Coupling mechanism still under investigationü NOT through SA towers

Importance of seism coherence analyses !

43.7 Hz 46.89 Hz

ONOFFDAQ ROOM AIR-COND SYS

Ground vibrations due tomotor devicesin Virgo CB

10÷100 Hz

SA performance during an earthquake shake

June 30th 2003 - Magnitude 4.1 Richter – 20 km West of Tirrenia

0.03

0.015

0

0.015

-0.03

acce

lera

tions

(m/s

2)

2 4 6 8 10

TIME(minutes)

X d

ispl

acem

ent (

mic

rons

)

Y displacement (microns)

(mic

rons

)

(mic

rons

)

TIME TIME

• ID control was active on NI tower

• ID control was active on NI tower

• Tower top stage displacement ≈ 400 µm peak-to-peak

• Tower top stage displacement ≈ 400 µm peak-to-peak

Ø Control was not lost !! Ø Control was not lost !!

Ø Usual RMS (≈ 0.3 µm) recovered in 3-4 minutes

Ø Usual RMS (≈ 0.3 µm) recovered in 3-4 minutes

SA performance during an earthquake shake

Summary

• Main seismic sources below 10 Hz: - oceanic micro seism, peaking at 0.3-0.4 Hz- local traffic within 1- 4 Hz

• Ground motion at Virgo site agrees with empirical law:A / f 2 with A = 10 -7 m/√Hz

• Seism daily variations due to human&natural sourcesoccur mainly below 10 Hz

• At Virgo CITF- seismic noise contributed to Dark Fringe noise below ≈1.5 Hz- mainly due to vertical seism,

that was not actively damped at low frequencies• Other (potentially dangerous) sources: motor devices

(AC, pumps, rack fans …)