Download
data conditioning and veto for tama burst analysis n.
Skip this Video
Loading SlideShow in 5 Seconds..
Data conditioning and veto for TAMA burst analysis PowerPoint Presentation
Download Presentation
Data conditioning and veto for TAMA burst analysis

Data conditioning and veto for TAMA burst analysis

139 Views Download Presentation
Download Presentation

Data conditioning and veto for TAMA burst analysis

- - - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript

  1. Data conditioning and veto for TAMA burst analysis Masaki Ando and Koji Ishidoshiro (Department of Physics, University of Tokyo) and the TAMA Collaboration GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  2. Main output Monitor signals Data conditioning (Whitening, freq.-band selection, Line removal) Systematic survey of all monitor channels Burst-event extraction (Burst filter: excess-power filter) Fake reduction even without understandings of noise mechanisms Coincidence analysis Yes No GW candidates Fake events Introduction (1) - Veto analysis - • Veto analysis by monitor signals GW detector is sensitive to external disturbances as well as real GWs Reject fake events using monitor signals recorded together with the main output of the detector GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  3. Introduction (2) - TAMA DT9 data - • TAMA data acquisition and analysis system Used data in this work: TAMA DT9 (Dec. 2003 – Jan. 2004) 200 hours of data (HDAQ 8ch, MDAQ 64ch) GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  4. Data Conditioning GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  5. Real data from detectors Non stationary behavior Drift of noise level Sudden excitations Frequency dependence Sensitivity degradation in high and low frequenciesLine noises DT6 DT8 DT9 Data conditioning Data conditioning (1) - Quality of data from detectors - • Data conditioning ‘ideal’ data (predictable behavior known statistics) Stationary noise White noise GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  6. Data conditioning (2) - TAMA conditioning filter - • Data conditioning for TAMA burst analyses Normalization of the data by averaged nose level  Remove time and frequency dependence Line removal Select frequency band to be analyzed In addition … Calibration : Convert v(t) to h(t) Resampling : Data compression Requirements Small loss in signal power Keep GW waveform GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  7. Data conditioning (3) - Data flow - • Data conditioning by FFT-IFFT Frequency selection, Whitening/Calibration, Resampling Frequency-band selection Overwrite ‘0’ for unnecessary frequencies Outside of the observation band AC line freq. , Violin-mode freq. Calibration peak Whitening Normalize by avg. spec. Calibration Convert to h(f) Fourier transform 72sec data Inverse Fourier transform Only below 5kHz Raw time- series data 20kHz samp. Whitened time- series data 5kHz samp. Calibrated time- series data 5kHz samp. GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  8. Data conditioning (4)- Results of conditioning - • Whitened data Time-series data  Waveform is maintained with resampling Without resamp. 5kHz resamp. Spectrum  Whitened Zoom GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  9. Data conditioning (5) - Calibrated spectrum - • Calibrated data Power spectrum of conditioned time-series data and calibrated spectrum of original data Blue: FFT of raw data  calib. Red: Conditioned data GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  10. Event extraction GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  11. Free selection of time-frequency window Previous works… Spectrogram  Averaged power Time-frequency resolution were limited by FFT length Event extraction (1) - Excess power filter - • Burst event trigger Excess-power filter Evaluate power in given time-frequency windows  Extract non-stationary events Data conditioning time-series data in a given frequency band Calculate power in a given time window Power in a given time-frequency window  SNR Threshold, Clustering  Burst events GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  12. Veto analysis GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  13. Veto analysis (1) - Concept - • Coincidence analysis External disturbances tend to appear in some monitor signals  Reject fake events Burst events Main output (SNR) Threshold Coincidence  Reject GW candidate Monitor (SNR) Threshold Time * Simulated data GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  14. Optimized to have… High fake-rejection efficiency Low accidental coincidence Threshold for monitor signals: Accidental ~ 0.1% Coincidence Time-frequency window: Highest coincidence rate selected from 50 (18) time-frequency combinations Coincidence ratio Accidental coincidence Playground data (~10%) are used for this optimization 0.1% Accidental coincidence: estimated by 1-min.time-shifted data SNR Threshold (monitor signal) Veto analysis (2) - Parameter optimization - • Parameter optimization Analysis parameters Time-frequency window Threshold Signals to be used for veto GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  15. Classify by the coincidence rate… <0.5 %  Do not use for veto 0.5 - 2 %  Use for veto > 2%  Intensive veto with lower threshold: accidental ~ 0.5% Selected signals MDAQ: 9 signals l- Err., l- FB, l+ Err., l+ FB, Bright Pow, Rec. Pow., EW Trans. Pow., SEIS center Z, Magnetic Field Y HDAQ: 3 signals Rec. Pow., L+ FB, Dark Pow. Veto analysis (3) - Signal selection - • Signal selection Even with small accidental rejection prob. (~0.1%) for each, many (~100)monitor signals may lose large amount of data Select signals to be used for veto Intensive fake-rejection with some monitor signals with strong correlations with the main signal Re-optimize the threshold GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  16. Veto analysis (4) - Intensity monitor - • Example of coincident events Intensity monitor in the power recycling cavity (time series data after data conditioning) Main output Monitor signal GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  17. Veto analysis (5) - Seismic motion - • Example of coincident events Seismic monitor – center room vertical motion (time series data after data conditioning) Main output Monitor signal GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  18. Veto analysis (6) - Magnetic field - • Example of coincident events Magnetic field – center room perpend. direction (time series data after data conditioning) Main output Monitor signal GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  19. Veto analysis (7) - Veto results - Veto results with 178 hours of data • Survival rate MDAQ 9 signals Accidental : 4.2% HDAQ 3 signals Accidental : 0.8% Total 12 signals Survival rate Accidental : 4.4% SNR Threshold GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  20. Summary GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  21. Summary • Summary Systematic survey of monitor signals Data conditioning filter Excess power burst filter Coincidence analysis between main and monitor signals Optimization of analysis parameters Analysis with TAMA DT9 data 200 hours data (10% are used as playground data) Correlations were found in 12 monitor signals (Intensity monitor, Seismic motion, Magnetic field) Reject 92% (SNR>10), 98% (SNR>100) fakes with 4.2% accidental rejection probability • Current tasks Hardware signal-injection test  Confirm the safety of veto (already done for some of the monitor signals) GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  22. End GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  23. Burst-wave analysis (3) - Data conditioning - • Data conditioning Line removal AC line, Violin mode peak, Calibration peak Without Line Removal With Line Removal Method FFT 72sec data Reject line freq. components Inverse FFT Normalization Track the drift of noise level Each spectrum is normalized by averaged noise spectrum 30min (5.6x10-4 Hz) Use 30min-averaged spectrum GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  24. Time series data (after conditioning) GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  25. TAMA300データ解析 (2) - パラメータの選択と解析結果 - • パワーフィルタでの解析結果 TAMA DT9, Run#140 (200時間分) を解析 以前のパラメータ Dt = 12.5 [msec] Df = 2260 [Hz] 時間幅 : 5通り (1.6msec – 25.6msec) 周波数帯 : 5通り (500Hz幅x3, 1kHz幅x2) SNRが最大のものを選び, 閾値と比較 青: veto前 赤: veto後 GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  26. TAMA300データ解析 (3) - 偽イベント除去 - • 強度モニタ信号によるフェイク除去 強度モニタ信号も解析, フェイク除去に用いる 時間幅 : 5通り (1.6msec – 25.6msec) 周波数帯 : 1通り (200-1500Hz)  重力波チャンネルと同時に閾値以上のものがあれば、イベント候補から除去 F.D: 1% に閾値を設定 Coincidence Off event (accidental) GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)

  27. 大きなイベントを 除去できていない 2桁程度低減 TAMA300データ解析 (4) - 偽イベント除去の解析結果 - • パワーフィルタでの解析結果 TAMA DT9, Run#140 (200時間分) を解析 時間幅 : 5通り (1.6msec – 25.6msec) 周波数帯 : 5通り (500Hz幅x3, 1kHz幅x2) 青: veto前 赤: veto後 GWDAW-10 (December 14, 2005, University of Texas at Brownsville, U.S.A.)