Transcription

1 THE FRENCH PERMANENT ACCELEROMETRIC NETWORK: TECHNOLOGICAL AND PARAMETRICAL CHOICES SCIENTIFIC GOALS AND RESULTS Catherine Berge-Thierry *, J.F. Fels, P. Dervin, P. Guéguen, S. Nechtschein, D. Hatzfeld, P. Volant, P.Y. Bard, F. Courboulex, P. Dominique, B. Feignier * Institut de Radioprotection et de Sûreté Nucléaire/berssin catherine.berge@irsn.fr g@ C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 1

2 Technicians and scientists that ensure the maintenance and improvment of the network: Institut de Physique du Globe de Paris (West Indies), Laboratoire de Géophysique Interne et de Tectonophysique, Laboratoire des Ponts et Chaussées, Institut de Radioprotection et de Sûreté Nucléaire, Géosciences-Azur, le Bureau de Recherches Géologiques et Minières, Ecole et Observatoire des Sciences de la Terre de Strasbourg, Commissariat à l Energie Atomique, Observatoire Midi Pyrénées, Centre d Etudes Techniques de l Equipement, Conseil Général de la Martinique, Ministère de l Environnement et du Développement Durable, Direction de la Recherche et des Affaires Scientifiques et Techniques, Institut t Nti Nationaldes Si Sciences de l Univers Université de Bretagne Occidentale Université de Clermont Ferrand C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 2

3 Strong motion stations in the 80 s: ~ the NPP map 0.01g threshold Analog instruments (SMA1, SMA2 essentially) Very few events detected not sensitive enough inadequacy for a SM database construction C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 3

4 Birth of the French Accelerometric Network Aims of the Network: Provide accurate data for scientists and engineers Sensitive threshold, without saturation for strong motion Instruments developed by a French company Testing phase in most of the seismic areas in France (metropolitan): Nice, Grenoble, Alps instrumented, Emphasized significant site effects in sedimentary valleys (Grenoble) Demonstrated the feasability and accuracy of instruments The Ministry decided sponsoring the RAP, with an objective of 100 stations C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 4

5 91 «RAP» stations installed in Number of stations from 1995 up to 2002 C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 5

6 RAP Network = RAP stations + Associated SM stations IRSN, BRGM, LDG and Martinique General Council (metropolitan and French West indies) Instrumentation and record processing RAP stations C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 6

7 RAP strong motion Database, end of 2003 Evolution of recorded events Evolution of records seismic events component records C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 7

8 RAP station: instrumentation and data processing Sensors: Kinemetrics (Episensor T), Guralp, Acquisition system: Titan developed by Agecodagis Weight/dimension: 4 Kg, 260 x 160 x 120 Power requirements: Volt, 2.5 Watt Input voltage full scale +/- 5 Volt +/- 1g Input impedance: Digitizers: Overall digital noise: 25 Kohm Crystal CS5321+CS5322, sigma-delta, 24 bits. 8 counts p-to-p. System clock stability: External time reference: Data time stamp: Data storage: Communication: DCF radio receiver. internal time + external time. trigger records saved on internal 8 Mb Flash memory serial RS232, to modem or PC C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 8

9 Trigger mode (STA/LTA) Time constant for LTA Time constant for STA Triggering threshold Pre-eventevent Trigger record minimum duration Trigger record maximum duration 30 s 1s 4 or 6 depending on site 10 s 30 s 100 s Sampling: 125 sps FIR filters (time delays removed during processing) «Titan» format (serial type) Data stored on a 8 Mb Flash Memory memory is full within 4 ou 5 days (~ 50 records) Operations performed «on site» (regional RAP center) C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 9

10 FTP servers Regional Centers RAP Central Site Regional Center Seismic catalogues Interogating stations & retrieving records Maintain stations RAP Offset Removal (SAC) Distribution through the Central Site Storage, checking- validation RAP WEB site C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 10

11 The IRSN Associated Network The Moyenne Durance Fault Network: Velocimetry + Accelerometry Grenoble Tricastin Durance C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 11

12 IRSN stations acquisition parameters K2/Etnas (k kinemetrics s) Internal FBA23 Characteristic ti frequency: 50Hz, damping 70%, dynamic >135dB from 0.01 to 50hz Internal Episensor Frequency range 0 to 200Hz, dynamic >145dB Sampling 200 sps Frequency range usable for data 40% of the sampling using digital anti-aliasing filters, ie [0 to 80Hz] for 200 sps. Final anti aliasing i filter no causal, llinear response for the phase, pure delay Triggering with a threshold Filter linked to the trigger 0.5 to 1mg depending on the stations IIR Bandpass filter : classic strong [0.1Hz-12.5Hz] with 200sps C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 12

13 IRSN strong motion data processing Standard Kinemetrics method : Strong Motion Analysis for computing corrected accelerograms V2 format Opening of the event file recorded by a K2 or Etna acquisition system (.EVT), Writing IRSN raw data in CDMG format (V1 File) Removing the continuous component (offset evaluated on the entire event file), Instrumental correction Integrating (including a high pass filter) velocity and displacement time series Writing IRSN corrected data in CDMG format (V2 File). C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 13

14 IRSN processing: TRIFUNAC Branch IRSN strong motion data processing 1st decimation factor (instrumental correction) [1] keeping the maximum number of samples for corrected data Half length of filter operator: low pass filter for offset correction, [10] Acceleration decimation factor: high pass filter of acceleration during [10] the offset correction, Velocity decimation factor: high pass filter of velocity for the high-pass [10] filter step, Displacement decimation factor: high pass filter of displacement for the [10] high-pass filter step, Terminal frequency: final frequency of the Ormsby low pass filter [50] (instrumental correction) Roll-off width: width of the Ormsby low pass filter (instrumental [3] correction) Corner frequency : corner frequency for Ormsby high pass filter (offset [0.25] correction and high pass filtering of velocity and displacement) Roll-off width - Ormsby high pass filter - (offset correction and high pass filtering of velocity and displacement) [0.1] Corrected IRSN data (under RAP format) RAP Central Site Storage in the database Distribution via RAP WEB site C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 14

15 3 other associated networks maintained by: - BRGM (16 in French West Indies) - LDG (2) - Martinique General Council (28) Geosig instruments (processing not described in this contribution) C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 15

16 RAP «weak motion» records accelerometer 15 m alluvion 40 m sand 556 m 447 m Clay Significant amplifications (duration, amplitude) in sedimentary valleys 24 m limestone accelerometer C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 16

17 The Mw=5,1 «Besançon» event (23/02/2004) recorded by RAP Network compute Moment magnitude, compare recorded acceleration with predicted one, C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 17

18 Conclusions Until the 90 s strong motion recorded associated to detection (NPP site safety) French Accelerometric Network: scientific and engineering aims - beginning of structural instrumentation, - improvement of geological site characterization Moderate Metropolitan Seismicity conducted the choice of low threshold for RAP stations (trade-off): - this strategy allows the recording of weak motion (site effects) - necessity to increase the Flash memory to avoid any loss of data in case of communication failures (aftershocks) Federation of the scientific and user communities to improve the network and the data distribution C. Berge-Thierry et al., May 2004, Strong Motion Record Processing 18