EXAMPLES OF USE OF THE SM-ROM-GL DATABASE IOAN SORIN BORCIA - PhD, National Institute for Research and Development URBAN-INCERC and European Centre for Buildings Rehabilitation – ECBR, e-mail: isborcia@yahoo.com IOLANDA-GABRIELA CRAIFALEANU - Associate Professor PhD, National Institute for Research and Development URBAN-INCERC and European Centre for Buildings Rehabilitation – ECBR and Technical University of Civil Engineering, Bucharest, Romania ELENA-ANDREEA CALARASU - PhD, National Institute for Research and Development URBAN-INCERC and European Centre for Buildings Rehabilitation – ECBR NICOLETA-FLORENTA TANASE - PhD, National Institute for Research and Development URBANINCERC and European Centre for Buildings Rehabilitation – ECBR IOAN-CONSTANTIN PRAUN – Eng., National Institute for Research and Development URBAN-INCERC and European Centre for Buildings Rehabilitation – ECBR Abstract: The SM-ROM-GL database includes data obtained from the processing of records performed at ground level by the Romanian seismic networks, namely INCERC, NIEP, NCSRR and ISPH-GEOTEC, during recent seismic events with moment magnitude Mw ≥ 5 and epicenters located in Romania. All the available seismic records were re-processed using the same basic software (from Kinemetrics, Inc.), the same procedures and options (filtering and baseline correction), in order to obtain a consistent dataset. The database stores computed parameters of seismic motions, i.e. peak values: PGA, PGV, PGD, effective peak values: EPA, EPV, EPD, control periods: TB, TC, TD, spectral values of absolute acceleration, relative velocity and relative displacement, as well as of two alternative instrumental intensities. The paper presents two examples of use of the SM-ROM-GL database, for the ground motions recorded obtained in Banat region in 1991 and for the Vrancea earthquake of October 27, 2004. Keywords: seismic records, ground motion database, earthquakes, instrumental seismic intensity. 1. Introduction The SM-ROM-GL database [1] contains results obtained by the processing of earthquake ground motions, recorded by the seismic networks of Romania (INCERC - now part of the National Institute for Research and Development URBAN-INCERC, NIEP - National Institute for Research and Development in Earth Physics, NCSRR - now part of the National Institute for Research and Development URBAN-INCERC and ISPH-GEOTEC). The records were obtained at ground level, during strong seismic events with moment magnitude Mw ≥ 5 and with epicenters situated in the Vrancea and Banat seismogenic zones. New numerical processing methodologies were developed and applied to the records of these earthquakes (i.e. instrumental intensity spectra averaged over different period intervals, response spectrum-based intensities, Arias intensities, destructiveness spectrum-based intensities), these being used in defining the seismic action for building research and design and for a better understanding of the structural behavior of buildings during strong Vrancea and Banat earthquakes. 2. Advanced processing of seismic records obtained at ground level First, the initial processing of the recorded accelerograms was performed, i.e. the time-histories of acceleration, velocity and displacement were obtained. The corrections were made by using an Ormsby filter with 0.16 Hz as corner frequency and 32 Hz as terminal frequency. Then, the peak acceleration, pga, peak velocity, pgv and peak displacement, pgd, for the records obtained at ground level were determined, and the secondary processing of accelerographic information was performed. Thus, response spectra (absolute acceleration response spectra, Saa, relative velocity response spectra, Svr, and relative displacement response spectra, Sdr) were computed. Based on the above spectral quantities, the “effective” values were estimated, i.e. the effective peak ground acceleration, epa, effective peak ground velocity, epv, effective peak ground displacement, epd. These parameters were determined, as specified in the Romanian seismic design code, P100-1/2013, according to the following relations: epa  ( S aa averaged on 0.4s ) max / 2.5 epv  ( S vr averaged on 0.4s ) max / 2.5 (1) epd  ( S dr averaged on 0.4s ) max / 2.5 (3) (2) These quantities were obtained by averaging the response spectra computed for a damping ratio n=5%, where (0.4 s) represents the moving average for a 0.4 s time window, performed on the 0.1 s … 4.0 s range. The effective values were used to compute the control periods, TC and TD, as follows: TC  2epv / epa  TD  2epd / epv  (4) (5) Additionally, instrumental intensity measures were computed [2, 3], as shown in the following. 1. Global instrumental intensities 1.1. The global intensity based on response spectrum, IS, a measure of ground motion severity, is defined by using the following parameters: EPAM (m / s 2 )  max T S aa (T ,0.05) / 2.5 (6) EPVM (m / s 2 )  max T S va (T ,0.05) / 2.5 (7) EPDM (m / s 2 )  max T S dr (T ,0.05) / 2.5 (8) where: - Saa(T,n) is the absolute acceleration response spectrum and Sva(T,n) is the absolute velocity spectrum, both expressed as functions of period and damping ratio; - Sdr(T,n) is the relative displacement spectrum, expressed as a function of period; n is the damping ratio, and maxT is the maximum spectral value, for periods, T, between 0.0625 s and 4.0 s. I S  log 6 ( EPAM  EPVM )  8.0 (9) 1.2. The Arias-type intensity: I A  log 6  [ w g ] 2 dt  7.05 (10) where wg(t) is the ground acceleration on a horizontal direction. 2. The following quantities were computed, as well, for intensities depending on the frequency φ (Hz). 2.1. Response spectrum-based intensity, is(φ): i s   log 6 [ S aa (,0.05)  S va (,0.05)]  7.75 (11) 2.2. Destructivity spectrum-based intensity, id (φ), determined from the (absolute) accelerogram wa(t, φ, 0.05), for a pendulum with the natural (undamped) frequency φ and a damping ratio of 0.05, i d ()  log 6 [  wa2 (t , ,0.05) dt ]  6.25 (12) 3. Intensities based on the application of the averaging rule on a specified frequency band (φ', φ") were computed as well, using the following expressions. 3.1. For the response spectrum-based intensity, is(φ):   1 / ln(" / ' )   [ S aa (,0.05)  S va (,0.05)d / ]  7.75 (13)   ' "  i s* (' , " )  log 6  3.2. For the destructivity-based intensity, id(φ):   i d* (' , " )  log 6   "  '  1 / ln(" / ' )   [( wa2 (t , ,0.05) dt )d / ]  6.25 (14) 4. Averaging rules for the two horizontal orthogonal directions were also provided. The following notations were used: Id1 = id*(0.25 Hz, 16.0 Hz), for averaging on the whole interval (0,0625 sec – 4,0 sec), Id31 = id*(0.25 Hz, 1.0 Hz), Id32 = id*(1.0 Hz, 4.0 Hz), Id33 = id*(4.0 Hz, 16.0 Hz) for averaging on 3 period intervals (i.e. (1 - 4 sec), (0.25 - 1 sec) and (0.0625 - 0.25 sec)). Similar notation conventions were used for averaging on 6 and 12 intervals, respectively. 3. Database structure The adopted codifications and the fields associated with various information in the database are given in the following. - Seismic events (source parameters) are given in table SM-ROM-GL Earthquakes, which contains information on the eleven earthquakes with Mw ≥ 5 that occurred in Romania between 1977 and 2009, of which eight with epicenters located in the Vrancea seismogenic zone and three with epicenters located in the Banat seismogenic zone). - Seismic stations (geographical coordinates and station codes) are given in table SM-ROM-GL Stations, which contains information on the 166 stations that provided at least one record from one of the earthquakes listed in table SM-ROM-GL Earthquakes. - Seismic records are given in table SM-ROM-GL Records, which contains information on the 307 records obtained from the networks in Romania, Bulgaria and Moldova. The records were obtained in the above seismic stations during the mentioned earthquakes. For these records, charts are provided for absolute acceleration response spectra, relative velocity spectra and displacement spectra, as well as for instrumental intensity spectra. - Representative parameters of seismic motion (peak values and "effective" values of acceleration, velocity and displacement, global instrumental intensities and intensities averaged on various period intervals and spectral values for periods whit ΔT = 0.05 s) are given in table SM-ROM-GL Components, which contains information on the 920 components of the records in the table SM-ROMGL Records. 4. Examples of use of the SM-ROM-GL database A summary of instrumental data on the strong Vrancea earthquakes between 1977 and 1990 is presented in [4]. The examples of use of the database concern the ground motion records obtained during the earthquake sequence that occurred in 1991 in the Banat seismogenic zone (Table 1) and to the spatial distribution of the seismic parameters (PGA and PGV) recorded during the Vrancea earthquake of October 27, 2004 (Mw = 6.0) Table 1 Table SM-ROM-GL Earthquakes (excerpt) Earthquake LatN Banat-Banloc Banat-TopletHerculane Banat-Voiteg Vrancea LongE CodeEarthq h (km) Date Mw 45.38 21.05 19911 11 1991.07.12 5.6 44.90 22.35 19912 12 1991.07.18 5.6 45.45 45.79 21.12 26.71 19913 20041 9 99 1991.12.02 2004.10.27 5.5 6.0 4.1. Analysis of ground motion records obtained in 1991 in Banat region The Banat region, located in the south-west of Romania, is characterized by an important seismic activity observed in the last years (for example in 1991 and 2014). The main records obtained in the INCERC seismic network (Table 2) during the Banat earthquakes of 1991 (Table 1), are listed in Table 3. The units used for the quantities in the table are [m/s2] for pga and epa, [m/s] for pgv and epv and [m] for pgd and epv. The parameters Is and Id are the global instrumental intensities, given by Eq. 9 and Eq. 10, respectively. Table 2 INCERC seismic Stations (analog accelerographs) in Banat region, Romania No. Station Address Lat.N Long.E Code 1 Banloc Town Hall 45.397 21.131 BNL1 2 Timisoara Bd. Dambovita, RAT Timisoara 45.747 21.236 TIM1 3 Timisoara Calea Buziasului 26, EAM SA 45.745 21.237 TIM2 Table 3 Table SM-ROM-GL Components (excerpt) Code Earthq 19911 19911 19911 19913 19913 19913 19913 19913 19913 19912 19912 19912 19913 19913 19913 Code Station TIM1 TIM1 TIM1 TIM1 TIM1 TIM1 TIM2 TIM2 TIM2 BNL1 BNL1 BNL1 BNL1 BNL1 BNL1 Code Axis l t v l t v l t v l t v l t v pga pgv pgd epa epv epd TC TD Is Ia 0.3378 0.3926 0.1111 0.2086 0.1641 0.0799 0.1977 0.1634 0.0970 0.3074 0.6328 0.2803 0.8458 1.2890 1.2863 0.0436 0.0342 0.0321 0.0284 0.0245 0.0206 0.0311 0.0244 0.0265 0.0148 0.0354 0.0114 0.1177 0.1990 0.0311 0.0176 0.0184 0.0176 0.0194 0.0120 0.0156 0.0179 0.0168 0.0178 0.0031 0.0033 0.0048 0.0190 0.0468 0.0083 0.3143 0.4897 0.0299 0.0362 0.0167 0.0184 0.60 0.46 3.52 3.18 5.70 6.08 5.68 5.86 0.1916 0.1917 0.0303 0.0225 0.0210 0.0130 0.99 0.74 4.35 3.62 5.38 5.18 5.15 5.12 0.1532 0.1700 0.0277 0.0266 0.0139 0.0108 1.14 0.99 3.15 2.55 5.31 5.32 5.30 5.30 0.2009 0.4892 0.0099 0.0204 0.0055 0.0040 0.31 0.26 3.48 1.23 4.86 5.74 4.83 5.39 0.7591 1.0085 0.0854 0.1490 0.0199 0.0313 0.71 0.93 1.46 1.32 6.64 7.12 6.54 6.87 The accelerometric records obtained by INCERC in Timişoara and Banloc starting with the July 12, 1991 earthquake have the following characteristics: relatively short durations (9…24 s); significant spectral values generally in the 0.1…0.3 s range, extending up to 0.7 s for the Timişoara record and up to 1.2 s for Banloc; maximum pga’s of the horizontal components of 0.3926 m/s2 for Timişoara and of 1.289 m/s2 for Banloc; rather high vertical pga’s for Banloc (1.2863 m/s2 for the December 2, 1991 earthquake). Time histories for the two main records obtained in 1991 in Banat (19911TIM1 and 19913BNL1) are shown in Fig. 1. Fig. 1 - Acceleration time histories Response spectra and averaged intensity spectra, as well as Is6 and Id6 intensities, computed for 6 dB intervals, are presented in Fig. 2, for the two main records obtained during 1991 Banat earthquakes. Fig. 2 - Response spectra (n=5%), Amplitude Fourier and intensities spectra (Is - response spectrum-based intensity (Eq. 13) and Id - destructivity-based intensity (Eq. 14)) 4.2. Spatial distribution of peak ground accelerations and velocities recorded during the earthquake Vrancea earthquake of October 27, 2004 Due to the inclusion of the geographical coordinates of seismic stations, the spatial distribution of the values provided in the database can also be easily obtained. Examples are given in Fig. 3, where the map of peak ground acceleration values (PGA) for the Vrancea earthquake of October 27, 2004 is shown, and in Fig. 4, where the map of peak ground velocity values (PGV) is presented, for the same earthquake. The characteristics of the earthquake are given in Table 1. Fig. 3 - Map of peak ground acceleration values (PGA) for the Vrancea earthquake of October 27, 2004 (EW components) [1] Fig. 3 - Map of peak ground velocity values (PGV) for the Vrancea earthquake of October 27, 2004 (EW components) [1] Recent studies based on the SM-ROM-GL database addressed the potential of an enhanced use of the information it contains, by including records obtained in SMBRS-DU-type stations in Romania (similar to COSMOS classification [5]), and which were not previously included in studies on the spatial distribution of significant ground motion parameters [6]. 5. Concluding remarks The maintenance of the SM-ROM-GL database and the addition of new records to be obtained in the future will be provided by the developers. The database, along with the users’ guide, will be soon available on the BIGSEES project website (http://infp.infp.ro/bigsees/ default.htm), thus becoming accessible to any interested user. Acknowledgements Funding for this research was provided by the Romanian Ministry of National Education, UEFISCDI Agency under the Contract Number 72/2012, Project BIGSEES. This support is gratefully acknowledged. References [1] Borcia I.S., Craifaleanu I.G., Calarasu E.A., Tanase N.F. & Praun I.C. (2014) „ SM-ROM-GL (Strong Motion ROMania Ground Level) database”, Second European Conference on Earthquake Engineering and Seismology, Istanbul, Turkey, August 24-29, Paper No 487 (full paper accepted for presentation to the Conference) [2] Sandi H. & Borcia I.S. (2011a) “Intensity spectra versus response spectra. Basic concepts and applications”, Pure and Applied Geophysics, 168(1-2):261-287, DOI 10.1007/s00024-010-0158-1 [3] Sandi H. & Borcia I.S. (2014) “An attempt to recalibrate instrumental criteria for intensity assessment”, Second European Conference on Earthquake Engineering and Seismology, Istanbul, Turkey, August 24-29, Paper No 182 (full paper accepted for presentation to the Conference) [4] Sandi H. & Borcia I.S. (2011b) “A summary of instrumental data on the recent strong Vrancea earthquakes, and implications for seismic hazard”, Pure and Applied Geophysics, 168(3-4): 659-694, DOI 10.1007/s00024010-0157-2 [5] COSMOS (2001) “Guidelines for Installation of Advanced National Seismic System Strong-Motion Reference Stations”, Consortium of Organizations for Strong-Motion Observation Systems, Richmond, CA [6] Craifaleanu, I.G., Borcia, I.S. (2014). “An analysis of the usability of strong ground motion records obtained from stations located in densely urbanized areas in Romania”. Second European Conference on Earthquake Engineering and Seismology, Istanbul, Turkey, August 24-29, Paper No. 499 487 (full paper accepted for presentation to the Conference)