Sharma, Nitin (2013) Ground Motion Prediction Equations For Time-dependent Seismic Hazard Analysis in Geothermal Areas. [Tesi di dottorato]

[img]
Preview
Documento PDF
Nitin_Sharma_2013.pdf

Download (49MB) | Preview
[error in script] [error in script]
Item Type: Tesi di dottorato
Lingua: English
Title: Ground Motion Prediction Equations For Time-dependent Seismic Hazard Analysis in Geothermal Areas
Creators:
CreatorsEmail
Sharma, Nitinsharma@na.infn.it
Date: 2 April 2013
Number of Pages: 290
Institution: Università degli Studi di Napoli Federico II
Department: Fisica
Scuola di dottorato: Scienze fisiche
Dottorato: Rischio sismico
Ciclo di dottorato: 25
Coordinatore del Corso di dottorato:
nomeemail
Zollo, Aldoaldo.zollo@unina.it
Tutor:
nomeemail
Zollo, Aldoaldo.zollo@unina.it
Convertito, Vincenzovincenzo.convertito@ov.ingv.it
Date: 2 April 2013
Number of Pages: 290
Uncontrolled Keywords: The Geysers, Seismic Hazard, Induced Sesimicity, Inter-event uncertainty, Intra-event uncertainty
Settori scientifico-disciplinari del MIUR: Area 02 - Scienze fisiche > FIS/06 - Fisica per il sistema terra e il mezzo circumterrestre
Date Deposited: 06 Apr 2013 09:13
Last Modified: 23 Jul 2014 07:27
URI: http://www.fedoa.unina.it/id/eprint/9394
DOI: 10.6092/UNINA/FEDOA/9394

Abstract

My thesis deals with detail understanding and analysis of ground-motion prediction equations in geothermal areas with special emphasis on “The Geysers” geothermal area. The interest for this specific area is because it is the largest productive vapor-dominated enhanced geothermal system in the world. No ground-motion prediction equations (GMPEs) were available, specifically for “The Geysers” region before this work. Non-linear mixed-effect regression analysis (NLMRA) is applied to compute ground-motion prediction equations for The Geysers geothermal area for PGV, PGA and SA(T) at T = 0.2s, 0.5s and 1.0s. NLMRA allows to split the total uncertainty of ground motion into two i.e. inter-event and intra-event uncertainties. It is found the our GMPEs are more robust when ground motion predictions have to made for induced earthquakes (smaller earthquakes). Two step method is used to compute GMPEs. First, a model is retrieved, which does not include station effects and represent the reference model. Second, a first-order correction is introduced for station effects based on the Z-test of the residual distribution at each station with respect to reference model. It is found that if information about local geology is not available, the method of introducing site/station effect correction is effective and shows significant improvement in the model. It is also demonstrated that the two components of standard deviation (i.e., inter-event and intra-event) are related to source effect and local site effects, respectively. The GMPEs are also computed for datasets of induced and natural seismicity from six different geothermal areas (from Basel, Geysers, Hengill, Roswinkel, Soultz and Voerendaal) . It is shown that ground motions from induced and natural earthquakes cannot be distinguished. The combined data sets leads to unrealistic high intra-event uncertainty. It is inferred that it is due to the mapping of epistemic uncertainty into aleatory component. However, a reduction in total standard error is observed over mid-to-long period range, but an increase in total standard error is observed in short period range. It is also shown that focal depth plays an important role for shallow earthquakes at shorter distances. An application to time-dependent seismic hazard is also discussed. It is observed that seismic hard is not constant with time and space. It is controlled by industrial activities in The Geysers geothermal area. The proposed technique of site specific seismic hazard analysis can be used for monitoring purposes and can be used in the same respect of the traffic light system. to control human activities such as mining, drilling, injection and extractions of fluid in geothermal areas etc. The concept of time-dependent ground-motion prediction equation is also discussed. It is emphasized that ergodic assumption should be avoided while computing seismic hazard analysis when induced seismicity is concern, because in case of induced seismicity the exposure times are longer than earthquake return period and this can lead to overestimation of ground motion hence the seismic hazard . A cyclic correlation between seismicity and industrial activities is observed. It is shown that the ground-motion recordings are varying with time. It is concluded that source, path and site effect are also changing with time varying seismicity.

Actions (login required)

View Item View Item