Clotaire Michel

Time–Frequency Analysis of Small Frequency Variations in Civil Engineering Structures Under Weak and Strong Motions Using a Reassignment Method

The analysis of strong motion recordings in structures is crucial to understand the damaging process during earthquakes. A very precise time-frequency representation, the reassigned smoothed pseudo-Wigner-Ville method, allowed us to follow the variation of the Millikan Library (California) and the Grenoble City Hall building (France) resonance frequencies during earthquakes. Under strong motions, a quick frequency drop, attributed to damage of the soil-structure system, followed by a slower increase is found. However, in the case of weak earthquakes, we show that frequency variations come from the ground motion spectrum and cannot be interpreted in terms of change of the soil-structure system.The analysis of strong and weak motion recordings in existing structures is critical for understanding both the damaging process during earthquakes and their structural behavior. The time—frequency representation is one of the existing methods to get information on the frequency variations in buildings, which may indicate either damage or degradation. In this article, we use standard time—frequency methods with amplitude normalization and reassignment method in order to observe smaller variations in earthquake recordings in buildings. The method used in this article is first validated on strong recordings from the R. Millikan Library (Pasadena, California) and later applied to weak earthquakes recorded in the Grenoble City Hall (France). This method detects torsion activation and very small frequency variations under weak motion. We show that these variations are probably due to variations of the input motion and cannot be attributed to variations in the parameters of the structural system.

Assessment of Site Effects in Alpine Regions through Systematic Site Characterization of Seismic Stations

Abstract In the framework of the renewal project of the Swiss Strong Motion Network (SSMNet), a procedure for site characterization has been established. The aim of the procedure was to systematically derive realistic 1D velocity profiles at each station. It is mainly based on the analysis of surface waves, particularly from passive experiments, and includes cross checks of the derived amplification functions with those obtained through spectral modeling of recorded earthquakes. The systematic use of three component surface‐wave analysis, allowing the derivation of both Rayleigh and Love dispersion curves, also contributes to the improvement of the quality of the retrieved profiles. The procedure is applied to the 30 SSMNet stations installed on various site types within the project, covering different aspects of seismic risk. The characterization of these 30 sites gives an overview of the variety of possible effects of surface geology on ground motion in the Alpine area. Such effects ranged from deamplification at hard‐rock sites to amplification up to a factor of 15 in lacustrine sediments with respect to the Swiss reference rock velocity model. The derived velocity profiles are shown to reproduce observed amplification functions from empirical spectral modeling. Although many sites are found to exhibit 1D behavior, the procedure allows the detection and qualification of 2D and 3D effects. The sites are therefore classified with respect to the occurrence of 2D/3D resonance and edge‐generated surface waves. In addition to the large and deeply incised alpine valleys of the Rhone, the Rhine, and the Aar, smaller structures such as local alpine valleys and alluvial fans are shown to exhibit 2D/3D behavior.

The Current State of Strong Motion Monitoring in Switzerland

The next generation Swiss Strong Motion Network has recently been funded by the Swiss Government: in the next 8 years the Swiss Seismological Service expects to install 100 new 24-bit broadband freefield stations in predominantly urban locations across the country with realtime, continuous data transmission at high sampling rates. This infrastructure will compliment the existing 30 comparable realtime stations installed over the last 3 years, and replace the original ~70 strong triggered dial-up network installed in the early 1990s. The introduction of these new stations provides an opportunity to reassess how strong motion data is used in Switzerland, for routine network operations, emergency response and scientific purposes. The strong motion data will be acquired in parallel with the existing broadband network, and will be processed together with broadband data for earthquake early warning, triggering and locations, near real time ShakeMaps , and moment tensor inversion. Challenges arise on how to archive and provide this type of data to the scientific and engineering communities. Metadata maintenance needs to parallel efforts for the broadband network. Although permanent online archival of large volumes of data is rapidly becoming more affordable, this size of the new dataset dwarfs the existing broadband data currently being generated. We discuss the optimal strategies to permanently archive the continuous data, both within Switzerland, and via the existing European data infrastructures.

Simplified non-linear seismic displacement demand prediction for low period structures

The prediction of non-linear seismic demand using linear elastic behavior for the determination of peak non-linear response is widely used for seismic design as well as for vulnerability assessment. Existing methods use either linear response based on initial period and damping ratio, eventually corrected with factors, or linear response based on increased equivalent period and damping ratio. Improvements to the original EC8 procedure for displacement demand prediction are proposed in this study. Both propositions may be graphically approximated, which is a significant advantage for practical application. A comparison with several other methods (equal displacement rule, EC8 procedure, secant stiffness and empirical equivalent period methods) is performed. The study is based on non-linear SDOF systems subjected to recorded earthquakes, modified to match design response spectra of different ground types, and focuses on the low frequency range that is of interest for most European buildings. All results are represented in the spectral displacement/fundamental period plane that highlights the predominant effect of the fundamental period on the displacement demand. This study shows that linearized methods perform well at low strength reduction factors but may strongly underestimate the displacement demand at strength reduction factors greater than 2. This underestimation is an important issue, especially for assessment of existing buildings, which are often related with low lateral strength. In such cases, the corresponding strength reduction factors are therefore much larger than 2. The new proposals significantly improve the reliability of displacement demand prediction for values of strength reduction factors greater than 2 compared to the original EC8 procedure. As a consequence, for the seismic assessment of existing structures, such as unreinforced masonry low-rise buildings, the current procedure of EC8 should be modified in order to provide accurate predictions of the displacement demand in the domain of the response spectrum plateau.

Long-period surface motion of the multipatch Mw9.0 Tohoku-Oki earthquake

We show that it is possible to capture the oscillatory ground motion induced by the Tohoku-Oki event for periods ranging from 3 to 100s using Precise Point Positioning (PPP). We find that the ground motions of the sedimentary basins of Japan were large (respectively > 0.15m/s and >0.15m/s2 for velocity and acceleration) even for periods larger than 3s. We compare geodetic observables with a Ground Motion Prediction Equation (GMPE) designed for Japan seismicity and find that the Spectral Acceleration (SA) is well estimated for periods larger than 3s and distances ranging from 100 to 500km. At last, through the analysis of the displacement attenuation plots, we show that the 2011 Tohoku-Oki event is likely composed of multiple rupture patches as suggested before by time-reversal inversions of seismic data.

Probabilistic mechanics-based loss scenarios for school buildings in Basel (Switzerland)

Developing earthquake scenarios for cities in areas with a moderate seismicity is a challenge due to the limited amount of available data, which is a source of large uncertainties. This concerns both the seismic hazard, for which only recordings for small earthquakes are available and the unknown earthquake resistance of the majority of structures not designed for seismic loading. The goal of the present study is to develop coherent probabilistic mechanics-based scenarios for a mid-size building stock including a comprehensive analysis of the uncertainties. As an application, a loss assessment for the school buildings of the city of Basel is performed for different scenarios of historical significance, such as the 1356 event, and from the deaggregation of the Swiss Probabilistic Seismic Hazard Model of 2015. The hazard part of the computations (i.e. ground motion estimation) is based on this model, a regional microzonation and recordings of small earthquakes on a dense strong motion network to compute site-amplification factors. The school buildings, which are mainly unreinforced masonry or reinforced concrete shear wall buildings, have been classified according to a specifically developed taxonomy. Fragility curves have been developed using non-linear static procedures and subsequently, vulnerability curves in terms of human and financial losses are proposed. The computations have been run with the OpenQuake engine, carefully propagating all the recognized uncertainties. Scenarios before and after retrofitting measures show their impact on the earthquake safety. A sensitivity analysis shows that the largest uncertainties come from the ground motion prediction although an improvement of all parts of the model is necessary to decrease the uncertainties. Although improved data and models are still necessary to be developed, probabilistic mechanics-based models outperform the capabilities of deterministic and/or empirical models for retrieving realistic earthquake loss distributions.

Deriving fragility functions from bilinearized capacity curves for earthquake scenario modelling using the conditional spectrum

The development of fragility curves to perform seismic scenario-based risk assessment requires a fully probabilistic procedure in order to account for uncertainties at each step of the computation. This is especially true when developing fragility curves conditional on an Intensity Measure that is directly available from a ground-motion prediction equation. In this study, we propose a new derivation method that uses realistic spectra instead of design spectral shapes or uniform hazard spectra and allows one to easily account for the features of the site-specific hazard that influences the fragility, without using non-linear dynamic analysis. The proposed method has been applied to typical school building types in the city of Basel (Switzerland) and the results have been compared to the standard practice in Europe. The results confirm that fragility curves are scenario dependent and are particularly sensitive to the magnitude of the earthquake scenario. The same background theory used for the derivation of the fragility curves has allowed an innovative method to be proposed for the conversion of fragility curves to a common IM (i.e. spectral acceleration or PGA). This conversion is the only way direct comparisons of fragility curves can be made and is useful when inter-period correlation cannot be used in scenario loss assessment. Moreover, such conversion is necessary to compare and verify newly developed curves against those from previous studies. Conversion to macroseismic intensity is also relevant for the comparison between mechanical-based and empirical fragility curves, in order to detect possible biases.

Fault zone signatures from ambient vibration measurements: a case study in the region of Visp (Valais, Switzerland)

Investigations of tectonic features, such as faults, are important challenges for geologists and engineers. Although direct investigational methods, such as boreholes and trenches, have the potential to provide accurate data, these direct methods are usually expensive and time consuming, and give only punctual insights into subsurface structures. Geophysical methods, for example electric surveys and ground penetrating radar, are less expensive and faster to implement. However, these geophysical methods may be difficult or sometimes even impossible to apply in regions with rough topography or regions which are highly urbanized. In this study, we propose an easy-to-use and affordable method to detect fault zones based on ambient vibration observations. We apply this method in the region between Visp and Unterstalden (canton Valais, Switzerland) on a small fault branch, which has no explicit surface expression, and which is linked to a major fault zone, the Simplon Fault Zone. The assumption is that the fault of interest is surrounded by damage zone consisting of fractured rock, and that this results in lateral changes of both seismic velocity and attenuation. The objective was, first, to identify such lateral changes in the observed seismic wave-field, and second, to map any anomalies and combine them with the available geological information. In this way, we were able to follow the fault trace even without a clear surface expression of the fault. Our observations showed the existence of a signature in the power spectra of the seismic noise that may correspond to a damage zone. Such signature is observed along the trace of the expected fault.ZusammenfassungUntersuchungen von tektonischen Einheiten, wie zum Beispiel Verwerfungen, stellen wichtige Herausforderungen für Geologen und Ingenieure dar. Obwohl direkte Untersuchungsmethoden, wie beispielsweise Bohrlöcher oder Grabungen das Potential in sich bergen, genaue Daten zu liefern, sind sie doch gewöhnlich sehr teuer, zeitintensiv und geben nur einen punktuellen Einblick ins Erdinnere. Geophysikalische Methoden wie elektrische Bodenuntersuchungen oder das Georadar sind zwar weniger teuer und einfacher in der Handhabung, diese Methoden sind aber oft nur sehr schwer oder teilweise gar unmöglich einzusetzen, vor allem im schweren Gelände oder in stark überbautem Gebiet. In dieser Studie präsentieren wir eine einfach anwendbare und vor allem kostengünstige Methode, um Verwerfungen mittels seismischer Bodenunruhe detektieren können. Wir wenden diese Technik in der Region zwischen Visp und Unterstalden an einer Nebenverwerfung der Simplon Linie an, die keinen expliziten Oberflächenausdruck aufweist. Dabei gehen wir von der Vermutung aus, dass die zu untersuchende Verwerfung das Gestein in seiner unmittelbaren Umgebung stark zerbrochen hat, was wiederum eine laterale Modifikation der seismischen Geschwindigkeit und Dämpfung zur Folge hätte. Das Ziel war es, eine solche laterale Veränderung des beobachteten Wellenfeldes zu ermitteln, diese Anomalie zu kartieren und in einen geologischen Kontext zu setzen. Auf diese Weise konnten wir der Spur der Verwerfung auch ohne klaren Oberflächenausdruck folgen. Unsere Beobachtungen zeigen die Existenz einer Signatur im seismischen Spektrum des Hintergrundrauschens, die der gesuchten Verwerfung entsprechen könnte. Eine solche Signatur konnte über eine grössere Distanz hinweg identifiziert werden. Diese Beobachtungen, zusammen mit geomorphologischen und geologischen Hinweisen, stützen die Hypothese einer Verwerfung im entsprechenden Untersuchungsgebiet.RésuméLes enquêtes sur les unités tectoniques, tels que des failles, sont des défis importants pour les géologues et les ingénieurs. Bien que les méthodes d’investigation directe, tels que les forages ou les tranchées aient le potentiel de fournir des données exactes, elles sont généralement très coûteuses, chronophages et ne donnent qu’un aperçu ponctuel des sous-sols. Les méthodes géophysiques telles que les sondages électriques ou les enquêtes géoradar sont moins chères et plus faciles à manipuler, mais ces méthodes sont souvent très difficiles, voir même impossibles à utiliser, surtout en régions urbanisées ou présentant un fort relief. Dans cette étude, nous proposons une méthode qui nous permet de détecter des failles au moyen du bruit de fond sismique. Nous utilisons cette technique dans la région entre Viège et Unterstalden sur une faille mineure, probablement associée au système extensif de la ligne de faille du Simplon, et qui n’a pas d’expression explicite en surface. Nous supposons que la faille à analyser est entourée de roches fracturées, ce qui aurait pour conséquence une modification latérale de la vitesse et de l’atténuation sismique. Le but est d’identifier une telle modification latérale du champ d’ondes observé, cartographier cette anomalie et de la replacer dans le contexte géologique relatif à la faille cherchée. Ainsi nous avons pu suivre la trace de la faille cherchée sans expression claire à la surface. Nos observations montrent l’existence d’une signature dans le spectre du bruit de fond sismique qui pourrait correspondre à la zone endomagée de la faille cherchée. Une telle signature a pu être identifiée le long de la faille supposée. Ces observations et les indices géomorphologiques soutiennent l’hypothèse d’une faille dans la région analysée.