Directionality and soils’ effects on the seismic hazard and risk. Applications to ground motion big data sets and to urban environments

Resumen   Abstract   Índice   Conclusiones

Pinzón Ureña, Luis Alejandro

2021-A
Descargar PDF  


Resumen

Los terremotos se definen como una «sacudida violenta de la corteza y el manto, producida por fuerzas que actúan en el interior de la Tierra». En la mayoría de los casos, estas fuerzas son causadas por procesos relacionados con la liberación de energía, producido por el contacto entre placas tectónicas. Otros terremotos menos frecuentes, son los inducidos por actividades antrópicas o los de origen volcánico. En cualquier caso, la energía se libera en forma de ondas multidireccionales, que llegan a la superficie, causando efectos diversos. Las intensidades causadas por un terremoto, sin embargo, no son uniformes en todas las direcciones de su propagación. A menudo, el movimiento fuerte está polarizado, debido al tipo falla y/o a la proximidad al hipocentro, de forma que, dependiendo de la geometría y dinámica de la ruptura, las aceleraciones son más intensas en determinadas direcciones. Este fenómeno se conoce como efecto de directividad. Por otra parte, la intensidad y la forma de la onda varían dependiendo del medio de propagación. En este sentido, los modelos predictivos del movimiento fuerte del terreno tienen en cuenta, también, los efectos debidos a la propagación, desde la fuente al lugar de interés, de la energía liberada. Los efectos locales, principalmente los debidos a la topografía y la geología, también son importantes. Los medios rígidos, como los suelos rocosos y compactos, no tienden a amplificar el movimiento sísmico, mientras que los suelos blandos sí amplifican frecuencias específicas, dependiendo de las propiedades del subsuelo y de las características del movimiento. En cuanto a los efectos de direccionalidad, estos hacen referencia al movimiento fuerte en un lugar específico. Esta tesis aborda dos temas importantes en relación con estos efectos. El primero se refiere a la orientación de los sensores que registran las acciones sísmicas y el segundo está relacionado con el daño esperado en edificios concretos, dependiendo de la orientación de sus ejes principales. Vale decir que, hoy en día, la mayor parte de las normativas sísmicas aún no consideran los efectos de direccionalidad. En esta tesis, se presta especial atención tanto a los efectos de direccionalidad como a los efectos de suelo.
Desde el año 2008, se han reportado 360,000 muertos, aproximadamente, por causa de terremotos. Este hecho pone en evidencia la necesidad de desarrollar más y mejores herramientas para evaluar y prevenir el riesgo sísmico. Por consecuente, el objetivo principal de esta tesis ha sido identificar y evaluar los efectos de direccionalidad del movimiento fuerte y los efectos de suelo que influyen en la peligrosidad y el riesgo sísmico, con aplicaciones en la microzonificación sísmica de suelos en entornos urbanos y en grandes bases de datos de acelerogramas. La tesis se divide en tres bloques: I) los efectos de direccionalidad, II) los efectos de suelo, clasificación de sitios y otros temas relacionados con el riesgo sísmico, y III) estudios de casos relevantes relacionados con los dos bloques anteriores. En el primer bloque, se han considerado los efectos de direccionalidad en las acciones sísmicas esperadas; por lo que, se han estimado medidas de intensidad utilizando las bases de datos de Italia y Costa Rica. Además, en la evaluación del daño previsible en edificios, utilizando análisis dinámicos no lineal, se ha propuesto un nuevo planteamiento que permite considerar los efectos de direccionalidad de forma simplificada. En el segundo bloque, se realizan estudios de microzonificación y de resonancia (suelo-edificio) a la Ciudad de Barcelona; además, se realiza una clasificación sísmica de los emplazamientos de las estaciones acelerométricas de la red española; el análisis de interacción suelo-estructura, considerando los efectos de direccionalidad, y la propuesta de una nueva medida de intensidad, altamente correlacionada con la deriva (drift) en edificos, aparecen como temas complementarios en este bloque. Finalmente, en el tercer bloque, fueron incluidas tres contribuciones relevantes que completan esta disertación. Los resultados demuestran, 1), que los efectos de direccionalidad en las acciones sísmicas esperadas son significativos y hay que considerarlos en los estudios probabilísticos de amenaza sísmica, así como en las evaluaciones de riesgo sísmico y confirman, 2), la relevancia que los efectos de sitio tienen, tanto en los estudios de peligrosidad sísmica como en la evaluación del daño esperado en estructuras. Esta tesis doctoral pretende ser un paso adelante hacia la evaluación, prevención y reducción del riesgo sísmico.

Palabras claves: ingeniería sísmica; sismología para la ingeniería; efectos de direccionalidad; efectos de sitio; evaluación del daño; clasificación de sitio; movimientos del terreno; medidas de intensidad; interacción suelo-estructura; microzonificación; amenaza sísmica; riesgo sísmico; cocientes espectrales H/V; periodo fundamental; deriva entre piso; efectos de suelo.

Códigos UNESCO: 2507.05; 3305.32.


Abstract

Earthquakes are defined as a «violent shaking of the Earth’s crust and mantle, caused by forces acting inside the Earth». In most cases, these forces are caused by an energy release process generated from the contact of the Earth’s tectonic plates. Other less common causes are the human-induced earthquakes or those generated through volcanic activity. In either case, the energy is released in the form of multi-directional waves, which reach the surface, causing different effects. However, the intensity of an earthquake is not uniform in all its propagating directions. Many times, the motion is polarized due to the type of fault and/or the proximity to it, causing higher intensities in specific directions, depending on the dynamics and geometry of the rupture. This is what is known as the directivity effect. Furthermore, both the intensity and the shape of the wave vary depending on the propagation medium. Ground motion prediction models deal with the spread of the released energy from source to site. Local site effects, both soil effects and topographical effects, are also important. Rigid media, such as rocky and stiff soils, do not tend to amplify the seismic motion, while soft soils amplify specific frequencies depending on local sub-soil geology and on the motion characteristics. Directionality effects refer to the strong motion in a specific site. This thesis deals with two important issues related to directionality. The first one refers to the orientation of the sensors recording the seismic actions; the second one refers to the expected damage in buildings depending on the directions of their strong and weak main axes. It is worth to mention that nowadays, directionality effects are not considered in most structural regulations. In this thesis, special attention is paid to the directionality and soil effects.
Since 2008, around 360,000 earthquake fatalities have been reported. This evidence demonstrates the need to develop more and better ways to assess and to prevent seismic risk. Therefore, the main objective of this thesis is to identify and evaluate the strong-motion directionality and the soils’ effects on the seismic hazard and risk, with applications to strong-motion data sets and soils’ in urban environments. This thesis is divided into three principal blocks: I) directionality effects, II) Soils effects, site classification and other seismic risk-related issues and, III) relevant case studies related to the previous two blocks. In the first block, directionality effects have been considered in the expected strong seismic actions, through the estimation of intensity measures using databases from Italy and Costa Rica. Also, in the assessment of the expected damage of buildings through non-linear dynamic analyses, a simplified approach has been proposed to consider directionality effects. In the second block, microzonation and soil-building resonance effects in the city of Barcelona are studied. In addition, a seismic site classification is defined for the Spanish strong-motion network. The dynamic soil-structure interaction, considering directionality effects and, the proposal of a new drift-correlated intensity measure, appeared as supplementary subjects in this block. Finally, in the third block, other relevant contributions were included to complement this dissertation. The results demonstrate 1) that directionality effects in expected seismic actions are significant and should be considered in Probabilistic Seismic Hazard Analysis (PSHA) and in seismic risk assessments; and 2) they confirm the relevance that site effects (soil effects), has both in seismic hazard studies and in the assessment of the expected damage. This PhD thesis wants to be an additional step towards the assessment, prevention, and reduction of the risk due to earthquakes.
Keywords: earthquake engineering; engineering seismology; directionality effects; site effects; damage assessment; site classification; ground-motions; intensity measures; soil-structure interaction; microzonation; seismic hazard; seismic risk; horizontal-to-vertical spectral ratios; fundamental periods; inter-story drift; soils’ effects; strong motion site characterization.
UNESCO codes: 2507.05; 3305.32.


Índice

1  Introduction 1
1.1  Background 1
1.2  Objectives 4
1.2.1  Block I. Directionality effects 4
1.2.2  Block II. Soils’ effects, site classification and other seismic risk related issues 5
1.2.3  Block III. Other relevant contribution 6
1.3  Methdology 6
1.3.1  Block I. Directionality effects 6
1.3.2  Block II. Soils effects, site classification and other seismic risk related issues 7
1.3.3  Block III. other relevant contributions 8
1.4  Thesis layout 9
1.5  Brief summary of chapters 12
2  Directionality models from ground motions of Italy 15
2.1  Abstract 15
2.2  Introduction 15
2.3  Time histories 16
2.3.1  As-recorded time histories 16
2.3.2  Response time histories 16
2.4  Directionality analysis 18
2.4.1  Rotated (Rot) time histories 18
2.4.2  Peak values and IMs 18
2.4.3  Rotation dependent (RotD) and independent (RotI) 20
2.4.4  Main IMs 21
2.4.5  IMs comparison 24
2.5  Directionality in Italy 24
2.5.1  Data 25
2.5.2  Results  26
2.5.3  Model 29
2.6  Summary, discussion and conclusions 31
3  A Strong-Motion Database of Costa Rica: 20 Years of Digital Records 33
3.1  Abstract 33
3.2  Introduction 33
3.3  Strong-Motion Network 34
3.4  Strong-Motion Database 35
3.5  Data processing 36
3.6  Intensity measures 38
3.7  Conclusions 41
4  An efficient method for considering the directionality effect of earthquakes on structures 43
4.1  Abstract 43
4.2  Introduction 43
4.3  Methodology 44
4.3.1  Complete rotational method 44
4.3.2  Predictive percentile method 45
4.3.3  Spectral matching technique 47
4.4  Building’s application model 48
4.5  Ground motions selection  50
4.6  Results 50
4.6.1  Regular building 50
4.6.2  Irregular building 54
4.6.3  Performance in the non-linear range  60
4.6.4  Near- and far-fault earthquakes 60
4.7  Discussion and conclusions 64
5  Increased seismic hazard in Barcelona (Spain) due to soil-building resonance effects 67
5.1  Abstract 67
5.2  Introduction 67
5.3  Building period map 69
5.4  Soil period map 71
5.5  Crossing period maps 71
5.6  Discussion and Conclusions 75
6  Seismic site classification from the horizontal-to-vertical response spectral ratios: use of the Spanish strong motion database 79
6.1  Abstract 79
6.2  Introduction 79
6.3  Methodology 83
6.4  Data 83
6.5  Results  85
6.6  Conclusions 90
7  Dynamic soil-structure interaction analyses considering directionality effects 93
7.1  Abstract 93
7.2  Introduction 93
7.3  Case study 94
7.3.1  Input motions 95
7.3.2  Surface structure 97
7.3.3  Main features of the finite element model 98
7.4  Complete rotational approach – CRA 100
7.5  Simplified rotational approach – SRA 102
7.5.1  Free-field seismic actions computations 102
7.5.2  Intensity measures 103
7.5.3  Peak ground acceleration, peak ground velocity, and PGV/PGA ratio 105
7.5.4  Arias intensity-AI 106
7.5.5  Specific energy density-SED 106
7.5.6  Cumulative absolute velocity-CAV 108
7.5.7  Maximum spectral response 108
7.5.8  Spectral response for the fundamental period of the building-SATf 109
7.6  CRA and SRA comparison 110
7.7  Conclusions 114
8  A drift-correlated ground motion intensity measure: application to steel framebuildings 117
8.1  Abstract 117
8.2  Introduction 117
8.3  Intensity measures 118
8.3.1  Intensity measures based on ground motion time histories 119
8.3.2  Intensity measures based on peak responses 120
8.4  Structural models 120
8.5  Ground motion records 122
8.6  Correlation between the IMs and the MIDR 124
8.7  New IM 124
8.8  Correlation between IMs 128
8.9  Validation through probabilistic NLDA 128
8.10 Conclusions 134
9  Discussion and conclusions 135
9.1  Main results and discussion 135
9.2  Recommendations and future trends 139
9.3  Conclusions and final remarks 142
Bibliography 143
A Appendix. Chapter 4 tables. 161
B Appendix. A simplified approach to account for directionality effects on 2D dynamic soil-structure interaction analysis 167
B.1  Abstract 167
B.2  Introduction 167
B.3  Methodology 169
B.3.1  Complete rotational approach (CRA) 169
B.3.2  Simplified rotational approach (SRA) 169
B.4  Finite element model 171
B.5  Results 172
B.6  Conclusions 173
C Appendix.  Case study of a heavily-damaged building during the 2016 MW 7.8 Ecuador earthquake: Directionality effects in seismic actions and damage assessment 175
C.1  Abstract 175
C.2  Introduction 175
C.3  Directionality in ground motions 176
C.4  Directionality in the spectral response 182
C.5  Damage assessment of a heavily-damage building 183
C.5.1  Studied building 183
C.5.2  Parametric model 186
C.5.3  Park and Ang based damage index 188
C.5.4  Fragility curves and mean damage state 188
C.5.5  Seismic actions 191
C.5.6  Building-seismic action interaction 192
C.6  Conclusions 196
D Appendix. Seismic site classification of the Costa Rican Strong-Motion Networkbased on VS30 measurements and site fundamental period 199
D.1  Abstract 199
D.2  Introduction 199
D.3  Average shear-wave velocity of the top 30 m (VS30) 200
D.4  Fundamental period through HVSR 201
D.5VS30and site fundamental period relation 201
D.6  CRSM site characterization 203
D.7  Conclusions 205
E Appendix. Do Directionality Effects Influence Expected Damage? A Case Studyof the 2017 Central Mexico Earthquake 207
E.1  Abstract 207
E.2  Introduction 207
E.3  Mexico City and the 2017 Central earthquake 208
E.3.1  Seismic hazard in Mexico City 208
E.3.2  The 2017 Central Mexico earthquake 208
E.4  The case of the Tlalpan 1C building  208
E.4.1  Structural analysis 210
E.5  Phenomena related to anomalous performance of buildings 212
E.5.1  Geological soft soil (GSS) effects 212
E.5.2  Soil-structure interaction (SSI) 213
E.5.3  Site-city interaction (SCI) 214
E.6  Directionality effects 214
E.6.1  Effects on the 1C building 215
E.6.2  Ground motions and design spectra 216
E.6.3  Overall directionality effects 218
E.7  Conclusions 220
E.8  Data and Resources 221
F Appendix. Reply to “Comment on ‘Do Directionality Effects Influence ExpectedDamage? A Case Study of the 2017 Central Mexico Earthquake’  223
F.1  Introduction 223
F.2  Collapse of a building in the ISSSTE multifamily complex 223
F.3  Directionality effects and earthquake-resistant design 226
F.4  Closure 226


Conclusiones

Although, this PhD thesis has been structured in three main parts, or blocks, all of them concern to important issues related to seismic risk, which is the glue, joining them, in an integrated environment, to provide new insights, methods and tools for seismic risk assessments, guided to earthquake protection and prevention. The words “prevention” and “protection” have been taken from the nice book by Coburn and Spence (2002).
The main issues tackled are related to: I) directionality effects in expected seismic actions and in the expected damage in buildings, II) microzonation and soil-building resonance effects, seismic site classification of accelerometric stations and directionality effects in the dynamic soil-structure interaction. In this second part a new intensity measure which is well correlated with drift, has been proposed, too. Four case studies related respectively to soil-structure interaction, to directionality effects and to site classification of accelerometric stations are included in the part III of this dissertation. One of the reasons for this structure of the document is the fact that each chapter and appendix is supported by a publication, mainly journal paper, but also proceedings of relevant meetings and internal reports. Therefore, the document has been prepared in the format foreseen for article-based-PhD-thesis at the Polytechnic University of Catalonia.
In this last chapter, after a short review of, and a discussion on the main results obtained, the main conclusions, coming from the analyses done, are highlighted. Also, quite a few of remaining open questions are pointed out and some suggestions for future research are given.
Main results and discussion
Significant and valuable results concerning directionality effects in expected seismic actions (Chapters 2 and 3, Block I.1, and Appendix C, Block III), are the following:
– Directionality effects in Italy and in Costa Rica have been investigated using large and wide strong ground motion databases. New sensor orientation-independent IMs, which are based on the maximum values of the time history obtained by a previous combination the two horizontal components, rather than on the combination of the two maximum values of each component, have been proposed. These new IMs have been called LRotDpp for the Larger measure, and mpGM, mpGMRotDpp, and mpGMRotDpp for the IMs involving geometric mean values.
– Geometric mean of the peak values (GM) is conservative in comparison to the mpGM. Also, it has been found that vector composition of the time histories, mpVC, is identical to RotD100 and LRotD100. In fact, mpVC, RotD100, and LRotD100 can be considered as the maximum response, corresponding to the most unfavourable case. These IMs would be useful for the design or risk assessments of structures of special importance such as historical-cultural heritage buildings or other high-risk constructions.
– From statistical analyses of great databases of strong motion recordings, as the ones of Italy and Costa Rica, it has been shown that, in average, the impact of the directionality effects on existing GMPE developed without considering directionality effects, although significant, is not very high. More specifically, the ratios GMRotD50/GM GMRotI50/GM and RotD50/GM, take values between 1.01 and 1.02 in the range of periods 0.01-2 s but, the ratio RotD50/GM takes values around 1.015 in the range of periods 0.1-1.0 s and values between 1.02 and 1.04 in the range of periods between 0.1 and 2 s. Anyhow, ratios between other intensity measures as, for instance, Larger/GM and LRotD50/GM take values between 1.15 and 1.20, but LRotD100/GM takes values in the range between 1.23 and 1.30. For details on other ratios, please see Chapters 2 and 3, where the specific studies for Italy and Costa Rica are detailed.
– Thus, the sensor orientation-independent IMs, GMRotD50, GMRotI50, and RotD50 are generally larger than the as-recorded GM, and, although the differences between GMRotD50, GMRotI50, RotD50, and GM are small, orientation-independent measures should be used as they reduce epistemic uncertainties in the ground motion predictive equations (GMPE).
– The results obtained in this study are of significant interest in the improvement of the GMPEs in PSHA and in seismic design and risk assessments. In this respect, a simple theoretical model has been developed. This directionality model for Italy can be used to improve existing GMPEs as it allows us to include directionality effects, by means of simple ratios, into equations obtained without any consideration of directionality.
– Directionality effects are more important for low magnitude near events.
Concerning general analyses related to directionality effects on buildings (Chapter 4, Block I2. a), main results are:
– When one deals with a single accelerometric station, directionality may be very important, producing a great variability with the orientation of the two orthogonal horizontal components. This is why the expected damage in a specific building, submitted to a concrete two-component horizontal acceleration time histories, may have a great variability.
– A simplified approach to consider directionality effects on structures has been presented (see Chapter 4). The results indicate that neglecting the directionality effect could underestimate the actual demand about 30%. This fact reaffirms the importance of considering the directionality effect in the structural analysis.
– The predictive percentile method (PPM) provides a simplified alternative to obtain any percentile response without the need of performing hundreds of NLDA; this represents a decrease from 180, if the increment for rotating the records is 1º, to only two NLDA. The values obtained with the PPM are very close to the ones obtained through the complete rotational method. Noteworthy, the differences, in average, are less than 5%.
– Directionality effects should be regarded in structural seismic codes too, since it allows to improve the evaluation of the expected seismic damage in structures. For instance, for special facilities such as nuclear plants, hospitals, schools and other facilities that need to be operating after an earthquake, the RotD100 matched accelerograms should be used to estimate the seismic demand. This way, it will be guaranteed that the maximum seismic demand is considered in the structural analyses.
In reference to MW 7.8 Ecuador earthquake occurred on April 16, 2016 earthquake (Appendix C, Block III):
– Directionality effects have been evaluated, in both orthogonal directions of a seriously damaged building.
– The ground motion records at a close accelerometric station, located 5 km away from the building, has been used to estimate the expected damages. The directionality analysis has shown that probably the directionality effect was responsible for an increased damage because in spite of both main axes of the building showed similar strength, a strongest seismic action, hit one of them.
– A very good agreement has been found between the numerical analyses and the damage observed.
– This study, has allowed us to test the parametric model, and to verify the robustness and usefulness of the Park and Ang damage index.
– The importance and relevance of considering directionality effects in expected damage and risk assessments of specific building has been clearly highlighted.
– Directionality effects on the peak ground-motion produced by this event in the country, were also evaluated, obtaining nice and relevant results. This event reaffirms the need and importance of keeping developing research to reduce the seismic risk in Ecuador.
Regarding to the case study of a building severely damaged during the 19 September MW 7.1, 2017 Mexico earthquake (Appendixes E and F, Block III):
– The anomalous seismic performance of a specific building in a multifamily residential area, which was the only one collapsed in the complex, was mainly attributed to directionality effects. In this case, assuming seismic actions recorded close to the building, the direction producing the strongest seismic action was the same as the one of the weak axis of the building.
– Thus, in damage and risk assessments, the direction in which the strongest seismic actions hit the buildings, directionality, should be considered, as similar buildings, located in the same place, may suffer very different damage grades, depending only of their specific orientation.
– In addition, concerning seismic hazard, the results show how the response spectra predicted by the SASID A v4.0.2.0 application, are lower than those corresponding to the seismic actions produced by this earthquake. This fact confirms that it is important to incorporate the results of directionality studies into the GMPEs by means of sensor orientation-independent measures.
– Anyhow, the consideration of maximum seismic actions could lead to excessively conservative GMPE. Median values or specific percentiles should be considered.
– Regarding design spectra, seismic regulations in Mexico City have been improved in recent years. However, latter design spectra were still surpassed by several accelerograms recorded during this earthquake. Noticeable, these excesses were due to directionality effects.
– Thus, an important conclusion of this study is that directionality effects must be considered in PSHA, in damage assessments, and in design regulations. Specific studies on directionality effects should be performed in urban areas located in high seismic hazard zones. However, studies undertaken in different countries may be also useful as the ratios RotD100/GM and RotD100/Larger, found are similar among them.
Concerning microzonation studies in Barcelona city (Chapter 5, Block II.1a):
– Previous work done in the city has been revisited and reviewed. An updated microzonation map of soils have been obtained. Also, a detailed map of predominant periods of buildings has been depicted. This way, crossing both maps, areas where soil-building resonance phenomena are expected has been highlighted. These maps can help to prevent and to reduce the seismic risk, since special attention can be devoted to vulnerable structures located on areas where soil-building resonance effects are likely (See Chapter 5 for details).
– Old towns located in the present-day neighbourhoods of Gràcia, Poble Sec, Sant Antoni, El Raval, and Sants are the most affected in the city. buildings in these neighbourhoods were built at the same period of time and have similar structural typologies. Most of them were constructed in the late 19th century and the predominant typology is unreinforced masonry. This structural typology is one of the most commonly used buildings worldwide and the one that has caused the most damage and death, due to its high vulnerability to earthquakes.
– The microzonation map, based on predominant soil periods, can be implemented in structural regulations to develop improved design spectra based on the site-specific spectral response. Some structural codes do not consider SSI effects, since in most cases it is more conservative to use structural regulations. However, the effect of SSI may be prominent when structure and soil periods are similar.
In reference to the site classification of accelerometric stations (Chapter 6, Block II.1b and Appendix D, Block III):
– The Spanish strong-motion database has been used to classify the site of each accelerometric station, of the Spanish strong motion network, by means of the predominant period obtained using H/V ratios of the 5% damped response spectra. Directionality effects were considered too. This study strongly enforces the knowledge on the network which beforehand was based only on geological maps (see Chapter 6 for details).
– The sites of the accelerometric stations of the Costa Rica strong motion recording network has been also classified (see Appendix D).
– These classifications provide several advantages compared to others. The acquisition and processing of data are simple and automatic since network work continuously. This methodology may be easily exported to other networks, depending on the availability of strong motion databases. Moreover, this classification considers the real response of the site to actual seismic actions. This way, more realistic site conditions are obtained in comparison to the more general, simplified, and proxy classifications, which are used in seismic codes.
– Moreover, additional information on the likely anisotropy of both the strong motion and the soil response is also available, since directionality effects have been considered, and, for the same reason, epistemic uncertainties decrease because all the orientations are considered in the analysis.
– Finally, from an engineering point of view, this classification plays an important role in the evaluation and design of structures, which can be useful, for instance, when actual accelerograms are needed for deterministic and/or probabilistic dynamic analyses, thus improving the results for specific areas and contributing to reducing epistemic uncertainties.
With respect to dynamic soil-structure interaction (DSSI) analyses (Chapter 7 in Block II.2a and Appendix B in Block III):
– The importance of considering directionality effects of ground motions in 2D DSSI analyses has been highlighted. An important shortage in DSSI analyses, usually performed, is that only the as-recorded components of ground motions are considered, without any consideration to other rotated time-histories that could lead to more severe structural responses.
– It has been shown that directionality effects can be properly addressed by analysing the considered problem with a number of linear combinations of the input motions for different incidence angles. Then, the orientation producing the maximum response of the structure can be determined; this procedure was termed the CRA. If directionality effects are overlooked and only the as-recorded components are employed to derive the input motions for the analyses, the maximum relative roof displacements of the building can be underestimated as much as 50%.
– DSSI analyses are computationally expensive preventing the use of the CRA in most of the cases. Already established simplified procedures to incorporate directionality effects for structural analyses cannot be employed directly with the input motions for DSSI analyses, since site effects tend to modify the input. Thus, a simplified procedure has been proposed here to identify, in advance, the angle that produces the maximum response of a structure for a given seismic action and, therefore, allowing us to get good results with only one DSSI analysis. Site effects are incorporated by applying the CRA to free-field site response simulations, since they are quite inexpensive compared to full DSSI analyses. Then, the critical orientation is predicted from the maximum value of a given IM, as a function of the rotation angle, from the motions computed at the surface of the free-field simulations.
– Different IMs were evaluated, and the PGV resulted in critical angles and relative roof displacements closest to the CRA. Although the exact values were not exactly predicted, the mean relative differences with respect to the CRA results, in terms of roof displacements, were smaller than those obtained using the as-recorded directions.
– Although directionality effects should always be considered for earthquake-resistant design and seismic damage and risk assessments, it is important to mention that using the most unfavourable incidence angle, in each of the selected ground-motion records, could lead to over-conservative results.
Concerning drift-correlated intensity measures (Chapter 8, Block II.2b):
– From a correlation analysis between the maximum inter-story drift (MIDR) and a group of well-known IMs, related to the destructiveness potential of earthquakes, it was found an important correlation between MIDR, PGV and the significant duration.
– Then a new intensity measure, which is highly correlated with the MIDR, called I?-PGV, has been proposed.
– The building-to-building structural characteristic variations were considered by means of a probabilistic approach, varying the geometry and mechanical properties of the building models to perform 500 probabilistic NLDA. This way the new IM, I?-PGV, was validated. An important finding and one of the main advantages of I?-PGV is that this new IM is independent of the building fundamental period.
Recommendations and future trends
The knowledge on topics covered in this PhD thesis is, nowadays, relatively limited, so that these topics keep on having relevant interest in engineering seismology and in earthquake engineering, as well. Research never stops, each issue addressed opens the door to new exciting dares deserving to be investigated. This way, at the end of each research stage there emerge methods and procedures which are worthy of being improved and new questions deserving to be answered. This is why the present backgrounds, tools, resources and attainments are far from the ones which were available only several years ago. This section is devoted to point out and to highlight spots being the natural development and maturation of the job done and to propose new trends in the same road of the issues investigated. Also, several ideas and guidelines for future research are depicted.
Concerning directionality in expected seismic actions, several suggestions follow:
– The quantity, quality and availability of strong motion databases, is increasing day by day. It is very important to keep on this effort in the maintenance of such big-data.
– To analyse more databases would help to improve and to refine the lessons and conclusions coming of this type of studies.
– The impact on ground-motion predictive equations and on PSHA of the new intensity measures based on the prior-combination of the acceleration time histories, which in our opinion has more physical sense, should be investigated deeper. A criterion of the goodness of the fits may be the dispersion in the predictive equations.
– An interesting issue would be to model the directionality effects from source to site, including the fault rupture and directivity. It would be interesting to study the similarities/dissimilarities of directionality effects for normal-, strike-slip-, thrust type faults.
– The variation of directionality effects with the distance to the causing fault is considered worthy to be investigated too. Very likely, seismic waves responsible for the strong motion may be different at short and long distances.
– The influence of site effects on directionality is another interesting issue. In this study, directionality effects in different soil conditions have been analysed. However, the fact that results obtained by others did not considered soil’s effects advised us to do the comparison without considering them. Further comparative analyses considering site effects are recommended.
Concerning directionality effects on expected seismic damage on specific structures subjected to a specific strong motion, the following hints are highlighted:
– Concerning the simplified method to consider the directionality effects on structures (see Chapter 4), it is worth to mention that this approach should be proved with different structural typologies, building shapes and heights, among other properties, that may influence the dynamic behaviour. More work is needed to go towards more general conclusions.
– Few studies have been carried out on effects of the azimuthal orientation of buildings in the expected damage. It seems clear that if the strong action hits the weak axis the damage will be greater. However, it would be interesting to see how damage probability matrices corresponding to both main axes are combined to get a global damage probability matrix. A reasonable assumption could be to use the Bayes theorem concerning to prior knowledge of conditions related to the event. To check for this assumption, to combine 2D non-linear static analyses and 3D incremental dynamic analyses is suggested.
– In this sense, it is proposed to analyze different 3D building configurations, having high variability in the fundamental periods in the directions of their main axes, with the aim of evaluating their susceptibility to directionality effects. It would be interesting as well to examine and to compare the response of such buildings to different seismic demands, through nonlinear dynamic analyses, using, for instance, only one acceleration component and the two orthogonal components, simultaneously. It is also suggested to analyze more realistic situations considering the vertical acceleration in addition to the two horizontal ones.
Concerning microzonation studies:
– In many cities, located in low-to-moderate seismic hazard regions, like Barcelona is, the seismic risk may become high because of the huge population density, the great exposed value and the high vulnerability of the built environment. Seismic risk studies in those cities are strongly recommended. Moreover, in these very dense cites, it is important to determine likely soil-structure-resonant areas.
– In dense cities too, it is strongly recommended to analyse Soil-Structure-Interaction (SSI) and Soil-City-Interaction (SCI) effects. The SCI effect is defined as the global interaction between urban fabric and its subsoil through ground motion Bard et al. (2005). These multiple interactions may modify the response of the buildings and the one of the subsoils itself. If SCI effects turn out to be significant, one immediate consequence is that erecting or destroying a building, or a group of buildings, could modify the seismic hazard for the neighbourhood affected.
– Macau et al. (2015) found that H/V measures in shallow quaternary layers in the delta of the Llobregat river at the south-eastern part of Barcelona, exhibit a second peak at which significant amplification may occur. Additional research on this finding and on its influence on buildings’ response is suggested.
Concerning site classifications:
– The procedure for classification of the site of accelerometric stations, which has been applied to accelerometric networks of Spain and Costa Rica, can be easily implemented to other networks.
– It would be interesting to go further in the directionality related issues and, to study the anisotropy of both the strong motion and the soil response. It would be challenging also to evaluate this anisotropy in the non-linear range.
– The impact of this site classifications on the development of new GMPEs, in comparison to the more traditional ones based on more simplified approaches that consider the local site effects based on conventional rock/soil classifications, remains as a relevant and interesting issue.
Related to the Dynamic Soil-Structure Interaction (DSSI) analyses:
– The procedure presented in this study that incorporates directionality effects in DSSI analysis, requires further validation by means of new analyses under different conditions, such as different characteristics of the soil deposit, additional seismic actions, corresponding to different kinds of earthquakes, and, of course, different surface structures.
– More analyses should be performed considering 3D models and the non-linear properties and expected damage of the structures.
In reference to the new Drift-correlated intensity measure (I?-PGV):
– I?-PGV is a linear combination of two well-known seismic parameters: peak ground velocity (PGV) and significant duration. The use of more than one parameter to define the potential damage of earthquakes to structures is a noteworthy novelty. This finding suggests that alternative methods, such as for instance Artificial Intelligence and Neural Networks, can be used in future similar studies.
– I?-PGV was validated for 2D symmetric steel frame buildings. Further analyses, with other structural typologies and 3D asymmetric buildings, are needed to establish its broad soundness.
– As well, coefficients, a and ß, of the linear combination, should be recalibrated with other building types. If this intensity measure holds for different structural types, it would be very useful and effective for seismic hazard and risk assessments.
– One of the main advantages of I?-PGV is that it is independent of the fundamental period of the building. In future research, the potential of this IM should be compared with other IMs that consider the structural behaviour by means of the elongation of the buildings’ periods.
Conclusions and final remarks
There are many remarks and conclusions that can be inferred from the work done, mainly because the extensive number of issues tackled in this dissertation. For detailed conclusions and remarks on the different topics, the reader is invited to have a look into the corresponding paper, which forms each chapter. Here, only a short number of conclusions remarks are stressed.
– Directionality effects in expected seismic actions are relevant. For probabilistic seismic hazard analysis, predictive equations of strong ground motion can and should be improved.
– Other ways of combining maximum IMs or peak values, as for instance, the ones based on prior combinations of the time-histories, may also help to have more realistic estimates, and with more physical sense, of the expected seismic actions.
– Expected damage in a specific building strongly may depend on the azimuthal orientations of its main axes with respect to the strong motion in the site.
– Even when located in low-to-moderate seismic hazard regions, detailed risk assessments must be done in big cities, including detailed microzonation analyses, soil-building resonant zones detection, SSI and SCI studies. Because of the high population density, the high vulnerability of the built environment, and because of the high concentration of facilities, infrastructures and services, the seismic risk of these cities may become very high.
– With the availability of huge databases of strong motion recordings, site characterization of accelerometric stations can be done in an easy and expedite way. Additional value studies concerning anisotropy of the site response and the one of the seismic actions can also be tackled.
– The available big-data concerning to accelerograms and the increasing computational power of modern devices and tools, also allows us to search for more and more reliable IMs related to drift, so that they may become excellent forecasts of the seismic performance of buildings and of damage expected. Artificial intelligence emerges as an interesting means to search for new IMs, connecting highly correlated parameters of seismic actions to those of the structural performance and response.
Final thought
And that’s all, it is over…
Unfortunately, earthquakes are turning on great catastrophes, and they will be keeping on doing that.  Many are the reasons for this apparent failure of the present-day civil engineering community. But, as it has been shown in recent earthquakes, like the 2010 Haiti, MW 7.7, earthquake, the dimension of an earthquake disaster goes beyond technical aspects, involving many other socioeconomic and, even, political issues. Discrimination, poverty and societal inequalities greatly increase the catastrophic breadth of earthquakes. In any case, the main motivation of my PhD research has been to provide new insights in the right road towards more knowledge-based, sustainable, safe and resilient communities around the planet Earth, for sure, the most beautiful of our solar system but, also the one most needing special care to protect the wonderful life it shelters.