Seismic Vulnerability of Structures

By Philippe Gueguen

This book is focused on the seismic vulnerability assessment methods, applied to existing buildings, describing several behaviors and new approaches for assessment on a large scale (urban area).
It is clear that the majority of urban centers are composed of old buildings, designed according to concepts and rules that are inadequate to the seismic context. How to assess the vulnerability of existing buildings is an essential step to improve the management of seismic risk and its prevention policy. After some key reminders, this book describes seismic vulnerability methods applied to a large number of structures (buildings and bridges) in moderate (France, Switzerland) and strong seismic prone regions (Italy, Greece).

Contents

1. Seismic Vulnerability of Existing Buildings: Observational and Mechanical Approaches for Application in Urban Areas, Sergio Lagomarsino and Serena Cattari.
2. Mechanical Methods: Fragility Curves and Pushover Analysis, Caterina Negulescu and Pierre Gehl.
3. Seismic Vulnerability and Loss Assessment for Buildings in Greece, Andreas J. Kappos.
4. Experimental Method: Contribution of Ambient Vibration Recordings to the Vulnerability Assessment, Clotaire Michel and Philippe Guéguen.
5. Numerical Model: Simplified Strategies for Vulnerability Seismic Assessment of Existing Structures, Cédric Desprez, Panagiotis Kotronis and Stéphane Grange.
6. Approach Based on the Risk Used in Switzerland, Pierino Lestuzzi.
7. Preliminary Evaluation of the Seismic Vulnerability of Existing Bridges, Denis Davi.

About the Authors

Philippe Guéguen is a Senior IFSTTAR Researcher at ISTerre, Joseph Fourier University Grenoble 1, France

• Language: English
• Publisher: Wiley
• Release date: Mar 5, 2013
• ISBN: 9781118604007

Book preview

Introduction

Earthquakes are one of the natural phenomena that, for a long time, has significantly affected the imagination of human beings. Indeed, earthquakes are sharp and sudden, and in a few moments the victims can be counted in thousands. They bore an even more mysterious character as they would undermine the innate beliefs of man in an unmoving Earth. Even if the physical origins of earthquakes are better understood today, the power of their vibrations is still sometimes astonishing. Those who have experienced a moderate or strong earthquake in a reinforced concrete building are frightened by the ease with which the walls and floors oscillate; this disturbs the faith of modern man in the robustness of the constructions of reinforced concrete.

No other natural forces can cause, in such a short time span, as much damage and lead to as many victims as earthquakes do. In the most recent catastophic examples, such as Kobe (Japan, 1995, M = 7.3), Izmit (Turkey, 1999, M = 7.6), Boumerdès (Algeria, 2003 M = 6.7), Kashmir (Pakistan, 2005, M = 7.6), Sichuan (China, 2008, M = 7.9) and even Haiti (Haiti, 2010, M = 7.0), earthquakes demonstrate the weakness of urban environments relative to the destructive power of these events. Wherever we are, the same observations are made: the weaker buildings suffer a lot of damage, the old constructions made up of earth or masonry resist the least, schools often greatly suffer from the ground vibrations, the zones of destruction are very scattered without any clear geographic distribution and populations are often taken by surprise. However, there is a hidden logic behind these general and repetitive observations, which if better understood and controlled, could allow us to reduce the impact of these earthquakes on urban areas.

Already, in his lifetime, Rousseau had spotted the urban incoherence of Lisbon in 1755 by explaining that if we hadn’t gathered here the twenty thousand houses of six to seven storys and if the inhabitants of the large city had been more evenly spread, and not as heavily burdened, the damage would have been much smaller, and maybe absent. With this sentence, Rousseau sums up all the observations made since the Lisbon earthquake, for each event. He clearly expresses the anthropism of all hazards that are said to be natural. Avoiding an excessive rousseauization that would make man bear the responsibility for all natural disasters, we observe a strong link between the phenomenon, the action of man and the disaster; this is the classic relationship with which every presentation addressing the notions of risk, hazard and vulnerability generally begins: R = H.V.E, a definition of which was given on the occasion of the International Decade for Natural Disaster Reduction [DIP 92].

In the above formula:

– R represents the risk, in other words the mathematical expectation of loss in human lives, injuries, damage to goods and effects on the economic activity during a reference period and in a given area, for a particular hazard.

– H is the hazard, the threatening event or the probability of occurrence in a region and during a given period, of a phenomenon capable of causing damage.

– V represents the vulnerability that we represent in degrees of loss (from 0% to 100%) of an element at risk resulting from a phenomenon susceptible of causing casualties and material damage.

– E represents the elements exposed or the elements at risk, or the Population, the civil engineering constructions and structures, the economic activities, the public services and infrastructures, etc., exposed to a hazard.

We could debate these definitions and their outlines, according to whether we take sides with the seismologist, the engineer or the sociologist. However, Coburn and Spence [COB 02] remind us that over the last century, the cost of earthquakes, in terms of values in the year 2000, is in the order of a thousand billion dollars. Calculated per year, they observed that this value increased during the 20th Century, essentially due to the increase in and concentration of populations in large urban areas exposed to a strong seismic hazard.

Indeed, over the last century, the hazard or even the number of earthquakes per year has neither increased nor decreased. Only the vulnerability of environments has varied. Approximately 50,000 earthquakes occur on average per year, as a result of the motion of the Earth’s tectonic plates. Of these 50,000 events, a few are of potentially devastating magnitude. Despite the improvement of our knowledge since the confirmation of the motion of tectonic plates suggested in 1912 by Wegener and confirmed by Hess [HES 62] and Dietz [DIE 61], it is still impossible to know exactly where and when the next large earthquakes will occur. However, what we are aware of is the huge increase in urban population size situated along the active seismic faults, which increases the probability that the future disasters will go beyond that of San Francisco or Tokyo. Considering the time between occurrences of the largest earthquakes, Jackson [JAC 06] affirms that larger seismic catastrophes are yet to come, and there are hardly any urban zones which in their present configuration that have suffered from these major events. Not only the number of people exposed to high seismic hazard is higher than ever before, but also the concentration of the wealth and modern infrastructure in the megacities could lead to infinitely more devastating effects in economic terms than the Kobe earthquake in 1995, which caused $100 billion in loss, making it one of the most costly natural hazards of all time.

Regarding natural disasters, the year 2011 will have been perhaps for companies most costly in modern history. For the reassurance companies Munich Re and Swiss Re, the economic losses linked to tectonic hazards, including tsunamis and earthquakes, could reach €276–300 billion for that year. It was a record, which was amplified by the fact that they do not take into account the costs induced by industrial catastrophes such as Fukushima and, which could have been higher if Japan had been better insured against earthquakes. Indeed, the Japan earthquake contributed largely to this number, this mega-event having an unreal aspect to it. Nevertheless, the New Zealand disaster, more modest (M = 6.3), which devastated Christchurch, also adds to the sum, raising the losses up to €13 billion of, which, €10 billion were covered by insurance.

This earthquake, just like other recent earthquakes in L’Aquila in Italy and Lorca in Spain, reminds us that moderate earthquakes (a magnitude of around 6) can sometimes cause important damage and produce losses, even in moderate seismic prone regions. It is the case, for example, of the Au Sable Forks earthquake (New York, April 20, 2002, Mw = 5.0) that caused damage requiring repairs for which the cost has been estimated at $15 million [PIE 04]. In France, we are not safe from this, as shown by the Ossau-Arudy earthquake of 1980 (ML = 5.1) and the Annecy earthquake of 1996 (ML = 4.8) that, despite low magnitudes, respectively, caused around €3 million [MED 82] and €50 million of damage [AFP 96]. Of course, the major historic event of the 20th Century that strongly affected the rural region in the southeast of France must not be forgotten. In 1982, this earthquake, called Lambesc’s earthquake (magnitude around 6) served as the benchmark to assess losses and fatalities were it to occur again. This simulation concludes on major direct and indirect impacts of earthquakes on constructions, on the life of human beings and the economy, this region having suffered important socioeconomic transformations since 1909.

It is the collapse of structures that causes death, not the intensity of the earthquakes. Indeed, the relationships [physical damage/losses in human lives] shows this well: there is a strong correlation between the number of victims and that of damaged structures after an earthquake. It is then possible to build structures that resist earthquakes and therefore reduce losses: this is the role of earthquake engineering to understand what seismic ground motion we must protect ourselves against and how to adapt structures to this hazard. Of course, it is only important to define the seismic rule when confronted with earthquakes. For this reason, these rules first appear in the countries that are more exposed to seismic hazard, such as Japan and the United States, closely followed by Italy. Coburn and Spence [COB 02], however, note that the number of casualties placed into two periods of observations, before and after (1950), the approximate date of the first modern earthquake rules, does not show improvement. The greatest cause in all cases is the collapse of masonry structures, in other words the old buildings often built before the application of rules. Also, they also observe that even if there were fewer reinforced concrete buildings before 1950, the number of casualties due to the collapse of these constructions considerably increased after this period. Reinforced concrete, nevertheless, is a material that, combining the properties of concrete with that of steel, presents the best characteristics of resistance to an earthquake. Unfortunately, a wrong estimation of the seismic hazard, a wrong use of the materials, faults or mistakes in production can render these structures vulnerable. Whatever the material used, the existing common building remains very often the weakest link at the time of a seismic crisis.

After 2000, new earthquake regulation appeared on Europe on the definition of seismic hazard and earthquake design, that are found in the Eurocode 8 (EC8). According to EC8, the earthquake design for new construction is defined so as to guarantee the protection of human lives, the limitation of damage to structures and the operational continuity of the important structures for social security. Besides, Ohta et al. [OHT 86] observe that the economic loss following an earthquake will obviously depend on the level of seismicity but mostly on the quality of the constructions and the consented earthquake engineering investment, the latter as referred to in the adapted construction and in the definition of seismic hazards.

After having caused considerable damage and losses, the majority of large earthquakes (e.g. Northridge, 1994; Kobe, 1995; Ismit, 1999, Taïwan, 2001; Boumerdes, 2003; Bam, 2003; Haïti, 2010) remind the politicians and decision-makers that reducing the seismic risk is essential not only for the well-being and the safety of the local populations but also, equally, for the world’s financial and economic equilibrium. This reduction obviously occurs with the reinforcement of the existing structures and by the anticipation of earthquakes in crisis management. Despite this, existing structures, designed without the application of earthquake design rules, are present everywhere. This is, in particular, prevalent in the most part in the historic centers of the European cities, where the seismic damage is usually concentrated. This preoccupation has led to the development of methods for vulnerability evaluation, the main objective of which was to represent the capacity of structures to support the seismic ground motion. Indeed, apart from the economic aspect, the evaluation of the seismic risk requires not only the knowledge of the probable hazard but also a representation of the seismic quality of the structure: it is the objective of the seismic vulnerability assessment methods.

This assessment must yield an estimation of the predictable damage to people and goods, made through an earthquake scenario. This information allows us, for example, to quantify the real level of risk to which the population is exposed, which we can compare or not to other natural phenomena, more frequent and so more appreciable by the populations, particularly those in the countries with moderate seismicity [DUN 12]. This representation can guide the strategy for a community that will, according to its political will, deal with the natural, industrial or domestic risks, which it wishes to invest in and actively act to reduce them [BOU 10]. The representation from different perspectives of the damage over an urban zone also allows anticipating the actions and reactions of the different intervention bodies in the case of a crisis. The means of rescue to put in place right after an earthquake, their intervention and their deployment over the sectors that would be the most damaged rest on the damage simulation. We can then plan ways of improving their efficiency. Finally, the assessement of vulnerability also gives landlords (private or public) information on the most vulnerable constructions that should, as a priority, benefit from reinforcement. This decision is political and is not the result produced from the study of vulnerability that decides the strategy of reinforcement to be put in place. It gives, however, quantitative and/or qualitative elements to integrate like a set of decision processes, as any other constraints to which a landlord is subjected (e.g. keeping up with electrical norms, the treatment of asbestos, accessibility to public buildings for the disabled, etc.). The scale of the work can be on that of a country, of a region, of a town or of a housing area (e.g. the schools and the strategic buildings). A large-scale study of seismic vulnerability is, therefore, the first step toward the determination of the buildings and the networks requiring a detailed diagnostic and eventual reinforcement.

It is a perilous exercise and a gamble when we decide to work on a town as a whole. The amount of structures and the different types of construction found are difficult to deal with, in particular for European regions that have seen the expansion of their urbanization over several centuries. The knowledge of the behavior of an old building is often impossible. It is, indeed, difficult to assess the conception and the quality of the construction materials used at the time of its construction. It is even harder to try to model the behavior of an old structure not knowing the laws of behavior and the construction principles that were followed in the steps of its construction. Indeed, we know how to evaluate a structure that respects all the constructive dispositions and rules. This is a procedure of designing. To the contrary, how can we evaluate a structure for which the constructive disposition has not been respected completely (or partially) and that cannot be associated with a statutory model of behavior? This is the phase of seismic diagnostic or vulnerability analysis.

It is certainly to reduce these difficulties that many empirical methods have been developed, based on probabilistic or deterministic approaches [CAL 06]. In the majority of cases, these methods have been established on the basis of post-earthquake observations, identifying the levels of damage observed as a function of the typology of the construction. As a result, all these methods (e.g. HAZUS [FEM 03], GNDT [GND 93] and RiskUE [MOU 06]) have been published using data originating from countries with strong seismicity and having recently suffered important damage. The information gathered in situ allows us to define according to the building characteristics and the levels of damage suffered the link between the class of construction and the probable damage for a given level of seismic demand. This way of proceeding, detailed in Chapter 1 of this book, leads to the definition of the fragility curve or the fragility curve.

The high levels of hazard of the regions at the origin of the fragility curves have justified the implementation of methods for which the costs remain difficult to gather for regions at lower seismicity such as France, Spain, Portugal and Switzerland. In addition, the historical and cultural particularities of a region guide the modes of conception and the realization of the houses that we do not necessarily find in the empirical methods. To use them then introduces a strong uncertainty, due to the need to classify each building in an existing behavior model or to regroup each building into a group of generic behavior to deal with the very large number of constructions to be analyzed [SPE 03]. Due to the fact that it rests on historical data, seismic aggression is generally represented in macroseismic intensity, a global scale defining the ground motion and, here again, introducing a strong uncertainty into the estimation of the response of structures.

This uncertainty called epistemic is strong and we can reduce it using detailed mechanical methods presented in Chapter 2. These methods consist of establishing vulnerability functions for a given building model, introducing hypotheses regarding its dynamic response and behavior. The seismic demand is generally defined under the form of a response spectrum and the capacity of the structure to resist is expressed for a given level of performance (or level of damage) and the seismic demand. These methods lead to the establishment of functions of damage for typical structures that we find in our study zone.

In the case of a global scale study (i.e., city, region...), we often face a dilemna. The number of buildings is too great to work with mechanical methods, but the post-seismic information used for empirical methods are not numerous enough to establish the functions of vulnerability, particularly for the highest levels of damage. A hybrid solution is presented in Chapter 3, which combines the empirical approaches for the low levels of damage with the mechanical methods for the higher ones. This solution, applied in the city of Thessaloniki in Greece, is a solution to compensate for the lack of data.

In any case, the definition of a generic behavior of the structure is a critical element. How do we simply and precisely define the behavior of an existing building? How do we know its system of foundation, the quality of the connections between walls and floors, especially when we deal with a very heterogeneous housing area and for which little post-earthquake information is available? It is one of the difficulties with which the countries of moderate hazard are confronted, which despite their low level of hazard need to take into account. However, there exists a solution based on rapid experiments that can be made in buildings and that allow us to evaluate the dynamic characteristics and attribute a generic model of behavior to each tested building. Using the ambient vibrations, this detailed approach, which is discussed in Chapter 4, has the merit of considerably reducing the uncertainty linked to the definition of the model, without requiring hypotheses or approximations. With several examples, led in Grenoble, and presented in Chapter 4, we see the advantage and the way of using this approach, knowing that restrictions regarding its usage should be mentioned. For example, Chapter 5 shows how in situ data allows the validating or the refining of the modelization of the building’s response.

Indeed, for particular structures, numerical methods can be considered. Their cost and complexity reduce their use for buildings of primordial function: this can be the case of buildings that must remain operational in the case of a crisis (hospitals, fire stations, etc.), buildings that are selected to host victims or buildings of which the collapse could have a strong impact on a population. These methods, such as that described in Chapter 5 and applied on Grenoble’s city hall, allows us to represent the distribution of the seismic damage in the structure, to define the locations where it is stronger, then allowing us to accompany and guide, if in need of reinforcement or restoring.

It must not be forgotten that the vulnerability analysis is only worthwhile if we wish to make use of it. A possible action is to program the retrofitting of the most vulnerable structures, at least identified as such, using the different methods defined previously. Regarding the costs of reinforcement, a method has been proposed in Switzerland, which is supposed to help a decision-maker or a house owner to define a reinforcement program. This method, presented in Chapter 6, includes the hazard level, the quality of the structure, the value and the lifetime so as to economically justify the importance in working on it. It provides the estimation of vulnerability according to different levels, refining the priorities according to the hierarchy of the most vulnerable buildings.

Another element is essential for anticipating crisis management in the urban environment: the networks and pathways of communication. Several elements can intervene, from the blocking of streets by rubble to important bridges and road or highway structure, preventing access to the rescue teams to affected areas. In the last chapter of this book, we describe the vulnerability of bridges from the point of view of operational management, shown by two applications in the southeast of France.

Other elements could have been considered here, for example the research of information characterizing large-scale environments with remote sensing methods, the definition of a taxonomy of constructions allowing us to classify them according to the general characteristics, the relationship between the damage and cost of earthquake disasters along with the relationships between the exposed elements and the risk considered at the time of the systematic analyses. This domain is still rich and full of prospects to come, the evolution of which is accompanied by an always more important diffusion of the observed data during earthquakes.

Bibliography

[AFP 96] AFPS, Le séisme d’Epagny (Haute-Savoie) du 15 juillet 1996, Mission report, AFPS Ed., Paris, pp. 128, 1996.

[BOU 10] BOUDIS M., SAILLARD Y., GUÉGUEN P., DAVOINE P.-A., Modèle d’aide à la décision pour la prévention parasismique urbaine. Une approche multi-agent de la vulnérabilité du bâti, European Journal of GIS and Spatial Analysis, vol. 20, no. 3, pp. 279–302, 2010.

[CAL 06] CALVI G.M., PINHO R., MAGENES G., BOMMER J.J., RESTREPO-VELEZ L.F., CROWLEY H., Development of seismic vulnerability assessment methodologies over the past 30 years, Journal of Earthquake Technology, vol. 43, no 3, pp. 75–104, 2006.

[COB 02] COBURN A., SPENCE R., Earthquake Protection, 2nd ed., John Wiley & Sons, Ltd, pp. 420, 2002.

[DIE 61] DIETZ R.S., Continent and ocean basin evolution by spreading of the sea floor, Nature, vol. 190, pp. 854–857, 1961.

[DIP 92] DIPCN., Glossaire international multilingue agréé de termes relatifs à la gestion des catastrophes, Technical report, UN DHA, Geneva, pp. 83, 1992.

[DUN 12] DUNAND F., GUEGUEN P., Comparison between seismic and domestic risk in moderate seismic hazard prone region: the Grenoble City (France) test site, Natural Hazards and Earth System Sciences, vol. 12, pp. 511–526, 2012.

[FEM 03] FEMA, HAZUS-MH Technical Manual, Federal Emergency Management Agency, Washington, DC, 2003.

[GND 93] GNDT, Rischio sismico di edifici pubblici – Parte I: aspetti metodologici, Quasco Service Center, Bologna, 1993.

[HES 62] HESS H.H., History of ocean basins, in Engel, A.E.J, JALES, H.L. and LEONARD, B.F. (eds), Petrologic studies: A volume in honor of A.F. Buddington: New York, Geological Society of America, pp. 599–620, 1962.

[JAC 06] JACKSON J., Fatal attraction: living with earthquakes, the growth of villages into megacities, and earthquake vulnerability in the modern world, Philosophical Transactions of the Royal Society, vol. 364, no. 1845, pp. 1911–1925, 2006.

[MED 82] MEDD, Le risque sismique, Délégation aux risques Majeurs, Ministry of Environment, Paris, France, 1982.

[MOU 06] MOUROUX P., LE BRUN B., Presentation of the Risk-UE project, Bulletin of Earthquake Engineering, vol. 4, no. 4, pp. 323–339, 2006.

[OHT 86] OHTA Y., OHASHI H., KAGAMI H., A semi-empirical equation for estimating occupant ca- sualty in an earthquake, Proceedings of the 8th European Conference on Earthquake Engineering, Lisbon, Portugal, vol. 2–3, pp. 81–88, 1986.

[PIE 04] PIERRE J.-P., MONTAGNE M., The 20 April 2002, Mw 5.0 Au Sable Forks, New York, earthquake: a supplementary source of knowledge on earthquake damage to lifelines and buildings in Eastern North America, Seismological Research Letters, vol. 75, no. 5, pp. 626–635, 2004.

[SPE 03] SPENCE R., BOMMER J., DEL RE D., BIRD J., AYDINOGLU N., TABUCHI S., Comparing loss estimation with observed damage: a study of the 1999 Kocaeli earthquake in Turkey, Bulletin of Earthquake Engineering, vol. 1, pp. 83–113, 2003.

Introduction written by Philippe GUEGUEN.

Chapter 1

Seismic Vulnerability of Existing Buildings: Observational and Mechanical Approaches for Application in Urban Areas

1.1. Introduction

Past and recent earthquakes have shown the high level of seismic vulnerability of old and historic down-town areas: the 2009 L’Aquila earthquake is one of the latest dramatic examples, in which several historical centers (such as – besides L’Aquila – Onna, Castelnuovo and Villa Sant’Angelo) were severely affected, with heavy damage extended across whole built-up areas and the collapse of large portions (sometimes even in their totality) of many urban blocks. This follows the relevance of providing reliable vulnerability and risk analyses from the economic, cultural and human safety points of view.

As known, vulnerability represents the intrinsic predisposition of the building to be affected and suffer damage as a result of the occurrence of an event of a given severity. The main aims of a vulnerability analysis on a large scale – such as that of a town – are (1) to be aware of the impact of an earthquake to groups of buildings in the area; (2) to plan preventive interventions for the seismic risk mitigation; and (3) to help the management of the emergency after a major earthquake.

The main steps of a vulnerability analysis may be summarized as follows:

1) acquisition and examination of the data available in the area of interest, identification of building classes and definition of the related vulnerability models;

2) for each class, the definition of