Computer-Aided Pre- and Post-Earthquake Assessment of Buildings Involving Database Compilation, GIS Visualization, and Mobile Data Transmission

The aim of this article is to present a computer-aided comprehensive strategy for the rapid visual inspection of buildings and the optimal prioritization of strengthening and remedial actions that are necessary prior to, and after, a major earthquake event, respectively. Based on the visual screening procedures used in the United States and past experience in seismic assessment of buildings in Greece and Turkey (the two countries with the highest seismic risk in Europe), a building inventory is first compiled; then a vulnerability ranking procedure that is specifically tailored to the prevailing construction practice in Southeast Europe is implemented into a multi-functional, georeferenced computer tool, that accommodates the management, evaluation, processing and archiving of the data stock gathered during the preand post-earthquake assessment process, and the visualization of its spatial distribution. The methodology proposed and the computer system developed is then applied to the city of Düzce, Turkey, a city strongly damaged during the devastating 1999 earthquake. ∗To whom correspondence should be addressed. E-mail: asextos@ civil.auth.gr. 1 BACKGROUND AND OBJECTIVES OF THE STUDY As the socio-economic cost of a possible earthquake may be significant, a major effort has been made worldwide, mainly during the last two decades, toward the development of organized large-scale action schemes for the enhancement of the safety and serviceability of the building stock, the upgrade of the seismic performance of the infrastructure, as well as for the mitigation of potential environmental impacts and any other direct or indirect consequence. This effort, which is primarily made in countries exposed to high seismic risk (due to high seismic hazard and/or high exposure caused by dramatic increase in the size of the urban habitat), is related not only to the preand post-earthquake disaster preparation and management methods to be employed, but also to the subsequent technical, social, administrative, legal, and financial measures necessary for the implementation of the foreseen schemes. A major component of the above strategic planning concerns buildings, whose performance is expected to affect considerably the overall C © 2008 Computer-Aided Civil and Infrastructure Engineering. Published by Blackwell Publishing, 350 Main Street, Malden, MA 02148, USA, and 9600 Garsington Road, Oxford OX4 2DQ, UK. 60 Sextos, Kappos & Stylianidis earthquake loss in urban areas. As such, the (preand post-earthquake) assessment of buildings has attracted major scientific interest worldwide and it is materialized through two main engineering procedures: 1. Pre-earthquake Assessment of buildings, which is aimed at evaluating the safety level of public and/or important buildings against the maximum considered seismic action, and the prioritization of strengthening activities, to assure safety against collapse of the general building stock and serviceability of the most critical buildings after a major earthquake event. It is noted that, in general, preearthquake rapid visual inspection (RVI) is normally carried out starting from important and/or public buildings, and aims at assessing their seismic capacity, as well as setting the priorities for the various strengthening/upgrading schemes that should be implemented. The ranking of buildings is performed on the basis of qualitative, and sometimes quantitative, criteria, which involve the period of construction (or seismic code used), the structural system, the level of seismic hazard, and a number of structural characteristics. The evaluation of buildings on the basis of such criteria typically culminates into a “score”; poor seismic performance of the building is associated with low scores, indicating a high priority for strengthening. The framework for the pre-earthquake assessment has been set by recent United States standards and guidelines, such as ATC-13 (1985) and ATC-38 (2003), which primarily refer to California building types, the FEMA 154-ATC 21 (2002) and the FEMA 310 (1998) documents for rapid visual screening, and also by the UNDP/UNIDO (1985) document targeting the Balkan countries. Similar methods have been used in Japan (Japan Building Disaster Prevention Association, 1990) and Greece (Ministry of Public Works, 1997, 2000), although large scale application after major earthquake events of the country-dependent provisions has been made in the United States (ATC-38 project in California, 2003), in Greece (Penelis & Kappos, 1997; Kappos, 1997; Stylianidis et al., 2003a, b; Anagnostopoulos & Moretti, 2006a, b) and in Turkey (Spence et al., 2002; Ansal, 2003) among other countries. 2. Post-earthquake Inspection of buildings, which allows the local and central authorities to obtain a quick but fairly detailed statistical overview of the damage extent (in terms of spatial distribution, degree, and total number of buildings associated with a particular damage level), thus assisting decision-making regarding the required remedial and recovery measures. Such experience with post-earthquake inspection of buildings and gathering of damage data has been also gained in the United States, especially after the Loma Prieta and Northridge earthquakes, as well as in Southeast Europe from the recovery strategies adopted following a number of recent major earthquakes. In certain cases, the post-earthquake assessment is nowadays assisted by the processing of satellite images (Eguchi et al., 2000; Matsuoka & Yamazaki, 2004; Saito et al., 2005), however, so far only a macroscoping assessment of damage is feasible. Notwithstanding the development of various preand post-earthquake assessment methodologies, a number of issues are still open regarding their application, hence posing the challenge to propose more comprehensive solutions toward the data-based, quantitative, and computer-aided management of seismic risk. The most significant problem is the lack of a unified framework for conducting joint preand postearthquake assessment studies and making decisions based on the complete seismic history of buildings, a quantified estimate of their current vulnerability and the data regarding the damage observed after a potential future earthquake, inclusive of the strengthening/rehabilitation measures and the associated cost. Unfortunately, the current practice is to perform postearthquake assessment independently from the screening process prior to the earthquake, using different and incompatible methodologies that do not allow any data flow between the two procedures. Moreover, especially in terms of pre-earthquake assessment, the more detailed procedures that are used in the United Status are not directly applicable overseas, since the building types, construction processes, and materials used often differ substantially. Another important issue is that, in most cases, the visual inspection results are either not organized electronically (i.e., large scale inspections are performed onsite by local personnel using hard-copies of standardized forms) or their visualization in GIS is not a standard part of the procedure followed, while the process to review the data and ensure their reliability is rather long. As a result, huge effort and time may be required until the authorities obtain a reliable and clear overview of the structural vulnerability (prior to an earthquake) or of the extent of damage (after an earthquake), a fact that is of paramount importance, especially during an earthquake crisis. Along these lines, a knowledge-based expert system (KBES) is developed which utilizes state-of-the-art knowledge in earthquake engineering and information technologies (IT) to develop a tool that will serve the following five main goals: 1. Develop a comprehensive framework for the joint management of the data gathered during the Computer-aided preand post-earthquake assessment of buildings 61 preand post-earthquake assessment of buildings in any given city, and of a GIS-oriented database system that can assist the local authorities to make decisions to mitigate seismic risk, based on complete information regarding the building history at a city scale (regarding earthquake-induced damage, strengthening, and rehabilitation works) and a set of quantitative assessments for both the preparedness and recovery period. 2. Improve the existing methodologies for the preand post-earthquake assessment by proposing additional structural types and scoring matrices for the case of moderate to large size Southeast European cities based on expert knowledge and the prevailing construction and assessment practices in Greece and Turkey. 3. Establish the framework to combine the building vulnerability scoring values obtained (prior to the earthquake) with the, also quantitative, actual damage level observed (after the earthquake), to update the weighting factors proposed by the experts for the various structural parameters that affect the vulnerability evaluation. 4. Enhance the reliability of the collected data by introducing internal logical error trapping checks, programmed through the database, to spot out potential conceptual controversies in the data as they are input into the system and thus automatically increase their reliability. 5. Eliminate the time required to digitize the data gathered on-site with the aid of mobile clients (pocket PCs or new generation cellphones) that permit electronic filling-out of the forms, mobile synchronization with the main system database (planned to be directly accessible by the local authorities), sorting of the city buildings according to their relative vulnerability (prior to the earthquake) or damage score (after the earthquake), immediate filtering of potential errors (in the main database), and final visualization in GIS of the results almost in real time. It is noted that this combination of a computer-aided building assessment procedure and a GIS approach, as well as the joint management of preand postearthquake assessment, is deemed a valuable contribution; it is noted that even the latest version of FEMA154 (2002) considers it as “beyond its scope.” Moreover, the particular system has already been implemented in the city of Düzce in Turkey (a city strongly damaged during the devastating 1999 earthquake as described in Section 3), although several of its components are currently being implemented for the pre-earthquake assessment of buildings in Greece as well. The main aspects of the particular comprehensive methodology and the architecture of the system developed are presented in the following. 2 OVERVIEW OF THE INTERACTIVE GIS/DATABASE DECISIONMAKING FRAMEWORK 2.1 Database architecture and work flow The GIS-database developed has to be installed at an operating server located in an administration building or, in case of an earthquake, on a portable computer to control the data flow during a field campaign. Depending on the expertise of the local engineers responsible for assessing the buildings and the available hardware, either the quick inspection forms are filled out in hardcopy form or a PDA/cell phone is used for direct data input. The database is built using Microsoft Access and is internally programmed using the VISUAL BASIC for APPLICATIONS (VBA) programming language and STRUCTURED QUERY LANGUAGE (SQL) for data filtering and building scoring. The former is also used for structuring a framework of logical checks aiming to eliminate the possibility of inconsistent data input. The data are recorded into three relational tables (for preand post-earthquake assessment, and building a common profile) consisting of a large number (102 in the version presented herein) of data fields. As the most important information gathered for the assessed buildings is related to their structural integrity, it is apparent that the main interface of the electronic management computer system developed closely simulates the (preand post-earthquake) quick inspection forms adopted (and specifically adapted) for the area under study. Through this interface, or ideally by filling out data on-site using a PDA/cell phone, the building data are stored in the database and the building scores are automatically computed according to the predefined multi-parametric scoring schemes provided by experts (as is described in detail in the following Section 2.2). Finally, all buildings are ranked and classified in terms of their structural vulnerability. The electronic database automatically provides all the statistics of both the gathered and computed data, although it communicates directly with a GIS environment for visualization of the assessment results in space using MapInfo (2001) software and specific workspaces. The use of advanced GIS mapping tools for building assessment and seismic hazard data has also been presented in Turkey in the framework of the Istanbul Earthquake Master Plan (Ansal, 2003) and for the city’s Early Warning System (Erdik et al, 2003), as well as in other 62 Sextos, Kappos & Stylianidis countries (e.g., O’Rourke et al., 2001) among other applications. However, the database presented herein is considered to be an extension of the aforementioned pioneering developments, since it integrates the preand post-earthquake assessment data, their visualization in space through GIS, and the expert system for rating/scoring of buildings, all into a comprehensive computer framework. Another significant advantage of the particular system is that the GIS built-in database and the external database (separately developed to be fully programmable) share a common table, hence no import and export procedures are needed and the user can retrieve or import data using any of the two different interfaces. The overall system architecture is shown in Figure 1. 2.2 Overview of a comprehensive procedure for the preand post-earthquake assessment of buildings and ensuing algorithm An important part of the overall system developed is the building evaluation process, related to the (preand post-) earthquake assessment methodologies to be used. As mentioned previously, the idea was to adapt widely used existing methodologies to create a model process applicable to cities of moderate to large size, mainly in Southeast Europe that are exposed to high seismic risk. Regarding the pre-earthquake assessment of buildings, a simplified standard (hardcopy) form was initially adopted for data collection, mainly based on the inspection forms proposed in the Istanbul Master-plan study (Ansal et al., 1999). This hardcopy pre-earthquake inspection form (sheet) consists of five sections distinguished according to the nature of the information they refer to. More specifically: Section A contains the building identification data (i.e., ID, city section, address, ownership, authorities for inspection, number of residents etc); Section B the building’s technical characteristics (i.e., number of storeys and basements, plan area, total built area, year of construction, building importance, previous rehabilitation or repair inclusive of time and reason, engineers carrying out the inspection, etc); Section C the relevant seismological and geotechnical data (i.e., seismic hazard zone and soil classification, as seen in Table 2); Section D the data related to the structural system (Table 1); and Section E the structural characteristics of the building that affect its seismic performance (Table 3). It is noted that both the structural types and the scoring matrices summarized in Tables 1–3 that present the relative importance of the various structural, geotechnical, and seismotectonic parameters with respect to seismic risk, are proposed by the authors and the board of experts (acknowledged at the end of the article) for the case of Southeast European cities. These values are based on experience from the statistical processing of the damage observed in a number of past earthquakes (in Greece and Turkey), numerical analyses performed by the authors, monitoring data, and experimental results. As an example, Tables 1–3 correspond to the particular code (defining the seismic zonation and soil categorization) valid for the country where the large scale application was foreseen (i.e., Turkey and more specifically the city of Düzce). Clearly, though, as shown in Section 3.2, the overall methodology is of broader use. Based on the above, the adapted pre-earthquake assessment-rating scheme (basic scores for each typology, and score modifiers accounting for particular features affecting the seismic vulnerability) finally depends on a number of parameters as described in the following. The final score (Sfin) is derived according to the following simple expression: Sfin = SABSC + SZSE + Ssc + SVUL (1) where SABSC is the basic score and is a function of the structural type of each building (Table 1), SZSE depends on seismological parameters (i.e., the seismic zone factor, as displayed in Table 2), Ssc depends on geotechnical parameters (i.e., the soil factor as displayed in Table 2) and SVUL depends on structural vulnerability parameters (i.e., seismic code used, modification of use, damage by previous earthquakes, inadequate maintenance, pounding possibility, presence of soft storey or short columns, irregularly arranged infills, significant overall height, irregularities in plan and height and possibility of torsion, as summarized in Table 3). The overall procedure, illustrated in detail in Figure 2, is followed for all buildings assessed, and a final score is given to each one to prioritize (through appropriate sorting) the buildings with the lowest score (i.e., the most vulnerable ones). Despite the expertise background required, the above calculation process related to the specific scoring matrices of Tables 1 to 3, is intentionally kept simple enough to be applied quickly on-site by the local personnel. The post-earthquake quick inspection procedure, on the other hand, that is presented herein, is based on the inspection form issued by the Turkish Ministry of Public Works and Settlement-–General Directorate of Disaster Affairs (1999). It consists of six quantitative criteria to decide on structural safety (in terms of settlement, inclination, pounding with adjacent buildings, percentage of moderate or severe damage in the columns of the most heavily damaged storey) and six qualitative criteria to assess the safety of non-structural elements. Each procedure runs separately for two main categories of damage type (structural and non-structural) and ranks the buildings on the scale of “Inspected” (or “safe” or “checked”), “Limited Entry,” and “Unsafe” (Figure 2). The system developed computes the quantitative factors, examines the various criteria, and provides a final Computer-aided preand post-earthquake assessment of buildings 63 Fig. 1. GIS-Database structure for data management as a part of the preand post-earthquake assessment process. 64 Sextos, Kappos & Stylianidis