Undamaged Buildings Research Articles

Monitoring the changes in the dynamic properties of an RC building experiencing column loss

In this study, an existing six-storey reinforced concrete building with an asymmetric structural plan and soft storey irregularity was used as a test specimen and subjected to three-step progressive structural damages to detect the variations in its dynamic properties. Mode shapes and dominant frequencies of the undamaged building were determined by the ambient vibration survey (AVS) and it was seen that its first three modes were torsion coupled. Besides, soft storey irregularity was evident due to the lack of masonry infill walls on its ground floor. Later on, three-step progressive damages were applied to the building. The first step targeted three columns and one beam of the building, located on a corner region of its ground floor to peel off their clear covers. The second step razed two adjacent corner columns which were already moderately damaged in the first step, while the third step knocked the third moderately damaged column down. After each damage step, AVS was repeated with the same details as applied for the undamaged building. The obtained dynamic properties for the four phases of the building were evaluated with the sustained damage. Numerical analyses with the finite element model of the building representing its four different phases were also performed and the unique responses due to damage effects on the structure were investigated numerically. As a result of induced damage, the quantified amount of frequency change in modes and the new mode observed after particularly column loss scenarios can be utilized for efficient structural health-monitoring strategies of plan-asymmetric buildings and post-earthquake assessment of partially damaged buildings where timely objective assessment is important.

Improved seismic intensity measures and regional structural risk estimation models

Seismic intensity measures are essential quantitative parameters for evaluating and predicting engineering structures’ seismic risk and vulnerability. Various regional earthquake analysis and vulnerability models with evaluation and prediction capabilities have been developed using macroseismic and instrumental intensity measures. These results have significant positive implications for evaluating large-scale regional earthquake resilience and predicting earthquake risks. However, the traditional determination of seismic intensity measures is based on given seismic station detection data and macroscopic intensity standards, and the developed seismic risk models are relatively outdated, resulting in relatively low accuracy in assessing the vulnerability of regional structures. This paper proposed an improved instrumental intensity calculation equation and updated the model using 216,000 acceleration records monitored by nine typical seismic stations in China during the Jiuzhaigou earthquake on August 8, 2017. The proposed calculation model was combined with the latest version of China’s macrointensity standard, and undamaged building samples were sufficiently considered. The vulnerability level of 1,499 buildings surveyed in the Jiuzhaigou earthquake was assessed. This paper innovatively proposed a membership index calculation model to estimate the impact relationship between different vulnerability levels. A comparative curve model was generated with the updated vulnerability databases of four typical regional structures, considering different vulnerability membership indices. Using the improved seismic intensity measures and nonlinear regression model, an improved vulnerability fitting model was developed to assess the seismic risk of regional structures. The model was analysed and verified using field survey data from the Jiuzhaigou earthquake.

Automated image-based generation of finite element models for masonry buildings

To predict the response of masonry buildings to various types of loads, engineers use finite element models, specifically solid-element and macro-element models. For predicting masonry responses to seismic events in particular, equivalent frame models—a subcategory of macro-element models—are a common choice because of their low computational cost. However, an existing bottleneck in modeling pipelines is generating the geometry of the model, which is currently a slow and laborious process that is done manually using computer-aided design tools. In this paper, we address this by automating the modelling process using recent advancements in computer vision and machine learning. We present an image-based end-to-end pipeline that automatically generates finite element meshes for solid-element and equivalent-frame models of the outer walls of free-standing historical masonry buildings. As the input, our framework requires RGB images of the buildings that are processed using structure-from-motion algorithms, which create 3D geometries, and convolutional neural networks, which segment the openings and their corners. These layers are then combined to generate level of detail models. We tested our pipeline on structures with irregular surface geometries and opening layouts. While generating the solid element mesh from the level of detail model is straightforward, generating equivalent frame models required algorithms for segmenting the façade and the meshing. Experts in the field analyzed the generated equivalent frame models and determined them to be useful for numerical modeling. These finite element geometries will be invaluable for future predictions of the seismic response of damaged and undamaged buildings. The codes and dataset are publicly available for future studies and benchmarking (https://github.com/eesd-epfl/FEM_buildings and https://doi.org/10.5281/zenodo.8094306).

Unpaired image-to-image translation of structural damage

Condition assessment of civil infrastructure from manual inspections can be time consuming, subjective, and unsafe. Advances in computer vision and Deep Neural Networks (DNNs) provide methods for automating important condition assessment tasks such as damage and context identification. One critical challenge towards the training of robust and generalizable DNNs for damage identification is the difficulty in obtaining large and diverse datasets. To maximally leverage available data, researchers have investigated using synthetic images of damaged structures from Generative Adversarial Networks (GANs) for data augmentation. However, GANs are limited in the diversity of data they can produce as they are only able to interpolate between samples of damaged structures in a dataset. Unpaired image-to-image translation using Cycle Consistent Adversarial Networks (CCAN) provide one means of extending the diversity and control in generated images, but have not been investigated for applications in condition assessment. We present EIGAN, a novel CCAN architecture for generating realistic synthetic images of a damaged structure, given an image of its undamaged state. EIGAN has the capability to translate undamaged images to damaged representations and vice-versa while retaining the geometric structure of the infrastructure (e.g, building shape, layout, color, size etc). We create a new unpaired dataset of damaged and undamaged building images taken after the 2017 Puebla Earthquake. Using this dataset, we demonstrate how EIGAN is able to address shortcomings of three other established CCAN architectures specifically for damage translation with both qualitative and quantitative measures. Additionally, we introduce a new methodology to explore the latent space of EIGAN allowing for some control over the properties of the generated damage (e.g., the damage severity). The results demonstrate that unpaired image-to-image translation of undamaged to damaged structures is an effective means of data augmentation to improve network performance.

Empirical fragility curves for masonry buildings struck by the 2016 Central Italy earthquake

The prevention of seismic risk at urban scale can be pursued through the estimate of the probability to reach or exceed a certain damage grade given the seismic input. In this framework, seismic fragility curves are nowadays of large interest as they express this probability in a synthetic way, also extended to large-scale applications. Real damage data are crucial in making more reliable predictions of damage occurrence, although they can be influenced by a proper definition of the structural types and the completeness of observations. The paper shows the empirical fragility curves obtained for a sample of 2263 masonry buildings located within 19 historical centers struck by the 2016 Central Italy earthquake. The damage grade was evaluated according to the European Macroseismic Scale (EMS-98), also considering undamaged buildings, at the end of the sequence that spanned between August and October. The buildings largely underwent several repairs and strengthening actions with reinforced concrete elements starting from the 1980s. The systematization of the structural features led to a taxonomy for strengthened and original buildings, which, based on the observed damage patterns, was matched to the EMS-98 vulnerability classification. The sample ranges from class A (worst behavior) to D (best behavior). Class A was typically assigned to original buildings (without interventions) or ill-advisedly tampered ones, i.e., those in which interventions had an unfavorable contribution to their seismic behavior. Class D described buildings with properly designed strengthening interventions, classes B and C intermediate situations. Fragility curves were obtained per each vulnerability class, as a function of the highest peak ground acceleration (PGA) observed in the sequence from ShakeMaps. The results were then compared to other empirical fragility models.

Influences of a debris flow disaster chain on buildings in remote rural areas, Southwest China

The magnitude and frequency of mountain hazards will continue to increase because of climate change especially in rural mountainous areas, which have not received much attention. In this article, a debris flow disaster chain in Southwest China caused damage to buildings along the river bank was investigated. The disaster chain included a debris flow, river blockage, lake formation, and an outbreak flood, and this study further investigated the disaster-related losses in downstream villages. Based on the characteristics of failed buildings, three structural types were identified, while the functions of undamaged buildings were affected by flooding or siltation by sediments. In addition, the building layout and riparian vegetation zone influenced the blockage effect, which resisted the flood impact and provided protection to rear buildings. Moreover, this disaster changed the river topography, the studied reach changed from a V shape to a wide–shallow shape, the appreciably affected river length was 6270 m, and the gully bed was noticeably raised 15–20 m. Furthermore, the flood control standard of a 50-year return period was recommended for reference, despite the limitations of these suggestions, we hope that this study attracts the attention of researchers on disaster prevention and mitigation in remote rural areas.

Hurricane Damage Detection From Satellite Imagery Using Convolutional Neural Networks

Hurricanes are one of the most disastrous natural phenomena occurring on Earth that cause loss of human lives and immense damage to property as well. For assessment of this damage, windshield survey is commonly used, which is an error-prone and time-consuming method. For solving this problem, computer vision comes into the picture. In this paper, a convolutional neural network-based architecture has been proposed to classify the post-hurricane satellite imagery into damaged and undamaged building classes accurately. The model consists of five convolutional and five pooling layers followed by a flattening layer and two dense layers. For this, a dataset of Hurricane Harvey has been considered having 23000 satellite images each of size 128 X 128 pixels. With the proposed model, the author has achieved an accuracy of 92.91%, F1-score of 93%, sensitivity of 93.34%, specificity of 92.47%, and precision of 92.65% at a learning rate of 0.0001 and 30 epochs. Also, low false positive rate of 7.53% and false negative rate of 6.66% were obtained.

Typological fragility curves for RC buildings: influence of damage index and building sample selection

The recent earthquakes occurred in the last years in Italy registered a great number of casualties and large economic losses. In this perspective, the effective seismic risk assessment has become crucial for the Government mitigation policies that have to allocate the limited economic sources for the risk reduction.In this paper, data collected after the 2012 Emilia and the 2009 L’Aquila, Italy, earthquakes were used to derive novel fragility curves for RC frame buildings classes. Compared to previous studies, different aspects affecting the fragility curves parameters were investigated. First of all, data are completed accounting for the presence of undamaged buildings that are generally uninspected after the seismic events, considering alternative criteria to complete the database. Then, the influence of the building damage index definition was considered by selecting four damage indices available in the literature that allow converting the damage data collected with the post-earthquake form into one of the five damage states, according to the European macroseismic scale. Separate sets of fragility curves, as a function of peak ground acceleration, were developed for the building stock of the two regions. The results reveal that the database correction with the uninspected buildings is relevant to avoid the overestimation of the exceedance probability, especially at lower peak ground acceleration values, resulting in less steep curves. Then, the building damage indices accounting for the damage of both structural and non-structural elements led to a higher exceedance probability with respect to damage indices based on damage of only structural elements, especially for the lower damage states.

Transfer Learning-Based Automatic Hurricane Damage Detection Using Satellite Images

After the occurrence of a hurricane, assessing damage is extremely important for the emergency managers so that relief aid could be provided to afflicted people. One method of assessing the damage is to determine the damaged and the undamaged buildings post-hurricane. Normally, damage assessment is performed by conducting ground surveys, which are time-consuming and involve immense effort. In this paper, transfer learning techniques have been used for determining damaged and undamaged buildings in post-hurricane satellite images. Four different transfer learning techniques, which include VGG16, MobileNetV2, InceptionV3 and DenseNet121, have been applied to 23,000 Hurricane Harvey satellite images, which occurred in the Texas region. A comparative analysis of these models has been performed on the basis of the number of epochs and the optimizers used. The performance of the VGG16 pre-trained model was better than the other models and achieved an accuracy of 0.75, precision of 0.74, recall of 0.95 and F1-score of 0.83 when the Adam optimizer was used. When the comparison of the best performing models was performed in terms of various optimizers, VGG16 produced the best accuracy of 0.78 for the RMSprop optimizer.

Fragility Curves and Economic Loss Assessment of RC Buildings after L’Aquila 2009 Earthquake

ABSTRACT A buildings stock of 12,016 Reinforced Concrete (RC) buildings is investigated to propose vulnerability, fragility, and loss curves, as well as an Expected Annual Loss (EAL) assessment. The buildings data are collected within Da.D.O. database, considering the observed damage due to L’Aquila earthquake (2009). The adopted methodology allows us to add undamaged and not-surveyed buildings, not included within the database, by considering the typological distribution of two different reference municipalities. The obtained results highlight that the completed database allows a more reliable seismic risk assessment, even if resulting EAL is poorly influenced by reference municipality assumed for undamaged buildings estimation.

Convolutional neural network based hurricane damage detection using satellite images

Hurricanes are tropical storms that cause immense damage to human life and property. Rapid assessment of damage caused by hurricanes is extremely important for the first responders. But this process is usually slow, expensive, labor intensive and prone to errors. The advancements in remote sensing and computer vision help in observing Earth at a different scale. In this paper, a new Convolutional Neural Network model has been designed with the help of satellite images captured from the areas affected by hurricanes. The model will be able to assess the damage by detecting damaged and undamaged buildings based upon which the relief aid can be provided to the affected people on an immediate basis. The model is composed of five convolutional layers, five pooling layers, one flattening layer, one dropout layer and two dense layers. Hurricane Harvey dataset consisting of 23,000 images of size 128 × 128 pixels has been used in this paper. The proposed model is simulated on 5750 test images at a learning rate of 0.00001 and 30 epochs with the Adam optimizer obtaining an accuracy of 0.95 and precision of 0.97. The proposed model will help the emergency responders to determine whether there has been damage or not due to the hurricane and also help those to provide relief aid to the affected people.

Census-Based Typological Damage Fragility Curves and Seismic Risk Scenarios for Unreinforced Masonry Buildings

Seismic risk assessment has become a crucial issue for optimal management of economic resources allocated to mitigation. For this purpose, in the last decades, several research activities were aimed to update hazard, exposure, and vulnerability models that contribute to seismic risk assessment. From this perspective, the present work focuses on developing new empirical damage fragility curves for census-based typological unreinforced masonry buildings. In particular, damage data observed after the 2009 L’Aquila earthquake, Italy, related to almost 57,000 residential buildings, were used to calibrate the fragility functions. These data were complemented with the census data with the aim of obtaining an accurate estimation of the number of undamaged buildings. Damage fragility curves were identified for typological building classes, defined considering parameters present in both post-earthquake observations and census data with the aim of extending the results to the whole national territory. Six typological classes were defined considering the categories of the construction timespan and of the state of repair parameters. Then, a further distinction of the typological classes considering the number of stories parameter was included where relevant. The fragility curves were defined as a function of peak ground acceleration for five damage states, defined according to the European macroseismic scale. The results confirmed that older buildings are more vulnerable than newer ones and highlighted the crucial role of the state of repair on the damage fragility curves. Finally, the new set of damage fragility functions was uploaded in the Italian Risk Maps information technology platform, used by the Civil Protection Department for risk evaluation, as an exemplification of the potential application of the fragility curves.

Predicting road blockage due to building damage following earthquakes

Transportation infrastructure supports the social and economic activities of communities. One of the impacts of roads’ disruption is the obstruction of emergency services (e.g., ambulance, firefighting, evacuation). Furthermore, the recovery process of a community following an extreme event (e.g., a natural hazard) depends on the functionality of the transportation infrastructure. Therefore, conducting a risk and resilience analysis of transportation infrastructure is critical to help communities minimize the initial impact of hazards and promote a rapid recovery. Current approaches model the probability of road blockage due to building damage using high-resolution optical satellite images and aerial photographs collected after past events. However, the data used by these methods are limited, and few data have been collected before 2010. Besides, data may not be available for specific regions that have not experienced recent earthquakes. Thus, a probabilistic predictive method is needed for risk and resilience analysis of roads. This paper proposes a probabilistic model using the data from the 2010 Haiti Earthquake and calibrated by Bayesian approach to predict the debris distance from undamaged buildings (e.g., the distance debris can reach from the footprint of the undamaged building). The model is then used to construct fragility curves that give the conditional probability of road blockage at a given road section for a given seismic intensity. The proposed model considers the relevant factors affecting the road blockage probability, including building types, damage level, and road characteristics. The probability of road blockage at a given road section is estimated for the four general road section types, considering buildings on only one side of the road or both sides, and with or without a raised traffic median. The probability of road blockage for an entire road is then calculated by system and parallel reliability analysis. The proposed models apply to any general urban area without the dependence on historical data from past earthquakes.

A framework for assessing impaired seismic performance as a trigger for repair

AbstractUncertainty about when post‐earthquake repairs are needed delays decision making, and can lead either to unnecessary demolition or repair, or to inadequate repair actions. This study proposes a framework for assessing the effect of earthquake damage on the future seismic performance of a building. In this framework, we first assess drift demands during a maximum considered earthquake (MCER) level ground motion in an undamaged building, and then compare the drift demands in the same building during the same motion when it has been damaged by a prior shaking. We assess each building for a range of prior shaking events and 15 MCER‐level motions. From these analyses, we determine the relationship between damage in the first motion—quantified as peak story drift demand—and the MCER‐level motion drift demands—quantified as the ratio of peak drift demands in the damaged building to those in the undamaged building. This trend is used to indicate the level of damage at which performance is impaired, and repair is needed due to structural safety concerns. We apply this framework to a set of single‐degree‐of‐freedom structures and reinforced concrete moment frame buildings representative of construction in the United States. The results show that code‐conforming structures do not see significant drift amplifications in the damaged building when story drifts in the damaging motion do not exceed 2%, indicating they would not require major repairs for structural safety. Structures with less deformation capacity saw larger amplifications, and were affected when story drift demands in the damaging motion were lower. We also found that stiffer structures tend to see larger amplifications of drift demands.

Damaged building extraction from post-earthquake PolSAR data based on the Fourier transform

ABSTRACT Due to the casualties and economic losses that can be caused by building collapse after an earthquake, building damage assessment is a very important part of earthquake disaster evaluation. As a result of the advantages of polarimetric synthetic aperture radar (PolSAR) imaging, both the accuracy and speed of the damage assessment can be guaranteed in the process of building damage identification using a single post-earthquake PolSAR image. The number of collapsed buildings is usually overestimated when adopting the traditional method, which is because the scattering mechanism is similar to that of intact oriented buildings. In view of this, the variable coefficient of angle domains based on the Fourier amplitude spectrum parameter (CV_AFI) is proposed to ameliorate the damaged building overestimation. By combining the improved Yamaguchi four-component decomposition method with the CV_AFI parameter, the proposed method can be used to extract building damage information from a single PolSAR image. With the Ms 7.1 Yushu earthquake in China as the research case, compared with the traditional method, the producer’s accuracy of the proposed method for undamaged buildings was increased by 38.87%. The proposed method can obtain a higher accuracy of building damage information extraction and significantly reduce building damage overestimation.

Performance Evaluation of Reinforced Concrete Buildings During the Sivrice-Elazığ Earthquake (Mw=6.8, January 24, 2020) in Accordance with Turkish Earthquake Code

A great number of reinforced concrete building structures were damaged or collapsed during the earthquake that hit Sivrice district of Elazığ city located in the southwest of the Eastern Anatolia Region of Turkey on January 24, 2020. Magnitude of the earthquake were announced as [Formula: see text] and [Formula: see text] by Kandilli Observatory & Earthquake Research Institute (KOERI) and Earthquake Department of the Disaster and Emergency Management Presidency (AFAD), respectively. More than a thousand of aftershocks with the magnitude of 1.2–5.1 occurred between January and February 2020. The magnitude of the structural damage was relatively high compared with the peak value of recorded ground acceleration (2.85[Formula: see text]m/s2) at the epicenter (Sivrice district) of the earthquake. It was determined that 584 buildings were demolished, 6845 were severely damaged, 1207 were moderately damaged, 14,389 were slightly damaged and 14,317 were undamaged. Also, there were 235 buildings that required urgent demolition. In this paper, the performance of the reinforced concrete buildings during the Sivrice-Elazığ earthquake is presented along with time-histories of ground motion records and response spectrums. Observed design and construction practices were carefully evaluated and compared with Turkish Earthquake Code. It was seen that the most significant causes of the failures were due to the poor concrete quality, the use of non-ductile detailing, the poor project design and construction quality. According to field inspections of our technical team, it is concluded that performances of the undamaged buildings following earthquakes do not prove that they have adequate seismic safety.

Direct identification of reduced building models based on noisy measurements for performance based earthquake engineering

Structural health monitoring (SHM) represents a useful tool in preventing structural failures. SHM can help to identify deviations between the structural properties at the design stage and those of the actual construction, and to detect aging effects as well as accumulated damage, which can contribute to uncertainty related to the seismic performance. Developing a method that can effectively be applied without the need for expensive high precision monitoring devices represents a key issue in order to broadly install this type of technology in the built environment. An identification method is therefore proposed that takes into consideration the noise level of the monitoring instrumentation as an input and can lead to relatively accurate results even for high levels of noise. Formulating a reduced numerical model of the structural system represents one major component of the proposed method; the identified reduced structural model is then used in order to predict the structural performance levels in accordance with seismic assessment regulations in cases of damaged and undamaged building structural systems

Damage correlation in flexural walls with a displacement approach method for boundary detailing

The 2010 Chile earthquake of magnitude Mw 8.8 showed the importance of using correct detailing at the wall boundary. Most design methodologies for wall boundary special detailing have been compared to testing, but at the building level have not been validated with evidence from actual earthquakes. For this purpose, two buildings that exhibited flexural wall damage were studied. An additional building that was not damaged, standing less than 100 m from one of the damaged buildings, was also included. The buildings were analyzed using linear–elastic models, as is common practice among designers, and the walls were analyzed in terms of compressive strain demands resulting from roof displacement. The strain demands were determined using a concentrated plasticity model, as well as a model that included the elastic and plastic components of deformation. In this last case, the elastic component was determined by pushing the building model with a pattern consistent with the first mode of lateral vibration until the first wall reached its moment capacity, including a reduced stiffness in the walls and slabs. Results indicate that the damaged walls were expected to reach moment capacity earlier than the other walls, obtaining compressive strain values greater than 0.003. Thus, the walls damaged during the 2010 earthquake correlated well with the estimations of compressive strains, since these were also the first walls to reach maximum capacity. On the other hand, the undamaged building presented generally lower levels of compressive strain values in the walls than those of the damaged buildings.

Seismic damage assessment in Sarpole-Zahab town (Iran) using synthetic aperture radar (SAR) images and texture analysis

The synthetic aperture radar SAR system with the capability of imaging during the night, day, and the all-weather conditions has a high potential in change detection on the ground surface. In this research, we used three SAR images of ALOS-2 satellite over Sarpole-Zahab town in the west of Iran that had an earthquake with the magnitude of 7.3 on November 12, 2017. The effects of speckle noise on the accuracy of the results were assessed based on noise reduction filters. Correlation coefficient, difference of intensity (in five window sizes), and difference of coherence and texture (in six window sizes) of the pre- and post-event images were calculated, and the output parameters were extracted. Then, the damage assessment was carried out based on four machine learning classifiers, containing the random forest (RDF), the support vector machine, the naive Bayes classifier, and K-nearest neighbor. The RDF showed an overall accuracy of 86.3%. Seventy percent of the dataset was used for training, and 30% of it was used for the prediction purpose (~ 300 buildings). Based on the training dataset, the total number of structures in the study area was predicted (approximately 9200 buildings). Finally, a discriminant analysis was carried out among the damaged and undamaged buildings.

Assessment of a house affected by ground movement using terrestrial laser scanning and numerical modeling

This study analyzed the internal mechanical state of houses disturbed by underground coal mining, especially those that are free from damage in subsidence areas. Traditional surveying and terrestrial laser scanning (TLS) were utilized as the external monitoring methods to obtain the outside displacement of the disturbed houses. The results were then used to estimate the internal mechanical state of houses using simulation software. Three high-quality houses in Paifang Village, China were scanned using TLS in their pre- and post-mining stages. The magnitude of movement and deformation was highly related to the location of the building in the panel, which was enlarged near the center of the subsidence basin. ANSYS was used to compute the internal state of the houses with a focus on the structural columns, ring beams, and walls of the houses under changing tensile conditions. Results showed that the structural column and ring beam can resist tensile deformation, whereas the window corner was the principal point of stress concentration with potential risk of cracking. Mastering the internal stress state of undamaged buildings in mining-affected areas can help to strengthen the premaintenance of houses and guide the construction of new anti-deformation houses. The results of this work may benefit the construction of anti-deformation buildings in mining regions.