A data-driven framework for seismic fragility and performance degradation assessment of historical masonry pagodas: integrating multi-task machine learning and interpretable uncertainty quantification

Seismic response prediction and fragility assessment of high-speed railway bridges using machine learning technology

A procedure for the seismic fragility assessment of nuclear power plants by applying ground motions compatible with the conditional probability distribution of a conditional spectrum (CS) is presented with a case study of a containment building. Three CSs were constructed using different control frequencies to investigate the influence of the control frequency. Horizontal component-to-component directional variability was introduced by randomly rotating the horizontal axes of the recorded ground motions. Nonlinear lumped mass stick models were constructed using variables distributed by Latin hypercube sampling to model the uncertainty. An incremental dynamic analysis was performed, and seismic fragility curves were calculated. In addition, a seismic input based on a uniform hazard response spectrum (UHRS) was applied to the seismic fragility assessment for comparison. By selecting a control frequency dominating the seismic response, the CS-based seismic input produces an enhanced ‘high confidence of low probability of failure’ capacity and lower seismic risk than the UHRS-based seismic input.

Gaussian process regression driven rapid life-cycle based seismic fragility and risk assessment of laminated rubber bearings supported highway bridges subjected to multiple uncertainty sources

New regulatory requirements in Japan have strengthened the mitigation of damage caused by natural hazards, such as earthquakes, and the operational guide for safety improvement evaluation recommends the use of a probabilistic risk assessment (PRA) as the evaluation method in Japan. In the PRA of an earthquake, also known as the seismic PRA, one of the most important issues is the realistic assessment of the structural seismic response and the conditional damage probability (fragility) assessment using a realistic response assessment of nuclear buildings and equipment. Accordingly, we conducted this study both on the analysis methods used for realistic seismic response and assessment methods of seismic fragility to ensure the seismic safety of nuclear buildings and equipment. In this study, we use a three-dimensional (3D) structural model of a reactor building to conduct a nonlinear seismic response analysis for input ground motions beyond a design basis. In addition, we identify the failure mode of the structural components of the reactor building associated with the equipment and assess the seismic fragility based on the 3D behavior of the reactor building. The local response and detailed damage progress of the reactor building obtained through seismic response analysis are reported, along with the results of the seismic fragility assessment.

Seismic fragility and resilience assessment of bridge columns with dual-replaceable composite link beam under near-fault GMs

Earthquakes can cause serious damage to traffic infrastructures, among which the impact on bridge structure is the most important. Therefore, in order to assess bridges serviceability, it is important to master their damage mechanism and to analyze its probability of occurrence under a given seismic action. Various uncertainties, like the location of epicentre of future earthquakes and their magnitudes, make this task quite challenging. We are also required to consider different earthquake scenarios and the damaged states of bridge components associated with those earthquakes. To suppress these difficulties, this study proposed a new method of performance‐based seismic fragility and risk assessment for bridges. The proposed method included three steps: (1) performance‐based seismic fragility estimation of a five‐span continuous rigid frame bridge, (2) seismic hazard analysis for locations of the bridge, and (3) seismic risk analysis of the bridge. The proposed method that considered the performance of the bridge and the uncertainty in the location of the earthquake epicentre and magnitudes can provide valuable references for seismic‐resistant design of multispan continuous rigid frame bridges in the future.

Restrainers, being of relatively low cost and easy to install, are often used to prevent unseating of bridge spans. The potential of using superelastic shape memory alloy (SMA) restrainers in preventing such failure has been discussed in the literature; however, the impact of such smart restrainers with optimized configurations in reducing the failure probability of bridge components and system as well as the long-term economic losses given different earthquake scenarios has not been investigated yet. This study presents a probabilistic seismic fragility and long-term performance assessment on isolated multi-span simply-supported bridges retrofitted with optimized SMA restrainers. First, SMA restrainers are designed following the displacement-based approach and their configuration is optimized. Then, seismic fragility assessment is conducted for the bridge retrofitted with optimized SMA restrainers and compared with those of the original bridge and the bridges with elastic restrainers (steel and CFRP). Finally, long-term seismic loss (both direct and indirect) are evaluated to assess the performance of the retrofitted bridges in a life-cycle context. Results showed that among three considered restrainers, SMA restrainers make the bridge less fragile and help the system lower long-term seismic loss. The design event (DE, 2475-year return period) specified in Canadian Highway Bridge Design Code (CHBDC, CSA S6-14 2014) may underestimate the long-term seismic losses of the highway bridges. Under DE, the damage probability of the bridge retrofitted with optimized SMA restrainers experiencing collapse damage is only 0.7%. Under the same situation, its expected long-term loss is approximate 17.6% of that with respect to the unretrofitted bridge.

In this paper, a comparison between probability density evolution method (PDEM)-based approach and linear-regression-based approach in seismic fragility assessment is performed, and an application into low-rise frame building is conducted for implementation. The principles of the two fragility approaches are introduced first. The PDEM-based fragility approach has solid foundations in the reliability field and is expressed as a non-parametric form without predefined fragility shapes, while the linear-regression-based fragility approach is expressed as a classic parametric form, commonly under the assumption of lognormal distributions. Then a 3-span-4-storey reinforced concrete frame (RCF) building, which represents the existing buildings with large quantities in China, is used for seismic assessment by the two approaches. In general, the application shows the similar tendency of fragility curves for the two methods, and verifies the effectiveness of the non-parametric PDEM-based approach for seismic performance assessment. Comparatively, the PDEM-based approach omits the heavy computing burden as Monte Carlo simulation, and reflects the fragility characteristics with considerable accuracy as linear-regression-based approach, which turns on a new path for the fragility assessment scheme in the performance-based earthquake engineering.KeywordsSeismic fragilityEarthquake engineeringPerformance assessmentProbability analysisReinforced concrete framesPdem

The seismic isolation system makes a structure isolated from ground motions to protect the structure from seismic events. Seismic isolation techniques have been implemented in full-scale buildings and bridges because of their simplicity, economic effectiveness, inherent stability and reliability. As for the responses of an isolated structure due to seismic events, it is well known that the most uncertain aspects are the seismic loading itself and structural properties. Due to the randomness of earthquakes and uncertainty of structures, seismic response distributions of an isolated structure are needed when evaluating the seismic fragility assessment (or probabilistic seismic safety assessment) of an isolated structure. Seismic response time histories are useful and often essential elements in its design or evaluation stage. Thus, a large number of non-linear dynamic analyses should be performed to evaluate the seismic performance of an isolated structure. However, it is a monumental task to gather the design or evaluation information of the isolated structure from too many seismic analyses, which is impractical. In this paper, a new methodology that can evaluate the seismic fragility assessment of an isolated structure is proposed by using stochastic response database, which is a device that can estimate the seismic response distributions of an isolated structure without any seismic response analyses. The seismic fragility assessment of the isolated nuclear power plant is performed using the proposed methodology. The proposed methodology is able to evaluate the seismic performance of isolated structures effectively and reduce the computational efforts tremendously.

Seismic fragility of RC shear walls in nuclear power plant Part 1: Characterization of uncertainty in concrete constitutive model

It is a key step to determine the structural capacity at different limit states for a seismic fragility assessment. However, limit state definition is highly dependent on engineering practices, which includes strong fuzziness. Therefore, it is necessary to conduct a comprehensive study on the influence of fuzziness at limit states in a seismic fragility assessment. This study performs a seismic fragility assessment by considering fuzziness at limit states, where ten types of membership functions are considered with varying fuzziness levels. Four reinforced concrete frame structures with varying heights and fortification levels are taken as the study cases. The fuzziness-probability integration is adopted to derive seismic fragility functions corresponding to different types of membership functions. The seismic fragility results with and without considering fuzziness at limit states are compared to extract the effect of limit state fuzziness with considering varying membership functions and different fuzziness levels. Moreover, the modified seismic fragility function considering limit state fuzziness is derived. The analysis results show that different membership functions show a significant effect on seismic fragility results considering limit state fuzziness. As the promotion of fuzziness level, the difference in seismic fragility results considering limit state fuzziness is growing accordingly. The proposed modified fragility function considering limit state fuzziness can well represent the effect of fuzziness at limit states on the seismic fragility analysis.

Machine learning-based seismic fragility and seismic vulnerability assessment of reinforced concrete structures

Finite element models used in industrial studies for seismic structural reliability analysis are in general very complex and computationally intensive. This is due to the important number of degrees of freedom as well as due to advanced damage models and failure modes that have to be simulated. In consequence, reliability studies are feasible only by means of simpler surrogate models able to represent essential physics. Then the choice of an accurate surrogate or metamodel is crucial for uncertainty propagation and sensitivity analysis. The construction of a pertinent metamodel is however a not simple task when random uncertainties, not explained by the model parameters, have to be accounted for. This is the case for transient seismic analysis where the ground motion, an intrinsically random phenomena, is modelled by a stochastic process that cannot be entirely described by a set of parameters. In this paper, we construct a versatile metamodel based on analysis of variance (ANOVA) decomposition. The ANOVA decomposition provides a convenient framework allowing both for parametric uncertainties and for stochastic variability introduced by seismic load. We then compute fragility curves and perform sensitivity analysis.

Bridges are the most vulnerable elements in regional transportation networks. Seismic fragility assessment has significant implications regarding the potential damage of regional bridges and their seismic maintenance operations. Unlike individual bridges, assessing bridges at the regional scale involves greater diversity and distinct ground motions. This paper proposes a Bayesian parameter estimation methodology to assess the seismic fragilities of regional beam bridges and schedule their maintenances. The proposed multi-parameter models contain several structural parameters and intensity measures to comprehensively consider the regional diversity. The selected intensity measures represent the diversity of regional ground motions at the peak effect, spectrum intensity, and the duration of strong motion. To demonstrate the applicability of the proposed methods, this paper analyzes the seismic damage results of 13,500 bridge-ground motion pairs generated to train the fragility models. Results of the proposed models are similar to the seismic fragility probability derived from Monte-Carlo simulation results. This study proposes a two-level strategy to effectively manage regional bridge maintenance. These priority maintenance and secondary maintenance levels are determined based on the results of regional seismic fragility analysis. Additionally, significant sequences of structural parameters and intensity measures are identified using the leave-one-attribute-out analysis method. This process helps bridge designers and managers better understand the sensitive parameters pertaining to different components. This study also tests the proposed models on a virtual transportation network to assess the seismic fragility of its associated regional bridges, resulting in a two-level regional maintenance plan for that specific system.

Seismic risk assessment of unreinforced masonry (URM) buildings is an important process for seismic retrofit of essential facilities located in the central and southern United States (CSUS), as more than 30 % of facilities there are low-rise URM buildings. Although HAZUS, the current loss estimation package for natural hazards, provides a set of fragility curves for such structures as an essential tool for conducting seismic risk assessment, seismic performance level variation due to geometric characteristics is not explicitly considered. This study investigates the effect of geometric variation of low-rise URM structures on seismic fragility assessment. Utilizing URM building inventory information within the CSUS region, variables that describe the physical shape of URM structures are identified. A simplified composite spring model developed for URM structures is then utilized to monitor nonlinear seismic behavior. Finally, seismic fragility curves corresponding to various shape configurations of URM structures are developed and compared. The analysis confirms that the length of out-of-plane walls and the number of stories in URM buildings have significant effects on seismic risk. An increase in the wall length or the number of stories makes URM buildings more vulnerable. On the other hand, the perforation ratio does not significantly affect seismic performance. It is suggested that using a single set of fragility curves is not adequate for seismic risk assessment of low-rise URM buildings unless geometric variation is considered explicitly. In addition, comparing the fragility curves developed in this study with HAZUS data, it is clear that the seismic vulnerability of low-rise URM is underestimated in HAZUS for lower limit states and overestimated for higher limit states.