Seismic fragility assessment of historical Khan structures in Izmir (Türkiye)

Seismic fragility assessment of historical Khan structures in Izmir (Türkiye) P. Usta a , S. Karimzadeh b,* , P.B. Lourenço b a Department of Civil Engineering, Isparta University of Applied Sciences, Isparta, 32200, Türkiye b Department of Civil Engineering, ISISE, ARISE, University of Minho, Campus de Azur´ em, Guimar˜ aes 4800-058, Portugal ARTICLE INFO Keywords: Real Ground Motion Records Masonry Structures Numerical Simulations Fragility Analysis Khan Izmir (Türkiye) ABSTRACT The province of Izmir in Türkiye is recognized as a city renowned for its rich cultural history. The province boasts numerous Khan masonry monuments, particularly within the Kemeralti commercial district. These structures, which carry cultural significance crucial to architectural heritage, showcase distinctive construction techniques that result in varied seismic behaviors. Conducting seismic assessments of these historic monuments is vital for risk mitigation and preserving them. The initial phase of risk assessment studies involves the development of fragility curves for historical structures. Fragility curves reveal the intricate relations between seismic forces, building features, and potential damage, providing valuable information for risk assessment studies. This paper develops seismic fragility curves for historical Khan masonry monuments through data from visual field observations, structural analysis, and numerical simulations, focusing specifically on the Kemeralti commercial district of Izmir, Türkiye. Using peak ground acceleration (PGA) as the primary intensity measure, the study examines the extent of damage at various PGA levels across different limit states, including Immediate Occupancy (IO), life safety, and collapse prevention. The analysis is conducted in both the Xand Y-directions of the structure using linear time history analysis. Results show higher probabilities of exceedance (PoE) in the Y-direction (perpendicular to the main façade) of the structures compared to the X-direction, which is attributed to the configuration of the structural elements. Furthermore, a strong alignment between the predicted and observed damage patterns during the comparison for the 2020 Samos earthquake (M w =7.0) is noted, indicating an IO performance level. This suggests that the Khan structures sustained only minor non-structural damage, underscoring their resilience in maintaining their structural integrity under low seismic demands. These findings emphasize the necessity of incorporating directional vulnerabilities and observed damage patterns into seismic risk assessments, advocating for targeted retrofitting strategies specifically designed to address the unique structural configurations of historical Khan monuments, thereby enhancing their resilience to higher seismic demands. 1. Introduction Historical structures have a significant role in carrying out the cultural inheritance of a country, and they are one of the most valuable pieces of cultural accumulation [1]. There are many historical buildings, religious monuments, and ruins of our ancestors [2]. The seismic behavior of ancient masonry buildings is particularly difficult to characterize and depends on several factors, namely the properties of the materials, the geometry of the structure, the connections between structural and non-structural elements, the stiffness of the horizontal diaphragms, and building condition [3]. Moreover, many historical buildings are quite vulnerable because they were built with low-resistance materials. These buildings also have insufficient connections among the various construction parts, such as masonry walls and floors [4,5]. Such problems in historical masonry buildings lead to an overturning collapse of the perimeter walls under horizontal earthquake records [6]. Although historic structures are highly susceptible to earthquake damage, the methods for assessing their seismic demand and capacity remain insufficient [7] Evaluating seismic risk in masonry buildings poses a significant engineering challenge because of the diverse variables influencing their structural behavior and the inherent uncertainties that increase with earthquake intensity [8,9]. The use of fragility curves is an effective strategy for probabilistically characterizing the seismic response of masonry structures [10-12]. Numerous applications of this approach to assess masonry structures can be found in the literature, underscoring its * Corresponding author. E-mail address: [email protected] (S. Karimzadeh). Contents lists available at ScienceDirect Structures journal homepage: www.elsevier.com/locate/structures https://doi.org/10.1016/j.istruc.2025.110208 Received 9 August 2024; Received in revised form 17 August 2025; Accepted 11 September 2025 Structures 81 (2025) 110208 Available online 16 September 2025 2352-0124/© 2025 The Author(s). Published by Elsevier Ltd on behalf of Institution of Structural Engineers. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). significance and efficacy [13–25,12,26–30]. In particular, Lourenço and Roque [31] provided insights into the seismic vulnerability of ancient masonry structures and offered simplified methods for evaluating seismic risks, making it useful for calibrating models to assess ancient buildings. Rota et al. [12] discussed both classical and advanced approaches to structural analysis of historic masonry structures and provided model calibration techniques that can aid in accurate seismic assessment of ancient structures. Betti and Vignoli [32] emphasized the seismic behavior and numerical model calibration of an ancient basilica, providing insights that apply to various historic buildings. Valente and Milani [33] used Finite Element Method (FEM) and simplified approaches to assess the seismic vulnerability of masonry towers, focusing on calibration techniques and comparing results with observed damage. Karimzadeh et al. [22] and Karimzadeh et al. [23] examined the fragility analysis of masonry buildings in Erzincan, Türkiye, using both simulated and real ground motion records. These authors developed fragility curves to show the probability of different damage states occurring at various seismic intensity levels in the buildings. The study by Ural and Çelik [34] developed a 3D numerical model to examine the structural performance of the Tahtani Mosque in Türkiye and determined the structural damages and their causes in the body walls and vaults of the mosque. Couto et al. [16] examined the seismic performance of historic residential buildings in Lisbon city center, focusing specifically on the effect of ground settlements, and developed fragility curves to evaluate the vulnerability of these structures under seismic loads. Marotta et al. [24] developed parametric seismic fragility curves for historical churches, taking into account various parameters, such as building geometry, material properties, and construction techniques, that affect their seismic response. They measured the probability of the occurrence of different damage states under different seismic intensities. Sandoli et al. [26] developed fragility curves for Italian unreinforced masonry (URM) buildings using a hybrid method that combined empirical data and numerical analysis. Biglari et al. [14] evaluated the structural integrity of historical mosques in Kermanshah, Iran, using field research and numerical modeling techniques. By developing fragility curves, the authors measured the probability of various damage states under different seismic intensities. Bernardo et al. [13] investigated the seismic vulnerability and fragility of pre-code masonry buildings in Portugal and conducted a comprehensive assessment using field investigations and numerical simulations to evaluate the structural integrity of these buildings under seismic loads. Usta and Bozda˘ g [28] developed fragility curves to conduct a seismic fragility analysis of traditional Himis structures in Türkiye and measured the probability of different damage states occurring under various seismic intensities. Bernardo et al. [20] evaluated the seismic performance and vulnerability of typological masonry buildings in the Azores, Portugal, by conducting detailed numerical simulations using simulated ground motion records. Follador et al. [17] investigated the effects of retrofit interventions on the seismic fragility of Italian masonry residential buildings. Using numerical simulations, they developed fragility curves to quantify the probability of different damage states occurring under seismic loads before and after retrofitting. Cima et al. [15] emphasized the importance of strengthening to increase the durability of historic and traditional masonry structures by developing and analyzing fragility curves and showed how different strengthening interventions can reduce the probability of damage at various seismic intensity levels. Prajapati et al. [25] used detailed numerical simulations to evaluate the structural performance of the Nepalese pagoda temple under seismic loads and assessed its seismic fragility. Cima et al. [15] developed fragility curves for the seismic evaluation of masonry buildings in historical centers prone to out-of-plane failure modes, taking into account the specific damage mechanisms that occur out-of-plane. Feizolahbeigi et al. [35] aimed to evaluate the seismic performance of a dome and surrounding minarets as secondary elements. Three-dimensional numerical models were developed according to the defined earthquake scenarios. Seismic analyses, including pushover analysis and boundary analysis, were performed to calibrate and evaluate the seismic behavior of the entire structure. The city of Izmir, situated in Türkiye’s highest earthquake hazard zone, lies within a highly seismically active area on the western edge of the Anatolian plate. It is a city well-known for its rich cultural heritage, featuring numerous historic masonry structures, including Khans (a type of inn that functioned as a trading center and hostel). Unexpected large earthquakes can occur frequently in this region and cause serious damage to buildings. Many earthquakes with surface magnitude M s ≥4.9 have occurred near Izmir between 1900 and 2005 [36]. More recently, three earthquakes with magnitudes M s =5.5, 5.9, and 5.9 occurred in the same region as the 2003 earthquake but did not cause any damage in Izmir [36]. In addition, on October 30, 2020, a severe earthquake of moment magnitude M w =7.0 occurred near the island of Samos, Greece, about 70 km southwest of Izmir [37]. The earthquake, which was strongly felt in Izmir, caused minor structural damage to some buildings, while 618 buildings suffered major damage, 660 buildings suffered moderate damage, and 7671 buildings suffered minor damage. 9 buildings were demolished, and there was a loss of life and property due to these demolitions [38]. In this earthquake, 118 people lost their lives, 1032 people were injured, and 9 buildings collapsed. Although the Samos earthquake was not in Izmir and was a distant earthquake, it caused damage to the structures. This event highlighted the seismic risk in the region and the importance of continued vigilance and preparedness [38]. Recent studies investigating the 2020 Samos event in the Izmir region have reported the hazard level and risk of structures in the region. Lekkas et al. [39] examined the preparedness and emergency response efforts after the 2020 Samos earthquake and evaluated the effectiveness of emergency actions in reducing the impact of the earthquake. Kiratzi et al. [40] analyzed the seismic characteristics of this event. Their study contributed to a broader understanding of seismic activity in the Aegean Sea and provided insights into the tectonic processes driving earthquakes in the region [40]. Altunisik et al. [41] reviewed the damage from this earthquake, considering the Turkish building earthquake design regulation [42]. Onat et al. [43] presented a field investigation and structural assessment of the damage caused by this earthquake. The authors emphasized the significant influence of local geological conditions on seismic damage and the need to incorporate these factors into building codes and disaster preparedness plans to mitigate earthquake risk. Askan et al. [44] presented a comprehensive analysis of the earthquake, focusing on the strong motion data collected during the event. The authors analyzed recorded ground motions to understand the characteristics of the earthquake and its impact on the surrounding region. The study examined the spatial distribution of seismic intensity and evaluated the response of various structures to earthquakes [44]. Cetin et al. [45] investigated site effects in the Izmir Bay area during this event. The authors analyzed the impact of local soil conditions on the seismic response in the study area. The results highlighted significant variations in ground shaking across different sites, emphasizing the importance of considering local geotechnical conditions in earthquake engineering and hazard mitigation. This study makes a significant contribution by addressing the gap in the literature in seismic fragility analyses of the historical Khan structures in the Kemeralti district of Izmir. Unlike previous regional studies that have predominantly focused on geotechnical and seismological investigations or seismic damage to other building types [7,12,16,23, 31], this research specifically evaluates the seismic performance of Khan structures, incorporating both Xand Y-directional responses. By doing so, it offers valuable insights that not only enhance the understanding of these structures’ resilience but also guide targeted preservation and retrofitting strategies, playing a critical role in future risk assessments and restoration efforts. A conceptual flowchart illustrating the analysis procedures adopted in this study is presented in Fig. 1. Within this framework, field P. Usta et al. Structures 81 (2025) 110208 2 observations are conducted to gather data on the structural characteristics of the selected Khans. Next, the finite element models are developed, followed by time history analysis using SAP2000 software [46]. Then, fragility curves are generated using PGA as the primary intensity measure. The results are analyzed using the impact of different PGA levels on the damage potential of the Khan structures in different structural directions (X and Y). Finally, the results are validated by comparing the predicted damage level at the intensity level of the 2020 Samos earthquake with the observed damage to various structures from the field observation. 2. Case study: historical Khans Türkiye, which bears the traces of many civilizations and has a deeprooted history, has a significant historical and cultural heritage in terms of its geographical location and characteristics. The built cultural heritage is of great importance in transferring historical values to future generations. There are many important historical buildings in Türkiye, and some of these buildings are located in Izmir, which was chosen as the study area. The Khans examined in this article are in the Kemeralti region, Izmir. This region is characterized by soft sedimentary layers that can significantly influence seismic responses. The soft sediments can amplify ground motion, making the structures more susceptible to seismic damage and posing unique challenges for seismic risk assessment of the Khans. Fig. 2 illustrates the spatial distribution of the six Khans considered in this study, offering insights into how their positioning on soft sediments may affect their seismic vulnerability. The figure includes an earthquake hazard map that highlights the seismicity of the region, a regional map showing the specific locations of the Khans, and individual images of each Khan. The listed Khans are Çakalo˘ glu, Abdurrahman, Büyük Karaosmano˘ glu, Demir, Keten, and Fazlıo˘ glu. The map details their precise positions within the Kemeralti commercial district. The typologies, material properties, and dimensions of each Khan are critical for evaluating their structural integrity and behavior under seismic loading, as well as identifying potential retrofitting needs. The features and plans of the Izmir historical Khan structures examined in this article are depicted in Fig. 3. This figure provides further details on the construction dates and materials of these historical Khans. Specifically, these Khans were built in the 18th and 19th centuries using various combinations of cut stone, rubble stone, and brick. Such information is vital for understanding their geometric characteristics and construction techniques, which influence their seismic performance. 3. Dynamic analysis of historical Khans Three primary methodologies are used for numerical modeling of masonry structures, depending on the system’s scale and the precision required: simplified micro modeling, detailed micro modeling, and macro modeling [48]. In detailed micro modeling, the material properties of masonry units and mortar are evaluated separately. Simplified micro modeling extends the masonry units to include half the mortar layer’s width, neglecting the mortar itself. In macro modeling, masonry is treated as a composite material, ignoring the distinction between units and mortar, which significantly reduces analysis time for large systems [48-50]. In the case of inadequate interconnection or interlocking of structural elements within a historical building, the potential for localized failure mechanisms to emerge in specific areas is heightened. These mechanisms often involve out-of-plane displacements, where walls or other elements move perpendicular to their plane, compromising the structural integrity and affecting the accuracy of the analysis. The distribution of mode shapes provides crucial insights into the behavior of the structure. By examining these mode shapes, it becomes possible to evaluate the performance of surrounding walls, assess the torsional rigidity (resistance to twisting), and understand how the structure responds to both in-plane (within the plane of walls) and out-of-plane deformations. This analysis helps pinpoint the structural directions or components that are either particularly vulnerable or resilient, allowing for the identification of areas that need reinforcement or are more likely to experience significant deformation during seismic events. In this study, the preliminary assessment of the seismic vulnerability of historic masonry Khans is based on the macro-modeling approach, a convenient and widely adopted method in the literature (e.g., [51-57]). In this approach, masonry is treated as a homogeneous and isotropic continuum, representing the combined behavior of stone units and lime Fig. 1. Conceptual flowchart of the fragility analysis procedures adopted in this study. P. Usta et al. Structures 81 (2025) 110208 3 mortar, without modeling the interface joints explicitly. This simplification is particularly suitable for large-scale historical structures where detailed material data are limited, as it significantly reduces computational demand while maintaining sufficient accuracy for preliminary seismic assessments. In the finite element implementation, the primary load-bearing walls are modeled using 8-node solid (hexahedral) elements capable of capturing three-dimensional stress states, whereas the roof systems are modeled with 4-node shell elements to represent membrane and bending behavior in thin components. This combination allows the model to differentiate the structural response of walls and Fig. 2. Hazard map of the study area, including the location of 6 selected Khans [47]. P. Usta et al. Structures 81 (2025) 110208 4 Fig. 3. Features and plans of the Izmir historical Khan structures (dimensions are in meters). P. Usta et al. Structures 81 (2025) 110208 5 roof diaphragms while preserving computational efficiency. The adopted strategy ensures that the essential stiffness and mass distribution of the Khans are accurately represented, enabling reliable estimation of their seismic demand and directional vulnerability. Among analysis approaches, nonlinear analysis is widely regarded as the most accurate method, but it poses challenges during the preliminary assessment stages because it necessitates detailed structural information. In situations where detailed experimental and field data on structural information are limited, linear analysis can be readily applied for the preliminary evaluation of masonry structures [58]. It requires less computational power, making it a straightforward and intuitive tool for practitioners, albeit with a certain level of uncertainty. In this study, given the limited information on the Khans, we opted for a linear time history analysis approach to obtain preliminary results on the seismic demand of masonry structures. When modeling historical structures using this approach, it is crucial to accurately reflect the geometric properties of the structure in the model. To develop three-dimensional models of the historical Khans, SAP2000 V23 software was used while linear analysis in the time domain was performed [59]. When creating the three-dimensional model, the geometrical survey of the historical structure was used, and the cross-section, plan, and dimensions were transferred to the model through available drawings [60]. During the modeling of the historical structures, the accuracy of the three-dimensional models was also verified using photographs taken from different angles and sections of the buildings. A macro modeling approach is employed herein in which the wall material (stone and mortar) has been modeled together as a single material property. The model consists of joints, restraints, solid, and shell elements. The primary load-bearing masonry walls were modeled using 8-node solid (hexahedral) elements, which can capture threedimensional volumetric stress and strain states. These elements are Fig. 3. (continued). P. Usta et al. Structures 81 (2025) 110208 6 particularly effective in simulating the behavior of thick masonry walls, especially where out-of-plane responses are critical. In contrast, the roof systems were modeled using 4-node shell elements, which are designed to represent both membrane and plate actions in thin structural components. This combination of solid and shell elements allowed for an accurate and computationally efficient representation of the distinct structural behaviors observed in masonry wall and roof components. The supports, representing the structural foundation, are assumed to be fixed for the analysis. The damping ratio is set to be 5 % [61]. The FEM of the Khans and the number of elements used to model the Khans are presented in Fig. 4. The size and shape of the mesh used are critical for accurately predicting stress and/or strain values in the FEM. A higher number of nodes in the FEM can lead to excessively long computation times. Therefore, mesh sizes were adjusted to achieve a suitable mesh concentration, and analyses were repeated for various scenarios. In modeling, the Sweep mesh was selected as it improves measurement distribution throughout the FEM and requires fewer nodes, as stated in the study by Karalar and Yes¸il [62]. Mesh selection was tested with various mesh sizes, adjusted along the length and perpendicular to the length in width. A mesh sensitivity study was conducted by systematically varying element sizes along both the longitudinal and transverse directions, ranging from 0.80 m to 0.20 m in 0.10 m increments. As shown in Fig. 5, the natural frequencies stabilize with finer mesh resolution, indicating convergence and confirming the adequacy of the selected element size for dynamic analysis. Accordingly, the coarsest mesh yielding converged results was adopted for each structure to ensure reliable simulation outcomes without excessive computational demand [62]. To provide transparency regarding the computational resources and efficiency of the analyses, details of the software, hardware, and runtime are reported as follows: All finite element analyses and mesh sensitivity studies were performed using SAP2000 v23 on a desktop computer with an Intel Core i7–12700 CPU, 32 GB RAM, and Windows 11 OS. The mesh sensitivity analysis required approximately 4–6 h per model, depending on the number of elements, while the linear time history analyses for the full set of 30 ground motions required 6–10 h per Khan structure. The materials used in the construction of the historic Khans of Izmir demonstrate a variety of masonry techniques that reflect regional practices and resource availability. The primary materials are cut stone, rubble, and brick, which are used in different combinations. These stones serve as the main load-bearing elements, providing compressive strength, while the bricks often contribute to infill and decorative details. The mortar used in these structures is generally lime-based, which was widely used in historic masonry because of its flexibility and compatibility with stone materials. Lime mortars, unlike modern cement-based mortars, allow for better movement accommodation and moisture exchange, which is essential for the longevity and structural integrity of masonry in seismic regions. The specific characteristics of these materials, including the type and quality of the stone and the composition of the lime mortar, significantly influence the seismic behavior of the Khans. In this study, the material definition was simplified to reflect the dominant structural components more clearly. Although the historical Khans include both cut and rubble stone, these materials were treated as a single stone type in the model due to their identical mechanical properties. Brick elements, observed in small quantities and mainly used for infill or decorative purposes, were not modeled separately. A macro modeling approach was adopted, where the masonry walls were considered as homogeneous materials. Accordingly, the analysis focused on the mechanical properties of the main load-bearing material, which is stone. An extensive literature review was conducted to determine the material properties without causing any damage to the historical texture. The properties of the covering materials used by Hrasnica and Medic [63] and Türkmen and Bilgin [64] were among the main sources for this study. Considering these references, the model types and material properties of the elements used in Khan’s structural model are presented in Table 1. It is noted that, since a linear elastic material model was adopted in SAP2000, compressive and tensile strength values were not required for the analyses. However, upon reviewing the available references, it was found that for stone masonry, a uniaxial compressive strength of 5 MPa and a null uniaxial tensile strength were reported, while no specific information was provided for the other materials (see Table 1). The loads acting on these historical Khans are divided into two categories: gravity loads and lateral loads. Gravity loads include dead loads, while lateral loads include earthquake loads. First, a threedimensional model was created, and gravity loads were applied. Then, predefined lateral earthquake loads were applied to the historical Khans. In the dynamic analysis of the structures, the ground was considered as a fixed support. 30 earthquake records covering different seismological characteristics and intensity levels are selected from the PEER NGAWest2 database recorded in soil classes A, B, and C [65]. Regarding the selection of accelerograms, we acknowledge that soil type A (rock) is uncommon in the vicinity of the Khans. However, this choice was made to ensure a diverse distribution of PGA values derived from global records. While many of these records correspond to soil classes B and C, those from soil class A generally exhibit lower PGA levels (approximately <0.1 g). Such lower PGA levels imply that the structures would primarily exhibit linear elastic behavior under these conditions, aligning with the linear time history analysis employed in this study. In this low PGA range, amplifications of softer soils at longer periods have minimal impact on the structural response. Subsequently, a linear time history analysis was performed on historical Khan structures for 30 different records in the x (parallel to the main façade) and y (perpendicular to the main façade) directions. The x and y directions of the historical Khans can be seen in the finite element model in Fig. 4. The characteristics of the ground motion data are summarized in Table 2. The PGA of the selected records ranges within 0.01–0.82 g, while the PGV ranges within 2.6–79.3 cm/s. Historical structures generally exhibit linear behavior within certain limits. Analyses conducted within these limits can adequately determine the load-carrying capacity and behavior of structural elements [66]. Linear analysis can be sufficient for an initial assessment and a quick, general evaluation of the current state of historical structures. This is particularly useful for identifying the overall stability and critical areas of the structure [67]. Although nonlinear approaches provide more realistic results in the seismic evaluation of historical structures, these approaches require detailed structural models and computational effort. In the literature, several authors, e.g. Betti et al. [68], Parisi et al. [69], H¨ okelekli et al. [70], Valente [56], Pouraminian [71], Cattari et al. [72], Valente [57], Xu et al. [73], and Ademovi´ c et al. [74] utilized nonlinear methods with detailed finite element models to evaluate the seismic response of historic masonry structures. On the other hand, linear analysis offers a simpler and faster calculation process [75]. Many regulations require the use of linear analysis to initially assess the safety of structures. Linear analysis helps carry out initial evaluations and determine reinforcement needs in compliance with regulations [76]. In conclusion, using linear analysis in SAP2000 for modeling and reinforcing historical structures allows for quick and effective initial assessments. This approach stands out with its compliance with regulations, computational efficiency, and sufficient accuracy within specified timeframes. Therefore, in this study, the linear analysis method was used to rapidly evaluate historical Khan structures. 4. Model verification In general, national codes provide simplified equations for estimating the fundamental frequency of structures. These formulas are particularly useful for applications such as static seismic analyses, where the distribution of static lateral forces is determined in a simplified manner, avoiding the need for modal linear analysis [77]. Moreover, these formulas can also serve as a benchmark for model validation by enabling frequency value comparisons, ensuring that numerical models P. Usta et al. Structures 81 (2025) 110208 7 Fig. 4. Finite element modeling of the historical Khan structures. P. Usta et al. Structures 81 (2025) 110208 8 accurately represent the dynamic behavior of structures. In this study, for assessing fundamental frequency, empirical formulas by ASCE 07–16 [78], C-NTC-2019 [79,80] are utilized to facilitate a simplified and pragmatic evaluation of the structural behavior of Khan structures, thereby enabling the numerical model’s accuracy. We note that the most recent update of the Italian Design Code NTC-2018 [81] introduces a relationship for estimating the natural frequency of a structure based on the lateral elastic displacement at its highest point under seismic loading. This approach assumes that the displacement is determined through, at minimum, an equivalent static seismic analysis, enabling the differentiation of oscillation frequencies along the two principal directions of the structure. While this method provides a more detailed framework, this study adopts the provisions of C-NTC-2019, as they directly relate to the geometric characteristics of the analyzed structures and ensure compatibility with the other formulas and methodologies considered. Fig. 6 provides a comparison of natural frequencies (in Hz) for several historical Khan structures calculated using four different approaches: ASCE 07–16, C-NTC-2019, and the numerical model. The empirical methods (ASCE 07–16 and C-NTC-2019) show strong alignment, with closely matching frequency values across all structures, indicating their consistency and reliability. However, the numerical model results exhibit some deviations, particularly for certain Khans, such as Abdurrahman Khan and Karaosmano˘ glu Khan, where significantly lower or higher frequencies are observed. These discrepancies highlight the ability of the numerical model to capture detailed structural characteristics, including material heterogeneity, irregular geometries, and specific boundary conditions, which are not fully addressed by simplified empirical formulas. The frequency variation among the Khans reflects differences in structural properties, with Fazlıo˘ glu Khan showing the highest frequencies (~8.4–8.6 Hz) due to its smaller height and increased stiffness, while Abdurrahman Khan and Demir Khan exhibit lower frequencies (~4–5 Hz), indicating more flexible structural behavior. Overall, the comparison underscores the importance of combining empirical and numerical approaches to provide both quick estimations and detailed insights into the seismic performance of masonry structures. Finally, to assess the accuracy of the numerical models, the fundamental vibration frequencies were additionally compared with values reported in the literature. The computed frequencies range from 4.2 to 8.8 Hz and are generally consistent with those reported for similar historical masonry structures in state-of-the-art studies. According to Aksoy and Aydo˘ gmus¸ [82], S ¸ahin [83], Günaydin et al. [84], Akın and Alag¨ oz [85], and Kılıç Demircan [86], typical fundamental frequencies for such structures fall within the range of 1.0–7.2 Hz, depending on parameters such as structural stiffness, span length, number of stories, and plan geometry. It is important to note that relatively few numerical modeling studies have specifically focused on historical Khan-type buildings. Fig. 5. Fundamental frequency mesh convergence analysis for historical Khan structures. Table 1 Material properties [63,64]. Structural element Element type Model type Elastic modulus (MPa) Density (kN/ m 3 ) Poisson’s ratio Uniaxial Compressive Strength (MPa) Uniaxial Tensile Strength (MPa) Wall Stone Solid 3000.00 20.00 0.20 5.00 0 Roof Covered Shell 13000.00 2.24 0.16 - - P. Usta et al. Structures 81 (2025) 110208 9 [45] Cetin KO, Altun S, Askan A, Akgün M, Sezer A, Kıncal C, ¨ Ozda˘ g ¨ OC, ˙ Ipek Y, Unutmaz B, Gülerce Z, ¨ Ozacar AA. Site effects in izmir bay during the October 30, 2020, M7.0 samos earthquake. Soil Dyn Earthq Eng 2022;152:107051. [46] SAP2000 V23. Integrated structural analysis and design software. Berkeley, California: Computers and Structures, Inc.; 2014. [47] UNDRR, 2024. Access: https://www.preventionweb.net/news/forensic-analysisreveals-causes-building-damage-izmir-oct-30-aegean-sea-earthquake, Access date: 26.07.2024. [48] Lourenço PB. Computational strategies for masonry structures (Doctoral dissertation, Delft University of Technology). Delft University Press; 1996. [49] Lourenço, P.B., & Cali` o, I. (2018). Macro-Element Nonlinear Dynamic Analysis for the Assessment of the Seismic Vulnerability of Masonry Structures. [50] Marques R, Lourenço PB. Unreinforced and confined masonry buildings in seismic regions: validation of macro-element models and cost analysis. Eng Struct 2014; 64:52–62. https://doi.org/10.1016/j.engstruct.2014.01.003. [51] Da Silva LC, Milani G. A FE-based macro-element for the assessment of masonry structures: linear static, vibration, and non-linear cyclic analyses. Appl Sci 2022; 12(3):1248. [52] Funari MF, Silva LC, Savalle N, Lourenço PB. A concurrent micro/macro FEmodel optimized with a limit analysis tool for the assessment of dry-joint masonry structures. Int J Multiscale Comput Eng 2022;20(5). [53] Gürbüz M, Kocaman ˙ I. Enhancing seismic resilience: a proposed reinforcement technique for historical minarets. Eng Fail Anal 2024;156:107832. [54] Pant` o B, Cannizzaro F, Caddemi S, Cali` o I. 3D macro-element modelling approach for seismic assessment of historical masonry churches. Adv Eng Softw 2016;97: 40–59. [55] Sangiorgio V, Uva G, Adam JM. Integrated seismic vulnerability assessment of historical masonry churches including architectural and artistic assets based on macro-element approach. Int J Archit Herit 2021;15(11):1609–22. [56] Valente M. Seismic behavior and damage assessment of two historical fortified masonry palaces with corner towers. Eng Fail Anal 2022;134:106003. [57] Valente M. Earthquake response and damage patterns assessment of two historical masonry churches with bell tower. Eng Fail Anal 2023;151:107418. [58] Lombardi L, De Luca F. Derivation of fragility curves at design stage through linear time-history analysis. Eng Struct 2020;219:110900. [59] Wilson EI, Habibullah A. Structural analysis programs. Berkeley, CA: Computers and Structures, Inc.; 1992. [60] DGF (2005). Direction General of Foundations. Izmir. [61] Hoveidae N, Fathi A, Karimzadeh S. Seismic damage assessment of a historic masonry building under simulated scenario earthquakes: a case study for ArgeTabriz. Soil Dyn Earthq Eng 2021;147:106732. [62] Karalar M, Yes ¸il M. Structural investigation of a historical masonry arch bridge under Far-Fault earthquakes. Tehˇ cki Vjesn 2023;30(5):1443–50. https://doi.org/ 10.17559/TV-20220711123816. [63] Hrasnica, M., & Medic, S. (2012). Seismic strengthening of historical stone masonry structures in Bosnia Herzegovina. 15WCEE, Lisbon. [64] Türkmen, M., & Bilgin, H. (2016). Structural behavior of domed roof systems in traditional architectural buildings. Retrieved on April 27, 2016, from [website URL]. [65] Ancheta TD, Darragh RB, Stewart JP, Seyhan E, Silva WJ, Chiou BS-J, Wooddell KE, Graves RW, Kottke AR, Boore DM, Kishida T, Donahue JL. NGAWest2 database. Earthq Spectra 2014;30(3):989–1005. [66] Kappos AJ, Manolis GD, Moschonas IF. Seismic assessment and design of R/C bridges with irregular configuration, including SSI effects. Eng Struct 2002;24 (10):1337–48. [67] Chopra AK. Dynamics of structures. Pearson Education India; 2007. [68] Betti M, Galano L, Vignoli A. Time-history seismic analysis of masonry buildings: a comparison between two non-linear modelling approaches. Buildings 2015;5 (2):597–621. [69] Parisi F, Balestrieri C, Varum H. Nonlinear finite element model for traditional adobe masonry. Constr Build Mater 2019;223:450–62. [70] H¨ okelekli E, Demir A, Ercan E, Nohutçu H, Karabulut A. Seismic assessment in a historical masonry minaret by linear and non-linear seismic analyses. Periodica polytechnica. Civ Eng 2020;64(2):438–48. [71] Pouraminian M. Multi-hazard reliability assessment of historical brick minarets. J Build Pathol Rehabil 2022;7(1):10. [72] Cattari S, Calderoni B, Cali` o I, Camata G, de Miranda S, Magenes G, Saetta A. Nonlinear modeling of the seismic response of masonry structures: critical review and open issues towards engineering practice. Bull Earthq Eng 2022;20(4): 1939–97. [73] Xu D, Xie Q, Hao W. Seismic damage evaluation of historical masonry towers through numerical model. Bull Earthq Eng 2024;22(4):2235–66. [74] Ademovi´ c N, Hadzima-Nyarko M, Zagora N, Stojnovi´ c V. Various numerical modeling procedures of XIX-century masonry building. Eng Struct 2024;301: 117361. [75] Bathe, K.J. (2006). Finite element procedures. Klaus-Jurgen Bathe. [76] FEMA 356 (2000). "FEMA 356 Prestandard and Commentary for the Seismic Rehabilitation of Buildings", ASCE for the Federal Emergency Management Agency, Washington, D.C., November 2000. [77] Lopez S, D’Amato M, Ramos L, Laterza M, Lourenço PB. Simplified formulations for estimating the main frequencies of ancient masonry churches. Front Built Environ 2019;5:18. [78] ASCE 07-16 (2017). Minimum Design Loads and Associated Criteria for Buildings and Other Structures American Society of Civil, Engineers. Virginia, NV: Reston. [79] Circolare (2019). Ministero delle Infrastrutture e dei Trasporti. (2019). Istruzioni per l ′ applicazione dell’aggiornamento delle “Norme Tecniche per le Costruzioni” di cui al decreto ministeriale 17 gennaio 2018 (Circolare 21 gennaio 2019, n. 7). Gazzetta Ufficiale della Repubblica Italiana. [80] NTC-2008 (2008). Norme Tecniche per le Costruzioni. D.M. 14/01/2008, S.O. n. 30 of the Official Gazette of the Italian Republic 2008. [81] NTC-2018 (2018). Norme Tecniche per le Costruzioni. Gazzetta Ufficiale della Repubblica Italiana, Supplemento Ordinario n. 8 del 20 febbraio 2018. [82] Aksoy, E., & Aydo˘ gmus¸, F. (2017). Tarihi yapıların deprem analizi ve Kargı Han ¨ orne˘ gi. Uluslararası Katılımlı 6. Tarihi Yapıların Korunması ve Güçlendirilmesi Sempozyumu, 411-419. [83] S¸ahin, E. (2019). Structural analyses and assessment of historical Çardak Caravanserai in Denizli (Master’s thesis, Izmir Institute of Technology (Turkey)). [84] Günaydin M, Demirkir C, Altunis¸ik AC, Gezer ED, Genc AF, Okur FY. Diagnosis and monitoring of historical timber velipas ¸a han building prior to restoration. Int J Archit Herit 2023;17(2):285–309. [85] Akın S, Alag¨ oz A. Structural behavior of historical obruk inn under different earthquakes. Chall J Struct Mech 2024;10(1):21–33. [86] Kılıç Demircan R. Determination of earthquake resistance of historical masonry inn with finite element analysis. Bull Earthq Eng 2025:1–25. [87] Cuadra C, Saito T, Zavala C. Diagnosis for seismic vulnerability evaluation of historical buildings in Lima, Peru. J Disaster Res 2013;8(2):320–7. [88] D’Ayala D, Speranza E. Definition of collapse mechanisms and seismic vulnerability of historic masonry buildings. Earthq Spectra 2003;19(3):479–509. https://doi.org/10.1193/1.1599896. [89] Da Silva LC, Milani G, Lourenço PB. Probabilistic-based discrete model for the seismic fragility assessment of masonry structures. Structures 2023;52:506–23. [90] Negulescu C, Ulrich T, Baills A, et al. Fragility curves for masonry structures submitted to permanent ground displacements and earthquakes. Nat Hazards 2014;74:1461–74. https://doi.org/10.1007/s11069-014-1253-x. [91] Tarque N, Manchego A, Lov´ on H, Blondet M, Varum H. Experimental in-plane behaviour and drift-based fragility assessment of typical Peruvian confined masonry walls. Constr Build Mater 2022;341:127893. [92] Vicente, R. (2008). Estrat´ egias e metodologias para intervenç˜ oes de reabilitaç˜ ao urbana. Avaliaç˜ ao da vulnerabilidade e do rísco sísmico do edificado da Baixa de Coimbra. PhD Thesis, Universidade de Aveiro, Portugal. [93] Park J, Towashiraporn P, Craig JI, Goodno BJ. Seismic fragility analysis of lowrise unreinforced masonry structures. Eng Struct 2009;31(1):125–37. [94] Caicedo D, Tomi´ c I, Karimzadeh S, Bernardo V, Beyer K, Lourenço PB. Collapse fragility analysis of historical masonry buildings considering in-plane and out-ofplane response of masonry walls. Eng Struct 2024;319:118804. [95] Zentner I, Gündel M, Bonfils N. Fragility analysis methods: review of existing approaches and application. Nucl Eng Des 2017;323:245–58. [96] Yang J, Wang J, Zhuang H, Zhu L, Shen Y. Seismic fragility analysis for underground subway station structures considering ground motion combinations based on IDA and cloud. Soil Dyn Earthq Eng 2025;195:109386. [97] Mackie KR, Stojadinovi´ c B. Comparison of incremental dynamic, cloud, and stripe methods for computing probabilistic seismic demand models. Structures Congress 2005. Metropolis and Beyond; 2005. p. 1–11. [98] Nassirpour A, Song B, D’Ayala D. IDA & cloud method for fragility assessment of bare & infilled steel frame structures. 16th World Conference on Earthquake Engineering, 16. National Information Centre of Earthquake Engineering; 2017. p. 1–12. [99] Lombardi L, De Luca F, Macdonald J. Design of buildings through linear timehistory analysis optimising ground motion selection: a case study for RC-MRFs. Eng Struct 2019;192:279–95. [100] Karimzadeh S, Askan A, Erberik MA, Yakut A. Seismic damage assessment based on synthetic ground motion dataset: a case study for Erzincan, Turkey. Nat Hazards 2018;92:1371–97. P. Usta et al. Structures 81 (2025) 110208 16