Seismic Assessment and Strengthening of a Load-Bearing Masonry Structure Considering SSI Effects

Abstract

This article examines the seismic assessment and strengthening of a traditional load-bearing masonry structure subjected to strong motion data, with particular emphasis on the effects of soil–structure interaction (SSI). The case study is the Archaeological Museum of Lemnos (AML)—a three-storey building with a composite load-bearing system of timber-framed stone masonry. Over time, the structure has undergone irreversible modifications, primarily involving reinforced concrete (RC) interventions. The building’s seismic performance was evaluated using two finite element models developed in the SAP2000 software (v. 25.3.00). The first model simulates the original structure, strengthened by grout injections, while the second represents the current condition of the structural system following RC additions. Soil–structure interaction was also investigated, given that the local soil is classified as Category D according to Eurocode 8 (EC8). Each model was analyzed under two different support conditions: fixed-base and SSI-inclusive. A suite of appropriate accelerograms was applied to both models, in compliance with Eurocode 8 using the SeismoMatch software, and linear time-history analyses were conducted. The results underscore the significant impact of SSI on the increase of peak tensile stress and interstorey drift ratios (IDRs), and highlight the influence of different strengthening techniques on the seismic response of historic load-bearing masonry structures.

1. Introduction

Although relatively rare, destructive earthquakes can cause extensive damage or even lead to the partial or total collapse of historical load-bearing masonry buildings that have not been retrofitted in accordance with modern seismic codes.

These structures typically exhibit inherent structural weaknesses, such as inadequate connections between individual load-bearing components and the low mechanical strength of materials, which render them particularly vulnerable to seismic forces. Failure mechanisms—such as in-plane and out-of-plane bending of masonry walls, leading to wall overturning, disaggregation and/or collapse of external wall corners [1]—have been repeatedly observed in recent major earthquakes, including those in Lesbos (2017) [2] and Turkey (2023) [3,4], thus confirming the need for further investigation into the seismic performance of such structures.

To this end, this paper examines the Archaeological Museum of Lemnos—a three-storey load-bearing stone masonry building with a timber-tiled roof, located in the island’s capital, Myrina. The structure features a regular floor plan and a distinctive composite system of timber-framed load-bearing masonry. Analogous construction typologies are found in the traditional architecture of various regions, including Lesvos [5], Lefkada [6], Turkey [4], the Himalayas [7], Chile [8], and Portugal (Pombalino System) [9].

Generally, in timber-framed load-bearing masonry buildings, the timber framing functions as a secondary seismic response mechanism (substructure). Whether implemented as floor-level ties, reinforcement at critical masonry points (e.g., around openings), or as a timber space truss embedded within the masonry—often in combination with internal timber-framed partition walls (bagdadi)—this system enhances the compressive strength of the masonry (by up to 20%), prevents vertical cracking, and reduces overall cracking (by up to 50%) [10]. It also increases the shear capacity and deformability of the masonry prior to failure [10]. Notably, despite the large deformations they may undergo, such structures have demonstrated the ability to withstand major seismic events [7]. In addition, timber-framed interior partition walls contribute—among other things—to increasing the stiffness and overall strength of the structure.

It should be noted that significant interventions have been carried out on the building under study, including the addition of reinforced concrete (RC) elements (an RC slab at the first-floor level and a bond beam at the roof level), which have altered its original structural system at the diaphragm levels [11].

Regarding the seismicity of the broader study area—the northeastern Aegean Sea, where the western end of the North Anatolian Fault extends—significant earthquakes have been recorded, including the Agios Efstratios event on 20 February 1968 (Mw 7.0–7.4) and the Lemnos earthquake on 24 May 2014 (Mw 6.9). However, seismic acceleration data are only available for the latter, with a recorded peak ground acceleration (PGA) of 0.11 g [12].

According to a seismic hazard assessment study for the North Aegean region [13], PGAs of up to 0.3 g are estimated in areas near Lemnos under earthquake scenarios with magnitudes of Mw 7.0 or higher. Notably, during the 12 June 2017 earthquake on the neighboring island of Lesvos, accelerations reached values as high as 0.2 g—approaching the design PGA value specified by the Eurocode 8 response spectrum for the broader region [2].

To assess the seismic response of the AML to strong ground motions, three major seismic events were selected: the Northridge earthquake (Pacific Palisades—Sunset station, Los Angeles, CA, USA, 1994, Mw = 6.7) [14]; the Kahramanmaraş earthquake (Turkey, 2023, Mw = 7.7) [15]; and the Kocaeli earthquake (Düzce station, Turkey, 1999, Mw = 7.51) [14]. To adapt these seismic records to the target spectrum calculated for the AML, the procedure outlined in EC8 was followed. The response spectra of the selected ground motions were scaled to match the EC8 design spectrum within the period range of 0.2 T to 2 T, where T is the fundamental period of the system in each seismic direction.

Additionally, in order to compare the original and the existing structural systems, two finite element models were developed, one excluding and one incorporating RC elements. It should be noted that interventions involving reinforced concrete (e.g., slabs and bond beams) can have a significant impact on load-bearing masonry buildings, depending on factors such as the adequacy of slab-to-masonry connections, the proper use of reinforcement [4], and whether the intervention is accompanied by masonry strengthening [2].

Finally, given that the building is founded on class D soft soil, as defined by EC8, soil–structure interaction (SSI) effects were also investigated for each model. The results of the analyses are compared with those of the study by Genç et al., 2023 [16] which analyzed the seismic behavior of a historic building under both fixed-base conditions and soil–structure interaction (massed and massless) across three different soil categories: hard, medium, and soft. The results are also compared with the study of [17]. Also, Requena-Garcia-Cruz [18] examined the impact of soil-structure interaction (SSI) effects on the seismic analysis of cultural heritage buildings, such as the Mosque-Cathedral of Córdoba in Spain. An investigation of Özmen & Sayin [19] was conducted into the variations in the seismic response of a historical masonry church, utilizing four different SSI models in addition to a fixed base model that did not consider SSI. Furthermore, the research of Altiok and Demir [20] examines the seismic response of the historical Lala Mehmet Pasha minaret by taking into account Soil Structure Interaction (SSI). The influence of SSI on the out-of-plane performance of an ancient structure in Iran, known as Arge-Tabriz has been investigated by Fathi et al., [21]. Tzanakis et al., [22] explored the seismic performance of St. Titus Church located in Heraklion, Crete, Greece, along with the necessity for its seismic retrofitting. Additionally, the impacts of soil-structure interaction have been considered.

This research tackles significant deficiencies in the seismic evaluation of historic masonry structures by offering a thorough analysis that concurrently takes into account original construction methods, contemporary strengthening measures, and the effects of soil-structure interaction, elements that are seldom analyzed together in the current literature. The engineering significance of these structures is underscored by their dual function as cultural heritage assets and operational public buildings, rendering their seismic safety crucial for both the protection of lives and the preservation of heritage. Historic load-bearing masonry edifices, such as the Archaeological Museum of Lemnos, are especially susceptible due to their age, material deterioration, and construction techniques that predate modern seismic regulations. Nevertheless, they must continue to fulfill modern roles while preserving their historical authenticity. The originality of this study arises from its direct comparison of two strengthening approaches, traditional grout injection against modern reinforced concrete solutions, utilizing the same analytical methods and realistic seismic inputs that adhere to Eurocode 8 standards [23]. By integrating soil-structure interaction effects for Category D soils and concentrating on quantitative metrics like interstorey drift ratios and peak tensile stress, this research offers valuable insights for engineers and conservators striving to reconcile structural safety with heritage preservation needs, while also establishing a replicable framework for evaluating similar composite timber-framed stone masonry systems globally.

2. Case Study

2.1. Historical Background

The Archaeological Museum of Lemnos is located in Myrina, on the coastal front of the “Romeikos Gialos” area, between the cape of the Meteorological Station and the Castle peninsula (Figure 1). Settlement remains from the Late Neolithic period have been discovered in this area, while the earliest settlement of the “city” of Myrina was established on the Castle peninsula during the Early Iron Age. The museum is situated near the prehistoric settlement of Myrina, and together the two sites form an important archaeological complex. Positioned directly adjacent to the Ecclesiastical Museum of the Holy Metropolis of Lemnos, the Myrina Gymnasium (Figure 2), and the Pantelideion Building (formerly the Girls’ School), the AML constitutes an integral part of the architectural ensemble of Romeikos Gialos, developed between the 19th and early 20th centuries.

The building was constructed in the late 19th century to serve as the seat of the Turkish Administration, while the idea of housing the Archaeological Museum was first proposed in the 1930s. The plan, however, was not realized until several years later, following the end of World War II. In 1956, the building was repaired, and the first exhibition of the AML was held in 1961, following the repatriation of the island’s archaeological finds, which had been transferred for safekeeping during the war to the Museum of Mytilene (Lesvos) and the National Archaeological Museum of Athens. Thirty years later, in 1991, a new exhibition was held following additional repair work on the building (Figure 3a).

In May 2014, the island was struck by a strong earthquake that caused damage to both the museum and its exhibits (Figure 3b and Figure 4). Although the building’s repairs were completed in October 2014, it was decided not to reopen the exhibition on the second floor, as the museum required renovation to meet modern museological standards.

In 2019, the Greek Ministry of Culture approved a project entitled “Modernization of the Archaeological Museum of Lemnos”. The scope of the project included the strengthening of the building’s foundations and stone masonry, floor repairs, replacement of wooden window frames and door frames, application of new coatings, and the installation of a lift for people with disabilities. The re-exhibition of the museum’s antiquities is scheduled for summer 2025.

2.2. Intervention Phases

Over the course of its history, the building has undergone five distinct phases of structural and functional transformation. The first construction phase dates to the 19th century and corresponds to its original function as the seat of the Turkish Administration. The original architectural form of the building remains unchanged to this day [25].

The second construction phase refers to the period associated with the building’s first museum exhibition in 1961. In preparation for this exhibition, extensive and largely irreversible interventions were carried out in 1956. These included the replacement of the original timber floors at the ground and first levels with new reinforced concrete (RC) slabs, and the strengthening of the second-level floor with the addition of metal beams.

The third construction phase took place thirty years later, in 1991, during the re-exhibition of the Archaeological Museum’s collection. At that time, several significant and mostly reversible interventions were made to the building, including roof repairs, the construction of new sanitary facilities, the installation of electrical and mechanical systems, and the addition of new wooden flooring.

The fourth construction phase followed the earthquake of 24 May 2014, which had a magnitude of Mw 6.9 [12]. During this phase, a reinforced concrete bond beam (senaz) was constructed to strengthen the masonry walls, and both roof waterproofing and roof tile repairs were carried out.

The fifth and final construction phase is currently underway and is set to be completed in time for the opening of the museum’s new exhibition, scheduled for the summer of 2025. It involves significant and irreversible interventions, including the demolition of exterior walls on the southwest elevation to accommodate a lift for people with disabilities; construction of a new reinforced concrete floor at ground level; foundation strengthening through grout injection; repointing of the stone masonry walls and application of new traditional-style renderings; reconstruction of the external staircases; replacement of the wooden staircase; installation of new timber flooring on the second floor; and general renovation works in the interior spaces.

2.3. Architectural Documentation

The Archaeological Museum is an impressive three-storey structure comprising a low ground floor and two upper levels, with a total area of approximately 606 square meters and a height of 13.09 m. The plot includes outdoor storage areas totaling approximately 210 square meters, along with a courtyard that surrounds the building. Three of the building’s elevations are exposed, while the fourth—on the northeastern side—borders the courtyard of the Myrina Gymnasium. The main façade faces the seafront to the northwest and is fronted by an open space enclosed by a low fence.

Although the museum building is not officially designated as a protected monument, it is a representative example of its period, featuring elements of restrained neoclassical architecture typical of urban buildings of the late 19th and early 20th centuries. All three levels share a similar spatial configuration, with a symmetrical layout that consists of a large central hall flanked by rooms on either side. The primary exhibition spaces are located on the upper floors, while auxiliary rooms occupy the ground level. The floors are internally connected by a staircase located to the southwest of the hall, and externally by staircases positioned at ground level on both main façades [25].

Morphologically, the building adheres to the eclectic architectural style prevalent in Myrina—and particularly in the Romeikos Gialos area—at the end of the 19th century. Its design blends restrained neoclassical elements with features of local vernacular architecture. Numerous openings characterize the building’s façades, which follow symmetrical design principles, most rigorously applied to the two principal elevations—the northwestern and southeastern. The fundamental neoclassical tripartite division into base (lower ground floor), main storeys (two floors) and crown (roof) is retained. Distinctive features of the building’s form include the corner quoin stones and two monumental, semi-circular external stone staircases, symmetrically positioned at the centre of the main façades, with the second-floor balconies located directly above them [25].

2.4. Structural Description

2.4.1. Ground Floor

The exterior and interior walls of the ground floor are made of load-bearing stone masonry, except for one wall constructed with clay bricks (optoplinthi). The external load-bearing masonry is 65 cm thick, the internal is 50 cm, and the clay-brick wall 30 cm. The external masonry is built as three-leaf rubble construction with lime mortar applied on all wythes. Large dressed stones are positioned at the corners of the walls, while semi-dressed stones and clay bricks are used as infill between them, contributing to the stability and integrity of the masonry. The core is filled with rubble stone and mortar, forming a solid mass between the outer wythes [5]. The internal stone masonry is constructed as two-leaf rubble masonry (Figure 5a).

The stone used in the masonry, of volcanic origin and gray in color, was sourced from quarries near the settlements of Thanos and Romanou in Lemnos. According to the geotechnical study of the Castle of Myrina [26] and after the identification of the building stones, it was found that they belong to the categories of lavas found in the local quarries (dacites or ignimbrites) with an average strength of fbc = 50 MPa.

The mortar used in the building is cohesive and contains lime and large dark-colored coarse aggregates (brown and gray), collected from streambeds where volcanic rock erosion products accumulate. A small amount of ceramic material grains was also identified. The plaster applied to the ground-floor masonry is a lime-rich mortar of medium consistency, containing medium-grained and coarse-grained aggregates. The tensile strength of the interior masonry plaster was estimated at 78 kPa [27].

No timber lacing was found in the ground floor masonry; however, iron ties and anchor bars—visible on the façades (Figure 5b)—were used at floor level to strengthen the walls. These ties—metal bars running along the exterior of the masonry—are not connected to the wooden elements of the floor or the timber framing present on the second floor. Timber ties were found embedded within the internal stone masonry, oriented transversely across the wall at door openings (Figure 5a).

2.4.2. First Floor

Regarding the horizontal load-bearing elements, the original timber floor at the first level was replaced in 1956 (second construction phase) with a 15 cm thick reinforced concrete slab, supported by 35 cm wide horizontal RC beams. The ground floor connects to the first floor via an internal staircase, made of reinforced concrete and ceramic tiles (second construction phase), or externally through the central entrance and via the external circular staircase on the main façades.

The exterior masonry of the first floor consists of 50 cm thick load-bearing stonework and appears to be more carefully constructed than that of the ground floor, incorporating a specific technique involving clay bricks [5]. On the interior side of the masonry, above the openings, there are arched spandrels made of clay bricks (Figure 6a). The interior partition walls are made of 30 cm thick clay brick masonry, added during the second construction phase following the installation of the reinforced concrete slab. The plaster on the interior face of the first-floor masonry is 1 cm thick and contains dark-colored aggregate in a 1:1 weight ratio, along with a significant amount of plant fibers [27]. The first level is connected to the second by a wooden staircase (Figure 6b).

2.4.3. Second Floor

At the second level, the timber floor was reinforced with metal IPE beams during the second construction phase, while an additional timber layer (floorboards) was installed during the third phase. The floor system comprises simply supported timber joists with a cross-section of 10 × 10 cm, spaced at intervals of 40–50 cm. Floorboards and ceiling boards, each approximately 2.4 cm thick, are nailed above and below the joists, respectively.

At the this level, the building features a distinctive hybrid system that combines stone masonry with integrated timber framing. This construction technique has been examined and thoroughly documented by N. Karydis (2001) in his study of the traditional settlement of Eressos in Lesvos [5]. The walls consist of 50 cm thick two-leaf masonry, reinforced on the inner face with a dense timber framework, locally known as “friggia”. This system consists of vertical timber posts, typically with a cross-section of 10 × 10 cm, spaced 40–100 cm apart depending on the wall configuration (Figure 7a). In areas requiring increased stiffness—such as between openings—horizontal members and diagonal braces of similar cross-section are installed. N. Karydis (2001) [5] (pp. 6–7) notes that vertical posts are generally aligned with the corners of openings, while horizontal ties correspond to lintels and sills. The timber framework is self-supporting and structurally capable of carrying roof loads independently of the masonry, while also connecting both wall faces transversely at openings [5] (p. 8).

The interior walls of the second floor are of the baghdati type—timber framed partitions with a total thickness of 17 cm (Figure 7b). These lightweight walls complete the aforementioned timber frame system and contribute to the overall stiffness of the structure. The plaster used is of the same composition and strength as that of the first floor [27].

The building has a four-pitched timber roof of a distinctive structural configuration, covered with Roman-style clay tiles. Several additional timber members were added during the second, third, and fourth construction phases. However, the primary structural elements of the roof today include vertical posts, principle rafters and tie beams (14 × 14 cm cross-section), along with diagonal braces (7 × 14 cm cross-section).

The structure is founded on stone foundation strips aligned with the ground-floor load-bearing masonry walls. These foundations reach a depth of 82 cm from the axis of the ground level floor and feature a cross-sectional widening of 20 cm on either side of the masonry in the final 60 cm of depth.

As described in detail in this section, the building’s structural system is of particular interest. It has a rigid base formed by load-bearing masonry with relatively small openings. The first floor is reinforced with arched spandrels to improve load distribution onto the piers surrounding the openings, while the second floor has two mechanisms for absorbing seismic forces: the stone masonry itself and the system of timber elements (timber wall frames, baghdati-type walls and timber flooring). These systems collectively facilitate the transfer of roof loads to the lower levels. In the event of masonry failure during an earthquake, the loads are transferred to the timber frame system.

Table 1. Repair Methods.

	Model 1	Model 2
Three-leaf rubble masonry (ground floor)	Grout injections	Repointing works
Two-leaf rubble masonry (1st and 2nd floors)	Grout injections	Repointing works
1st level’s floor	Double layered timber flooring —timber joists	15 cm thick reinforced concrete slab supported by 35 cm wide horizontal RC beams
2nd level’s floor	Double layered timber flooring —timber joists	Triple layered timber flooring—timber joists
Openings	-	Demolition of exterior walls to accommodate a lift for people with disabilities
Roof beams	Wooden tie beams	Reinforced concrete bond beam (senaz)

depicts the repair methods of the structure under study.

The steel MRF study of [17] indicates that the incorporation of infill masonry walls (IMWs) can lead to a reduction in IDR__Med by as much as 32.26% and enhance collapse performance by 12.48%, thereby illustrating the efficacy of wall-based retrofitting methods. The present research’s comparison of grout injection strengthening with RC interventions offers a complementary analysis of various retrofitting philosophies applicable to masonry structures. While the steel study emphasizes the addition of infill walls to frame systems, the investigation of the authors explores how different strengthening methods (traditional versus modern) influence the seismic response of load-bearing masonry. Both studies enhance the understanding that the effectiveness of retrofitting is significantly influenced by the interplay between the chosen strengthening technique and the original structural framework.

3. Materials and Numerical Modeling

3.1. FE Models

The structural composition of the Archaeological Museum of Lemnos comprises a complex system that incorporates both traditional materials and later reinforced concrete additions. Each floor level exhibits different masonry types and construction techniques, which were carefully taken into account in the development of the numerical models.

The ground floor is constructed with three-leaf rubble masonry, while two-leaf masonry is used for the interior walls. On the first floor, the exterior walls are also constructed with two-leaf masonry, whereas the interior partitions are made of clay brick masonry. The second floor features a dual construction system: two-leaf masonry integrated with a traditional embedded timber frame system (friggia). Although structurally significant, this system was excluded from direct modelling due to its complexity and the variability in the performance of timber elements, which depends on factors such as age and the quality of joinery. Timber components were, however, included in the model to represent floor joists, ties, and wall plates. Subsequent intervention phases introduced reinforced concrete (RC) elements into the structure, including a floor slab with beams, RC bond beams, and IPE steel beams supporting the second-storey floor.

The numerical modeling and structural analysis were performed using SAP2000 [28]. For modeling purposes, the following material and section typologies were identified and assigned geometrical and mechanical properties:

• Three-leaf rubble masonry (ground floor exterior walls).
• Two-leaf rubble masonry (ground floor interior walls and first-floor exterior walls).
• Two-leaf rubble masonry with timber frames (second-floor exterior walls, modeled with adjusted unit weight to reflect embedded timber content).
• Clay brick masonry (ground floor and first-floor interior walls).
• Steel beams (IPE).
• Timber elements (tie beams, floor joists, and roof wall plates).
• Reinforced concrete elements (slab with beams, bond beam).

Two structural models were simulated to evaluate the seismic behavior of the building (Figure 8):

The first model (Model 1 or M1) represents the building with its original structure (wooden floors), assuming that the load-bearing masonry has been strengthened through grout injection. Although constructed at a later phase, the first-floor interior walls are modeled as clay brick masonry, to support the overlying timber joists.

The second model (Model 2 or M2) represents the current state of the building (fifth construction phase). In this case, a 20% increase in masonry strength is applied [29] to account for improvements due to repointing works.

In both models, masonry is treated as a homogeneous, isotropic material that combines the properties of its constituent materials (stone and mortar). The rationale for adopting an isotropic approach probably arises from practical modeling factors and the constraints of existing material property data for historical masonry. Assessing the complete anisotropic material properties of century-old masonry would necessitate comprehensive and possibly destructive testing, which may not be practical for a functioning museum structure. Furthermore, the intricate geometry and diverse construction methods employed throughout the building could complicate the establishment of uniform directional properties. However, the mechanical contribution of the embedded timber frames is not explicitly modeled; only their self-weight is taken into account in the structural analysis. The geometric data for both models were derived from the architectural plans provided in [25].

To facilitate the modeling process and reduce computational complexity, a series of simplifications and assumptions were adopted:

• Architectural details were simplified and elements (e.g., openings) were moved by 5–10 cm in order to reduce the number of structural nodes. Decorative elements on the façades were not simulated.
• The roof was not modeled, but its loads were applied at the points of contact between the roof trusses and the masonry.
• Interior and exterior staircases, as well as permanent operational loads associated with the building’s function as a museum, were not included in the simulation.
• The marble balcony on the main façade was not modeled; however, its loads were applied as pairs of forces and moments at its four support points.
• The masonry walls were modeled with finite shell elements. The eccentricity of the masonry at each floor was taken into account—since the ground floor walls are thicker than those of the upper floors—and was modeled using the ‘Area Thickness Overwrites’ command.
• All timber elements (roof and floor joists and ties), steel IPE beams and the RC bond beam at roof level were modeled as linear frame elements, with frame releases applied at their ends to reflect realistic connection behavior.
• The ‘Insertion Point’ command was used on all frame elements to accurately model their eccentric positioning relative to the global geometry.
• The reinforced concrete floor slab was modeled using shell finite elements.
• Timber floorboards were not modeled; however, the weight of the timber floors was applied as distributed loads on the joists.
• The discretization of shell elements was performed using the ‘Divide Areas’ command with a maximum element size of 0.5 × 0.5 m. Frame elements were meshed using Assign Automatic Frame Mesh.

3.2. Mechanichal Properties

The mechanical properties of all structural elements were calculated according to [1,23,26,27,29,30,31,32,33,34,35,36,37] and they are presented in

Table 2. Mechanical properties of masonry.

Section	Model 1				Model 2
	fc (MPa)	E50% (MPa)	ft (MPa)	γ (ΚΝ/m3)	fc (MPa)	E50% (MPa)	ft (MPa)	γ (ΚΝ/m3)
Three—leaf Masonry	2.06	800	0.185	23	2.48	900	0.22	22
	[32]	[22]	[31]		[22]	[22]	[31]
Two—leaf Masonry	4.93	2500	0.44	23	6.42	3210	0.58	22
	[32]	[28]	[31]		[22]	[28]	[31]
Clay Brick Masonry	1.76	879	0.12	18	1.76	879	0.12	18
	[28]	[28]	[31]		[28]	[28]	[31]

,

Table 3. Mechanical properties of concrete materials.

Concrete	fck (MPa)	γ (ΚΝ/m3)	E (Gpa)
C25/25 [30]	20	25	30
B160 [26]	9	24	27

,

Table 4. Mechanical properties of steel materials.

Steel	fyk (MPa)	fu (MPa)	E (GPa)
Sthl I [30]	240	340	200
B500C [25]	500	650	200
S235 [29]	215	360	210

and

Table 5. Mechanical properties of wood.

Wood	γ (ΚΝ/m3)	E1//(MPa)	E2 ┴ (MPa)	E3 ┴ (GPa)
D30 [30]	6.4	8000	640	640

.

In the above tables, f_c stands for compressive strength, f_t for tensile strength, f_yk and f_u (steel) for characteristic yield strength and ultimate tensile strength respectively, E denotes the modulus of elasticity, and γ the unit weight. The Poisson’s ratio is assumed to be 0.2 for all materials, except for wood and steel S235, for which a value of 0.3 is adopted.

3.3. Loads

Dead loads are assumed to be 0.5 kN/m² for wooden floorboards and 1.2 kN/m² for ceramic tiles [38]. In addition, the self-weight of structural elements has also been taken into account. Live loads are assumed to be 5.0 kN/m² for the floors and 0.5 kN/m² for the roof [38].

3.4. Soil Properties and Soil–Structure Interaction (SSI)

According to the geotechnical report carried out for the AML, the wider coastal zone of Romeikos Gialos is covered by Holocene Quaternary (al) deposits, coastal deposits and dunes. The deposits consist of alluvial clay, argillic materials, sands, and weathered products of older sedimentary and volcanic rocks [39]. Τhe subsoil of the building’s foundations is classified as category D: Deposits of loose-to-medium cohesionless soil (with or without some soft cohesive layers), or of predominantly soft-to-firm cohesive soil [23].

In general, soil–structure interaction (SSI) can significantly affect structures built on soft or saturated soils, as it can lead to an increase in their fundamental period and consequently alter their dynamic behavior. According to [40], taking SSI into account is particularly important for structures founded on soils with high groundwater levels and high plasticity (PI > 40), typically corresponding to soil categories C or D, for which reduction factors to the shear wave velocity (Vs) and the shear modulus (G) of the soil are recommended.

For the purposes of the present study, the soil–structure interaction effect was investigated by using formulas [41], which model the foundation-soil system as a discrete system consisting of frequency independent springs, dashpots and masses. More specifically, for each model examined in this study, an additional case was analyzed. In these cases, the fixed supports (as defined in the initial models) (M1fix, M2fix) were replaced by link elements with specific mechanical properties (M1ssi, M2ssi).

To apply the aforementioned formulas, the length of the foundation strips was calculated and then divided by the number of masonry joints at the foundation level. The width of the foundation strips is 1.05 m, the calculated length is 0.27 m, so α = 0.27 m. The shear wave velocity was taken as 150 m/s, the soil density as ρ = 1900 kg/m³ [39], and the Poisson’s ratio as ν = 0.3. Based on these values, the shear modulus G was calculated as 42,750.00 KPa. Applying a reduction coefficient of 0.37 [40], the reduced shear modulus becomes 15,817.50 KPa, corresponding to a reduced shear wave velocity of 91.24 m/s. The values of static stiffness (K) and damping (C) of each link element are presented in

Table 6. Stiffness and damping coefficients for FSI models.

	Static Stiffness K	Damping C
Vertical	28,273.80 KN/m	66 KNs/m
Horizontal	22,788.90 KN/m	10.84 KNs/m
Rocking	1705.45 KNm/rad	2.99 KNms/rad
Torsion	2480.15 KNm/rad	0.92 KNms/rad

. The soil was assumed to be massless.

To ascertain the fundamental frequency of the site and evaluate possible resonance effects with the structure of the Archaeological Museum of Lemnos, calculations were conducted utilizing the established soil parameters. With a shear wave velocity of Vs = 150 m/s and presuming a typical soil layer depth of H = 30 m for Category D soils, the fundamental frequency of the soil deposit was computed using the quarter-wavelength approximation: f₀ = Vs/(4H) = 150/(4 × 30) = 1.25 Hz. The shear modulus G = 42,750.00 KPa was obtained from the equation G = ρVs², where ρ = 1900 kg/m³ and Vs = 150 m/s, confirming the relatively soft characteristics of the soil typical of Category D classification. The computed site frequency of 1.25 Hz is within the typical range for historic masonry buildings (0.8–2.0 Hz), indicating potential resonance conditions that could significantly enhance the seismic response. This frequency analysis lays the physical groundwork for understanding why the SSI effects led to increased peak tensile stress and interstorey drift ratios, as the structure-soil system likely underwent dynamic amplification when the building’s fundamental frequency neared the site frequency, especially following various strengthening interventions that may have modified the original structural frequency.

4. Type of Analysis Performed

4.1. Static Analysis

Based on the structural analysis using the load combination 1.35G + 1.5Q, the total weight of M1 was calculated as 14,778.3 kN, and that of of M2 as 16,056.92 kN. Under this load combination, no exceedance of the maximum compressive strength was observed. However, with regard to tensile strength, exceedance points were identified in Model 2, specifically at the connection points between the second-floor steel beams and the masonry wall.

4.2. Modal Analysis

In the modal analysis, the periods and participating mass ratios were calculated for each model. According to seismic codes [23,29], the cumulative effective modal mass of the modes considered in the spectral analysis must amount to at least 75% of the total mass of the structure, as specified in [29], and 90% as required by [23]. For the purposes of the present study, 200 modes were considered for each model (

Table 7. M1fix—modal participating mass ratios.

StepNum	Period	SumUX	SumUY	SumUZ
Unitless	Sec	Unitless	Unitless	Unitless
1	0.18602	0.00%	5.20%	0.00%
2	0.13667	12.09%	5.20%	0.00%
41	0.05178	87.83%	87.10%	21.49%
133	0.03032	90.03%	88.84%	84.80%
190	0.02547	90.49%	90.00%	86.03%
200	0.02468	90.56%	90.13%	86.17%

,

Table 8. M1ssi—modal participating mass ratios.

StepNum	Period	SumUX	SumUY	SumUZ
Unitless	Sec	Unitless	Unitless	Unitless
1	0.189562	0.02%	16.77%	0.00%
2	0.168865	88.18%	16.77%	0.00%
3	0.143698	88.18%	94.25%	0.01%
9	0.117733	91.98%	94.31%	0.03%
15	0.091661	93.43%	94.38%	98.23%

,

Table 9. M2fix—modal participating mass ratios.

StepNum	Period	SumUX	SumUY	SumUZ
Unitless	Sec	Unitless	Unitless	Unitless
1	0.287152	0.00%	0.00%	0.42%
3	0.156688	5.01%	3.73%	0.43%
5	0.150461	73.52%	3.74%	0.99%
8	0.122583	75.61%	57.18%	1.00%
59	0.05362	90.05%	89.25%	77.62%
81	0.045667	90.48%	90.03%	83.13%
200	0.029105	92.11%	91.54%	87.77%

and

Table 10. M2ssi—modal participating mass ratios.

StepNum	Period	SumUX	SumUY	SumUZ
Unitless	Sec	Unitless	Unitless	Unitless
1	0.287175	0.00%	0.00%	0.44%
2	0.272051	0.02%	5.90%	0.45%
3	0.194507	91.14%	5.90%	0.45%
4	0.166549	91.16%	93.56%	0.46%
17	0.102015	93.04%	94.23%	84.39%
200	0.031548	97.52%	97.82%	87.40%

):

The following results are extracted from Table 7, Table 8, Table 9 and Table 10:

• Compared to M1fix and M1ssi, the fundamental period increases by 54.37% for M2fix and 51.49% for M2ssi, respectively.
• The fundamental period of M1ssi is 1.90% higher than that of M1fix, while M2ssi exhibits only a 0.01% increase relative to M2fix, indicating no significant change in the fundamental natural period in either case.
• The ratios T₁M1ssi/T₁M1fix = 1.02 and T₁M2ssi/T₁M2fix = 1.00 are both ≤ 1.08; therefore, the structures can be analyzed as if founded on rigid soil [42]. Nevertheless, the effect of SSI was still examined in order to evaluate its influence on the maximum stress values.

4.3. Time History Analysis

For the linear time history analysis of the structure, three strong motion time histories were selected based on the guidelines in [23], and for each time history, three accelerograms were used: two for the horizontal directions and one for the vertical direction.

The seismic records selected were those from the Northridge earthquake (Pacific Palisades—Sunset station, Los Angeles, CA, USA, 1994, Mw = 6.7) [14], the Kahramanmaraş earthquake (Turkey, 2023, Mw = 7.7) [15], and the Kocaeli earthquake (Düzce station, Turkey, 1999, Mw = 7.51) [14]. The process of selecting earthquake records for the seismic evaluation of the Archaeological Museum of Lemnos illustrates a methodical strategy that integrates historical seismic information with ground motions pertinent to the region. However, a more robust site-specific justification would enhance the analysis. The incorporation of the Northridge earthquake (1994, Mw = 6.7) offers a significant historical context as one of the most thoroughly documented seismic occurrences, providing high-quality accelerometric data that acts as a standard for comparison and guarantees the inclusion of well-analyzed ground motion characteristics. Moreover, the choice of two Turkish earthquakes—Kahramanmaraş (2023, Mw = 7.7) and Kocaeli (1999, Mw = 7.51)—demonstrates the geographic closeness and seismotectonic resemblance to the Lemnos site. Both the Turkish seismic zones and the North Aegean region are affected by the intricate tectonic interactions between the Eurasian and African plates, including the westward extension of the North Anatolian Fault system into the Aegean Sea. The Kahramanmaraş earthquake supplies contemporary ground motion data utilizing modern recording techniques, while the Kocaeli earthquake provides insights from a significant strike-slip event that shares similar fault mechanisms with those anticipated in the North Aegean region. Nevertheless, the authors ought to offer clearer justification concerning tectonic similarity, the appropriateness of source-to-site distance, the compatibility of original recording site conditions with the Category D soils at Lemnos, and whether the spectral characteristics remain representative after scaling to align with Eurocode 8 requirements, which would bolster the credibility of employing these records for evaluating this vital cultural heritage structure. The ground motion records were modified to match the target response spectrum specified in [23], within the period range of 0.2T to 2T for each direction, following the methodology outlined therein. Spectral matching as shown in Figure 9 was performed using the Seismomatch software (v. 2025) [43].

According to the spectrum defined in [23], the building is classified as Importance Class III, with an importance factor of γΙ = 1.2. A behavior factor of q = 1.5 was adopted, appropriate for masonry structures. The seismic zone is classified as Zone Z2 with a design ground acceleration of Ag = 0.24 g, and a vertical design acceleration of Av = 0.9 × Ag. The site soil is categorized as Type D (S = 1.35), and a constant damping ratio of ζ = 5% was assumed in the analyses.

For each earthquake record, two separate load cases were created by reversing U1 and U2 (Load Case Data—Linear Direct Integration History) for each horizontal accelerogram. The following load cases were created:

• Kahramanmaraş1 (Kah1) and Kahramanmaraş2 (Kah2),
• Northridge1 (Nor1) and Northridge2 (Nor2) and
• Kocaeli1 (Koc1) and Kocaeli2 (Koc2)

4.4. Analysis Challenges

According to [44], the combination of actions for seismic design situations applied to the building under study is G + 0.3Q + E. However, it was not possible to extract this load combination from the analysis software, nor was it possible to obtain envelope stresses, i.e., the maximum (SMAX) and minimum (SMIN) stress values for each elevation of the building.

For this reason, time series were extracted for selected shell elements on each elevation, in order to determine the approximate time points at which maximum stress values occur on each elevation. Additionally, time series were obtained for two shell elements in each model—one on the ground floor and one on the second floor—on the elevation that exhibits the highest SMAX exceedance. This enables direct comparison of the stresses developed across the models. To assess the interstorey drift ratio (IDR), four joints were selected at each floor level along the same vertical axis on the same elevation. It should be noted that the load combination G + 0.3Q does not result in any stress exceedances.

5. Results

5.1. Maximum SMAX Stress Values per Elevation

The distribution of maximum tensile stresses (SMAX) was evaluated across all four elevations for each model. The results offer an overview of the seismic stress demands imposed on the building and facilitate a comparative assessment between fixed-base and SSI conditions, as well as between the two different strengthening methods (

Table 11. SMAX for shells 7342 (M1) and 6724 (M2) on ground floor.

M1	SMAX			M2	SMAX			% Change
Shell 7342	At (s)	Value (kPa)	% Change	Shell 6724	At (s)	Value (kPa)	% Change	M1 vs. M2
1fkah1	11.22	323.40	14.97%	2fkah1	11.52	394.20	−19.13%	21.89%
1skah1	11.23	371.80		2skah1	12.14	318.80		−14.25%
1fkah2	12.42	272.00	47.32%	2fkah2	12.74	474.00	−12.81%	74.26%
1skah2	11.23	400.70		2skah2	11.55	413.30		3.14%
1fnor1	11.33	231.70	79.11%	2fnor1	11.67	467.40	55.03%	101.73%
1snor1	9.38	415.00		2snor1	10.22	724.60		74.60%
1fnor2	5.86	283.70	57.14%	1fnor2	13.62	447.90	25.92%	57.88%
1snor2	7.86	445.80		2snor2	7.89	564.00		26.51%
1fkoc1	8.28	262.80	92.85%	2fkoc1	11.11	465.70	41.51%	77.21%
1skoc1	9.96	506.80		2skoc1	11.13	659.00		30.03%
1fkoc2	8.98	264.50	52.82%	2fkoc2	9.95	480.30	33.06%	81.59%
1skoc2	9.38	404.20		2skoc2	10	639.10		58.11%

and

Table 12. SMAX for shells 7587 (M1) and 6724 (M2) on ground floor.

M1	SMAX			M2	SMAX			% Change
Shell 7587	At (s)	Value (kPa)	% Change	Shell6939	At (s)	Value (kPa)	% Change	M1 vs. M2
1fkah1	11.93	1024.00	123.24%	2fkah1	19.22	1014.00	17.46%	−0.98%
1skah1	11.94	2286.00		2skah1	19.21	1191.00		−47.90%
1fkah2	11.04	777.40	79.70%	2fkah2	11.09	1019.00	1.08%	31.08%
1skah2	11.02	1397.00		2skah2	11.09	1030.00		−26.27%
1fnor1	8.96	1078.00	78.01%	2fnor1	11.22	1088.00	30.97%	0.93%
1snor1	9.34	1919.00		2snor1	10.67	1425.00		−25.74%
1fnor2	8.96	1091.00	80.66%	1fnor2	11.22	1137.00	31.84%	4.22%
1snor2	9.34	1971.00		2snor2	10.67	1499.00		−23.95%
1fkoc1	9.31	976.00	77.36%	2fkoc1	9.13	1112.00	32.19%	13.93%
1skoc1	9.31	1731.00		2skoc1	10.18	1470.00		−15.08%
1fkoc2	9.21	1133.00	88.08%	2fkoc2	9.34	941.50	49.65%	−16.90%
1skoc2	11.26	2131.00		2skoc2	11.3	1409.00		−33.88%

). Figure 10 and Figure 11 illustrate the SMAX stress distribution for the Kahramanmaraş1 event at selected elevations.

5.2. Interstorey Drift Tables and Residual Interstorey Drift Tables

On the same elevation, four joints were monitored for displacements along the axis where maximum deformations occur (Figure 12).

Table 13. Peak IDR_U2 (out-of-plane deformation) for the top selected joint on elevation 4.

Model - Event	Max IDR ‰	Time of Max IDR (s)	Model - Event	Max IDR ‰	Time of Max IDR (s)
M1fix - Kah1	−3.137	11.82	M2fix - Kah1	−7.639	19.08
M1fix - Kah2	2.355	11.04	M2fix - Kah2	−7.456	11.27
M1ssi - Kah1	6.806	11.94	M2ssi - Kah1	−9.530	19.08
M1ssi - Kah2	4.191	11.02	M2ssi - Kah2	7.385	11.09
M1fix - Nor1	3.198	8.96	M2fix - Nor1	−7.778	10.54
M1fix - Nor2	−3.274	10.20	M2fix - Nor2	−8.229	10.54
M1ssi - Nor1	−6.313	10.21	M2ssi - Nor1	−10.124	12.47
M1ssi - Nor2	−6.512	10.21	M2ssi - Nor2	−10.409	12.47
M1fix - Koc1	−3.306	9.40	M2fix - Koc1	−7.477	10.05
M1fix - Koc2	3.383	9.21	M2fix - Koc2	6.248	11.29
M1ssi - Koc1	−6.232	10.00	M2ssi - Koc1	9.668	10.18
M1ssi - Koc2	6.354	11.26	M2ssi - Koc2	−9.633	11.41

presents the height distributions for peak interstorey drift ratios (IDR) calculated with reference to the top joint. For this assessment, the deformation limits adopted in [29] were used: 4‰ for in-plane and 8‰ for out-of-plane deformations.

It should be noted that U1 and U2 refer to the X and Y axes, respectively, while elevation 4 is oriented towards the XZ plane. Therefore, U1 refers to in-plane deformations and U2 refers to out-of-plane deformations.

6. Discussion

The analyses conducted under both static and dynamic loading conditions reveal critical differences in the structural behavior of Models 1 and 2, and highlight the influence of soil–structure interaction (SSI) on their seismic performance.

With regard to the static analysis, soil–structure interaction results in varying tensile stress distributions across the building elevations, thereby preventing the derivation of reliable conclusions. However, a comparison of compressive stresses within the building’s external masonry indicates that models incorporating SSI effects exhibit lower stress levels.

The two models, M1 and M2, were subjected to particularly high seismic loads. Elastic dynamic time history analysis of the selected events revealed significant damage across all four elevations of both models. This damage includes the ramming of the external masonry by floor joists and transverse masonry, and pertains to the influence of roof loads on the external walls, as well as to both in-plane and out-of-plane failure mechanisms. These mechanisms indicate potential wall collapse, as evidenced in the SMAX tables for Elevation 4.

However, the observed failure pattern of the second floor does not accurately reflect its actual structural behavior, as the wooden elements were not simulated. Consequently, the influence of the embedded timber frames (friggia) and the timber-framed internal walls was not taken into account, rendering the second-floor masonry more vulnerable to seismic actions.

Regarding SSI influence, the conclusions according to Table 11 and Table 12 are summarized below:

• Ground floor:

• In Model M1, the maximum SMAX value is observed in case M1ssiKoc1, exhibiting a 93% increase compared to M1fKoc1. Stress increases are observed across all M1ssi models, ranging from 15% to 93%.
• In Model M2, the maximum SMAX value is observed in case M2ssiNor1, exhibiting a 55% increase compared to M2fNor1. The Kah1 and Kah2 events exhibit stress reductions under SSI conditions. Overall, the observed increases range from 26% to 55%, while the reductions range from 13% to 19%.
• A comparison between Models M1 and M2 reveals an increase in stress across all M2 cases, with the exception of M2ssiKah1, which presents a 14.25% decrease relative to M1ssiKah1. Additionally, all M2 cases exhibit a delay in the timing of peak stress occurrence compared to M1.

• 2nd floor:

• In Model M1, the maximum SMAX value is observed in case M1ssiKah1, exhibiting an increase of 123.24% compared to M1fKah1. All cases show increases ranging from 78% to 125%, indicating that stress increases on the first floor are significantly higher than those on the ground floor.
• In Model M2, the maximum SMAX value is observed in case M2ssiNor2, exhibiting an increase of 32% compared to M2fNor2. However, the greatest increase occurs in the Koc2 event, reaching 50%. Unlike the ground floor, no event exhibits stress reduction under SSI conditions. Overall, stress increases in this model range from 1% to 50%.
• A comparison between Models M1 and M2 shows that all M2fix models exhibit increased stress, except for cases M2fkah1 and M2fKoc2. In contrast, M2ssi models show a reduction in stress compared to M1ssi. Finally, similar to the ground floor, all M2 cases—except for M2fKoc1—exhibit a delay in the timing of peak stress occurrence compared to M1.

The findings indicate that the influence of soil–structure interaction (SSI) on stress amplification is more significant in Model M1, which exhibits a shorter fundamental period due to increased stiffness resulting from masonry consolidation via grout injection. Model M2, by comparison, which features greater mass, reduced masonry stiffness, and a longer fundamental period than M1, also experiences stress increases under SSI (with the exception of the Kahramanmaraş earthquake), albeit to a lesser extent.

It is also important to note that, on the second floor, soil–structure interaction results in stress reduction across all combinations for Model M2. Interestingly, the second-floor masonry in Model M1 is reinforced with timber laces at both the floor and roof levels, yet exhibits greater stiffness and strength compared to M2. In contrast, in Model M2 there is a reinforced concrete beam at roof level. Nevertheless, stress exceedances are observed in both models, while the structural system of M2 appears to be more vulnerable under the seismic scenarios examined in this study.

Concerning interstorey drift ratios (IDRs), SSI increases the relative floor displacements—typically by nearly double—across most seismic combinations and structural models. The sole exception is case M2Kah2, in which the maximum IDR value remains unchanged.

Regarding in-plane deformation, no IDR value exceeded the 4‰ threshold, with drift ratios remaining low (around 0.2‰). However, out-of-plane deformation exhibited significantly higher values on the second floor, particularly in cases M2ssiKah1, M2ssiKoc1, M2ssiKoc2, M2fixNor2, and M2ssiNor1 and Nor2, where the 8‰ threshold was surpassed. The IDR differences for each seismic scenario are listed below in detail:

• Model 1

• In the Kahramanmaraş earthquake, the maximum IDR on the second floor is recorded in case M1ssiKah1, reaching 6.8‰—the highest IDR value among all M1 seismic cases, in agreement with the values reported in Table 11. The increases observed in the SSI models are 117% for the Kah1 event and 78% for Kah2.
• In the Northridge earthquake, the maximum IDR on the second floor is recorded in case M1ssiNor2, reaching 6.5‰. The increases observed in the SSI models are 97% for the Nor1 event and 99% for Nor2.
• In the Kocaeli earthquake, the maximum IDR on the second floor is recorded in case M1ssiKoc2, reaching 6.35‰. The increases observed in the SSI models are 89% for the Koc1 event and 88% for Koc2.

• Model 2

• In the Kahramanmaraş earthquake, the maximum IDR on the second floor is recorded in case M2ssiKah1, reaching 9.53‰. With regard to the SSI models, the Kah1 event exhibits a 25% increase, while SSI has virtually no effect in the Kah2 event, as the IDR increases by only 1%.
• In the Northridge earthquake, the maximum IDR on the second floor is recorded in case M2ssiNor2, reaching 10.41‰—the highest IDR value among all M2 seismic cases, in agreement with the values reported in Table 11. The increases observed in the SSI models are 30% for the Nor1 event and 27% for Nor2.
• In the Kocaeli earthquake, the maximum IDR on the second floor is recorded in case M2ssiKoc1, reaching 9.69‰ (Koc2 exhibits a nearly identical value of 9.63‰). The increases observed in the SSI models are 29% for the Koc1 event and 54% for Koc2.

A comparison between models M1 and M2 regarding the effects of SSI on IDRs leads to the same conclusion previously drawn concerning maximum stresses—namely, that SSI has a greater influence in model M1. Overall, the increase in IDRs observed in the SSI models is in line with [16], where the incorporation of soil–structure interaction (SSI) into the analysis of historical load-bearing masonry structures revealed significant increases in drift ratios.

Finally, comparing models M1 and M2 in terms of IDR values, it is observed that, for the same seismic combination, maximum relative floor displacements exhibit significant increases on the second floor in M2. It is also observed, that the largest percentage increases are recorded in fix–based models. Specifically, the greatest percentage change is recorded in the M1fixKah2 and M2fixKah2 models (217%), with a range of values for all combinations spanning from 85% to 217%. The SSI models show an average percentage change in IDR of approximately 60%, with a range between 40% and 76%. It should be noted that this analysis refers to percentage changes; however, as mentioned above, the highest values are recorded in the Mssi models. The detected concentration of maximum displacements on the second floor in M2 could indicate that the RC interventions have led to an irregular distribution of stiffness, fostering soft-story behavior. This alarming finding suggests that the modern strengthening techniques may have inadvertently created new vulnerabilities. This analysis emphasizes the critical need to consider both the strengthening approaches and the soil-structure interaction effects when reviewing retrofit strategies for historic masonry structures, as the combination of these factors can profoundly impact the evaluation of structural performance and safety.

The study of [17] and the present study underscore the vital significance of accurate soil classification in seismic analysis. The steel MRF study reveals that “assuming different soil types can influence the results of IDR__Med and RDR__Med, which can subsequently affect the performance of the building,” with soil type C increasing maximum IDR__Med by 66.76% in comparison to soil type B for a 5-story MRF under DBE loading. In a similar vein, the present investigation of the Archaeological Museum indicates substantial SSI effects for Category D soils, where the soft soil conditions greatly enhance peak tensile stress and interstorey drift ratios. This parallel observation bolsters the reliability of the approach in explicitly modeling SSI effects, as both studies validate that neglecting proper soil characterization can lead to significant underestimation of seismic demands.

The comprehensive approach of the steel MRF study [17], which utilizes 3456 performance curves derived from nonlinear dynamic analyses (NDAs) and incremental dynamic analyses (IDAs), establishes a benchmark for the validation level of analyses that enhances confidence in the research outcomes. In contrast, the present study’s dependence on linear time-history analyses employing equivalent linear soil modeling (with a 37% reduction in stiffness) could gain from a similar validation process through nonlinear analyses, especially considering the recognized limitations in addressing “gapping, sliding, and uplift” effects that are crucial for Category D soils. The methodology of the steel study indicates that a more thorough validation of the linear spring approach might reinforce the conclusions regarding the relative effectiveness of various strengthening strategies for the Archaeological Museum and this constitutes a limitation.

7. Conclusions

The findings of the present study indicate the critical role of SSI in the seismic behavior of load-bearing masonry buildings. Significant increases in both stresses and displacements were recorded in the models with elastic supports (link elements), clearly demonstrating the influence of SSI on the seismic response. Additionally, most cases exhibited a delay in the timing of peak structural response, indicating a dynamic coupling between the soil and the superstructure. On the contrary, excluding SSI effects from the seismic assessment of historical masonry buildings could result in a significant underestimation of seismic demand and, consequently, of the necessary strengthening interventions.

Finally, with regard to retrofitting interventions, the model that includes reinforced concrete (RC) additions (Model 2) showed higher stress concentrations on the story beneath the RC diaphragm (i.e., the ground floor), as well as significantly higher interstorey drifts on the second floor. Despite the potential benefits of RC-based interventions—such as improved box–type and diaphragm behavior etc.—their effectiveness depends on construction quality, proper detailing, and, most importantly, on adequate mechanical connection with the masonry. These measures should always be accompanied by direct strengthening of the masonry itself.

This study highlights the importance of this distinction by comparing two strengthening strategies: one focused on masonry strengthening (Model 1), and another emphasizing diaphragm strengthening through reinforced concrete (RC) intervention (Model 2). Based on the analysis results, the masonry-centered approach (Model 1) demonstrated better seismic performance, emphasizing its suitability for the seismic protection of historical masonry structures.