Experimental and numerical investigation of vibrations induced by manual demolition work in masonry buildings

Abstract

Manual demolition activities in urban environments can generate significant vibrations that propagate to neighboring buildings, potentially affecting occupant comfort and structural integrity, yet limited research exists on vibrations induced by hand-operated demolition tools. This study investigates the dynamic response of a three-story masonry residential building subjected to vibrations from manual demolition work using different tools (small hammer, large hammer, pneumatic hammer, and circular saw). Experimental measurements were conducted using piezoelectric accelerometers positioned across three building levels, while a three-dimensional finite element model was developed in SAP2000 and calibrated against experimental data through modal analysis. The experimental campaign identified dominant vibration frequencies between 11 and 14 Hz, corresponding to the first six structural modes of the building. An accurate Finite Element Model is calibrated on the experimental results. The validated numerical procedure, calibrated through experimental modal data, enables the assessment of vibration levels induced by manual demolition activities and facilitates parametric studies for different scenarios, providing a practical tool for structural engineers.

Introduction

Building vibrations represent a significant concern in urban environments, particularly affecting occupant comfort and structural integrity^1,2. The human body response to vibrations depends on multiple factors including frequency content, amplitude, exposure duration, and transmission pathways. Understanding these effects is important for ensuring adequate living conditions in densely populated areas where construction and demolition activities are frequent³. Modern urban environments present multiple structural safety challenges that require integrated assessment approaches, considering various loading conditions and environmental factors that can affect building performance and occupant safety^4,5.

The human perception of vibrations varies significantly across different frequency ranges. Research has established that vibrations in the 1–2 Hz range are associated with motion sickness symptoms including nausea, dizziness, and general discomfort^6,7. Low-frequency vibrations between 2 and 20 Hz, commonly generated by construction equipment and demolition activities, can cause more severe physiological effects including respiratory difficulties at 1–4 Hz, balance disorders at 4–8 Hz, and visual acuity reduction at 1–10 Hz⁸. These low-frequency vibrations are particularly concerning as they can alter normal biological and psychophysiological responses, potentially leading to osteoarticular disturbances, cardiovascular problems, and digestive system disorders⁹.

Recent research has revealed significant gaps in current vibration assessment standards. Darby et al.¹⁰ demonstrated through the development of the VSimulators facility that current vibration serviceability assessment criteria for wind-induced vibrations in tall buildings, based largely on human “perception” thresholds, are not directly translatable to human “acceptability” of vibrations, defined as the subjective perception and tolerance of vibration by building occupants and workers as quantified in international standards^9,11. The researchers highlighted that vibrations can be perceptible but still acceptable, not adversely affecting performance, while conversely, long-term exposure to sub-perceptible motion can be potentially unacceptable due to adverse effects on human health and wellbeing including sopite syndrome. This has led to a paradigm shift toward acceptability-based criteria rather than traditional perception-based thresholds^12,13.

Current regulatory frameworks, including UNI 9614:2017⁴ and UNI ISO 2631 standards¹¹, provide guidelines for assessing building vibrations and their effects on human occupants. The Italian standard UNI 9614:2017 establishes specific acceleration limits for residential environments: 7.2 mm/s² during daytime, 3.6 mm/s² during nighttime, and 5.4 mm/s² during holiday daytime periods. However, international standards show substantial variations, with DIN 4150 being generally more conservative than other standards, while British standards prove most permissive^14–16. The ISO 2631-2 standard is currently undergoing revision, reflecting ongoing developments in the field¹⁷. Various international standards and guidelines have been developed to protect structures from construction-induced vibrations, though recommended limits vary widely and are often presented without appropriate explanation or scientific basis, particularly for historic buildings^18,19. Despite the number of guidelines for protection of historic buildings from external vibrations, there is no commonly accepted standard, and vibration limits to protect artwork and fragile objects within historic buildings are generally few addressed in the literature²⁰.

However, several studies focus on vibrations generated by heavy mechanical equipment or controlled demolition techniques involving explosives. Manual demolition work, characterized by hammering and hand-tool operations, represents a fundamentally different excitation source in terms of frequency content, amplitude, and spatial distribution, yet has received limited attention in scientific literature.

The combined effects of vibration and noise emerge as particularly significant, with Nering²¹ showing that simultaneous vibration and noise stimuli cause more annoyance than individual stimulus. These finding challenges current standards that assess vibration and noise separately, suggesting need for integrated assessment approaches. Similar advances in structural response modeling have been achieved in wind engineering, where time-dependent aeroelastic responses and shape-dependent pressure statistics have provided insights transferable to vibration analysis of complex building geometries^17,21.

While extensive research has been conducted on vibrations induced by heavy construction machinery, railway traffic, and road vehicles^22–25, limited attention has been given to vibrations generated by manual demolition activities. Manual demolition tools, including hammers, pneumatic equipment, and cutting devices, generate transient vibrations that can propagate through building structures and affect neighboring buildings. The dynamic characteristics of these vibrations differ significantly from those produced by continuous sources, requiring specific investigation methodologies^26–28. This gap is particularly significant as manual demolition represents a substantial portion of urban renovation work in historic centers where mechanical equipment access may be restricted due to logistical constraints or heritage preservation requirements²⁹. Furthermore, existing vibration guidelines primarily address mechanical demolition scenarios and may not adequately capture the distinct excitation characteristics of manual operations²⁰.

Recent advances in vibration transmission modeling have improved understanding of demolition-induced effects. Colaço et al.³⁰ developed comprehensive numerical models using finite element method-perfectly matched layer (FEM-PML) approaches, demonstrating significant potential for vibration prediction in urban environments where demolition activities affect nearby buildings. Subsequent work by the same authors³¹ validated axisymmetric FEM-PML approaches through experimental assessment, showing that open trench construction near impact sources serves as an efficient mitigation measure. Nevertheless, continuum-based FE approaches may face limitations when dealing with material discontinuities, large block displacements and post-peak softening behavior. Complementary to these methods, distinct element methods (DEM) provide a discrete representation of the masonry fabric and an explicit description of contact and fracture mechanisms, making them particularly suitable for the analysis of complex geometries and failure processes^32,33. In this context, DEM has been successfully employed to assess the impact of construction sequence on the structural response of complex masonry curved shells³⁴ and to investigate fracture initiation and propagation in masonry elements and assemblies^35,36, highlighting the central role of DEM and fracture mechanics in capturing large-displacement, post-failure scenarios that are difficult to reproduce with standard FE models.

Recent studies have emphasized the importance of combined experimental and numerical approaches for vibration assessment. Moschioni et al.³⁷ compared different metrics for evaluating disturbance from building vibrations, highlighting the complexity of human response assessment. The RIVAS project³⁸ and Cargovibes project³⁹ provided comprehensive guidelines for vibration evaluation, emphasizing the need for validated numerical models to predict vibration transmission and assess comfort criteria.

Masonry buildings, particularly those constructed in the early-to-mid 20th century, present unique challenges for vibration assessment due to their complex dynamic behavior and material heterogeneity. The interaction between masonry walls and reinforced concrete floors creates coupled dynamic systems that require sophisticated modeling approaches for accurate prediction of vibration transmission and amplification^40,41. Recent advances in finite element modeling of masonry structures have achieved unprecedented sophistication through automation and hybrid approaches^42,43. Model updating techniques based on experimentally identified modal parameters have proven essential for achieving reliable predictions, with recent applications demonstrating that elastic modulus variations of ± 10–20% can result in frequency shifts of 5–15%^44,45.

Experimental methods for masonry buildings have reached new sophistication levels through integration of traditional and modern monitoring techniques. Ramos et al.⁴⁶ established foundational work demonstrating that vibration-based damage detection in unreinforced masonry structures can be satisfactorily performed using modal parameters including natural frequencies, mode shapes, and damping ratios. Modern approaches include operational modal analysis⁴⁷ and MEMS-based acquisition systems⁴⁸, enabling cost-effective monitoring solutions suitable for heritage preservation applications. Operational modal analysis has been successfully applied to various masonry typologies including historic towers, palaces, and residential buildings for seismic assessment and structural identification^49,50, though its application to characterize vibrations during ongoing demolition activities remains limited in the literature.

The integration of machine learning approaches represents a significant advancement in civil engineering field and in particular for vibration assessment⁵¹. Abouelmaty et al.⁵² developed surrogate modeling frameworks using Artificial Neural Networks and Extreme Gradient Boosting, enabling rapid assessment of vibration risks in construction planning. Similarly, Işık et al.⁵³ developed Random Forest, Support Vector Regression, and K-Nearest Neighbors algorithms for seismic vulnerability assessment, demonstrating integration of ML techniques with traditional vibration analysis.

Despite the extensive body of research on construction-induced vibrations and dynamic characterization of masonry structures, a significant gap exists in the literature regarding the specific case of manual demolition activities in masonry buildings. While numerous studies have addressed vibrations from heavy mechanical equipment and others have investigated the dynamic behavior of masonry structures through operational modal analysis for seismic assessment, the characterization of vibrations induced by manual demolition work remains largely unexplored. This gap is particularly relevant given that manual demolition represents a considerable portion of renovation work in historic urban contexts where mechanical equipment is impractical or prohibited. Furthermore, existing vibration transmission models and protection guidelines, primarily developed for mechanized activities, may not adequately address the distinct frequency content and spatial distribution characteristics of manual hammering operations. The present study addresses this gap by combining experimental vibration monitoring with operational modal analysis and finite element modeling to characterize vibrations induced by manual demolition in traditional masonry buildings.

This study aims to investigate the dynamic response of a masonry residential building subjected to vibrations from manual demolition activities through a combined experimental and numerical approach. The primary objective is to characterize and quantify vibrations transmitted from manual demolition tools to building structures through comprehensive accelerometer measurements. This experimental investigation represents the core contribution of the research, providing essential data on vibration transmission mechanisms and their effects on occupant comfort in masonry buildings subjected to demolition activities in adjacent structures.

To support and extend the experimental findings, the study develops and validates a three-dimensional finite element model capable of reproducing the observed dynamic behavior. This numerical model serves to interpret experimental results and provides the foundation for future predictive studies, though such applications remain beyond the scope of the current work. The impact of measured demolition-induced vibrations on occupant comfort is assessed based on established national and international standards, enabling practical interpretation of the results for engineering applications and regulatory compliance.

The integration of experimental measurements with numerical validation delivers a comprehensive understanding of vibration transmission in masonry buildings during manual demolition activities, establishing a methodological framework that can be extended to predictive applications in future research. A flowchart of the applied methodology used in this study is reported in Fig. 1.

Fig. 1.

Methodology flowchart illustrating the experimental and numerical approach.

The novelty of this research lies in providing the first systematic experimental characterization of manual demolition-induced vibrations in masonry buildings, addressing the research gap identified in the literature review. The study establishes quantitative acceleration thresholds and practical guidelines for safe demolition practices, directly bridging the gap between structural engineering knowledge and construction site practices. These contributions advance both the scientific understanding of vibration propagation in masonry structures and the practical tools available for managing demolition activities in sensitive historical contexts.

The article is structured as follows:  “Case study description and experimental methodology” describes the case study building and experimental setup;  “Finite element model” presents the numerical modeling methodology and calibration procedure; “ Model calibration and validation” discusses the experimental results and model validation and  “Conclusions” provides conclusions and recommendations for future research.

Case study description and experimental methodology

The investigated structure is a three-story masonry building located in the San Cataldo district of Bari, Italy (Via Caprera n.2). The position of the building and the territorial context is shown in Fig. 2. The building, dating from the early 20th century, presents typical characteristics of southern Italian residential construction of that period and shows signs of weathering and material deterioration consistent with its age and exposure to Mediterranean climatic conditions, characterized by hot, humid summers (average temperatures 25–35 °C), mild, dry winters (average temperatures 8–15 °C), and moderate annual precipitation (500–700 mm), with occasional intense rainfall events and salt-laden winds from the nearby Adriatic Sea. These environmental conditions, typical of the Apulia region in Southern Italy, contribute to material degradation through wetting-drying cycles, salt crystallization, and thermal expansion-contraction effects on masonry materials.

Fig. 2.

(a) Territorial framework showing building location in Bari, Italy. Map data: Google Maps, Google LLC, Mountain View, CA, USA. Available at: https://www.google.com/maps (accessed September 2025). (b) Building identification and surrounding area. Alt text: Aerial view and satellite imagery showing the location of a three-story masonry building in the San Cataldo district of Bari, Italy. Part (a) shows the territorial context with street layout, while part (b) provides a closer view of the building and its immediate residential surroundings.

The structural system consists of load-bearing masonry walls constructed with irregular stone elements (tuff and calcarenite), typical of the local construction tradition. Based on visual inspection and comparison with similar structures of the same period, the masonry composition corresponds to what the Italian building code NTC 2018 classifies as “irregular soft stone masonry.” The floor system is assumed to be reinforced concrete with clay block infills (latero-cemento), a common solution adopted in Italy during the mid-20th century renovations of older buildings. Details of the structure and its walls and floors are shown in Fig. 3.

Fig. 3.

(a) Building current state, (b) masonry walls section, (c) floor section. Alt text: Three-part technical illustration showing: (a) photograph of a weathered three-story masonry building with visible external walls deterioration, (b) cross-sectional diagram of irregular soft stone masonry wall construction with tuff and calcarenite blocks, and (c) technical drawing of reinforced concrete floor system with clay block infills typical of mid-20th century Italian construction.

The building configuration includes three levels: ground floor (located 80 cm below street level), first floor, and second floor. Each level presents a similar layout with two rooms symmetrically arranged around a central stairwell that serves as the main entrance hall. The ground floor features two rooms, each with a window facing Via Caprera, separated by the central stair core. The first floor maintains an identical layout, while the second floor consists of two terraces separated by the stairwell tower structure. The ground floor plan and the main section of the building are shown in Figs. 4 and 5.

Fig. 4.

Ground floor plan. Alt text: Architectural floor plan drawing showing the ground floor layout of the masonry building with two symmetrical rooms separated by a central stairwell, entrance hall, and windows facing Via Caprera. Dimensions and structural elements are indicated with technical drafting symbols.

Fig. 5.

Main façade. Alt text: Technical elevation drawing of the building’s main façade showing the three-story structure with regular window openings, architectural details, and dimensional annotations indicating the building height and floor levels.

Key structural dimensions include approximate floor plan dimensions of 8.0 m × 6.0 m, with wall thicknesses varying between 40 and 50 cm depending on their structural function. The total building height is approximately 10.5 m including the stairwell tower. The foundation system, partially below street level, provides adequate stability for the superstructure while creating specific boundary conditions that influence the dynamic response of the building.

Experimental setup and instrumentation

The experimental campaign was designed following the guidelines established in UNI 9614:2017 for vibration measurements in buildings⁶. The investigation focused on characterizing vibrations transmitted from simulated manual demolition activities to the target building structure through systematic accelerometer measurements^54,55.

Vibration sources

Manual demolition activities were simulated using four different tools commonly employed in building renovation and demolition works: small hammer (2.5 kg), large hammer (5.0 kg), pneumatic hammer (electric-powered), and circular saw (motor-driven). These tools were selected to represent the range of manual demolition equipment typically used in urban renovation projects, each generating distinct vibration characteristics in terms of frequency content, amplitude, and duration. All the instruments used to generate vibrations are shown in Fig. 6.

Fig. 6.

Manual demolition tools used in the experimental campaign: (a) Small hammer, (b) large hammer, (c) pneumatic hammer, (d) circular saw. Alt text: Four photographs showing demolition tools used in the vibration study: (a) small 2.5 kg manual hammer, (b) large 5.0 kg manual hammer, (c) electric-powered pneumatic hammer, and (d) motor-driven circular saw. Tools are shown against work surfaces and represent typical equipment used in building renovation projects.

The demolition activities were conducted within the target building itself to simulate realistic scenarios where renovation work occurs in one part of a building while other areas remain occupied. Four different demolition scenarios were implemented: partial partition wall demolition using small and large hammers, simulated floor slab demolition using the pneumatic hammer, and masonry cutting using the circular saw. Each activity was performed for predetermined durations to ensure consistent measurement conditions across all tests. The position of the demolition activities in the adjacent building are shown in Fig. 7.

Fig. 7.

Demolition work scenarios and vibration source positions. Alt text: Schematic diagram showing the three-dimensional layout of the building with marked positions where different demolition activities were conducted. The diagram indicates four demolition scenarios: partition wall demolition with hammers, floor slab demolition with pneumatic hammer, and masonry cutting with circular saw. Small picture-in-picture shows the demolition activities.

Accelerometer configuration

Vibration measurements were conducted using piezoelectric accelerometers (IEPE type) strategically positioned across the three building levels. The sensors included uniaxial, biaxial, and triaxial configurations to capture the complete three-dimensional response of the structure. Accelerometer placement followed UNI 9614:2017 guidelines, focusing on floor-mounted sensors in habitable areas while excluding non-occupied spaces such as storage areas, bathrooms, and corridors⁶.

The monitoring campaign employed 26 accelerometers distributed across the three floors of the building, with different sensor configurations optimized for the expected vibration patterns. Specifically, uniaxial accelerometers measuring vertical acceleration (Z direction) were positioned at points #1 and #3 (ground floor); biaxial configuration was used to measure horizontal accelerations (X and Y directions) at points #2, #4, #6, #8, #10, and #12; and triaxial configuration was adopted to capture all three spatial directions (X, Y, and Z) at points #5, #7, #9, and #11.

The sensor layout was designed to monitor two adjacent rooms on each floor simultaneously. Each monitored room has a floor area of approximately 24 m². Within each room, measurement points were positioned at inter-distances of approximately 2 m to capture the spatial distribution of vibrations across the floor span. The two rooms monitored on each floor are separated by approximately 7 m, measured center-to-center.

Manual demolition activities, consisting of hammering on masonry vaults and walls, were conducted in a building directly adjacent to the monitored structure. The two buildings share structural connections through the masonry walls, allowing vibration transmission from the demolition site to the instrumented building. This configuration represents a typical scenario in urban renovation projects, where demolition work in one building can induce vibrations in adjacent structures. The distance from the demolition work area to the nearest measurement points was approximately 2 m.

Figure 8 shows the detailed sensor layout in plan view for each floor level, indicating the position of measurement points, sensor types, and the location of demolition activities in the adjacent building. The three sensors configurations are shown in Fig. 9.

Fig. 8.

Accelerometer positioning plan showing measurement points across building levels. Alt text: Technical floor plan with overlaid measurement grid showing the strategic placement of piezoelectric accelerometers across three building levels. Nine measurement points are marked with symbols indicating triaxial, biaxial, and uniaxial sensor configurations positioned to capture critical structural response locations.

Fig. 9.

Accelerometer configurations. (a) uniaxial; (b) biaxial; (c) triaxial. Alt text: Three photographs showing different types of piezoelectric accelerometers configurations used in the experimental setup: (a) single-axis sensor for measuring vibrations in one direction, (b) dual-axis sensor for two-dimensional measurements, and (c) three-axis sensor for complete spatial vibration characterization.

The measurement positions were selected to capture critical structural response points, including corner locations and central spans where maximum vibration amplification typically occurs⁵⁶. A total of nine measurement points were established: three points per floor level, with each point equipped with accelerometers capable of measuring vibrations in the vertical (z-axis) and two horizontal directions (x and y axes). The coordinate system was aligned with the building’s principal structural axes, with the x-axis corresponding to the building’s length, y-axis to its width, and z-axis to the vertical direction.

Data acquisition system

The data acquisition system consisted of a modular conditioning unit with integrated DAQ capabilities, powered by battery to minimize electrical noise interference. The system featured individual conditioning modules for each accelerometer channel, converting the high-impedance charge output from the piezoelectric sensors to low-impedance voltage signals suitable for digital acquisition.

Six separate measurement sessions were conducted, each lasting 30 min to capture both the active demolition periods and background vibration levels. Data were sampled at 1000 Hz and stored in LabVIEW Measurement format for subsequent processing. Each measurement session was synchronized with the demolition activities to enable correlation between specific tools and measured vibration responses.

The conditioning system provided anti-aliasing filtering and appropriate gain settings to optimize the dynamic range for the expected vibration levels. Sensor sensitivity calibration was performed before the measurement campaign using a reference calibrator to ensure measurement accuracy and traceability to international standards.

Measurement protocol

The experimental protocol was designed to capture both the transient nature of manual demolition-induced vibrations and the background vibration environment. Each 30-minute measurement session included periods of active demolition work interspersed with quiet periods to establish baseline conditions. This approach enabled separation of demolition-induced vibrations from ambient building vibrations caused by other sources such as traffic, wind, or building services.

Quality control measures included real-time monitoring of signal levels to prevent saturation, verification of sensor mounting integrity, and documentation of all demolition activities with precise time stamps. Environmental conditions including temperature, humidity, and wind speed were recorded to account for potential influences on the measurement results.

The measurement campaign yielded a comprehensive dataset comprising vibration time histories for each combination of demolition tool, building level, and measurement direction. This dataset formed the foundation for subsequent frequency domain analysis, modal parameter identification, and finite element model calibration.

Finite element model

A three-dimensional finite element model was developed using SAP2000 software to reproduce the dynamic behavior of the masonry building and validate the experimental observations. The modeling approach followed established procedures for masonry structures while incorporating specific considerations for the building construction characteristics and boundary conditions identified during the experimental campaign.

The finite element method (FEM) provides a numerical framework for solving the complex system of differential equations governing structural dynamics. For a multi-degree-of-freedom system, the governing equation of motion is expressed as:

1

where M, C, and K represent the mass, damping, and stiffness matrices respectively, u(t) is the displacement vector, and f(t) represents external forces. For modal analysis (free vibration without external loads and neglecting damping), this reduces to the eigenvalue problem:

2

where ω represents natural frequencies and φ the corresponding mode shape vector.

Masonry walls modelling

Masonry walls were modeled using shell elements with subsequent extrusion to solid brick elements to capture the three-dimensional behavior accurately. This approach enables representation of both in-plane and out-of-plane response characteristics essential for dynamic analysis of masonry structures^57,58. The main properties of the materials used are summarized in Table 1.

Table 1.

Mechanical properties of masonry materials based on NTC 2018 guidelines for irregular soft stone masonry.

Masonry type	fc [MPa]	t0 [MPa]	fv0 [MPa]	E [MPa]	G [MPa]
Regular blocks masonry of soft stone (tuff, calcarenite, etc.)	2 ÷ 3.2	0.04 ÷ 0.08	0.10 ÷ 0.19	1200 ÷ 1620	400 ÷ 500

Material properties were assigned based on NTC 2018 specifications for “irregular soft stone masonry,” corresponding to the visual assessment of the building. The elastic modulus and specific weight values were selected from the reference ranges for this masonry typology, providing a conservative approach for the initial model configuration.

The elastic modulus is a critical parameter that significantly affects structural stiffness and natural frequencies. Literature on FEM model updating of masonry structures indicates that natural frequencies typically exhibit sensitivity to elastic modulus variations in the range of 5–15% for modulus changes of ± 10–20%^44,45.

The meshing strategy employed a systematic division of wall surfaces to create a uniform finite element discretization. The mesh density was selected to balance computational efficiency with accuracy requirements, ensuring adequate representation of the dynamic modes of interest while maintaining reasonable analysis times.

Floor system representation

The reinforced concrete floor system was modeled using a combination of frame elements for beams and shell elements for the slab portion. This approach accurately represents the composite behavior of the hollow clay brick floor system typical of the building construction period.

Beam elements were positioned with appropriate spacing to represent the actual joist layout, with material properties corresponding to C25/30 concrete class as typically used in mid-20th century construction. The slab thickness was modeled using shell elements with membrane and bending thickness parameters calibrated to represent the composite action of concrete and clay block infills. Figure 10 report the complete FE model.

Fig. 10.

Complete structural model. Alt text: Three-dimensional finite element model rendering showing the complete structural representation of the masonry building. The model displays shell elements for walls, frame elements for beams, and shell elements for floor systems, rendered in a wireframe or solid view showing the building’s geometric complexity.

Boundary and load conditions

Boundary conditions were applied to represent the actual constraint conditions observed in the building. Fixed supports were assigned to foundation elements and walls in contact with adjacent buildings, while roller supports were used for the interface wall with the adjacent building where demolition activities were simulated. The boundary conditions in the FE environment are shown in Fig. 11.

Fig. 11.

Constrained structural model. Alt text: Three-dimensional finite element model with applied boundary conditions and constraints highlighted. The model shows fixed supports at foundation level, roller supports at interfaces with adjacent buildings, and other constraint conditions used to represent realistic structural behavior in the numerical analysis.

The roller support configuration allows displacement in the x and z directions while constraining movement in the y direction, representing the realistic condition where some relative movement may occur between adjacent buildings while maintaining vertical support.

The floor-to-wall connections were modeled using roller supports at the interface between beams and masonry walls. This boundary condition represents the actual construction detail observed during in-situ inspection: the beams simply rest on masonry wall seats without mechanical anchorage, bolts, or ties. The roller supports allow vertical load transfer while providing minimal rotational restraint and no constraint against horizontal displacement, thus accurately representing the unanchored beam-on-wall connection.

Permanent loads were applied to represent the building’s service conditions, including a distributed load of 1.20 kN/m² on floor beams to account for floor finishes, partitions, and typical live loads. This load value is consistent with typical floor finishing systems in early 20th-century buildings and complies with recommendations in Italian building codes for existing structures (NTC 2018, Circolare 2019)^59,60.

The load was modeled as a linear load distributed along the beam elements, accurately representing the mass distribution that influences the dynamic characteristics of the floor structure.

Self-weight was automatically calculated by the software based on material densities and element geometries, with a self-weight multiplier of 1.0 applied to include the complete dead load effects in the analysis.

Modal analysis

The mass matrix for modal analysis was derived from the applied loads and self-weight, following standard procedures for dynamic analysis of building structures^61,62. The mass source definition included both dead loads and appropriate portions of live loads to represent realistic building mass distribution.

Modal analysis was configured to extract sufficient modes to capture the dynamic characteristics relevant to the frequency range of interest (1–30 Hz). The analysis considered modes with participating mass ratios greater than 5% and ensured that the total participating mass exceeded 85% in each principal direction, following recommendations in NTC 2018.

Quality control criteria were established to validate the modal analysis results:

• Modes with participating mass > 5% were considered significant.

• Total participating mass > 85% was required for adequate representation.

• Translational mass participation > 75% was targeted for primary modes.

The finite element analysis generated natural frequencies, mode shapes, and participating mass ratios for comparison with experimental data. Results were processed to identify modes corresponding to the frequency ranges observed in the experimental measurements. The main results of the modal analysis are shown below, in Fig. 12.

Fig. 12.

Modal analysis main results. Alt text: Screenshot of SAP2000 software interface displaying modal analysis results in tabular format. The table shows natural frequencies, periods, participating mass ratios, and other dynamic characteristics for the first several structural modes of the building, with numerical values and percentages clearly visible.

The numerical model provided the foundation for calibration against experimental data and subsequent validation of the modeling approach for masonry buildings subjected to manual demolition-induced vibrations.

Model calibration and validation

The experimental data collected during the measurement campaign were processed using MATLAB to extract frequency domain characteristics and identify the dominant vibration modes of the building structure. Power Spectral Density (PSD) analysis was performed on the time history data to determine the frequency content of vibrations induced by different demolition tools^63,64. The results of the PSD analysis are shown in Fig. 13.

Fig. 13.

Power Spectral Density plots for different demolition tools and building levels: (a) ground floor response, (b) first floor response, (c) second floor response. Alt text: Three graphs showing Power Spectral Density analysis results with frequency (Hz) on x-axis and PSD magnitude on y-axis. Each graph contains multiple overlaid curves representing different demolition tools (hammers, pneumatic hammer, circular saw) and shows dominant frequency peaks in the 11–14 Hz range across ground floor, first floor, and second floor responses.

The PSD represents the distribution of signal power across frequency components and is particularly useful for identifying structural resonances and dominant response frequencies. For the discrete time signals collected during the experimental campaign, PSD estimation was performed using Welch method to reduce noise effects and improve frequency resolution.

The experimental analysis revealed dominant frequency peaks in the range of 11–14 Hz across all building levels, with variations depending on the specific demolition tool employed and measurement location. These frequencies correspond to the fundamental vibration modes of the building structure and provide the target values for finite element model calibration.

Modal parameter identification

From the experimental data analysis, the dominant structural frequencies were identified by examining the peak responses in the PSD plots. The frequency range of 11–14 Hz consistently appeared across different measurement points and demolition scenarios, indicating that these frequencies correspond to the fundamental dynamic characteristics of the building rather than tool-specific excitation frequencies.

The experimental campaign successfully identified the primary vibration modes that control the building dynamic response to manual demolition activities. These modes represent the coupling between the masonry wall system and reinforced concrete floors, creating a complex three-dimensional dynamic behavior characteristic of mixed structural systems.

Finite element model calibration

The initial finite element model, constructed with material properties derived from NTC 2018 code provisions for irregular soft stone masonry, was subjected to modal analysis to determine its dynamic characteristics. The analysis identified the first six structural modes with significant participating mass ratios in the frequency range relevant to the experimental observations. The main structural modes for the first model are shown in Fig. 14.

Fig. 14.

Initial FEM model results showing natural frequencies, periods, and participating mass ratios for modes with frequencies near experimental observations. Alt text: Screenshot of finite element analysis software displaying a results table with modal analysis data. The table shows mode numbers, natural frequencies in Hz, periods in seconds, and participating mass ratios as percentages for the first six structural modes, with values highlighted for comparison with experimental measurements.

Model calibration was performed by comparing numerical natural frequencies with the experimentally identified dominant frequencies. The calibration process focused on adjusting material properties, particularly the elastic modulus of masonry, to achieve better agreement between numerical predictions and experimental observations.

The calibration error was quantified using the percentage difference between experimental and numerical frequencies:

3

where f_exp represents the experimentally identified frequency and f_num represents the corresponding numerical frequency from the finite element model.

The finite element model was developed to replicate the structural configuration observed experimentally, with particular attention to the floor system that exhibited the dominant dynamic response. Modal analysis of the complete building structure revealed the presence of mixed modes, including torsional components, particularly at higher frequencies and more prominently in the second floor. However, many of these modes exhibited very low participating mass ratios and were associated with the relatively stiff masonry walls, which showed minimal response to the demolition-induced excitation.

The experimental evidence indicated that the degraded floor structure was significantly more sensitive to vibrations than the stocky masonry walls. Consequently, the model calibration focused on the vertical flexural modes of the floor system, which dominated the measured response and are most relevant for assessing vibration levels during manual demolition operations. This modeling approach reflects the actual structural behavior, where the more flexible and degraded floor elements exhibit higher dynamic sensitivity compared to the massive masonry walls.

Calibration results

The calibration process revealed varying levels of agreement between experimental and numerical results for different structural modes. The analysis focused on the first six modes of vibration, which showed frequencies in the range of interest (11–14 Hz) and significant participating mass ratios. A summary of the calibration procedure in terms of frequencies is reported in Table 2.

Table 2.

Calibration error analysis showing experimental frequencies, numerical frequencies, and percentage errors for significant modes.

Floor [n.]	Sensor [n.]	Mode [n.]	fs [Hz]	fm [Hz]	ε% [%]
P2	11	1	13	13.69	-5%
P2	9	2	11	13.7	-20%
P2	11	3	13	14.5	-10%
P1	5–7	4	14	14.66	-5%
P1	5–7	5	14	14.69	-5%
P2	9	6	11	14.77	-26%

The calibration results demonstrated that:

• Mode 1: 5% error (acceptable for engineering applications).

• Mode 4: 8% error (acceptable range).

• Mode 5: 12% error (acceptable range).

• Modes 2, 3, 6: 15–26% error (requiring model refinement).

The calibration process identified several sources of uncertainty that contribute to the differences between experimental and numerical results. The primary source of uncertainty stems from material property estimation, as the masonry material properties were determined based on visual inspection and code provisions rather than direct material testing. The actual elastic modulus and density of the irregular soft stone masonry may vary significantly from the assumed values, particularly given the age of the structure and potential variations in construction quality throughout the building.

Connection modeling represents another significant source of uncertainty, as the interface between masonry walls and reinforced concrete floors was modeled assuming fully rigid connections. Some flexibility likely exists at these interfaces due to construction tolerances, material compatibility, and potential deterioration over time. This simplified assumption may lead to overestimation of structural stiffness and consequently higher predicted natural frequencies.

The boundary condition idealization also contributes to modeling uncertainties, particularly regarding the interface with adjacent buildings. The numerical model simplified these conditions using roller and fixed supports, while the actual constraint conditions may involve complex soil-structure interaction and varying degrees of connection with neighboring structures. Similarly, the geometric approximations inherent in the finite element model represent an idealized version of the building that may not capture all irregularities present in the actual construction, including variations in wall thickness, material discontinuities, and construction imperfections accumulated over the building’s operational life.

The experimental validation confirms that manual demolition activities effectively excite these fundamental structural modes, leading to building-wide vibration transmission that can be felt by occupants throughout the structure. This finding has important implications for construction planning and occupant comfort assessment, as it demonstrates that even manual demolition work can generate significant vibrations that propagate through the entire building. The validated model provides a reliable tool for predicting the magnitude and distribution of vibrations under various demolition scenarios, enabling informed decisions regarding construction scheduling and protective measures for sensitive operations or vulnerable occupants.

The experimental results highlight a significant differential dynamic response between structural elements. The masonry walls, characterized by large cross-sections and high stiffness-to-mass ratios, exhibited minimal response to the demolition-induced excitation. In contrast, the floor system showed pronounced sensitivity to vibrations, with clearly identifiable resonance peaks corresponding to vertical flexural modes. This behavior can be attributed to two main factors: the structural flexibility of the floors compared to the masonry walls and the degraded condition of floor materials, which reduces stiffness and damping capacity.

The validated model could provide a reliable starting point for assessing the building dynamic response to manual demolition activities and forms the basis for comfort evaluation and practical engineering applications.

Conclusions

This study investigated the dynamic response of a masonry residential building to vibrations induced by manual demolition activities through a combined experimental and numerical approach. The research successfully characterized vibration transmission mechanisms and assessed their impact on occupant comfort, providing new insights into an underexplored area of building vibration analysis.

The experimental campaign revealed that manual demolition activities using common hand tools (hammers, pneumatic equipment, circular saws) generate significant building-wide vibrations with dominant frequencies in the 11–14 Hz range. These frequencies correspond to the fundamental structural modes of the masonry building, indicating effective excitation of the primary dynamic characteristics. Measured acceleration levels reached 0.1–0.25 m/s² during active demolition periods, exceeding established human perception thresholds by factors of 2–5 and significantly surpassing regulatory comfort limits defined in UNI 9614:2017.

The vibration transmission exhibited building-wide propagation characteristics, with measurable responses recorded across all three building levels regardless of the specific demolition tool employed. These finding challenges common assumptions that manual demolition work produces only localized effects and demonstrates the importance of considering vibration transmission in renovation planning for occupied buildings.

The three-dimensional finite element model successfully reproduced the dominant frequency characteristics observed experimentally, achieving calibration errors of 5–12% for the primary structural modes. This level of accuracy is acceptable for engineering applications and validates the modeling approach for masonry buildings with reinforced concrete floor systems. The calibration process identified material property estimation as the primary source of modeling uncertainty, highlighting the importance of experimental validation for existing masonry structures where direct material testing may not be feasible.

The measured vibration levels significantly exceed current regulatory limits, with peak accelerations reaching values 10–35 times higher than the 7.2 mm/s² daytime limit specified in UNI 9614:2017. The frequency content (11–14 Hz) falls within the range known to cause physiological disturbances, including interference with desk work activities and potential motion sickness during prolonged exposure.

However, the intermittent nature of manual demolition work (approximately 9 min of high-level exposure per 30-minute period) suggests that cumulative effects may be manageable through appropriate work scheduling and occupant notification protocols.

The study was conducted on a single masonry building with specific construction characteristics, and the quantitative results may not directly transfer to other building types without appropriate model adjustments. Material properties were estimated based on visual inspection and code provisions rather than direct testing, introducing uncertainty in the numerical predictions.

Future research should expand the experimental database to include different masonry building types and investigate vibration mitigation strategies specifically applicable to manual demolition activities. Development of real-time monitoring systems and integration with construction management protocols could provide more sophisticated approaches to vibration control in urban renovation projects.

The research establishes that effective assessment of manual demolition-induced vibrations requires integrated experimental and numerical approaches that account for the complex dynamic behavior of masonry buildings and the intermittent nature of demolition activities.

Author contributions

A.L.S. wrote the main manuscript, prepared all the figures and reviewed it.

Funding

This research was funded by Prof. Dora Foti with the Scientific Research Programs of Relevant National Interest 2022 P.N.R.R. (P2022FSXEP): " Vibrations Induced on Buildings by natural and anthropic sources for the definition of Reduction And mitigation strategies (VIBRA)” Scientific coordinators Prof. D. Foti (Polytechnic University of Bari) and Head of Unit Prof. D. Foti (Polytechnic University of Bari).

Data availability

The data that support the findings of this study are available from the corresponding author, A.L.S. , upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.