Analysis of Seismic Responses and Vibration Serviceability in a High-Rise Timber–Concrete Hybrid Building

Abstract

Timber–concrete hybrid structures are commonly employed in multi-story timber buildings; however, further research is necessary to fully understand the seismic performance of these structures as well as the dynamic properties of the floor. The two dynamic concerns, seismic effects and the vibration of floors in hybrid structures, are key issues, in view of which this study aimed to investigate the small-seismic-response spectra and elastic time histories in a high-rise timber hybrid building, specifically the medical technology building of Jiangsu Provincial Rehabilitation Hospital in China. The dynamic characteristics of a localized cross-laminated timber (CLT) floor were tested in situ, and the impacts of human-induced vibration were quantified. Comprehensive theoretical analysis results reveal that the basic vibration pattern of the structure was mainly translational in nature and that the period ratio, inter-story displacement angle, and shear-to-weight ratio all met the demands of the Chinese timber building design code. The experimental test results show that the vertical natural frequency of the CLT floor was about 15.96 Hz and thus met appropriate requirements with respect to natural frequency. However, peak floor acceleration was found to be high under the conditions of a single person walking quickly, a single person trotting, and multiple persons walking randomly. In light of these findings, the floor should be paved with a fine-grained concrete building surface, according to design requirements, so that its serviceability might be improved. Overall, the relevant analytical methods presented in this paper provide guidance and practical reference for the seismic analysis of timber hybrid structures, as well as vibration serviceability analysis for CLT floors.

1. Introduction

Timber buildings exhibit a number of advantages in terms of carbon reduction, appearance, speed of erection, and seismic performance. Because of this, a developing trend in recent years has been the use of renewable and sustainable wood resources in the construction of high-rise buildings [1]. Systems commonly used today for multi-story timber buildings include timber-frame braced structures, timber-frame shear-wall structures, cross-laminated timber (CLT) shear-wall structures, top and bottom hybrid timber structures, and concrete-core timber structures [2]. Examples of such buildings may be readily given. The 18-story, 85.4 m tall Mjøstårnet building in, Brumunddal, Norway, features an internal CLT cylinder system and external glued laminated timber frame supports [3]. The 18-story, 53 m tall University of British Columbia (UBC) Brock Commons building in Vancouver, British Columbia (BC), Canada, employs a hybrid timber structural system consisting of a glued laminated timber frame with concrete tubes [4]. The 25-story, 86.6 m tall Ascent building in Milwaukee, Wisconsin (WI), USA, is formed of an upper 19-story timber frame and a lower 6-story concrete frame. In order to fully utilize the advantages of each material, hybrid constructions consisting of both concrete and wood are increasingly commonly used for high-rise timber buildings.

In timber hybrid structural systems, the overall seismic performance is affected by the construction and mechanical characteristics of the floor [5]. Additionally, CLT floors are used on account of the better structural performance, fire resistance, and stiffness that they exhibit, compared with wood-based structural panel floors [6,7,8,9]. These advantages derive from the orthogonal laminate lay-ups used for gluing during the production process. Every floor ought to meet the necessary standards for strength and stiffness, as well as serviceability metrics; however, because wood is characterized by low density and a low modulus of elasticity, CLT floors are more prone to vibration than conventional concrete floors [10,11,12], and this affects serviceability. When timber floors are considered for building purposes, special attention must be paid to serviceability under human-induced vibration [13,14,15]; nevertheless, there is currently no established design technique for the serviceability of timber flooring among the building rules that apply in China. In order to achieve serviceability standards by restricting the natural frequency of the floor, methods for evaluating the vibration performance of lightweight floors are currently employed in China and elsewhere. According to Eurocode 5 [16], the natural frequency of the floor should not be any lower than 8 Hz. In ASIC, it is specified that the natural frequency of lightweight floors must not be lower than 9 Hz [17]. Finally, the Chinese technical standard for human comfort of floor vibration, JGJ/T 441-2019 [18], stipulates that the first-order vertical natural frequency for floors with a walking excitation dominant must not be lower than 3 Hz. A number of studies on timber floors have been published in recent years. In field experiments, Jarnerö et al. [19] found that floor vibration performance was greatly affected by boundary conditions. Huang et al. [20] used finite element simulation to determine that supporting-beam stiffness is crucial to mitigating floor vibration. Wang et al. [21] used numerical simulations and experiments to assess and forecast the human-induced vibration of a CLT floor under multi-person loads; they found that the vibration response induced by multiple persons was almost twice as much as that induced by a single person and was more likely to cause discomfort. Lou et al. [22] suggested a relationship between serviceability and peak floor acceleration; based on this, they suggested a peak acceleration limit for floors subjected to human-induced vibration, with peak acceleration being classified as follows: less than 0.0125 m/s² was considered comfortable; between 0.0125 m/s² and 0.15 m/s² was considered slightly comfortable; between 0.15 m/s² and 0.8 m/s² was considered uncomfortable; and greater than 0.8 m/s² was considered unbearable.

However, many studies have focused on timber hybrid structures with reinforced concrete cores as the lateral resistance system [23,24,25,26,27,28]. For multi-story timber hybrid structures, the core is prone to stiffness concentration and structural torsion and can be replaced by a concrete frame instead of a core, but there are few relevant studies. Meanwhile, studies on the serviceability of CLT floors involving field measurements of human-induced vibration have been relatively rare, both in China and across the world. Given this, this paper takes as its focus an actual timber–concrete hybrid project in which the core was changed to a concrete frame as an example to analyze the overall seismic performance and investigate the dynamic performance and human-induced vibration of a local CLT floor. Serviceability conditions for different configurations of the CLT floor were also analyzed, and design recommendations to enhance the serviceability of the timber floor were proposed in light of the results obtained. The findings presented in this paper provide a practical reference for those engaged in seismic analysis and vibration serviceability of high-rise timber hybrid buildings.

2. Project Overview

For the present work, a construction project for a rehabilitation hospital in China was studied; this facility had a total construction area of about 208,000 m², divided into an outpatient medical building and an inpatient building. In pursuit of the study objectives, the main focus was on the outpatient medical building. This building had one underground layer (with a height of 6.4 m) and eight floors above ground. The first floor had a height of 5.4 m, the second to fifth floors each had a height of 4.5 m, and the sixth to eighth floors each had a height of 3.7 m, giving a total above-ground height of 34.5 m. An architectural rendering of the project is shown in Figure 1. Modern engineered wood materials were used for the seventh and eighth floors, as well as for the roof frame. For the beams and columns, glue-laminated timber was utilized, while for the composite floors, cross-laminated timber was employed.

The project used glulam, in line with Chinese timber building design code GB 50005-2017 [29]. TC_T32 grade was selected with an elastic modulus of 9500 MPa and a flexural strength of 22.3 MPa.

Table 1. Sections of glulam beam and column members.

Member	Section Size (mm × mm)
Column	600 × 600
Beam	600 × 660, 400 × 660, 250 × 660

displays the component dimensions for glulam beams and columns. The floor and roof panels were constructed of CLT with strength grade C20; this was chosen in reference to the Chinese technical guide for CLT structures [30], and its performance parameters are shown in

Table 2. Properties of CLT material.

Standard Value (MPa)				Design Value (MPa)
fm,k	fc,k	ft,k	Ek	fm	fc	ft	E
19.3	15.5	12.0	6400	13.2	10.6	8.2	9500

. CLT laminates were made of spruce. Each floor slab consisted of five layers, each 140 mm in thickness; each CLT roof slab consisted of five layers, each 180 mm in thickness. The timber structure area of the project is shown in Figure 2. Plate connectors were made of Q235B steel; bolts were ordinary grade 4.8 and hot-dip galvanized; pins were made of stainless steel.

In line with GB 50011-2010 [31], the structural safety level of the project was grade I, group I was the design seismic grouping, the seismic intensity was 7 degrees, and the basic seismic acceleration was 0.10 g. In addition, the building belonged to the class B seismic defense category, the soil category of the site was class II, the design level of the foundation was class B, and the design service life of the structure was 50 years.

3. Structural System and Arrangement

3.1. Structural System

For the outpatient medical building, considering the comprehensive needs of building function and structural stress, a combination of timber and concrete structures with vertical and horizontal hybrid methods was adopted, as shown in Figure 3 and Figure 4, respectively. A reinforced concrete frame structure was adopted for the first to fifth floors, which formed the outpatient medical technology area; a concrete elevated floor was adopted for the sixth floor, which was used as a roof garden and equipment room; and a hybrid system of timber and concrete was adopted for the seventh and eighth floors, which were used as a scientific research office area. Concrete frame structures were adopted for elevator and staircase areas. In addition, a local elevator room and a water tank room were located on the roof; for the purposes of our analysis, this was considered to be a ninth floor. In a seven-to-eight-story horizontal hybrid system such as this, a timber structure bears only the vertical load of the area where it is located, and there is no side-resisting wall; such a structure is, therefore, transparent and easy to use. Moreover, the concrete frames used for the core of the building and the stairwell areas take on all the horizontal action of the floor and the vertical loads of the area in which they are located at the same time; this fully utilizes the benefits of the performance of the two materials, i.e., concrete and timber. The system reported in this paper is the first of its kind in China, and so our work had no normative basis for reference purposes.

3.2. Floor Construction

To effectively transfer seismic effects, as well as horizontal wind loads, from portions of the timber frame to the concrete frame, CLT composite floors were used in timber frame areas. The tops of the CLT floors were lined with steel tie slats and fixed with self-tapping screws; a top layer of reinforced fine gravel concrete was then poured on top, to a thickness of 50 mm, to ensure that the various parts of the structure worked in concert under horizontal loads. The steel slat arrangement and composite floor construction are shown in Figure 5. At present, the concrete surface has not been constructed.

4. Structural Analyses

Because two different materials were used for the project structure and because different materials are subject to different degrees of damping, the strain energy factor method was used in the analysis to obtain the damping ratio for each type of vibration of the structure, i.e., the damping ratio of the hybrid construction was determined by taking a weighted average of the strain energy of the damping ratios of the different materials [32]. Based on the results of trial calculations, the damping ratio under conditions of frequently occurring earthquake was finally determined to be 0.045.

When calculating a seismic index for the whole structure, floors were considered simply to be rigid. As shown in Figure 6, the nodes between timber columns and concrete columns, as well as between timber beams and timber columns, were considered according to hinge joints.

Nodes between timber beams and concrete bearings were also considered according to the hinge joint, considering the difference of vertical deformation in the junction between the timber structure and the concrete area during construction and usage. The concrete columns on the first floor were regarded as embedded in the basement roof slab, and the bearings were regarded as fixed connections.

4.1. Response Spectrum Analysis under Frequently Occurring Earthquake

The analytical model was constructed by using YJK 3.0.1 software, which is extensively utilized in China for structural analysis and design, as seen in Figure 7.

The dynamic characteristics of the structure were analyzed by using the eigenvalue analysis approach.

Table 3. Main results of response spectrum under frequently occurring earthquake conditions.

Software	Period (s)			First-Floor Shear–Weight Ratio (%)			First-Floor Inter-Story Drift Angle
	T1	T2	T3	X	Y	Minimum	X	Y	Limit
YJK	1.23	1.20	1.10	3.23	3.27	1.60	1/1194	1/1218	1/550
MIDAS	1.37	1.32	1.20	2.64	2.47	1.60	1/1079	1/1115	1/550

displays the periods that correlate to the first three modes of vibration, while Figure 8 displays the vibration modes. It is evident that the first- and second-order vibration modes of the structure are dominated by translation, while the third-order vibration mode is dominated by torsion. The ratio of torsion period (T₃) to translation period (T₁) was calculated to be 0.89. Because this value was less than 0.9, the torsional stiffness of the structure met the requirements of the Chinese design standard for multi-story timber buildings, GB/T 51226-2017 [33].

Because the structure was a timber–concrete hybrid structure and hybrid forms are more complex, MIDAS was used for additional analysis. From Table 3, it can be seen that the results of the YJK and MIDAS (https://www.midasoft.com/) analyses are similar. In addition, the values obtained for the period ratio (T₃/T₁), the first-floor shear–weight ratio, and the inter-story drift angle were all within the limits of the Chinese timber building codes.

4.2. Elastic Time History Analysis

Two sets of artificial ground motions and five sets of historical ground motions, as summarized in

Table 4. Ground Motions.

ID	RSN	Earthquake	Year	Station	Magnitude	Rrup (km)
GM1	446	Morgan Hill	1984	APEEL 1E-Hayward	6.2	51.69
GM2	1681	Northridge-04	1994	Moorpark	5.9	15.43
GM3	58	San Fernando	1971	Cedar Springs Pumphouse	6.6	92.59
GM4	613	Whittier Narrows-01	1987	Covina-W Badillo	6.0	18.59
GM5	2959	Chi-Chi, Taiwan-05	1999	CHY055	6.2	97.72
GM6	ArtWave-RH2	artificial
GM7	ArtWave-RH3

, were selected to supplement the calculation of the structure for the purpose of carrying out elastic time history analysis under the conditions of frequently occurring earthquake. The selected ground motions met the requirements of spectral characteristics, effective peak value, and duration; at the same time, the selected records also met the requirements of base shear and higher-order vibration mode, with the maximum value of seismic acceleration being 35 cm/s². The seven earthquake spectra and the target spectrum are shown in Figure 9.

The elastic time history analysis results are presented in

Table 5. Results of base shear from elastic time history analysis.

Earthquake Effect		First-Floor Shear (kN)		Ratio to Response Spectrum Method
		X	Y	X	Y
Response spectrum method		32,996	33,063	—	—
Historical ground motions	GM1	30,747	31,259	0.93	0.95
	GM2	28,990	28,715	0.88	0.87
	GM3	28,772	29,196	0.87	0.88
	GM4	29,723	30,076	0.90	0.91
	GM5	29,825	31,234	0.90	0.94
Artificial ground motions	GM6	28,668	26,048	0.87	0.79
	GM7	30,505	33,309	0.92	1.01
Average value of each motion		29,604	29,977	0.90	0.91

and Figure 10. The results demonstrate that the average base shear of seven groups of ground motions was not less than 80% of that obtained by using the response spectrum approach, and that the base shear of each group of ground motions was not less than 65%. The selected ground motions, therefore, met the requirements of Chinese seismic design code GB 50011-2010 [31]. Figure 10 shows that under horizontal seismic action in the X- and Y-directions, the largest inter-story drift angle was found on the seventh floor, where timber and concrete were mixed horizontally, but the inter-story drift angle for all floors was less than the code limit of 1/550.

5. Serviceability Analysis of CLT Floor

Considering that the project has not yet been completed and the concrete top layer of the CLT composite floor has not yet been poured, in order to assess the serviceability of the CLT floor and to compare its performance with that of the combination floor after the concrete top layer is poured at a later stage, at this stage, the CLT floor underwent tests for dynamic features and human-induced vibration to ascertain the natural frequency and peak acceleration under human-induced vibration, respectively.

The test area had an area of 8.4 m × 8.4 m, as illustrated in Figure 11. The peripheral glulam beam (GLB1) in the test region had a cross-section measuring 600 mm by 660 mm. The side glulam beam (GLB2) had a cross-section of 400 mm by 660 mm, while the two secondary beams (GLB3) had a cross-section of 250 mm by 660 mm, and four CLT boards of 140 mm in thickness were continuously laid on areas (I), (II), (III), and (IV). The connections between the glulam beam and the CLT floor were secured with self-tapping screws.

5.1. Dynamic Characteristics of CLT Floor

Environmental excitation was used to stimulate the structure and obtain the vibration acceleration time history curve during the dynamic characteristics test of the CLT floor. By using a quick Fourier transform, the time history curve was converted into a frequency response function, from which the natural frequency of the CLT floor was derived. The measurement equipment included LC0108 acceleration sensors and a TST3828E acceleration acquisition system. Seven acceleration sensors were positioned, as depicted in Figure 12, after careful evaluation of the floor’s potential vibration patterns and the locations where a significant acceleration reaction was anticipated. Data for acceleration time history response under environmental excitation were collected for 300 s at each measurement point.

A time history curve and spectrogram are shown in Figure 13 and Figure 14, using two measurement points, 1 (at the center of the end span) and 6 (at the center of the floor span), as an example. From the spectrogram, it can be seen that the fundamental natural frequencies corresponding to the peak amplitude at measurement points 1 and 6 are about 16.98 and 15.96 Hz, respectively. At acceleration measurement points 1–7 on the CLT floor, frequencies between 13.04 and 16.98 Hz were measured, with a maximum at point 1 and a minimum at point 5. There was no successful acquisition at point 4. The frequency mean, standard deviation, and coefficient of variation were 15.49 Hz, 1.21 Hz, and 7.8%, respectively.

The floor in the project was laid in multiple continuous spans (see Figure 11b); however, the application of vertical natural frequency calculation theory to multi-span continuous plates has been relatively rare. For the purposes of the present study, for a plate supported with four beam sides, the vertical natural frequency could be calculated by using the Eurocode 5 [16] formula:

$$f_1={\displaystyle \frac{\mathsf{\pi }}{2L^2}}\sqrt{{\displaystyle \frac{{(EI)}_{\mathrm{L}}}{m}}}$$

(1)

where L is the floor span, (EI)_L is the bending stiffness in the direction of the strong axis of the floor, and m is the mass of the floor per square meter.

In this project, without considering the effects of the actual continuous laying of the floor, the plate-and-beam support surface at the screw connection, or any other favorable factors on the vertical vibration frequency contribution, the natural frequency of the floor was calculated, by using Formula (1), to be 8.34 Hz and thus satisfied the code requirement that it should not be less than 8 Hz.

5.2. Human-Induced Vibration of CLT Floor

The acceleration response of the CLT floor was assessed in the current study’s human-induced vibration test by mimicking the excitation that a person would typically experience when walking. According to the AISC [17], it is recommended that the deadweight of the tester, with a mean of 700 N and a standard deviation of 145 N, meet the requirements of a normal distribution. The human-induced vibration test was used to determine several important influence parameters, including movement frequency, walking path, number of pedestrians, and movement mode. The traveling excitation-related parameters are depicted in

Table 6. Traveling excitation-related parameters.

Traveling Excitation	Speed (m·s−1)	Stride Lengths (m)	Frequency (Hz)
SW	1.10	0.60	1.70
W	1.50	0.75	2.00
FW	2.20	1.00	2.30
TR	2.20	0.60	3.00

Note: SW—slow walking; W—walking; FW—fast walking; TR—trotting.

.

Throughout the examination, the moving frequency was regulated by a metronome, and distinct excitations were administered by stepping on corresponding colored markers. The acceleration sensor arrangement was consistent with the dynamic characteristics test. Two traveling paths were arranged close to the central axis of the test area; these were defined as Path A (P-A) and Path B (P-B), as shown in Figure 15.

Table 7. Traveling conditions.

Condition	Excitation	Traveling Path	Number of People
1	SW	P-A	1
2	W	P-A	1
3	FW	P-A	1
4	TR	P-A	1
5	W	P-B	1
6	RW		8

Note: RW—random walking.

lists each traveling condition, and Figure 16 shows the typical condition site.

5.2.1. Influence of Movement Frequency on Vibration Performance of CLT Floor

One person moving at various frequencies was used to test the vibration response of the CLT floor. Figure 17 depicts the peak accelerations recorded by each of the seven sensors for the traveling conditions from 1 to 4 (Table 7). These results demonstrate that the peak acceleration of the floor increased gradually with an increase in frequency. Taking measurement point 6 as an example, the peak acceleration of the CLT floor under the “W”, “FW”, and “TR” conditions increased by 63%, 257%, and 339%, respectively, compared with the corresponding values for the “SW” condition.

5.2.2. Influence of Traveling Paths on Vibration Performance of CLT Floor

A single individual followed Paths A and B in succession at a motion frequency of 2 Hz (traveling conditions 2 and 5, respectively) in order to examine the impact of traveling paths on the vibration response of the CLT floor. Figure 18 displays the peak acceleration at each CLT floor measuring location under various traveling paths. It is evident that for the CLT floor under various traveling pathways, the peak acceleration was larger in the center region of the floor and decreased with the increase in distance from the center. Comparing the peak acceleration of the CLT floor under different traveling paths, when a single person traveled over each path, it was discovered that the peak acceleration under Path A was less than that under Path B. This was because a glulam secondary beam constraint acted on both sides of the path in the direction of Path A and this decreased the peak acceleration at the measurement point under the action of traveling in this direction.

5.2.3. Influence of Random Walking by Multiple Individuals on Vibration Performance of CLT Floor

Eight persons were randomly selected to walk, which corresponds to condition 6, in order to test the vibration response of the CLT floor under the condition of several individuals walking at random. Under this condition, the maximum obtained value for the peak acceleration of the CLT floor occurred at measurement point 6 with a value of 0.51 m/s². Compared with condition 6, the maximum peak acceleration for conditions 1–4 all occurred at measurement point 5, with values of 0.08, 0.12, 0.23, and 0.31 m/s². The maximum peak acceleration for condition 5 occurred at measurement point 6, with a value of 0.15 m/s². A number of people walking on the CLT floor at random has a significant impact on its vibration performance.

5.3. Serviceability Evaluation of CLT Floor

The natural frequency requirement for the floor was met, as indicated by the test results, which show that the average natural frequency of the CLT floor in the test area was 15.49 Hz and the minimum value was 13.04 Hz. However, the peak acceleration of the CLT floor was higher than 0.15 m/s² under conditions 3 (single-person fast walking), 4 (single-person trotting), and 6 (multiple people random walking), pedestrians were prone to feel uncomfortable. Considering the two serviceability evaluation factors of natural frequency and peak acceleration under human-induced vibration, the field-tested CLT floor could not meet the requirements. Thus, pouring a certain thickness of concrete topping on the surface of CLT floors as proposed in the design documents is necessary to improve serviceability.

6. Conclusions

In this study, which involved a typical timber–concrete hybrid building under construction in China, the small-seismic-response spectra and elastic time history were analyzed, and the vibration performance of a localized CLT floor was examined. This work was undertaken to comprehensively explore the seismic performance indexes of a timber hybrid building and the serviceability of the floor. From in-depth studies of the building, the following main conclusions were obtained:

(1) When YJK and MIDAS were used to analyze the response spectra of the timber–concrete hybrid building under the condition of frequently occurring earthquake, similar results were obtained using each method. The main vibration mode of the structure was dominated by translation, and indexes such as the period ratio, the shear–weight ratio, and the inter-story drift angle all satisfied the Chinese code.

(2) The base shear of each set of ground motions was not less than 65% of that obtained by using the response spectrum approach; the average base shear of the ground motions employed in the elastic time history analysis was not less than 80% of that obtained by using the response spectrum method; and the largest inter-story drift angle of each earthquake was found for the seventh floor, where the timber and concrete were mixed horizontally, but the inter-story drift angle at all floors was less than the Chinese code limit of 1/550.

(3) The average value, standard deviation, and coefficient of variation of the natural frequency of the CLT floor were found to be 15.49 Hz, 1.21 Hz, and 7.8%, respectively, based on the results of the dynamic characteristics test; the average value was within the limit value for the natural frequency of the floor but was higher than that calculated according to Eurocode 5, most likely as a result of differences in boundary conditions.

(4) The examination of the human-induced vibration of the uncast concrete surface layer revealed that the peak acceleration of the floor rose with frequency when a single person moved across it; different traveling paths affected the vibration performance of the floor, so that the peak acceleration of vibration was inversely correlated with the degree of edge limitations on the floor. The random walking of multiple persons had the greatest effect on the vibration performance of the CLT floor, with peak acceleration reaching 0.51 m/s². In light of these results, pouring concrete topping on the surface of the CLT floor as proposed in the design documents later is necessary to improve its serviceability.

In summary, by using a reasonable and optimized structural design, a novel timber–concrete structural system with multiple horizontal and vertical combinations delivered improved structural performance compared with traditional designs and was more able to exploit the performance advantages associated with the component timber and concrete materials; such systems, therefore, offer a broad prospect for future applications.

To further investigate the vibration serviceability of the composite floor, a performance test of the CLT–concrete composite floor will be conducted in the next stage and compared with the findings of this paper. Additionally, an investigation will be conducted into the impact of the floor construction on the overall seismic performance of the structure as well as the in-plane stiffness.