Experimental study on the compressive performance of new sandwich masonry walls

Abstract

Sandwich masonry wall, namely, multi-leaf masonry wall, is widely applied as energy-saving wall since the interlayer between the two outer leaves can act as insulation layer. New types of sandwich walls keep appearing in research and application, and due to their unique connection patterns, experimental studies should be performed to investigate the mechanical behavior, especially the compressive performance. 3 new types of sandwich masonry wall were investigated in this paper, and 3 different technical measures were considered to guarantee the cooperation between the two leaves of the walls. Based on the compression tests of 13 specimens, except for some damage patterns similar with the conventional masonry walls, several new failure patterns are found due to unique connection construction details. Comparisons were made between the tested compression capacity and the theoretical one which was calculated according to the Chinese Code for Design of Masonry Structures. The results indicate that the contributions of the 3 technical measures are different. The modification coefficient (g) was suggested to evaluate the contribution of the technical measures on the compression capacity, and then a formula was proposed to evaluate the design compression capacity of the new sandwich masonry walls.

Keywords

sandwich wall / insulation wall / connection / compressive performance / compression test

Introduction

Since 1973, four energy crises (oil crises) have taken place worldwide [2]. On one hand, frequent occurrence of energy crises urges human beings to find new energy such as nuclear energy and solar energy; on the other hand, it turns energy conservation a worldwide issue. Some statistical and research results on generalized building energy consumption indicate that building energy consumption in a broad sense takes up more than 40% of the total energy consumption of human society [3]; as for building energy consumption in a limited sense, its proportion in the total energy consumption is 20%-27.6% according to some investigations in the mainland of China [4,5]. Undoubtedly, building energy conservation plays a more and more important role in future energy saving.

Exterior wall is the major component of building envelope. The heat loss through exterior wall is about 25% of building’s total heat loss [6]. As for the buildings in heating area, this proportion is as high as 50% [7]. Therefore, exterior-wall insulation is the key technology to realize the building energy conservation. Exterior-wall insulation is usually achieved by the application of an insulation layer, and the wall insulation construction can be classified into internal, external and sandwich insulation according to the relative position between the insulation layer and the wall.

For the internal insulation, it requires quite little for the waterproof, weathering and some other technical specifications of decorative material and insulation material; but the indoor effective area is reduced considerably by the insulation layer, and the ring beam, slab and tie column are hotspots for heat channel [6]. For the external insulation, the influence of heat channel is mostly eliminated, and structure is protected by insulation layer from external environment; however, the insulation layer damages easily and is inconvenient for strengthening and repairing [8]). The sandwich insulation combines both advantages of the above two technologies. Therefore, nowadays more and more civil engineers pay more attention to the development of new types of sandwich walls.

Sandwich masonry wall can also be referred to as multi-leaf masonry wall, it has been used since the ancient times, and it is a typology often found in some historical city centers worldwide [9-11]. A number of experimental investigations have been carried out to study the structural behavior, especially the compression capacity of multi-leaf wall [12,13]. Binda et al. [14] even proposed one simplified calculation method for the compression capacity. Sandwich insulation wall is accomplished by replacing the rubble material in the interlayer of multi-leaf wall with insulation material, and it has been commonly applied since the 1930s. In China, sandwich insulation wall has been widely applied in heating area since the late 1980s [15]. It should be noted that in the past, sandwich masonry wall only acted as non-bearing element. In the 21st century, some experimental studies have been completed to evaluate the bearing behavior of such sandwich walls. Li et al. [16,17] carried out tests on a couple of sandwich walls laid up by perforated bricks to investigate their failure patterns and mechanical response, the formulas for the cracking load and peak load of the walls were proposed. Zhang et al. [18] studied the cohesion behavior between the two leaves of sandwich walls by tests, the influence of connector and compression stress level on the seismic behavior of the walls were also focused on. Tang et al. [19,20] performed experiments on insulation walls that carried loads by their inner leaves to study the failure mechanisms of the specimens and to discuss the calculating method of their compression capacities. Shang et al. [21] carried out compressive tests on insulation masonry specimens from the aspect of material properties, and they worked out a formula for compressive strength of this kind of masonry. In fact, the compressive performance of sandwich walls, especially the cooperation between the two leaves, becomes a critical safety issue that needs further study. However, to the best knowledge of the authors, neither the mechanical behavior nor the structural performance of existing sandwich walls has been well understood; what’s more, because of the difference in construction details, most of the conclusions in structural performance for existing sandwich walls cannot be applied on new sandwich walls.

Motivated by the above analysis and application need, 3 brand-new types of sandwich walls were introduced in this investigation, and a total number of 13 specimens were constructed and tested under compressive loading. Some experimental results were obtained including failure patterns, compressive load-vertical strain curves, compression capacity, and etc. Monographic investigation was performed to study the contribution of the technical measures on compression capacity for the new walls. A compression capacity modification coefficient (g) was suggested in to evaluate this kind of contribution, and then a new formula for the design value of these new walls’ compression capacity was put forward.

Test program

Test specimens

The brand-new sandwich walls were classified into 3 groups denoted by A, B and C according to their masonry connection patterns and construction details, and they were laid up by 3 types of bricks respectively. Following technical measures were taken in the 3 groups of walls, respectively, to guarantee the cohesion between the two leaves. A header course was added every 2 or 3 stretcher courses in Group A, a prefabricated steel mesh (see Fig. 1) composed by two longitudinal bars connected by diagonal bars was embedded in mortar every 3 or 5 bed joints in Group B, and the bricks in Group C overlapped each other. Apparently, the header course in Group A and the steel mesh in Group B worked as transverse connectors, and the distinctive masonry cohesion pattern of Group C can help the two leaves work together as well.

A total of 13 specimens in 4 groups were constructed and tested under compressive loading. For Group A, there were A1 and A2 with different height-thickness ratios, aiming to study the influence of height-thickness ratio on the cooperation between the two leaves, which was the same for Group B. There were 3 specimens with same dimension for Group C, and they were loaded according to different in-plane eccentricities (0, 55 mm, 110 mm), aiming to study the influence of loading eccentricity on the compression capacity. The detail drawings for the 4 groups of walls are shown in Fig. 2 and some major parameters of the specimens are presented in Table 1. Group D is a traditional group which was used only for comparisons.

Test setup, instrumentation and procedure

The test setup is illustrated in Fig. 3. The compression load was applied by an actuator and a reaction frame. From the beginning of the tests, the compression load was applied step after step with an increment of 20 kN, which was smaller than 10% of the estimated compression capacities of the walls. After obvious cracking sounds were heard or the first crack was observed, the loading increment was reduced to 10 kN. Once the specimen was seriously damaged or the vertical deformation kept increasing even the compression load remained as a constant, the loading process was stopped.

The in-plane deformation and out-of-plane displacement of the wall were measured by a set of linear variable differential transducers (LVDTs). The arrangement of the LVDTs in the tests is displayed in Fig. 3. A static strain measuring system and a computer were used to record and monitor the load-deformation diagrams of the walls.

Material properties

Before carrying out tests of the walls, the mechanical properties of the materials, namely bricks and mortars were obtained through a set of compression tests. According to the Chinese codes, Fired Perforated Bricks (GB 13544, 2000), Fired Common Bricks (GB 5101, 2003) and Test Methods for Wall Bricks (GB/T 2542, 2003), the average values of the compressive strength for the three types of bricks in Table 1 were 10.64, 10.60, and 18.67 MPa, and their strength grades can be termed as MU10, MU10, MU15, respectively. During the construction of the walls, 9 cubes (3 groups) of mortar (70.7 mm × 70.7 mm × 70.7 mm) were also cast according to the Standard for Test Method of Performance on Building Mortar (JGJ/T 70, 2009). Compression tests were carried out on these cubes after 28 day’s curing age with the walls under the same ambient condition, and the average values for compressive strength and strength grades of the mortars were reported in Table 2.

Test results

Main results of compression load and deformation

The main results of the compressive loading tests are summarized in Table 2, including cracking load P_cr, peak load P_u, deformation corresponding to P_u, namely D_u, average value of compressive strength for brick f₁, average value of compressive strength for mortar f₂, design value of compressive strength for masonry f, etc. Further analysis on compression capacity will be conducted in Section 4.

Compression load-vertical strain curves and test process

The process of the tests and the performance of the specimens under compression loading are comprehensively reflected in the compression load-vertical strain curve, so the compressive loading-vertical strain curves obtained in the tests are presented in Fig. 4. The red dots in the curves indicate the appearance of the first vertical crack, which was along the loading direction.

The loading-strain curves of all the specimens could be divided into 2 stages.

1) Elastic stage: from the beginning to the appearance of the first crack. The curve kept linear, though no cracks could be seen during this stage, it should be noted that there was tensile stress in the directions perpendicular to the compressive stress. Since the tensile stress increased with the compression load increasing, the original micro fractures in the wall developed gradually. When the tensile stress in some point reached the tensile strength of the masonry, the first macro crack occurred and it developed into the first crack visible to naked eyes.

2) Non-elastic and failure stage: from cracking to the end. Pronounced curvature could be observed in the curve, indicating that the vertical rigidity of the wall decreased. With the increase of the compressive loading, the first crack developed downward and other vertical cracks kept appearing, meanwhile, cracking sounds could be heard clearly.

No obvious difference could be observed between the curves of Group D, namely the conventional masonry wall, and the curves of the other groups, reflecting that the brand-new sandwich masonry walls shared similar loading and failure process with the traditional ones.

Failure patterns

The detail drawings of the new types of sandwich walls are presented in Fig. 5, because of the presence of insulating materials, only the dashed areas could be taken as the effective area to carry the compression load. It should be noted that the protruding part in Fig. 5(a) didn’t carry load and the boundary between the effective area and the protruding part was weakened by the holes in the perforated bricks. From the observations of the tests, it could be found that the 3 brand-new types of sandwich walls not only shared some similar failure patterns with the traditional masonry walls, such as the development of vertical cracks, but were also damaged easily due to some distinct failure patterns which are presented as follows.

1) Failure pattern 1: Local failure

Initially, the sandwich wall was divided by the insulation layer into two leaves, once the wall began to crack, the two leaves of the wall were further divided by the vertical cracks into several small masonry columns. With the compression load and vertical deformation increasing, the load distribution between the small masonry columns tended to be non-uniform, so some of the small masonry columns which carried relatively larger load crushed easily during the non-elastic and failure stage. The failure patterns of Specimen A1-1 and B1-1 were depicted in Fig. 6.

2) Failure pattern 2: Protruding part peeling off

In this series of tests, the top of the walls in Group A were all stretcher courses (see Fig. 5(a)). During the loading process, the upper surface of the protruding part carried the compression load transmitted from the loading beam while the lower surface of it was free which caused shear stress in the “boundary” (see Fig. 5(a)). Meanwhile, as has mentioned before, the “boundary” was weakened by the holes in the perforated bricks. As a result, the protruding portion peels off easily during the non-elastic and failure stage. Figure 7 is a picture of Specimen A2-2 after its protruding part peeling off.

3) Failure pattern 3: cracking in the side view surface

Since the whole section of the walls in Group C could not be taken as the effective section as a whole, there were boundaries between the effective sections and the non-effective sections. These boundaries separated one brick into several parts and there were strain difference and shear stress between them, leading to the brick cracked easily along the boundaries during the non-elastic and failure stage. The cracking pattern in the side view surface of Specimen C-2 is shown in Fig. 8.

Analysis on compression capacity

The thickness of the effective sections for the brand-new sandwich walls could be got from the drawing details easily (see Fig. 5). Since technical measures were taken in the 3 types of sandwich walls to guarantee the cooperation between two leaves, the compression capacity should be larger than the capacity calculated by the effective section (P_cal1) while smaller than the capacity calculated by the gross section (P_cal2). To evaluate the contribution of the technical measures on compression capacity, a capacity modification coefficient (g) which represented the ratio between P_u and P_cal1 was introduced.

Compression capacity of the walls in Group A

The thickness of the effective section for the walls in Group A was 190 mm (see Fig. 5(a)), and the thickness of the gross section was 280 mm. The test results and calculated results of compression capacity for Group A1 and A2 are presented in Table 3 for comparisons.

It was shown by the results of Group A1 that the value of P_u for Specimen A1-1 was 1.25 times of P_cal1 which indicated that the header courses contributed quite a lot to the cooperation between the two leaves of the wall, so the compression capacity modification coefficient of Group A1 was calculated as 1.25.

The results of Group A2 indicated that the values of P_u for the walls in Group A2 were 1.16 and 1.19 times of P_cal1 and the average was 1.18. Obviously, the header courses in Group A2 were not as effective as the ones in Group A1. The compression capacity modification coefficient of Group A2 was set as 1.18.

The height-thickness ratios of the two walls in Group A2 were different, but their compressive capacities were pretty close, and the two walls both ended up with strength failure instead of stability failure, so the height-thickness ratio had little impact on the cooperation between the two leaves of the walls in Group A within the range of this series of tests.

Compression capacity of the walls in Group B

The thickness of the effective section for the walls in Group B was 190 mm (see Fig. 5(b)), and the thickness of the gross section was 240 mm. The test results and calculated results of compression capacity for Group B1 and B2 are presented in Table 4 for comparisons.

It could be observed in the results of Group B1 that the values of P_u for the walls in Group B1 were, respectively, 1.09 and 1.13 times of P_cal1 and the average was 1.11 which indicated that the steel meshes had some contribution in the cooperation between the two leaves of the walls. As a result, the compression capacity modification coefficient of Group B1 was defined as 1.10.

The results of Group B2 indicated that the values of P_u for the walls in Group B2 were 1.08 and 1.07 times of P_cal1 and the average value was 1.08. Obviously, the steel meshes in the walls of Group B2 did not contribute so much as the ones in Group B1 because the number of them decreased. Therefore, the compression capacity modification coefficient of Group B2 was 1.08.

Additionally, the height-thickness ratios of the two walls in Group B1 were different, but no obvious difference in compression capacity could be observed between the two walls, so the height-thickness ratio had no influence on the cooperation of the two leaves of the walls in Group B.

Compression capacity of the walls in Group C

The thickness of the effective section for the walls in Group C was 136 mm (see Fig. 5(c)), and the thickness of the gross section was 200 mm. The test results and calculated results of compression capacity for Group C are presented in Table 5 for comparisons.

What indicated by the results of Group C was that the values of P_u for the walls in Group C were, respectively, 1.11, 1.24 and 1.24 times of P_cal1 and the average was 1.20. The compression capacity modification coefficient of Specimen C-1 (axial loading) was smaller than that of Specimen C-2, and the values of g for Specimen C-2 and C-3 were both 1.24, which were unexpected, so this set of data couldn’t reflect the influence of in-plane eccentricity on the cooperation between the two leaves of the walls in Group C. Additionally, the compression capacity modification coefficient of the walls in Group C was set as 1.20.

A proposed formula for the compression capacity of the new sandwich walls

The modification coefficient g for compression capacity was introduced into the formula for the capacity of masonry compressive members in the Chinese Code for Design of Masonry Structures [1], and then we got a new formula for the compression capacity of the 3 types of sandwich walls:

(1)N\le \phi \gamma fA,

where N stands for the design value of axial force; j represents the influence coefficient for height-thickness ratio (b) and axial force eccentricity (e) to bearing capacity of compression member, f is the design value for compressive strength of masonry, j and f are got from the Chinese Code for Design of Masonry Structures [1] according to the strength grade of bricks and mortar; A indicates the net section area of sandwich masonry wall.

The values of g defined in the former three sections were taken into Eq. (1) for calculation and the design values of the compression capacity for the new sandwich walls (P_d) can be got. The ratio between the peak load (P_u) and P_d was figured out too, which could help to evaluate the safety redundancy of Eq. (1). The results are presented in Table 6.

The data in Table 6 indicated that most of the ratios between P_u and P_d were above 2, which was suitable for masonry elements, so the safety redundancy of Eq. (1) was satisfactory.

Conclusions

Based on the compressive tests and analysis of the new sandwich masonry walls, following conclusions could be presented:

1) Height-thickness ratio has little impact on the cooperation between the two leaves of the new sandwich walls within the range of this test investigation.

2) The header courses in the walls of Group A, the steel meshes in the walls of Group B and the overlapping of bricks in the walls of Group C all contribute to the cooperation between the two leaves of the walls, leading to some increase in the compression capacity. A modification coefficient (g) for compression capacity is introduced to evaluate the contribution.

3) Eq. (1) can be used to calculate the design value of compression capacity for the new sandwich walls.

4) Just like the conventional masonry walls, the development of vertical cracks in the front surface is the major reason for the failure of the walls in Group A and B; while the development of vertical fracture in side view surface is the main cause for the damage of the walls in Group C.

5) The distinctive connection patterns of the 3 types of sandwich walls also lead to some unique failure patterns: the protruding part on top of the walls in Group A peels off easily; the wall in Group A and Group B usually damage seriously because of local failure.

References

[1] GB 50003-2001. Chinese Code for Design of Masonry Structures. Beijing: Press of Chinese Building Industry, 2001

[2] Feng B, Wang X, Wang B C. A brief talk on the fourth energy crisis. Modern Economic Information, 2009, 15: 55–56 (in Chinese)

[3] Li Z J, Jiang Y. Pondering over the situation of domestic generalized building energy consumption. Architectural Journal, 2006, 7: 30–33 (in Chinese)

[4] Tu F X, Wang Q Y. Current situation and development of building energy conservation in China. New Building Materials, 2004, 7: 40–42 (in Chinese)

[5] Jiang Y. Current building energy consumption in China and effective energy efficiency measures. Heating Ventilating and Air Conditioning, 2005, 35(5): 30–40 (in Chinese)

[6] Zhang D L, Kou Q Q. Analysis on advantages and disadvantages of exterior wall insulation forms in China. Heilongjiang Science and Technology Information, 2011, 23: 306 (in Chinese)

[7] Qian X Q. Research on the relationship of building energy conservation technology. Zhejiang Construction, 2006, 23(9): 70–72 (in Chinese)

[8] Yang J. Perspective on technological system of exterior wall insulation on the exterior surface. Journal of Shandong Jianzhu University, 2010, 25(1): 70–73 (in Chinese)

[9] Binda L, Palma M, Penazzi D. Cautious repair of stone-masonry structures in seismic areas. In: Proceedings of the International Conference of the SFB315 Research Center of Karlsruhe University. Karlsruhe, 1999, 99–108

[10] Binda L, Saisi A, Messina S, Tringali S. Mechanical damage due to long term behaviour of multiple-leaf pillars in sicilian churches. In: Proceedings of the 3rd International Seminar on Historical Construction. Guimaraes, Portugal, 2001, 707–718

[11] Binda L, Saisi A, De Benedictis R, Tedeschi. Experimental study on the damaged pillars of the Noto Cathedral. In: Proceedings of the 8th Conference on Structural Studies, Repairs and Maintenance of Heritage Architecture. Halkidiki, Greece, 2003, 89–98

[12] Oliveira D V, Lourenco P B. Experimental Investigation on the structural behavior and strengthening of three-leaf stone masonry walls. Structural Analysis of Historical Constructions, 2006, 817–826

[13] Vintzileou E, Tassios T P. Three-leaf stone masonry strengthened by injecting cement grouts. Journal of Structural Engineering, 1995, 121(5): 848–856

[14] Binda L, Pina-Henriques J, Anzani A, Fontana A, Lourenço P B. A contribution for the understanding of load-transfer mechanisms in multi-leaf masonry walls: Testing and modelling. Engineering Structures, 2006, 28(8): 1132–1148

[15] Li S D, Zhou L H. Utilization and Problems of Compound Wall with Core Insulation. Block-Brick-Tile, 2010, 12: 29–31 (in Chinese)

[16] Li H N, Zhang J W, Liu L. Experiment and Analysis of Seismic Behavior of Hollow Brick Insulating Cavity Walls. Journal of Building Structures, 2001, 22(6): 73–80 (in Chinese)

[17] Liu L, Li H N, Zhang J W. Analysis on aseismic behavior of insulation cavity energy-saving walls. Earthquake Engineering and Engineering Vibration, 2002, 22(1): 79–84

[18] Zhang Y N, Li H, Liu M, . Interaction Working Behavior about the Internal and External Wall of Foam-in-Place Cavity Wall. Journal of Civil Architectural & Environmental Engineering, 2010, 32(3): 14–21 (in Chinese)

[19] Tang D X, Zhao Y. Compression capacity of compound wall made up of hollow bricks and mineral wool. Low Temperature Architecture Technology, 1994, 4: 17–19 (in Chinese)

[20] Li X A, Tang D X, Zhang S X. An experimental study on the cavity walls of the volcanic cinders concrete blocks. Journal of Harbin University of Civil Engineering and Architecture, 1995, 28(4): 33–39 (in Chinese)

[21] Shang J L, Fan J W, Fan F F. Experimental study on compression strength of sandwich energy saving composite wall masonry. New Building Materials, 2008, 11: 29–32 (in Chinese)

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg