- Email: [email protected]
Masonry elements with chases: Behaviour under compression
Construction and Building Materials 25 (2011) 4415–4425
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Masonry elements with chases: Behaviour under compression Nebojša Mojsilovic´ ⇑ Institute of Structural Engineering, Department of Structural, Environmental and Geomatic Engineering, ETH Zurich 8093, Switzerland
a r t i c l e
i n f o
Article history: Available online 12 January 2011 Keywords: Chases Clay blocks Load test Masonry compressive strength Unreinforced masonry
a b s t r a c t Accommodating electrical wiring and similar installations in masonry construction often necessitates the cutting of chases in masonry. In order to investigate the influence of chases on masonry behaviour compression load tests were performed on three series of unreinforced masonry elements with chases. Three different types of chase (horizontal, vertical and inclined) were cut into the masonry elements. After placing the ducts several different types of infill were applied to fill the chases. The compressive load was applied in a deformation-controlled manner up to the failure of the specimen. A considerable reduction in the masonry compressive strength was observed. The test results are discussed in detail and recommendations for practical application are given. In addition, a simple analytical truss model for predictions of failure load is proposed and discussed. Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction Accommodating electrical wiring and similar installations in masonry construction often necessitates the cutting of chases in masonry. These chases can considerably reduce the masonry compressive strength, which is the key parameter in the design of structural masonry. In general, the chases are cut vertically or horizontally, but as can be seen in many practical examples the chases in the walls can be cut quite arbitrarily, see Fig. 1. Moreover, the chases are usually not properly filled after placing the ducts and the material used for infill is often of unsatisfactory quality, especially in regard to its strength and stiffness. Previous research work in this area included tests on typical masonry elements with vertical and horizontal chases. However, the amount of the published work is rather modest. Fischer [1] tested 2.55 m high clay block masonry walls of three different thicknesses (102.5, 178 and 215 mm) and several combinations of mainly vertical 38 mm wide and 25 mm deep chases. A reduction in the masonry compressive strength of up to 28.5%, compared to the strength for walls without chases, was observed. Kirtschig and Metje [2] investigated the influence of chases and recesses on masonry compressive strength. They performed numerous tests on clay block masonry with vertical and horizontal chases. It was found, that in general, the reduction in masonry strength tends to be proportional to the reduction in cross-sectional area. As result of this experimental work, a table with dimensions of chases that are allowed without additional verification of the load-bearing
⇑ Tel.: +41 446333763; fax: +41 446331064. E-mail address: [email protected] 0950-0618/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.conbuildmat.2010.12.027
capacity has been proposed. Sahlin [3] discussed the tests results from [1] and proposed a damage index (ratio of cumulative chase length and wall thickness) as a measure for damage, i.e. strength reduction induced by chases. He related this index to the observed damage in tests and found a good correlation with test data. In addition, provisions for chases in masonry can be found in structural codes. Some codes, e.g. the Swiss Masonry Code SIA 266 [4], the Building Code Requirements and Specifications for Masonry Structures ACI 530-08/ASCE 5-08/TMS 402-08 [5] and the Australian Structural Masonry Code [6], only give general recommendations, whereas others, such as the European Masonry Code ENV 1996-1-1 [7], provide tables with values of depth and width of chases that are allowed without additional verification of the load-bearing capacity. Analogous provisions can be found in the Canadian Structural Masonry Code [8]. A research project is underway at the Institute of Structural Engineering of the ETH Zurich to investigate the structural behaviour under compression loading of typical Swiss unreinforced clay block masonry elements with chases. The main goal of the research project is to investigate the influence of chases on masonry compressive strength. In the first phase of the project load tests on the first two series of squat masonry elements with chases were completed in 2008, see Mojsilovic´ et al. [9]. Recently, a third series, which complemented previous two, has been performed. The present paper gives an overview of all the experimental work, discusses the results obtained, introduces a simple truss model for analytical analysis and provides recommendations for practical implementation. It should be noted here that the present investigation concentrates on squat masonry elements subjected to centric compression and does not investigate influence of second order effects.
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
4416
Fig. 1. Arbitrarily cut chases in masonry walls.
2. Test programme and masonry materials Load tests on the three series of masonry elements with chases have been completed. Each series consisted of seven unreinforced clay block masonry elements. The test specimen dimensions were 1200 (length bw) 1200 (height hw) mm. The thickness of the elements, tw, was 150 mm (Series A and C) and 175 mm (Series B). The test specimens were produced and the chases cut by a professional mason. Dry, factory-made standard cement mortar was mixed with water at the construction site to build the wall elements in running bond. Both the bed and the head joints were 10 mm thick and fully filled. Twenty-eight days after preparation 45–50 mm deep (ts) and 35 mm wide (bs) chases were cut in six elements of each series while the seventh one was left intact and served as a
a
control specimen, see Fig. 2. For the specimens in Series A and C the chase depth was 1/3 of tw and for those of Series B it was 2/7 of tw. Three different types of chase (horizontal, vertical and inclined), two each per series, were cut only on one side of the masonry elements except for one specimen of Series C which had horizontal chases on the both element sides. The slope of the inclined chases was tan a = 2/3. An overview of the test programme is given in Table 1 for Series A and B and in Table 2 for Series C. These tables also show the types of chase infill applied: gypsum, cement mortar and high-strength mending mortar. Note that some of the chases were left unfilled. Table 3 gives information on the properties of the different types of (hollow) clay block units that were used in the tests. fb
b
bs
hw
bw
ts
tw
c
d
α
bs bs
ts
ts
Fig. 2. Test specimens: (a) without chases; (b) horizontal chase; (c) vertical chase; (d) inclined chase.
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
4417
Table 1 Specimen designation for test programme for Series A and B. Wall thickness (mm)
Without chases
150 175
A B
Horizontal chase
Vertical chase
Inclined chase
Gypsum infill
Mortar infill
Gypsum infill
Mortar infill
Gypsum infill
Mortar infill
AH1 BH1
AH2 BH2
AV1 BV1
AV2 BV2
AS1 BS1
AS2 BS2
Table 2 Specimen designation for test programme for Series C. Wall thickness (mm)
Without chases
Horizontal chase Without infill
150 *
C
CH0
CB0*
Vertical chase
Inclined chase
Mending mortar infill
Without infill
Without infill
Mending mortar infill
CH3
CV0
CS0
CS3
Note: Test specimen CB0 had horizontal chases on both specimen faces.
Table 3 Properties of clay block units. Series
Unit
Dimensions (mm)
Void area (%)
fb (MPa)
A
B 15/19
Shape
290x190x150
45
32.1 (0.7)
B
B 17.5/19
290x190x175
46
28.6 (1.3)
C
B 15/19
290x190x150
46
42.0 (1.0)
fbtz (MPa)
1.4 (0.4)
Note: Value given in parenthesis next to the mean strength value is the corresponding standard deviation.
Table 4 Properties of mortar and chase infill. Mortar/infill type
Joint mortar
Joint mortar
Series Curing Number of specimens (ffm/fm) ffm (MPa) fm (MPa)
A, B Code 6/12 4.5 (4.1/4.8) 17.1 (16.4/17.8)
C Code 3/6 4.1 (3.8/4.2) 17.2 (16.7/17.8)
Gypsum infill Mortar infill A, B Air Code 3/6 3/6 2.8 (2.6/3.0) 2.2 (2.0/2.4) 10.0 (8.7/10.9) 5.9 (5.7/6.0)
A, B Code 3/6 4.7 (4.6/4.9) 17.2 (16.7/17.5)
Mending mortar infill C Air Code 3/6 3/6 6.2 (6.1/6.2) 6.6 (6.6/6.8) 19.9 (19.6/20.4) 57.1 (53.1/60.6)
Air 3/6 7.2 (6.8/7.5) 48.8 (44.3/51.5)
Note: Values given in parenthesis next to the mean strength value are the recorded minimum and maximum values.
denotes the unit’s compressive strength, fbtz its (net) tensile strength in the z-direction. Note that x and y are, respectively, directions perpendicular and parallel to the bed joints. The compressive strength was determined according to the European code EN 772-1 [10]. The tensile strength was determined only for Series C units using a specially developed set-up and procedure. The average compressive, fm, and flexural, ffm, strengths of the cement mortar used for the joints and for the three mortar types used for chase infill are given in Table 4. The mentioned strengths were determined according to the European code EN 1015-11 [11]. Note that two different curing procedures were applied (cf. Table 4): some mortar prisms were cured according to the code provisions [11] and some were air-cured in the laboratory, in the vicinity of the test specimens.
Fig. 3 shows typically the preparation phases for a specimen with a horizontal chase. First, a masonry saw was used to cut the chase. Afterwards, the inner-rib leftovers of the hollow clay brick were removed by means of a chisel. Finally, the installation duct was placed and the chase was filled with corresponding infill or left unfilled. Fig. 4 shows typically three specimens, CH3, AV1 and AS2 with chases filled with high-strength mending mortar, gypsum and cement mortar, respectively. Note that chases were cut on the south face of the specimens (cf. Fig. 5). After a further two weeks the specimens of Series A and B were transported to the test site and tested, whereas the specimens of Series C were produced on test site. It should be noted here that the cutting of the chases was carried out on unloaded specimens, whereas in practical applications some pre-compression due to dead load is present.
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
4418
Fig. 3. Preparation of the specimen with horizontal chase.
Fig. 4. Chases filled with mending mortar (CH3), gypsum (AV1) and cement mortar (AS2).
3. Test procedure and measurements Fig. 5 depicts the test set-up. The axial load was applied by means of three hydraulic jacks (2) placed between the support frame (1) and the upper spreader beam (3). The test specimen
(4) was placed between two spreader beams (3) and two sets of steel plates (6) that provided good contact with the specimen. In this way an unrestrained lateral deformation of the specimen was ensured. To achieve the exact position of the steel plates, two thin plaster layers (7) were applied to both the upper and the lower edges of the specimen. Additionally, except for Series C, a set of neoprene plates (5) was placed between the steel plates and the lower spreader beam, which was placed directly on the laboratory’s strong floor (8). These neoprene plates helped to achieve uniform load distribution over the specimen. Both spreader beams had a thin Teflon layer on the faces towards the test specimen. After setting-up, the specimens were subjected to an axial load, which was increased in several load steps up to failure of the test specimen. Apart from the applied load, the measurements included node displacements of the measurement nets on both specimen surfaces and the deflection at the middle of the specimen. The measurement net consisted of triangles whose nodes were aluminium bolts glued in the middle of the clay brick unit’s face and had two base lengths, sh = 300 mm and sd = 250 mm, see Fig. 6. This figure also
sd sh Fig. 5. Test set-up.
Fig. 6. Measurement net (south face) and its position relative to the chases.
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
Fig. 7. Measurement net and LVDTs for specimens CH0 (south face) and CB0 (north face).
shows the net’s position relative to the chases. Additionally, vertical shortening of the specimen was measured by means of four LVDTs with a nominal base length of 1000 mm; two on each surface (see Fig. 7). During the nodal displacements measurement, the deformation was kept constant. All measuring devices were connected to a personal computer, which processed the data in real time. In order to follow crack development and to measure crack widths, the specimens were painted white prior to testing. 4. Test results and structural behaviour Table 5 shows the values of the masonry compressive strength fx (perpendicular to the bed joints) obtained from all test series. Fmax denotes the maximum recorded compressive force, Aw is the gross cross-sectional area and D is the percentile change of the compressive strength (a negative value signifying a reduction) compared to the control specimen without chases (A, B or C).
Fig. 8. Force–deformation relationships for Series A.
4.1. Load–deformation characteristic Figs. 8–10 present the force–deformation characteristics for Series A, B and C, respectively. The deformation value shown in the diagram is an average of all four LVDT measurements (vertical shortening of specimen). As Figs. 8–10 reveal, the specimens
Table 5 Test results. Specimen
Fmax (kN)
Aw (m2)
fx (MPa)
D (%)
A AH1 AH2 AV1 AV2 AS1 AS2
849.10 666.10 637.20 616.60 689.60 666.60 551.40
0.18 0.18 0.18 0.18 0.18 0.18 0.18
4.72 3.70 3.54 3.43 3.83 3.70 3.06
–
B BH1 BH2 BV1 BV2 BS1 BS2
670.30 650.20 763.20 745.80 553.20 543.70 471.60
0.21 0.21 0.21 0.21 0.21 0.21 0.21
3.19 3.10 3.63 3.55 2.63 2.59 2.25
–
C CH0 CB0 CH3 CV0 CS0 CS3
1158.00 997.79 647.02 1365.03 1049.58 828.65 1493.24
0.18 0.18 0.18 0.18 0.18 0.18 0.18
6.43 5.54 3.59 7.58 5.81 4.60 8.30
–
21.6 25.0 27.4 18.8 21.5 35.1
Fig. 9. Force–deformation relationships for Series B.
3.0 13.9 11.3 17.5 18.9 29.6 13.8 44.1 17.9 9.4 28.4 28.9
Fig. 10. Force–deformation relationships for Series C.
4419
4420
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
Fig. 11. Typical crack patterns.
Fig. 12. Deformation field and crack detail (Specimen BS2).
behaved more or less linear-elastically almost up to failure reaching a maximum (recorded) deformation which ranged between 0.9 and 1.2 mm, 0.5 and 1.1 mm and 0.6–1.1 mm for Series A, B and C, respectively. It should be noted that the LVDTs were removed just before reaching the failure load in order to prevent any damage. The first cracks appeared at a load level of about 50–60% of the failure load. A large majority of cracks run vertically. The cracks were visible or (for those that develop within the element) audible. Typical crack patterns are shown in Fig. 11. For specimens with filled horizontal and inclined chases some cracks which developed orthogonally to the chase were observed, see Fig. 11. These cracks could be ascribed to different mechanical characteristics (Poisson’s ratio) of chase infill and block unit.
4.2. Deformations Strains and curvatures were calculated from the displacements of the measurement net. Firstly, the changes in distance between the net nodes were measured and, in order to correct measurement errors, the analogy between the principles of the minimum of the sum of squared errors and the minimum of elastic strain energy was used [12]. For this purpose, using standard FEM software the measurement net was modelled as a (statically indeterminate) plane truss and the truss member’s length changes were applied as loading. In this way nodal displacements could be obtained. Fig. 12 shows displacement fields for both the north and south faces of specimen BS2 for three load stages LS2, LS3 and LS4 with corresponding loads of 160 kN, 320 kN and 430 kN, respectively. The drawing on the right shows the south face (with chase). In addition, Fig. 12 presents a large crack opening (3 mm), which can also be detected in the deformation field.
Secondly, assuming a linear variation of the displacements over the area of the triangle, the constant average strains in each triangle were calculated from the nodal displacements [12]. Finally, the corresponding curvatures were calculated from the calculated strains on both specimens’ surfaces. Fig. 13 shows the Mohr’s circles of the calculated strains and curvatures for the test specimen AH1 for four load stages LS2, LS3, LS4 and LS5 with the corresponding loads of 180 kN, 270 kN, 350 kN and 500 kN, respectively. As can be seen, the strains were larger on the surface with a chase (south), i.e. the presence of the chase induced (small) curvatures, cf. also the Mohr’s circle of curvatures in Fig. 13c. 4.3. Behaviour at ultimate load The failure patterns observed during the tests were characteristic for masonry failing in compression. Subjected to vertical pressure, the specimens failed by exceeding the tensile strength in the directions orthogonal to the applied pressure. For the specimens of Series A failure planes developed perpendicular to the plane of the wall (x–y plane), thus dividing the specimen into several pillars, see, e.g. the failure pattern for the specimens AH2 and AV1 shown in Fig. 14. This figure also exhibits a failure pattern of the specimen AS1 that was typical of the failure of the specimen with an inclined chase. For those specimens, the damage at the transition between units and infill was visible and the relative movement (sliding) of the upper part of the element along the chase could be observed. For all specimens, with progressive crushing of the infill the influence of bending could be observed. Series B specimens showed somewhat different failure patterns. Specimens with horizontal chases, BH1 and BH2, exhibited behaviour similar to that of Series A, cf. also Fig. 15. On the other hand, for the specimens with vertical and inclined chases, the failure
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
a
b
4421
c
Fig. 13. Mohr’s circles (Specimen AH1): (a) strains (north face); (b) strains (south face); (c) curvatures.
Fig. 14. Failure patterns for Series A: specimens AH2, AV1 and AS1 (south faces).
Fig. 15. Failure patterns for Series B: specimens BH2, BV1 (north faces) and BS2 (east side).
plane developed mainly parallel to the plane of the wall, thus dividing the specimen into two parts, see, e.g. the failure pattern for specimens BV1 and BS2 shown in Fig. 15. Since the chase depth was somewhat smaller than for Series A the influence of bending was not so pronounced.
Series C specimens’ boundary conditions at the bottom of the masonry element differed from those of the other two series. As previously described, specimens of this series were placed on the lower spreader beam without neoprene plates and thus their lateral deformation was more restrained than in the previous
4422
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
Fig. 16. Failure patterns for Series C: specimens CH0, CB0 (east sides), CH3 (south face) and CS3 (west side).
two series. All elements, except for CV0 and CH3, failed in a similar way to the majority of Series B specimens: the failure plane developed mainly parallel to the plane of the wall, thus dividing the specimen into two parts, see, e.g. the failure pattern for specimens CH0, CB0 and CS3 shown in Fig. 16. This figure also shows the failure pattern for specimen CH3 that failed like those of Series A. This specimen developed mostly vertical cracks and in the vicinity of the chase (on the south face) the spalling of the block’s outer shells was observed. A failure of all specimens of all series was brittle and the failure load for control specimens A, B and C (especially for specimen B) was lower than expected considering the mechanical characteristics of the components, the unit and mortar and according to code provisions [4,7]. However, the present paper discusses only the relative change in the strength and not the absolute values, cf. also Table 5. 5. Discussion From Table 5 it is obvious that cutting chases in unreinforced masonry elements reduces the masonry compressive strength fx. For Series A specimens, in which the depth of chase was one-third of the element thickness tw, the reduction ranged between 18.8% and 35.1%. Generally, the reduction of fx was approximately equivalent to the reduction of the cross-sectional area. For the elements of Series B, in which the chase depth was two-sevenths of tw, a somewhat different behaviour was observed. For these specimens, with the exception of the elements with inclined chases, BS1 and BS2, a smaller reduction of fx was observed and two specimens, BH2 and BV1, were even stronger than the control specimen, B. This has been investigated further, and it seems that some damage to the control specimen during transport from production site to the test site could be responsible for the unexpected low strength of control specimen B. A different failure pattern for thicker elements (see Fig. 15) will also have to be considered. For specimens of Series C, in which the chase depth was again one-third of the element thickness tw, the strength reduction ranged between 9.4% and 44.1%. However, specimens that were repaired with high-strength mending mortar exhibited higher compressive strength than the control specimen without chases, C. This strength increase was 17.9% and 28.9% for specimens CH3 and CS3, respectively. 5.1. Influence of chase orientation and depth The overview of the change in compressive strength depending on chase orientation is shown in Fig. 17. The largest decrease in strength was observed for specimens with inclined chases. This decrease ranged from 18.9% to 35.1%. This fact should be taken into account when chases are cut in load-bearing walls designed to
transfer combined vertical and shear forces, i.e. an inclined resultant force. As a rule, the cutting of any kind of chases in such walls should not be permitted. As observed during testing, only when those chases are filled (after placing the ducts) with high-strength mending mortar a compressive strength can be preserved and even exceeded. However, it must be noted that in all tests with inclined chases a relative movement (slip) of the upper part of the specimen over the lower part was observed. A considerable decrease in strength has been also observed in tests on specimens with horizontal chases cut on one element face only, except for Series B. Excluding the specimens BH1 and BH2, the reduction ranged between 13.8 % and 25.0%. As expected, the largest reduction has been obtained for specimen CB0, which had chases on both element faces, and amounted to 44.1%. A somewhat unexpected decrease of compressive strength of specimens with vertical chases has been observed, especially in Series A, where this reduction was 18.8% and 21.5% for specimens AV2 and AV1, respectively, see also Fig. 17. A moderate decrease of 9.4% was observed for specimen CV0 without chase infill. Cutting the vertical chase demolished the strongest part of the unit (outer shell) and introduced a new plane of weakness, thus reducing the strength considerably. At failure, the specimens with a vertical chase were divided into two parts along the chase. In general, a larger ratio between the chase depth and element thickness, ts/tw, induced larger decrease of compressive strength. As previously mentioned, the chases cut into the elements reduced their gross cross-sectional area for 33% for Series A and C and 29% for Series B. This reduction for specimen CB0 (see Fig. 7), which had chases on both element faces, amounted to 60% because the core of the block was left undamaged, cf. also block shape shown in Table 3. Considering these values and the strength values given in Table 5 it can be concluded that the reduction of masonry compressive strength can be directly related to the reduction of the crosssectional area. Such a simple criterion lies on the safe side and can be used as a reasonable approximation when estimating the influence of chase depth on compressive strength.
5.2. Influence of infill type The overview of the change in compressive strength depending on chase infill is shown in Fig. 18. As could be expected, the lowest strength was observed in elements with chases without infill. The difference in strength between elements with typical infill, i.e. gypsum and cement mortar, is smaller than one would expect, especially for the results of Series A. It should be noted here that these two infill types are the most widely applied in practice. From the results of the present investigation it is clear that an infill that is at least as strong and stiff as the masonry unit should be applied in order to avoid the loss of strength of elements with chases. In the present case the strengths of high-strength mending mortar
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
4423
Fig. 17. Compressive strength change depending on chase orientation.
Fig. 18. Compressive strength change depending on chase infill.
(applied for the infill of Series C specimens CH3 and CS3) and the clay block unit of the same series were 57.1 MPa and 42.0 MPa, respectively. Both materials had, according to the respective product sheets, a similar modulus of elasticity of about 30 GPa. The application of such infill resulted in compressive strengths that were higher than those of the control specimen and thus can be recommended for practical application. 5.3. Truss model In order to perform an analytical analysis of the influence of the chases on the behaviour of masonry elements a truss model is applied, cf. Mojsilovic´ [13] and Mojsilovic´ and Marti [14]. Such a
model is able to give a deeper insight into the force flow in the masonry and allows for a better description of the structural behaviour. However, there is no unique truss model for a particular situation and engineering judgment is needed in each case. One possible solution for the transfer of the compressive loading through the part of the specimen with a (horizontal) chase is presented in Fig. 19. This solution is based on the elastic distribution of the stresses in specimen with one horizontal chase, which was obtained by applying a standard FEM procedure using 2D-quadrilateral elements. As can be seen the compressive forces diverge in the vicinity of the chase, i.e. a geometrical discontinuity. The inclination of the compressed truss member D is given by the angle h. In order to allow for this deviation, i.e. to maintain equilibrium
4424
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
Fig. 19. Truss model (note: all dimensions in mm).
the horizontal component of the diagonal compression D must be compensated by a tensile force T. Assuming that the total compressive force equals 3F and that one-third of this total compressive force (F) is transferred directly from the upper to
the lower boundary of the masonry element shown in Fig. 19, from equilibrium conditions in truss joints one obtains D = F/cos h and C = T = F tan h. From the truss model and considering failure patterns observed in tests it is clear that failure is governed by exceeding the tensile capacity of the masonry. It should be noted that the angle of inclination of the inclined truss member, h, i.e. the height of the diverging zone, h, is actually a free parameter of the chosen truss model. The variation of the tensile force T (represented by T/F = tan h), depending on the angle of the inclination, is given in Fig. 20. The shaded area represents the recommended range for h, whereas the elastic FEM solution delivers an angle of deviation of the principle compressive stresses from the vertical of ca. 36°. The next step in the analysis is determining the governing failure tensile stress rT that results from the assumed truss model, i.e. tensile force T. Since the truss model represents the resultants of the corresponding stress fields, the width of the corresponding tension field, b (see also Fig. 19), has to be determined. This width is yet another free parameter in the model. A reasonable value seems to be b = 1.5 h, thus delivering, in this particular case (cf. Fig. 19), a resulting tensile stress of rT = T/b = F tan2 h/(1.5 37.5). The variation of this stress, depending on the angle of the inclination, is given in Fig. 21. This figure also compares calculated ultimate tensile stresses for specimens CH0, CS0 and CB0 to the (net) tensile strength of the clay block of the same series, fbtz, cf. also Table 3. It can be seen that by increasing value of h the difference between rT and fbtz becomes larger. From this comparison it is obvious that the use of a clay block tensile strength in practical design would lead to an underestimation of the wallette strength, when considering inclinations of 35–40°. However, it should be kept in mind that the tensile resistance of tested elements is actually higher than that determined on units due to some restraint of lateral deformation in the masonry composite. Furthermore, the large scatter of the test results when determining the tensile strength should also be taken into consideration. It should be noted here that the proposed, simple truss model has been applied to elements with different orientations (horizontal and inclined) of the chase. The same truss model could also be applied to the specimens of Series A and B. This has not been done here, since no tensile tests for the units of these series were performed. However, it could be expected that similar results would be obtained. Failure loads in Series A and B were smaller than those in Series C, but the block compressive strength was also smaller; see also Tables 3 and 5. Thus, assuming that the compressive and tensile strengths of the block are directly related, similar results can be expected.
Fig. 20. Variation of the truss model tensile force T.
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
4425
Fig. 21. Variation of tensile stress rT for specimens CH0, CS0 and CB0.
6. Conclusions and outlook An investigation of the influence of chases in masonry elements subjected to compression is underway at ETH Zurich. Results of the tests on three series of elements (total of 21 specimens) were analysed and allow a number of conclusions to be drawn: Chases can reduce the masonry compressive strength considerably. In general, a reduction of the masonry strength proportional to the reduction of the cross-section area seems to be a reasonable approach in analysing the influence of chases on masonry compressive strength. The choice of the typically used material of chase infill (gypsum and cement mortar) has relatively small influence on the reduction of masonry compressive strength. It seems that an infill that is as strong and stiff as the masonry unit has to be applied in order to eliminate the loss in masonry strength. From the present investigation, the high-strength mending mortar could be recommended as suitable chase infill for practical application. In order to get a deeper insight into the structural behaviour of masonry elements with chases an analytical investigation based on a simple truss model has been performed. It was shown that such a model can predict the masonry strength of elements with horizontal and inclined chases. As a governing material parameter the tensile strength of the clay block unit was selected. Major issues deserving further research include tests on similar specimens with additional thicknesses (125 and 200 mm) as well as with different masonry units (calcium-silicate and concrete blocks) which could be of use to confirm/modify or extend the present findings. Further, to extend the presented research and to gain additional insight into the problem further testing on slender walls (second order effects, i.e. stability failure) and a subsequent theoretical investigation would be needed. Finally, the
numerical modelling, e.g. application of micro-modelling, deserves more study. Acknowledgements The clay bricks were supplied free of charge by Ziegelei Schumacher, Körblingen, and Keller Ziegeleien, Pfungen, as was the mortar by Maxit AG, Dätwill, and the high-strength mending mortar by Sika Schweiz AG, Baar. This support is gratefully acknowledged as well as that of graduate students S. Notz, B. Wäfler, J. Yu and D. Geisseler who were involved in performing the tests. References [1] Fischer K. The effects of chasing on the compressive strength of brickwork. In: Proc 3rd int brick/block masonry conf. Essen, Germany; 1973. p. 106–14. [2] Kirtschig K, Metje WR. Influence of chases and recesses on the strength of masonry. Proc Br Masonry Soc 1988;2:61–3. [3] Sahlin S. Fischer’s tests on chasing revisited the effects of chasing on the compressive strength of brickwork (3. IBMAC, Essen 1973). Masonry Int 2007;20(2):85–90. [4] SIA 266:2003. Structural masonry. Zurich: Swiss Society of Engineers and Architects; 2003. [5] ACI 530-08/ASCE 5-08/TMS 402-08. Building code requirements and specifications for masonry structures. Masonry Standards Joint Committee (MSJC), Boulder/Farmington Hills/Reston, USA; 2008. [6] AS3700:2001. Masonry structures. Sydney: Standards Australia; 2001. [7] EN 1996-1-1:2005. Design of masonry structures – part 1: general rules for reinforced and unreinforced masonry. Brussels: European Committee for Standardization; 2005. [8] CSA standard S304.1-04. Design of masonry structures. Canadian Standards Association, Mississauga, Ontario, Canada; 2004. [9] Mojsilovic´ N, Notz S, Waefler B, Yu J. Load tests masonry elements with chases. In: Proc 11th canadian masonry sympos, Toronto, Ontario; 2009. p. 379–88. [10] EN 772-1:2000. Methods of test for masonry units – part 1: determination of compressive strength. Brussels: European Committee for Standardization; 2000. [11] EN 1015-11:1999. Methods of test for mortar for masonry – part 11: determination of flexural and compressive strength of hardened mortar. Brussels: European Committee for Standardization; 1999. [12] Mojsilovic´ N, Marti P. Tests on masonry walls subjected to combined actions (in German). Report no. 203. Zurich: Institute of Structural Engineering; 1994. [13] Mojsilovic´ N. Masonry subjected to combined actions (in German). Report no. 216. Zurich: Institute of Structural Engineering; 1995. [14] Mojsilovic´ N, Marti P. Strength of masonry subjected to combined actions. ACI Struct J 1997;94(6):633–42.
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Masonry elements with chases: Behaviour under compression Nebojša Mojsilovic´ ⇑ Institute of Structural Engineering, Department of Structural, Environmental and Geomatic Engineering, ETH Zurich 8093, Switzerland
a r t i c l e
i n f o
Article history: Available online 12 January 2011 Keywords: Chases Clay blocks Load test Masonry compressive strength Unreinforced masonry
a b s t r a c t Accommodating electrical wiring and similar installations in masonry construction often necessitates the cutting of chases in masonry. In order to investigate the influence of chases on masonry behaviour compression load tests were performed on three series of unreinforced masonry elements with chases. Three different types of chase (horizontal, vertical and inclined) were cut into the masonry elements. After placing the ducts several different types of infill were applied to fill the chases. The compressive load was applied in a deformation-controlled manner up to the failure of the specimen. A considerable reduction in the masonry compressive strength was observed. The test results are discussed in detail and recommendations for practical application are given. In addition, a simple analytical truss model for predictions of failure load is proposed and discussed. Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction Accommodating electrical wiring and similar installations in masonry construction often necessitates the cutting of chases in masonry. These chases can considerably reduce the masonry compressive strength, which is the key parameter in the design of structural masonry. In general, the chases are cut vertically or horizontally, but as can be seen in many practical examples the chases in the walls can be cut quite arbitrarily, see Fig. 1. Moreover, the chases are usually not properly filled after placing the ducts and the material used for infill is often of unsatisfactory quality, especially in regard to its strength and stiffness. Previous research work in this area included tests on typical masonry elements with vertical and horizontal chases. However, the amount of the published work is rather modest. Fischer [1] tested 2.55 m high clay block masonry walls of three different thicknesses (102.5, 178 and 215 mm) and several combinations of mainly vertical 38 mm wide and 25 mm deep chases. A reduction in the masonry compressive strength of up to 28.5%, compared to the strength for walls without chases, was observed. Kirtschig and Metje [2] investigated the influence of chases and recesses on masonry compressive strength. They performed numerous tests on clay block masonry with vertical and horizontal chases. It was found, that in general, the reduction in masonry strength tends to be proportional to the reduction in cross-sectional area. As result of this experimental work, a table with dimensions of chases that are allowed without additional verification of the load-bearing
⇑ Tel.: +41 446333763; fax: +41 446331064. E-mail address: [email protected] 0950-0618/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.conbuildmat.2010.12.027
capacity has been proposed. Sahlin [3] discussed the tests results from [1] and proposed a damage index (ratio of cumulative chase length and wall thickness) as a measure for damage, i.e. strength reduction induced by chases. He related this index to the observed damage in tests and found a good correlation with test data. In addition, provisions for chases in masonry can be found in structural codes. Some codes, e.g. the Swiss Masonry Code SIA 266 [4], the Building Code Requirements and Specifications for Masonry Structures ACI 530-08/ASCE 5-08/TMS 402-08 [5] and the Australian Structural Masonry Code [6], only give general recommendations, whereas others, such as the European Masonry Code ENV 1996-1-1 [7], provide tables with values of depth and width of chases that are allowed without additional verification of the load-bearing capacity. Analogous provisions can be found in the Canadian Structural Masonry Code [8]. A research project is underway at the Institute of Structural Engineering of the ETH Zurich to investigate the structural behaviour under compression loading of typical Swiss unreinforced clay block masonry elements with chases. The main goal of the research project is to investigate the influence of chases on masonry compressive strength. In the first phase of the project load tests on the first two series of squat masonry elements with chases were completed in 2008, see Mojsilovic´ et al. [9]. Recently, a third series, which complemented previous two, has been performed. The present paper gives an overview of all the experimental work, discusses the results obtained, introduces a simple truss model for analytical analysis and provides recommendations for practical implementation. It should be noted here that the present investigation concentrates on squat masonry elements subjected to centric compression and does not investigate influence of second order effects.
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
4416
Fig. 1. Arbitrarily cut chases in masonry walls.
2. Test programme and masonry materials Load tests on the three series of masonry elements with chases have been completed. Each series consisted of seven unreinforced clay block masonry elements. The test specimen dimensions were 1200 (length bw) 1200 (height hw) mm. The thickness of the elements, tw, was 150 mm (Series A and C) and 175 mm (Series B). The test specimens were produced and the chases cut by a professional mason. Dry, factory-made standard cement mortar was mixed with water at the construction site to build the wall elements in running bond. Both the bed and the head joints were 10 mm thick and fully filled. Twenty-eight days after preparation 45–50 mm deep (ts) and 35 mm wide (bs) chases were cut in six elements of each series while the seventh one was left intact and served as a
a
control specimen, see Fig. 2. For the specimens in Series A and C the chase depth was 1/3 of tw and for those of Series B it was 2/7 of tw. Three different types of chase (horizontal, vertical and inclined), two each per series, were cut only on one side of the masonry elements except for one specimen of Series C which had horizontal chases on the both element sides. The slope of the inclined chases was tan a = 2/3. An overview of the test programme is given in Table 1 for Series A and B and in Table 2 for Series C. These tables also show the types of chase infill applied: gypsum, cement mortar and high-strength mending mortar. Note that some of the chases were left unfilled. Table 3 gives information on the properties of the different types of (hollow) clay block units that were used in the tests. fb
b
bs
hw
bw
ts
tw
c
d
α
bs bs
ts
ts
Fig. 2. Test specimens: (a) without chases; (b) horizontal chase; (c) vertical chase; (d) inclined chase.
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
4417
Table 1 Specimen designation for test programme for Series A and B. Wall thickness (mm)
Without chases
150 175
A B
Horizontal chase
Vertical chase
Inclined chase
Gypsum infill
Mortar infill
Gypsum infill
Mortar infill
Gypsum infill
Mortar infill
AH1 BH1
AH2 BH2
AV1 BV1
AV2 BV2
AS1 BS1
AS2 BS2
Table 2 Specimen designation for test programme for Series C. Wall thickness (mm)
Without chases
Horizontal chase Without infill
150 *
C
CH0
CB0*
Vertical chase
Inclined chase
Mending mortar infill
Without infill
Without infill
Mending mortar infill
CH3
CV0
CS0
CS3
Note: Test specimen CB0 had horizontal chases on both specimen faces.
Table 3 Properties of clay block units. Series
Unit
Dimensions (mm)
Void area (%)
fb (MPa)
A
B 15/19
Shape
290x190x150
45
32.1 (0.7)
B
B 17.5/19
290x190x175
46
28.6 (1.3)
C
B 15/19
290x190x150
46
42.0 (1.0)
fbtz (MPa)
1.4 (0.4)
Note: Value given in parenthesis next to the mean strength value is the corresponding standard deviation.
Table 4 Properties of mortar and chase infill. Mortar/infill type
Joint mortar
Joint mortar
Series Curing Number of specimens (ffm/fm) ffm (MPa) fm (MPa)
A, B Code 6/12 4.5 (4.1/4.8) 17.1 (16.4/17.8)
C Code 3/6 4.1 (3.8/4.2) 17.2 (16.7/17.8)
Gypsum infill Mortar infill A, B Air Code 3/6 3/6 2.8 (2.6/3.0) 2.2 (2.0/2.4) 10.0 (8.7/10.9) 5.9 (5.7/6.0)
A, B Code 3/6 4.7 (4.6/4.9) 17.2 (16.7/17.5)
Mending mortar infill C Air Code 3/6 3/6 6.2 (6.1/6.2) 6.6 (6.6/6.8) 19.9 (19.6/20.4) 57.1 (53.1/60.6)
Air 3/6 7.2 (6.8/7.5) 48.8 (44.3/51.5)
Note: Values given in parenthesis next to the mean strength value are the recorded minimum and maximum values.
denotes the unit’s compressive strength, fbtz its (net) tensile strength in the z-direction. Note that x and y are, respectively, directions perpendicular and parallel to the bed joints. The compressive strength was determined according to the European code EN 772-1 [10]. The tensile strength was determined only for Series C units using a specially developed set-up and procedure. The average compressive, fm, and flexural, ffm, strengths of the cement mortar used for the joints and for the three mortar types used for chase infill are given in Table 4. The mentioned strengths were determined according to the European code EN 1015-11 [11]. Note that two different curing procedures were applied (cf. Table 4): some mortar prisms were cured according to the code provisions [11] and some were air-cured in the laboratory, in the vicinity of the test specimens.
Fig. 3 shows typically the preparation phases for a specimen with a horizontal chase. First, a masonry saw was used to cut the chase. Afterwards, the inner-rib leftovers of the hollow clay brick were removed by means of a chisel. Finally, the installation duct was placed and the chase was filled with corresponding infill or left unfilled. Fig. 4 shows typically three specimens, CH3, AV1 and AS2 with chases filled with high-strength mending mortar, gypsum and cement mortar, respectively. Note that chases were cut on the south face of the specimens (cf. Fig. 5). After a further two weeks the specimens of Series A and B were transported to the test site and tested, whereas the specimens of Series C were produced on test site. It should be noted here that the cutting of the chases was carried out on unloaded specimens, whereas in practical applications some pre-compression due to dead load is present.
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
4418
Fig. 3. Preparation of the specimen with horizontal chase.
Fig. 4. Chases filled with mending mortar (CH3), gypsum (AV1) and cement mortar (AS2).
3. Test procedure and measurements Fig. 5 depicts the test set-up. The axial load was applied by means of three hydraulic jacks (2) placed between the support frame (1) and the upper spreader beam (3). The test specimen
(4) was placed between two spreader beams (3) and two sets of steel plates (6) that provided good contact with the specimen. In this way an unrestrained lateral deformation of the specimen was ensured. To achieve the exact position of the steel plates, two thin plaster layers (7) were applied to both the upper and the lower edges of the specimen. Additionally, except for Series C, a set of neoprene plates (5) was placed between the steel plates and the lower spreader beam, which was placed directly on the laboratory’s strong floor (8). These neoprene plates helped to achieve uniform load distribution over the specimen. Both spreader beams had a thin Teflon layer on the faces towards the test specimen. After setting-up, the specimens were subjected to an axial load, which was increased in several load steps up to failure of the test specimen. Apart from the applied load, the measurements included node displacements of the measurement nets on both specimen surfaces and the deflection at the middle of the specimen. The measurement net consisted of triangles whose nodes were aluminium bolts glued in the middle of the clay brick unit’s face and had two base lengths, sh = 300 mm and sd = 250 mm, see Fig. 6. This figure also
sd sh Fig. 5. Test set-up.
Fig. 6. Measurement net (south face) and its position relative to the chases.
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
Fig. 7. Measurement net and LVDTs for specimens CH0 (south face) and CB0 (north face).
shows the net’s position relative to the chases. Additionally, vertical shortening of the specimen was measured by means of four LVDTs with a nominal base length of 1000 mm; two on each surface (see Fig. 7). During the nodal displacements measurement, the deformation was kept constant. All measuring devices were connected to a personal computer, which processed the data in real time. In order to follow crack development and to measure crack widths, the specimens were painted white prior to testing. 4. Test results and structural behaviour Table 5 shows the values of the masonry compressive strength fx (perpendicular to the bed joints) obtained from all test series. Fmax denotes the maximum recorded compressive force, Aw is the gross cross-sectional area and D is the percentile change of the compressive strength (a negative value signifying a reduction) compared to the control specimen without chases (A, B or C).
Fig. 8. Force–deformation relationships for Series A.
4.1. Load–deformation characteristic Figs. 8–10 present the force–deformation characteristics for Series A, B and C, respectively. The deformation value shown in the diagram is an average of all four LVDT measurements (vertical shortening of specimen). As Figs. 8–10 reveal, the specimens
Table 5 Test results. Specimen
Fmax (kN)
Aw (m2)
fx (MPa)
D (%)
A AH1 AH2 AV1 AV2 AS1 AS2
849.10 666.10 637.20 616.60 689.60 666.60 551.40
0.18 0.18 0.18 0.18 0.18 0.18 0.18
4.72 3.70 3.54 3.43 3.83 3.70 3.06
–
B BH1 BH2 BV1 BV2 BS1 BS2
670.30 650.20 763.20 745.80 553.20 543.70 471.60
0.21 0.21 0.21 0.21 0.21 0.21 0.21
3.19 3.10 3.63 3.55 2.63 2.59 2.25
–
C CH0 CB0 CH3 CV0 CS0 CS3
1158.00 997.79 647.02 1365.03 1049.58 828.65 1493.24
0.18 0.18 0.18 0.18 0.18 0.18 0.18
6.43 5.54 3.59 7.58 5.81 4.60 8.30
–
21.6 25.0 27.4 18.8 21.5 35.1
Fig. 9. Force–deformation relationships for Series B.
3.0 13.9 11.3 17.5 18.9 29.6 13.8 44.1 17.9 9.4 28.4 28.9
Fig. 10. Force–deformation relationships for Series C.
4419
4420
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
Fig. 11. Typical crack patterns.
Fig. 12. Deformation field and crack detail (Specimen BS2).
behaved more or less linear-elastically almost up to failure reaching a maximum (recorded) deformation which ranged between 0.9 and 1.2 mm, 0.5 and 1.1 mm and 0.6–1.1 mm for Series A, B and C, respectively. It should be noted that the LVDTs were removed just before reaching the failure load in order to prevent any damage. The first cracks appeared at a load level of about 50–60% of the failure load. A large majority of cracks run vertically. The cracks were visible or (for those that develop within the element) audible. Typical crack patterns are shown in Fig. 11. For specimens with filled horizontal and inclined chases some cracks which developed orthogonally to the chase were observed, see Fig. 11. These cracks could be ascribed to different mechanical characteristics (Poisson’s ratio) of chase infill and block unit.
4.2. Deformations Strains and curvatures were calculated from the displacements of the measurement net. Firstly, the changes in distance between the net nodes were measured and, in order to correct measurement errors, the analogy between the principles of the minimum of the sum of squared errors and the minimum of elastic strain energy was used [12]. For this purpose, using standard FEM software the measurement net was modelled as a (statically indeterminate) plane truss and the truss member’s length changes were applied as loading. In this way nodal displacements could be obtained. Fig. 12 shows displacement fields for both the north and south faces of specimen BS2 for three load stages LS2, LS3 and LS4 with corresponding loads of 160 kN, 320 kN and 430 kN, respectively. The drawing on the right shows the south face (with chase). In addition, Fig. 12 presents a large crack opening (3 mm), which can also be detected in the deformation field.
Secondly, assuming a linear variation of the displacements over the area of the triangle, the constant average strains in each triangle were calculated from the nodal displacements [12]. Finally, the corresponding curvatures were calculated from the calculated strains on both specimens’ surfaces. Fig. 13 shows the Mohr’s circles of the calculated strains and curvatures for the test specimen AH1 for four load stages LS2, LS3, LS4 and LS5 with the corresponding loads of 180 kN, 270 kN, 350 kN and 500 kN, respectively. As can be seen, the strains were larger on the surface with a chase (south), i.e. the presence of the chase induced (small) curvatures, cf. also the Mohr’s circle of curvatures in Fig. 13c. 4.3. Behaviour at ultimate load The failure patterns observed during the tests were characteristic for masonry failing in compression. Subjected to vertical pressure, the specimens failed by exceeding the tensile strength in the directions orthogonal to the applied pressure. For the specimens of Series A failure planes developed perpendicular to the plane of the wall (x–y plane), thus dividing the specimen into several pillars, see, e.g. the failure pattern for the specimens AH2 and AV1 shown in Fig. 14. This figure also exhibits a failure pattern of the specimen AS1 that was typical of the failure of the specimen with an inclined chase. For those specimens, the damage at the transition between units and infill was visible and the relative movement (sliding) of the upper part of the element along the chase could be observed. For all specimens, with progressive crushing of the infill the influence of bending could be observed. Series B specimens showed somewhat different failure patterns. Specimens with horizontal chases, BH1 and BH2, exhibited behaviour similar to that of Series A, cf. also Fig. 15. On the other hand, for the specimens with vertical and inclined chases, the failure
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
a
b
4421
c
Fig. 13. Mohr’s circles (Specimen AH1): (a) strains (north face); (b) strains (south face); (c) curvatures.
Fig. 14. Failure patterns for Series A: specimens AH2, AV1 and AS1 (south faces).
Fig. 15. Failure patterns for Series B: specimens BH2, BV1 (north faces) and BS2 (east side).
plane developed mainly parallel to the plane of the wall, thus dividing the specimen into two parts, see, e.g. the failure pattern for specimens BV1 and BS2 shown in Fig. 15. Since the chase depth was somewhat smaller than for Series A the influence of bending was not so pronounced.
Series C specimens’ boundary conditions at the bottom of the masonry element differed from those of the other two series. As previously described, specimens of this series were placed on the lower spreader beam without neoprene plates and thus their lateral deformation was more restrained than in the previous
4422
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
Fig. 16. Failure patterns for Series C: specimens CH0, CB0 (east sides), CH3 (south face) and CS3 (west side).
two series. All elements, except for CV0 and CH3, failed in a similar way to the majority of Series B specimens: the failure plane developed mainly parallel to the plane of the wall, thus dividing the specimen into two parts, see, e.g. the failure pattern for specimens CH0, CB0 and CS3 shown in Fig. 16. This figure also shows the failure pattern for specimen CH3 that failed like those of Series A. This specimen developed mostly vertical cracks and in the vicinity of the chase (on the south face) the spalling of the block’s outer shells was observed. A failure of all specimens of all series was brittle and the failure load for control specimens A, B and C (especially for specimen B) was lower than expected considering the mechanical characteristics of the components, the unit and mortar and according to code provisions [4,7]. However, the present paper discusses only the relative change in the strength and not the absolute values, cf. also Table 5. 5. Discussion From Table 5 it is obvious that cutting chases in unreinforced masonry elements reduces the masonry compressive strength fx. For Series A specimens, in which the depth of chase was one-third of the element thickness tw, the reduction ranged between 18.8% and 35.1%. Generally, the reduction of fx was approximately equivalent to the reduction of the cross-sectional area. For the elements of Series B, in which the chase depth was two-sevenths of tw, a somewhat different behaviour was observed. For these specimens, with the exception of the elements with inclined chases, BS1 and BS2, a smaller reduction of fx was observed and two specimens, BH2 and BV1, were even stronger than the control specimen, B. This has been investigated further, and it seems that some damage to the control specimen during transport from production site to the test site could be responsible for the unexpected low strength of control specimen B. A different failure pattern for thicker elements (see Fig. 15) will also have to be considered. For specimens of Series C, in which the chase depth was again one-third of the element thickness tw, the strength reduction ranged between 9.4% and 44.1%. However, specimens that were repaired with high-strength mending mortar exhibited higher compressive strength than the control specimen without chases, C. This strength increase was 17.9% and 28.9% for specimens CH3 and CS3, respectively. 5.1. Influence of chase orientation and depth The overview of the change in compressive strength depending on chase orientation is shown in Fig. 17. The largest decrease in strength was observed for specimens with inclined chases. This decrease ranged from 18.9% to 35.1%. This fact should be taken into account when chases are cut in load-bearing walls designed to
transfer combined vertical and shear forces, i.e. an inclined resultant force. As a rule, the cutting of any kind of chases in such walls should not be permitted. As observed during testing, only when those chases are filled (after placing the ducts) with high-strength mending mortar a compressive strength can be preserved and even exceeded. However, it must be noted that in all tests with inclined chases a relative movement (slip) of the upper part of the specimen over the lower part was observed. A considerable decrease in strength has been also observed in tests on specimens with horizontal chases cut on one element face only, except for Series B. Excluding the specimens BH1 and BH2, the reduction ranged between 13.8 % and 25.0%. As expected, the largest reduction has been obtained for specimen CB0, which had chases on both element faces, and amounted to 44.1%. A somewhat unexpected decrease of compressive strength of specimens with vertical chases has been observed, especially in Series A, where this reduction was 18.8% and 21.5% for specimens AV2 and AV1, respectively, see also Fig. 17. A moderate decrease of 9.4% was observed for specimen CV0 without chase infill. Cutting the vertical chase demolished the strongest part of the unit (outer shell) and introduced a new plane of weakness, thus reducing the strength considerably. At failure, the specimens with a vertical chase were divided into two parts along the chase. In general, a larger ratio between the chase depth and element thickness, ts/tw, induced larger decrease of compressive strength. As previously mentioned, the chases cut into the elements reduced their gross cross-sectional area for 33% for Series A and C and 29% for Series B. This reduction for specimen CB0 (see Fig. 7), which had chases on both element faces, amounted to 60% because the core of the block was left undamaged, cf. also block shape shown in Table 3. Considering these values and the strength values given in Table 5 it can be concluded that the reduction of masonry compressive strength can be directly related to the reduction of the crosssectional area. Such a simple criterion lies on the safe side and can be used as a reasonable approximation when estimating the influence of chase depth on compressive strength.
5.2. Influence of infill type The overview of the change in compressive strength depending on chase infill is shown in Fig. 18. As could be expected, the lowest strength was observed in elements with chases without infill. The difference in strength between elements with typical infill, i.e. gypsum and cement mortar, is smaller than one would expect, especially for the results of Series A. It should be noted here that these two infill types are the most widely applied in practice. From the results of the present investigation it is clear that an infill that is at least as strong and stiff as the masonry unit should be applied in order to avoid the loss of strength of elements with chases. In the present case the strengths of high-strength mending mortar
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
4423
Fig. 17. Compressive strength change depending on chase orientation.
Fig. 18. Compressive strength change depending on chase infill.
(applied for the infill of Series C specimens CH3 and CS3) and the clay block unit of the same series were 57.1 MPa and 42.0 MPa, respectively. Both materials had, according to the respective product sheets, a similar modulus of elasticity of about 30 GPa. The application of such infill resulted in compressive strengths that were higher than those of the control specimen and thus can be recommended for practical application. 5.3. Truss model In order to perform an analytical analysis of the influence of the chases on the behaviour of masonry elements a truss model is applied, cf. Mojsilovic´ [13] and Mojsilovic´ and Marti [14]. Such a
model is able to give a deeper insight into the force flow in the masonry and allows for a better description of the structural behaviour. However, there is no unique truss model for a particular situation and engineering judgment is needed in each case. One possible solution for the transfer of the compressive loading through the part of the specimen with a (horizontal) chase is presented in Fig. 19. This solution is based on the elastic distribution of the stresses in specimen with one horizontal chase, which was obtained by applying a standard FEM procedure using 2D-quadrilateral elements. As can be seen the compressive forces diverge in the vicinity of the chase, i.e. a geometrical discontinuity. The inclination of the compressed truss member D is given by the angle h. In order to allow for this deviation, i.e. to maintain equilibrium
4424
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
Fig. 19. Truss model (note: all dimensions in mm).
the horizontal component of the diagonal compression D must be compensated by a tensile force T. Assuming that the total compressive force equals 3F and that one-third of this total compressive force (F) is transferred directly from the upper to
the lower boundary of the masonry element shown in Fig. 19, from equilibrium conditions in truss joints one obtains D = F/cos h and C = T = F tan h. From the truss model and considering failure patterns observed in tests it is clear that failure is governed by exceeding the tensile capacity of the masonry. It should be noted that the angle of inclination of the inclined truss member, h, i.e. the height of the diverging zone, h, is actually a free parameter of the chosen truss model. The variation of the tensile force T (represented by T/F = tan h), depending on the angle of the inclination, is given in Fig. 20. The shaded area represents the recommended range for h, whereas the elastic FEM solution delivers an angle of deviation of the principle compressive stresses from the vertical of ca. 36°. The next step in the analysis is determining the governing failure tensile stress rT that results from the assumed truss model, i.e. tensile force T. Since the truss model represents the resultants of the corresponding stress fields, the width of the corresponding tension field, b (see also Fig. 19), has to be determined. This width is yet another free parameter in the model. A reasonable value seems to be b = 1.5 h, thus delivering, in this particular case (cf. Fig. 19), a resulting tensile stress of rT = T/b = F tan2 h/(1.5 37.5). The variation of this stress, depending on the angle of the inclination, is given in Fig. 21. This figure also compares calculated ultimate tensile stresses for specimens CH0, CS0 and CB0 to the (net) tensile strength of the clay block of the same series, fbtz, cf. also Table 3. It can be seen that by increasing value of h the difference between rT and fbtz becomes larger. From this comparison it is obvious that the use of a clay block tensile strength in practical design would lead to an underestimation of the wallette strength, when considering inclinations of 35–40°. However, it should be kept in mind that the tensile resistance of tested elements is actually higher than that determined on units due to some restraint of lateral deformation in the masonry composite. Furthermore, the large scatter of the test results when determining the tensile strength should also be taken into consideration. It should be noted here that the proposed, simple truss model has been applied to elements with different orientations (horizontal and inclined) of the chase. The same truss model could also be applied to the specimens of Series A and B. This has not been done here, since no tensile tests for the units of these series were performed. However, it could be expected that similar results would be obtained. Failure loads in Series A and B were smaller than those in Series C, but the block compressive strength was also smaller; see also Tables 3 and 5. Thus, assuming that the compressive and tensile strengths of the block are directly related, similar results can be expected.
Fig. 20. Variation of the truss model tensile force T.
N. Mojsilovic´ / Construction and Building Materials 25 (2011) 4415–4425
4425
Fig. 21. Variation of tensile stress rT for specimens CH0, CS0 and CB0.
6. Conclusions and outlook An investigation of the influence of chases in masonry elements subjected to compression is underway at ETH Zurich. Results of the tests on three series of elements (total of 21 specimens) were analysed and allow a number of conclusions to be drawn: Chases can reduce the masonry compressive strength considerably. In general, a reduction of the masonry strength proportional to the reduction of the cross-section area seems to be a reasonable approach in analysing the influence of chases on masonry compressive strength. The choice of the typically used material of chase infill (gypsum and cement mortar) has relatively small influence on the reduction of masonry compressive strength. It seems that an infill that is as strong and stiff as the masonry unit has to be applied in order to eliminate the loss in masonry strength. From the present investigation, the high-strength mending mortar could be recommended as suitable chase infill for practical application. In order to get a deeper insight into the structural behaviour of masonry elements with chases an analytical investigation based on a simple truss model has been performed. It was shown that such a model can predict the masonry strength of elements with horizontal and inclined chases. As a governing material parameter the tensile strength of the clay block unit was selected. Major issues deserving further research include tests on similar specimens with additional thicknesses (125 and 200 mm) as well as with different masonry units (calcium-silicate and concrete blocks) which could be of use to confirm/modify or extend the present findings. Further, to extend the presented research and to gain additional insight into the problem further testing on slender walls (second order effects, i.e. stability failure) and a subsequent theoretical investigation would be needed. Finally, the
numerical modelling, e.g. application of micro-modelling, deserves more study. Acknowledgements The clay bricks were supplied free of charge by Ziegelei Schumacher, Körblingen, and Keller Ziegeleien, Pfungen, as was the mortar by Maxit AG, Dätwill, and the high-strength mending mortar by Sika Schweiz AG, Baar. This support is gratefully acknowledged as well as that of graduate students S. Notz, B. Wäfler, J. Yu and D. Geisseler who were involved in performing the tests. References [1] Fischer K. The effects of chasing on the compressive strength of brickwork. In: Proc 3rd int brick/block masonry conf. Essen, Germany; 1973. p. 106–14. [2] Kirtschig K, Metje WR. Influence of chases and recesses on the strength of masonry. Proc Br Masonry Soc 1988;2:61–3. [3] Sahlin S. Fischer’s tests on chasing revisited the effects of chasing on the compressive strength of brickwork (3. IBMAC, Essen 1973). Masonry Int 2007;20(2):85–90. [4] SIA 266:2003. Structural masonry. Zurich: Swiss Society of Engineers and Architects; 2003. [5] ACI 530-08/ASCE 5-08/TMS 402-08. Building code requirements and specifications for masonry structures. Masonry Standards Joint Committee (MSJC), Boulder/Farmington Hills/Reston, USA; 2008. [6] AS3700:2001. Masonry structures. Sydney: Standards Australia; 2001. [7] EN 1996-1-1:2005. Design of masonry structures – part 1: general rules for reinforced and unreinforced masonry. Brussels: European Committee for Standardization; 2005. [8] CSA standard S304.1-04. Design of masonry structures. Canadian Standards Association, Mississauga, Ontario, Canada; 2004. [9] Mojsilovic´ N, Notz S, Waefler B, Yu J. Load tests masonry elements with chases. In: Proc 11th canadian masonry sympos, Toronto, Ontario; 2009. p. 379–88. [10] EN 772-1:2000. Methods of test for masonry units – part 1: determination of compressive strength. Brussels: European Committee for Standardization; 2000. [11] EN 1015-11:1999. Methods of test for mortar for masonry – part 11: determination of flexural and compressive strength of hardened mortar. Brussels: European Committee for Standardization; 1999. [12] Mojsilovic´ N, Marti P. Tests on masonry walls subjected to combined actions (in German). Report no. 203. Zurich: Institute of Structural Engineering; 1994. [13] Mojsilovic´ N. Masonry subjected to combined actions (in German). Report no. 216. Zurich: Institute of Structural Engineering; 1995. [14] Mojsilovic´ N, Marti P. Strength of masonry subjected to combined actions. ACI Struct J 1997;94(6):633–42.