**2. Programme of Our Own Tests**

Three series (12 test models in total) with the same shape and dimensions were prepared and tested. T-shaped models were monosymmetric, with a web and flange length of ~89 cm. A vertical joint, whose structure varied intentionally, was formed between loaded and unloaded walls. A series of test models marked with **P** had traditional masonry joints between the web and the flange (Figure 1a). Those elements were regarded as reference models, whose mechanical parameters and behaviour at loading and failure were compared with results from other tests. In two other series, joints between webs and flanges were made with steel connectors (wall geometry acc. to Figure 1b). They were single punched flat profiles in series **B10** (Figure 1c), and modified flat profiles with a widened central part in series **BP10** (Figure 1d). This solution was proposed on the basis of our own tests [7] on perforated connectors. The widening of the central part was intended to increase the flexural capacity of the connector and its stiffness. The proposed shape is copyrighted via an application to the Polish Patent Office [8]. Joints made of galvanized perforated steel with a thickness of 1 mm were used in both series.

**Figure 1.** Geometry and details of test models and: (**a**) traditional masonry bond, P series, (**b**) walls with steel joints (B10 and BP10 series), (**c**) joining method with a punched flat bar, (**d**) joining method with a punched widened flat profile (mm).

Tests were conducted in a test stand specially designed for that purpose—see Figure 2. Models 1a and 1b with confining elements 3 and elements taking load 2 were put on the strong floor (panel 1b) and placed on a dynamometer 6, which with a resistor 4 acted as a fixed articulated support. Models were placed below a steel frame 8, to which a hydraulic actuator was fixed (with an operating capacity of 1000 kN), generating shearing at a constant displacement gain equal to 1 mm/min. The structure response was registered using an inductive force transducer with an operating capacity of 250 kN and reading accuracy of ±2.5 kN. Prestress of 0.1 MPa was exerted using reinforced concrete elements *3* and steel strands *7* to model the considerable length of a joined wall in panel *1b*. Models were loaded in one cycle until failure. Vertical load generating shear was transmitted linearly along the whole height of the wall through elements *2*. As a result, shear stresses on joints were distributed uniformly. The loading and displacement of a loaded wall against the unloaded one were continually registered during tests. Two independent types of software were used to register data. One side of the test model was monitored using ARAMIS—an optical sensor of displacements. Another side was monitored with inductive transducers of displacement of type PJX-10, with an operating capacity of 10 mm and an accuracy of ±0.002 mm.

**Figure 2.** A scheme and photo of the test stand (longitudinal wall (*1a*), transverse wall (*1b*), reinforced concrete column transferring shear load (*2*), reinforced concrete pillars limiting horizontal deformation (*3*), horizontal support (*4*), system of the hydraulic cylinder and the force gauge used to induce shear stress (*5*), force gauge, vertical reaction (*6*), horizontal tie (*7*), steel frame (*8*)).

Tests were performed on models made of AAC masonry units with system mortar M5 class for the thin joints and the unfilled head joint. The compressive strength of the masonry specified in the code PN-EN 1052-1:2000 [9] and presented in [10] was fc = 2.97 N/mm2, and the modulus of elasticity was Em = 2040 N/mm2. The initial shear value determined according to the code PN-EN 1052-3:2004 [11] and presented in [12] was fvo = 0.31 N/mm2.

The mean friction coefficient in joints without mortar was μ = 0.92 [13]. The shear modulus determined according to the code ASTM E519-81 [14] and presented in [15] was G = 329 N/mm2. Mortar for thin joints was used in the tested elements for the AAC blocks. This mortar is dedicated to the erection of AAC masonry walls Additional tests on steel connectors—see Figure 3—were conducted according to the standard [16]. Three elements were chosen randomly from each series of connectors and placed in the jaws of a testing machine. The basic mechanical parameters of the connectors were determined by controlling the displacement gain. The measurement of strains was non-contact with a video extensometer MEVIX 200. Strain was measured using a base with the length *L*e = 53.5 mm in standard connectors and *L*<sup>e</sup> = 75.0 mm in thickened connectors. Figure 3 illustrates (stress σ –strain ε) relationships. The stress–strain relationship of tested connectors was found to have no clear yield point. Therefore, results were approximated with a bi-linear relationship. A theoretical yield point fy was determined at the intersection of straight lines. The slope of the tangent straight line presented within the range of 0–fy was assumed as the mean initial modulus of elasticity Es. Tensile strength was determined at failure of specimens, and the tangent of the straight-line slope within the range of fy–ft was determined as the mean secant modulus of elasticity Et. Test results for connectors and the research programme are compared in Table 1.

**Figure 3.** Test tested connectors: (**a**) connectors B10, (**b**) connectors BP10.


**Table 1.** Programme of tests and basic characteristics of connectors.

Area of gross section (A), moment of inertia of gross section (I), mean yield stress (fy), tensile strength (ft), mean initial modulus of elasticity (Es), mean secant modulus of elasticity (Et).

## **3. Test results and Analysis**
