2013-2017 (PhD)

Out-of-plane instability of rectangular RC walls: Numerical simulation and experimental investigation

I started my PhD research at the University of Canterbury in 2013 and under supervision of Prof. Rajesh Dhakal and Prof. Stefano Pampanin.

a- Numerical simulation

Out-of-plane instability of rectangular structural walls was one of the failure modes observed in in the 2010 Chile and the 2011 Christchurch earthquakes (Wallace 2012). Figure 2a shows simulation of out-of-plane instability using solid elements of the finite element program ABAQUS. Although unconfined concrete properties could be assigned to the cover concrete by mesh discretization across the thickness, and a detailed prediction of the wall response could be therefore achieved, the number of element nodes generated in this approach for a decent mesh size was huge, making the analysis computationally demanding. Furthermore, due to the significant convergence issues of this modeling approach, the model could not be verified against experimental results of wall specimens under cyclic loading.

One of my main achievements during the first 18 months of my PhD research was proposing a shell-based numerical model that could successfully predict the out-of-plane instability observed in several wall specimens (Dashti et al. 2014a and 2014b). To the best of our knowledge, I was the first to model this failure mode of rectangular walls under in-plane cyclic loading, which was ultimately awarded by the New Zealand Concrete Society in 2014. Figure 2b indicates the predicted evolution and recovery of out-of-plane displacement during unloading and reloading stages for a rectangular wall specimen using this model. I conducted a comprehensive validation of this model (Dashti et al. 2017a, 2018a, 2018b, 2019a), which mainly focused on verification of the out-of-plane instability simulated by the model using results of several tested wall specimens as well as a parametric study. On top of this, the model was verified using a blind prediction practice to simulate the in-plane and out-of-plane responses of a singly reinforced flanged wall specimen. This validation was conducted within a collaboration with the research group of Professor Beyer in EPFL, Switzerland. The comparison of the predicted response with the experimental measurements (Dashti et al. 2018b), provided a further proof on capability of the model to predict the out-of-plane response of walls. This modeling approach has ever since been used by other researchers for investigation of this failure mechanism (Parra 2015, Scolari 2017, Rosso et al. 2017, Daza Rodríguez 2018).

Figure 2. Numerical simulation of out-of-plane instability in rectangular walls: (a) solid elements-ABAQUS; (b) curved shell elements-DIANA (Dashti et al. 2018a)

b- Experimental investigation

Although out-of-plane instability failure was observed in several wall experiments, its evolution and the controlling parameters have not been fully investigated. An experimental study on the parameters affecting out-of-plane response of doubly reinforced ductile walls was first conducted in the experimental portion of my PhD research (Dashti 2017, Dashti et al. 2017b and 2017c). The experimental program was designed based on a parametric study using the verified numerical model (Dashti et al. 2019b). The test specimens were half-scale models, representing the first story of multi-storey high walls. The test setup was thus designed to apply the lateral load as well as the axial load and bending moment coming from the upper stories producing a shear-span of 6.0 m. Figure 3 displays the configuration of horizontal and vertical actuators producing this loading pattern. As movements of the horizontal and vertical actuators were interdependent, a control program was designed to balance the actuators at each step through an iterative approach so that they complied with the above-mentioned loading conditions and satisfied the design shear-span ratio. The progression and recovery of out-of-plane deformation and development of the subsequent instability in one of the specimens are indicated in Figure 3.

Figure 3. Evolution of out-of-plane deformation and subsequent global instability-Experimental observation (Dashti et al. 2018c)

The experimental findings regarding the evolution of this mode of failure and the influential parameters were in good agreement with the numerical predictions. Also, the experimental findings were used to evaluate the assumptions made in the available analytical models used for prediction of wall instability failure (Dashti et al. 2018c). The experimental observations regarding the effects of different parameters on development of out-of-plane instability are compared with the predicted responses (Dashti et al. 2019c).

Two types of instability were observed (Dashti et al. 2018d), namely global and local modes of instability. Evolution of global out-of-plane instability in one of the specimens (Specimen RWL, Figure 3), which was not affected by any other failure modes is described in detail by Dashti et al. (2018c). Based on the observations made in this experimental campaign as well as the wall tests reported in the literature, the out-of-plane response of walls has been classified into five different modes and the likely forms of force-displacement and out-of-plane displacement history plots corresponding to each mode are discussed (Dashti et al. 2019d).