Monotonic and cyclic sway behaviour of 2-dimensional frames made of pultruded GFRP I-section profiles

This paper presents an experimental study on the structural response of 2-dimensional full-scale frames comprising pultruded GFRP I-section profiles under lateral quasi-static monotonic and cyclic loading, including four different frame configurations, to assess (i) the influence of using different beam-to-column connection systems (one series included cleats and other included cuff parts); (ii) the benefits of using bracings (one series included a cable bracing system); and (iii) the impact of using infill walls (one series included plasterboard panels). The beam-to-column connections and the bracing system were materialized exclusively with stainless steel parts. The results of the monotonic tests showed that the type of beam-to-column connection has considerable influence on the frames' response, namely on their lateral stiffness; additionally, the frame with bracings and the frame with infill walls exhibited the highest strength and stiffness, respectively, among all series. In the cyclic tests, all frames presented considerable pinching, limiting the energy dissipation; the results showed that the benefits of the infill walls should not be accounted for in design and that the bracing system used is not adequate for seismic areas, as both the plasterboards and the cable bracing system presented non-recoverable deformations/damage that limited their structural contribution from that point forward. The best performing frame series without bracings or walls was simulated using a simple finite element (FE) model. The profiles were modelled using frame elements and the beam-to-column connections were simulated using link elements with the Pivot hysteresis model, calibrated from previous experiments on the beam-to-column connection system. The FE model was able to predict the lateral response with reasonable accuracy, validating this numerical procedure. Then, the FE models were used to evaluate the effects, on the cyclic performance, of including GFRP bracings and a steel plate damper, leading to an improved hysteric response, especially in what concerns the energy dissipation capacity. Overall, the results show that (i) even with ductile connections, the energy dissipation capacity of simple pultruded frames is limited, (ii) although they present high drift capacity; (iii) simple models can be used for design purposes and (iv) they show that the introduction of material-adapted dissipation systems can improve the seismic performance of pultruded frame structures.