Dynamic in‐plane racking tests of curtain wall glass elements

Abstract

AbstractAttention has been paid recently to the potentially serious life safety and economic loss issues related to the seismic performance of ‘architectural,’ or ‘non‐structural’ building elements such as glass lites in curtain wall systems. In response, a pilot study was undertaken at the University of Missouri‐Rolla to investigate the dynamic racking performance of a 15·6 × 12·0 ft (4·56 × 3·68 m) section of curtain wall containing three 5 × 6 ft (1·52 × 1·84 m) glass panels. The curtain wall system was a ‘wide mullion’ design that had generous 1 in (25 mm) clearances between glass edges and the aluminium glazing pocket. Dynamic racking tests were performed totally in plane; no out‐of‐plane or torsional motions were included. Various types of glass specimens were tested, including annealed, heat‐strengthened and fully tempered glass in monolithic and laminated configurations. Dry glazed and two‐side structural silicone curtain walls were tested. Test results for dry glazed specimens showed that annealed and heat‐strengthened laminated glass experienced no fallout whatsoever. By contrast, annealed monolithic glass experienced frequent fallout in both small and large shards. Fully tempered monolithic glass experienced dicing, which resulted in occasional fallout of entire glass lites. Loss of rigidity in fully tempered laminated glass (when both glass plies were broken) occasionally caused entire lites to fall out. Polyester film (not anchored to the mullions) was applied to annealed monolithic glass; it prevented small shards from falling out, but sometimes contributed to entire lite fallout after the onset of severe glass cracking. Annealed laminated glass units with two‐side structural silicone glazing exhibited only very minor glass damage and no glass fallout.