Fatigue behavior and design of welded tubular T-joints with CHS brace and concrete-filled chord

Concrete-filled welded tubular structures have been increasingly used in large-span constructions such as truss bridges where fatigue failure is always a critical issue that should be focused on. This paper deals with fatigue behavior of CHS-CFSHS T-joints, which are made up of circular hollow section (CHS) braces and concrete-filled square hollow section (CFSHS) chords. Experimental results on stress concentration factor (SCF) were briefly summarized. Fatigue tests were conducted under cyclic tensile force in the brace, so that cracking patterns, failure modes and fatigue data of such joints were recorded. The tested CHS-CFSHS T-joints suffered from one of the following three failure modes: (i) brace-90 degrees-side failure, (ii) chord-90 degrees-side failure, (iii) chord-0 degrees-side failure. Based on the fatigue data in terms of hot spot stress range (S-r,S-hs) versus number of cycles (N), fatigue design S-r,S-hs-N curves were determined using the deterministic and least-squares methods, respectively. It was found that most of the existing S-r,S-hs-N curves are unsafe for fatigue design of CHS-CFSHS T-joints, except for the X S-r,S-hs-N curve recommended by American Petroleum Institute [32] and American Welding Society [36]. In addition, the fatigue data of the CHS-CFSHS T-joints were found comparable with CHS-CFCHS T-joints (T-joints with CHS brace and concrete-filled CHS chord), hence a new fatigue design S-r,S-hs-N curve was proposed for the both types of concrete-filled joints: A comparison between CHS-CFSHS T-joints and empty CHS-SHS T-joints showed that CHS-CFSHS T-joints generally have better fatigue behavior compared to their empty counterparts except for the case of extremely large ratio of brace diameter to chord width (beta).