Development of novel connection methods between precast concrete panels

• Main Author: Vella, Jean Paul
• Other Authors: Vollum, Robert
• Published: Imperial College London 2017
• Subjects: 624
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.733197

This thesis addresses the design and behaviour of narrow cast in-situ joints between precast concrete elements in which continuity of reinforcement is achieved through overlapping headed bars. The use of headed bars minimises the required reinforcement lap length and hence joint width, thereby, maximising the area of precast concrete units. Confining reinforcement in the form of transverse bars and vertical shear studs is installed within the joint. Improved understanding of the behaviour of such joints is obtained by means of experimental testing and numerical analysis. In total, 32 tensile specimens and five flexural specimens were tested. The main longitudinal reinforcement in all specimens consisted of 25 mm diameter headed bars with 70 mm square heads. The tests investigated the influence on joint performance of concrete strength, confining shear studs, transverse bar arrangement, headed bar lap length and spacing, and out-of-plane offset of precast planks. For tensile specimens, a lap length of 100 mm with four 16 mm transverse bars and concrete compressive strength of 28 MPa was sufficient for headed bar yield. Ductile flexural failure, with headed bar yield, was achieved with the same lap length, two 20 mm transverse bars and concrete compressive strength of 39 MPa. A nonlinear finite element model (NLFEM) is presented with results showing that it captures the behaviour of the joint well. Additional parametric studies carried out with the validated NLFEM provide otherwise unavailable insights into joint behaviour. In conjunction with an appropriate safety format, the model is considered suitable for the design of standard joint configurations. Analytical models based on the strut-and-tie method and upper bound plasticity are presented. The models were found to give reasonable predictions of joint strength, but are not capable of fully capturing the observed joint behaviour. An improved strut and tie design method is suggested and recommendations are made for optimising joint design.