Cunningham, Lee Scott
Automatic design of concrete structures using a strut & tie approach.
PhD thesis, University of Glasgow.
Full text available as:
The major part of the work presented in this thesis is an investigation of the strut and tie method for designing 2-D in-plane, reinforced concrete structures. Two important issues relating to this method are addressed. Firstly, the issue of visualising an appropriate strut and tie model is dealt with. In many situations it may be difficult to visualise an appropriate model for a given structural system. Here, a convenient method of visualising strut and tie models is presented. Using elastic finite element analysis, low stressed parts of a structure are removed in a step by step process until the main stress paths, which represent the ties and struts, are defined.
The second important issue to be addressed is that of serviceability of the designed structure because the strut and tie model naturally represents a great departure from the elastic stress distribution. Since the strut-tie model is used to design for the ultimate load situation, it is necessary to assess the suitability of the same model in relation to serviceability characteristics of the resulting design. It is important that ductility of the structure should be maintained at ultimate loads while avoiding excessive deflections and cracking at service loads. A wide variety of structures were designed, and to assess the performance of each design, non-linear finite element analysis was used. Verification of some of the numerical results was carried out through physical testing in the laboratory which also allowed the serviceability behaviour of the structures to be assessed. The test program comprised of three corbel joints and two frame corner joints.
It was concluded that design from the strut and tie method can produce adequate performance both at service and ultimate loads. In terms of ultimate load prediction, the strut-tie method can produce results of comparable accuracy to non-linear finite element analysis.
Actions (login required)