Soils, like several other materials, exhibit strong time-dependent behaviour which can be evidenced in terms of creep or strain-rate effects. The degree of this rheological behaviour varies with the type of soil, its structure, and with the stress history. This effect is exacerbated in pile load testing where the procedure duration tends to be shortened under increasing time pressures. The modelling needed to interpret the results therefore becomes more and more complex, including soil viscosity, wave radiation into the soil and other significant phenomena. The objective of the research reported herein is to refine the rheological parameters characterizing the influence of the loading rate within the framework of a relevant pile/soil interaction model fed with dynamic measurements acquired during pile Dynamic Load Tests (DLTs). The final goal is to predict and simulate the quasi-static pile load settlement curve. The pile/soil interaction system is described by a non-linear mass/spring/dashpot system supposed to represent the pile and the soil, with constitutive relationships existing within and between them. These relationships account for the static and the dynamic or rheologic behaviour. A back-analysis process based on a matching procedure between measured and computed quantities allows one to characterize the pile/soil interaction in terms of constitutive and rheologic parameters based on the dynamic measurements. After optimisation of the matching procedure, the parameters obtained are used to simulate the “static” load-settlement curve. The matching procedure is based on an automatic and stochastic parameter perturbation analysis. Since the parameters influence the system response with a relative weight, they are sorted in order to optimise all the parameters by successively retrieving the most influential ones and working on the remaining ones. The back-analysis performed on real dynamic measurements in this research leads to an improved pile/soil interaction model. The slippage between pile and soil along the pile shaft must be explicitly taken into account. This refinement increases the number of degrees of freedom needed to describe the pile/soil system but brings deeper insight into the behaviour of an interfacing zone of limited thickness surrounding the pile shaft.

A structure subjected to extreme load due to explosion or human error may lead to progressive collapse. One of the direct methods specified by design guidelines for assessing progressive collapse is the Alternate Load Path method which involves removal of a structural member and analysing the structure to assess its potential of bridging over the removed member without collapse. The use of this method in assessing progressive collapse therefore requires that the vertical load resistance function of the bridging beam assembly, which for a typical laterally restrained reinforced concrete (RC) beams include flexural, compressive arching action and catenary action, be accurately predicted. In this thesis, a comprehensive study on a reliable prediction of the resistance function for the bridging RC beam assemblies is conducted, with a particular focus on a) the arching effect, and b) the catenary effect considering strength degradations. A critical analysis of the effect of axial restraint, flexural reinforcement ratio and span-depth ratio on compressive arching action are evaluated in quantitative terms. A more detailed theoretical model for the prediction of load-displacement behaviour of RC beam assemblies within the compressive arching response regime is presented. The proposed model takes into account the compounding effect of bending and arching from both the deformation and force points of view. Comparisons with experimental results show good agreement. Following the compressive arching action, catenary action can develop at a much larger displacement regime, and this action could help address collapse. A complete resistance function should adequately account for the catenary action as well as the arching effect. To this end, a generic catenary model which takes into consideration the strength degradation due to local failure events (e.g. rupture of bottom rebar or fracture of a steel weld) and the eventual failure limit is proposed. The application of the model in predicting the resistance function in beam assemblies with strength degradations is discussed. The validity of the proposed model is checked against predictions from finite element model and experimental tests. The result indicate that strength degradation can be accurately captured by the model. Finally, the above developed model framework is employed in investigative studies to demonstrate the application of the resistance functions in a dynamic analysis procedure, as well as the significance of the compressive arching effect and the catenary action in the progressive collapse resistance in different designs. The importance of an accurate prediction of the arching effect and the limiting displacement for the catenary action is highlighted.