An Advanced Numerical Tool for the Prediction and Analysis of Spalling in Concrete Structures Exposed to Combined Thermal and Mechanical Loading

Lead Research Organisation: Newcastle University
Department Name: Civil Engineering and Geosciences

Abstract

Concrete is one of the most widely used construction materials in the world. Throughout this extensive usage many concrete structures are subjected to high temperatures either by consequence of their intended function, for example as in nuclear reactor vessels or aircraft runways, or unintentionally, often as a result of accidental or deliberate fires. For both of these scenarios the structural behaviour of concrete under highly elevated temperatures is clearly very important.High temperature is known to have a number of easily observable effects on concrete including loss of strength, loss of stiffness and spalling, i.e. the fracturing and loss of material from the surface of concrete elements. Spalling, which is the focus of this project, varies considerably in occurrence, extent and severity, and manifestations, ranging from minor and non-violent, to severe and explosive, have been regularly observed. In all five distinct forms of spalling have been identified. In order of increasing violence these are Post Cooling spalling, Aggregate spalling, Corner spalling, Surface spalling and Explosive spalling.Whatever the form, spalling can have significant structural and safety implications depending on the structure that is affected. Where structural members, such as columns and walls, are exposed to elevated temperatures, spalling can result in a loss of load bearing cross-section and the exposure of steel reinforcement, which is highly susceptible to heat damage, and can ultimately lead to the collapse of the entire structure. Even where collapse does not occur there are significant economic implications associated with the time and cost of repair and this extends to situations where collapse is not an issue, for example in aircraft runways, where even minor spalling can have safety or (at the very least) serviceability implications and hence associated economic consequences.Despite its common occurence there is a fundamental lack of understanding of the phenomenon of spalling and the processes that control it. This is demonstrated by the results of numerous studies and the analyses of several high profile incidents including the Channel Tunnel fire, in which a range of hypotheses have been presented but no real consensus as to the exact mechanisms underlying the observed spalling behaviour has emerged.The aim of this work is to develop an advanced numerical tool that, by accounting for the physical processes at work within concrete exposed to elevated temperatures, will ultimately be capable of predicting the occurrence, type and extent of spalling that may be expected in any particular structure subjected to any combination of thermal and mechanical loading. The model will thus have significant applications in the structural and safety assessment of either existing or proposed new structures under actual, historical or hypothetical thermal and mechanical loading scenarios.Furthermore, by applying this model in an extensive series of numerical experiments the fundamental processes underlying and controlling spalling may be explored and a better understanding of the phenomenon can be achieved.The improved understanding of spalling behaviour and the application of the model will allow types of concrete, concrete structures and remedial or protective techniques to be designed such that their performance under severe conditions can be specifically addressed and optimised, and hence the structural, economic and safety implications of structural exposure to elevated temperatures can be minimised.
 
Description The project set out to investigate numerically the fundamental underlying causes of spalling (the fracturing and loss of material) in concrete exposed to elevated temperatures (particularly fire). From the literature it was identified that two potential processes were most commonly considered to be the key cause of spalling; namely the build up of pore pressures inside the micro-structure of the concrete as liquid water is evaporates (forming gas) due to the increase in temperature and the development of thermal stresses resulting from differential thermal expansion as the concrete is heated from the outside in.

Numerous analyses were carried out using a fully coupled hygro-thermo-mechanical model and considered in detail the cases of a concrete wall and a concrete column exposed to fire on all sides. For all these cases spalling of concrete had been observed under experimental and real fire situations. The numerical model developed and employed in this work was able to capture damage behaviour indicative of spalling, both qualitatively in terms of the patterns of damage that occurred and quantitatively in terms of the time at which the damage occurred after the start of heating. An extensive series of numerical experiments, designed to consider situations where the concrete was exposed to thermally induced stress and a range of pore pressures, was carried out. The results of these experiments showed that, in all cases considered, thermally induced stressed were primarily responsible for the development of damage, and hence spalling, in the concrete. Pore pressures were found to have a negligible, or at most, a secondary effect.

The secondary part of the project was modified due to the findings of the work conducted to address the primary part. The project hypothesis was that pore pressures led to spalling, at least in some circumstances, due to their coupled interaction with discrete fractures that developed internal to the concrete. Due to the findings of the first part of the project, that the effect of pore pressures was at most secondary to the effect of thermally induced stresses, the implementation of a discrete fracture model coupled to pore pressure behaviour became unnecessary. Instead the project was modified to consider the use of a continuum damage model to capture spalling behaviour or to be used as an indicator for spalling behaviour (caused by thermally induced stressed).

Again various numerical experiments were employed to investigate the application of the continuum damage model to the prediction of spalling. The ability of the model to capture spalling behaviour was assessed as was its sensitivity to the choice of various parametric relationships that are required for the model. It was found that the damage model could be a useful indicator for spalling but that the selection of model parameters (relating to strength, stiffness, ductility, the initiation and development of damage etc.) must be carefully considered as a whole and not in isolation.
Exploitation Route Spalling of concrete in fire can have significant implications for safety, both in respect of the stability of fire damaged structures and for fire fighters subjected to concrete shrapnel, and can have enormous economic consequences due to the costs of repair and loss of revenue during building closures.

The numerical aspects of this work also go towards the development of design tools for concrete structures in fire. While the model developed and employed here is very complex and perhaps impractical for everyday use in design and better understanding has been developed of what factors need and need not be considered and hence simplified but appropriate design tools may be developed.

The design tools would be employed by practicing engineers in design consultancies.
Spalling of concrete in fire can have significant implications for safety, both in respect of the stability of fire damaged structures and for fire fighters subjected to concrete shrapnel, and can have enormous economic consequences due to the costs of repair and loss of revenue during building closures.

This work goes towards a better understanding of the fundamental processes at work in concrete when it is exposed to fire. Ultimately a better understanding will aid in the design of materials and structures that will be better able to resist fire and be less affected by spalling. This should then lead to an increase in safety and a reduction in the economic consequences of fire in concrete structures.

The numerical aspects of this work also go towards the development of design tools for concrete structures in fire. While the model developed and employed here is very complex and perhaps impractical for everyday use in design and better understanding has been developed of what factors need and need not be considered and hence simplified but appropriate design tools may be developed.
Sectors Construction