Thoughts about existing structures
The opportunity to respect in general terms the concept of sustainability in concrete structures implies the necessity to extend conceptual design substantially to durability aspects.
The availability of the recently published Model Code for Service Life Design (MCSLD),1 which includes four different approaches for service life design, helps the designer fulfil the durability performance criteria in relation to the importance of the structure under consideration.
Furthermore, the MCSLD may be used for the evaluation of the residual life of existing structures and for the generation of a priority list of interventions in the case, common in practice, that several structures belonging to the same set should be repaired while, at the same time, the overall functionality of the entire set should be maintained.
However, it should be pointed out that in this field some important innovations are necessary for modelling damaged structures and to reach a reliable evaluation of their actual safety level and its evolution in time. In the following the main aspects of the design procedure that need innovations will be analysed.
First of all, when we handle existing structures, we do not need to take into account the uncertainties related to the construction process, but we can directly evaluate the mechanical properties of the materials by means of a well established set of experimental tests. Therefore the γm coefficient may assume reduced values with respect to the ones used for new structures. A useful guide for this fundamental aspect may be found within the Probabilistic Assessment of Existing Structures published in 2001 by the Joint Committee on Structural Safety.2 Another aspect, related to the safety field, is the definition of material characteristics along the structure. If we can accept that in a newly built structure the same probability density function for material characteristics may be used all along the structure, this concept generally appears to be too conservative for damaged structures, in which generally the material properties should be described, also taking into account spatial variability. The random field approach may then produce numerical results close to the actual conservation state.
A second set of innovations regards the physical models able to describe the mechanical behaviour of damaged structures. Focusing our attention to the main phenomenon of degradation, that is corrosion of reinforcement induced by carbonation and/or chlorides, at least four different deterioration effects can be enumerated
cross section reduction of reinforcement
variation of yielding and tensile strength and ductility of reinforcement
reduction of concrete compressive strength due to the tensile stresses transversally acting to the compression direction and generated by the volume increase of corroded bars
bond strength reduction between reinforcement and concrete.
The first two effects may easily be evaluated on the bars and the current physical models may be used with the new geometrical-mechanical parameters. The third effect may be evaluated by the use of a resistance criterion for concrete crossed by reinforcement, stressed in a biaxial state, in agreement with recent proposals in this field. Then a modified compression strength may be used in the mechanical resisting model.
The fourth effect implies a more accurate description of bond, as a relative slip between concrete and reinforcement may occur along the beam. As a consequence, the generally accepted hypothesis of plane sections remaining plane is no longer valid and the compatibility of longitudinal deformations should be verified. Reinforcement bars work as bonded until a critical level of slip is reached. Subsequently, they can be considered as basically unbonded without further stress increase.
Finally a new model is necessary to describe the mechanical behaviour in regions in which a combination of bending and shear is able to produce a substantial slip of the reinforcement which is accompanied by a sudden increase of compressive concrete stresses, a large shift of the neutral axis and a reduction of the ultimate bearing capacity.
In conclusion, new physical models and new safety concepts, taking into account the different nature of model uncertainties, should be introduced in the next generation of design codes to help the designers in the evaluation process of existing structures: the new fib Model Code which is under preparation, can play an important role in this field.
