# Interpretation of Results

Most of the research results published until now provided a phenomenological description of the long-term behaviour of FRC elements in cracked conditions and identified some significant parameters. Analytical models were used to provide a description of the long-term behaviour ad it was shown that functions that are normally used to describe creep in compression can properly represent creep in bending [3, 44].

Detailed interpretation of results from long-term bending tests is still an open problem. In fact, the behaviour of the specimen, either in terms of CMOD or in terms of mid-span deflection, is affected by both the creep concrete matrix and the creep of the cracked part. This latter is the result on different phenomena that interact and whose individual contributions have not been yet clearly defined: creep of the concrete matrix in tension, creep of the fibre-to-concrete bond, micro-cracking [45], and, as far as MS fibres are concerned, creep of the fibres [46]. Concerning this latter contribution, it is worth noticing that currently there are no minimum requirements for MS fibres in terms of creep deformation.

In order to introduce creep in design it is mandatory to separate the tension behaviour from the one in compression. A viable approach seems to be constituted by inverse analysis. Buratti and Mazzotti [47], for example, proposed to derive compression creep form specific tests and to use a fibre based-model to provide a phenomenological description of the creep in tensions (using an equivalent continuum model). This model was then used to estimate creep in tension using an inverse analysis procedure on long-term flexural data. The model is being further extended using data from uniaxial tension tests for defining the creep model for the portion of the cross section in tension.