The current implementation takes into account only prestress losses
due to steel relaxation, other losses (e.g. slip at anchorage,
thermal dilation, friction, etc.) need to be treated separately. The same holds
for the stress transfer from prestressing reinforcement to concrete in
the region called *transmission length*. On the other hand,
losses due to sequential prestressing, elastic deformation and both
short-time and long-time creep and shrinkage are taken into account
automatically provided that a suitable material model is chosen for
concrete.

In the first approach the stress on the end of the time step is explicitly given by the current stress, prestressing level and the cumulative value of prestress losses. On the other hand, in Bazant's approach it is necessary to iterate on the material point level in order to reach equilibrium.

As a simplification the stress-strain diagram is in the current implementation assumed to be linear (no yielding), this should be sufficient for most cases.

Under a constant strain, the evolution of prestress loss is defined as

(232) |

where is the initial value of prestress reduced for losses during prestressing, is time after prestressing in hours, , is the characteristic strength of prestressing steel in tension, and finally , , and are material parameters determined by the relaxation properties of the reinforcement. For wires or cables with normal relaxation (class 1) , and , for cables or wires with reduced relaxation (class 2) , and , and for hot-rolled and modified rods (class 3) , and .

The prestress is not specified in the input record. It is initialized automatically at the time instant when stress differs from zero.

The material model has one internal variable which has a meaning of
cumulative prestress loss when *equivalent time* approach is
employed; otherwise its meaning is a cumulative strain caused by
relaxation.

2018-01-02