The development of the phase size dependent on the number of cycles, which resulted from the 
measurements, is represented Fig. 4. The average phase length decreased horizontally to the rolling 
direction (X direction), whereas the phases widened vertically to the rolling direction (y direction) 
with the vertical widening being in the inverse ratio to the horizontal decrease. 
160 phase size mj 
phase size x, ar) hase size y; ar 
140 [gphase size x; water g phase size y; water 
120 
100 
80 
60 CoE 
40 
20 4 — 
Of 
J 10 8 2 2 3 
number of cycles [1] i. 
Fig. 4: development of the phase size dependent on the number of 
thermo cycles between room temperature and 1050;C 
CHANGE OF THE INHERENT STRESS CONDITION 
The analysis of the phase-specific inherent stress condition was impeded by a superimposed, ther- 
mally-induced temperature stress distribution over the cross section of the sample. By concertedly 
measuring these stresses, their influence on the microscopic residual stresses which were to be de- 
termined in the ferrite and austenite was eliminated. The following results represent the microscopic 
residual stresses which were corrected according to these conditions. In order to analyse the influ- 
ence of the heat treatment on the inherent stress condition, the stress condition was first of all inves- 
tigated in its initial state. In the ferrite, the result was a corrected compression stress amount of 50 
MPa in the rolling direction. Simultaneously, a tensile stress of the same amount was measured in 
the austenite. Compared to the initial state, thermal alternating stress in the range between room 
temperature and 1050;C led to a stress increase of 60 MPa. The co-ordination between tensile 
stresses and compression stresses has continued to exist after the heat treatment. 
SIMULATION CALCULATIONS 
It is not possible to determine the influence of the phase length by means of the two-dimensional : 
models which have been employed so far (3,4). These models presuppose a generalised plain strain, : 
which appears if phase length 1 is much bigger than phase gauge d. Thus, a three-dimensional model N . 
of the microstructure was created (Fig. 5). Since the calculation of these models is considerably Tl 
. a . . ad c 
complicated they are, however, limited to small microstructural clips. or 
Phage 
306