362 Prakt. Met. Sonderband 30 (1999) 
Thus the Ar temperature is lowered with increase of the elastic stored energy during the sweeping 
effect (14). The dependence of the reversible deformation (transformation plasticity) on the number 
of performed working cycles is plotted in Figure 1. After some initial cycles (approximately 100 
cycles), the extent of TWSME is stabilized and a degradation of TWSME accompanied with 
lowering of transformation plasticity follows afterwards. The specimens B, which underwent “hard 
training procedure”, exhibited better performance. The reversible strain exceeded the critical value : 
of s = 2% during the entire testing period (up to 10 000 working cycles). 
p00 60 C000 ooo oe fx 
3 —aA— sample A 
—e&— sample B 
Phas aa a, —-— 
0 2000 4000 6000 8000 10000 
working cycles N 
Figure 1: Reversible strain € vs. number of working cycles N. 
In the case of A specimens (“soft training”), the observed number of stable cycles (€ > 2 %) ranged 
between 3000 and 5000 working cycles. The differences between A and B samples were ascribed to 
the higher stability of the transformation process in more work hardened B samples. These results 
are in a good correlation with results reported in (14) as well as with later investigations (12, 13) 
according to which optimum work hardening level prevents generation of new dislocations and 
subsequent decrease of reversible strain. The higher work hardening rate in the case of trained B 
specimens compared to the A samples was confirmed by the testing of mechanical properties. The 
applied forces vs. the length dependencies as obtained in bending tests performed at room 
temperature (below Ms) and at 135°C (above Af) are shown in Fig. 2. 
Zz 
wo 
32% 
9 fv 
a 
2. 
< 
J £ 
length [mm] 
Figure 2: Force F vs. change of length 1 dependencies of trained specimens as detected at room 
temperature (A;, B in martensitic state) and at 135°C (Az B2 in austenitic state)