|
Level 2 Model Improvement
Case Study: Oregon Steel
<Continue>
Flow Stress Learning Parameters
C3 and C4
To study medium values of C3 and C4
for all model grades in all the three temperature regions, a survey on the
existing C3 and C4 used in the Level 2 system was
conducted besides the study on Metal Data online flow stress collection [5]. C3
from all fits revealed widely scattered values that were mostly much higher than
the theoretical ones for hot forming (Table 1). The medium value for C3
was near 0.30. About 40% of the C3 were with a value of zero, and the
minimum value was -0.5. Among the non-zero records there were about 30% in each
of the ranges (0.3-0.4) and (0.2-0.3), among others. The experimentally verified
medium value published earlier for the hot forming was about 0.18 [6] from
Hensel/Spittle (Germany), or from other researchers (Japan) for hot rolling [7],
which was 0.21. Nearly 1% records were with the C3 value over 0.6
while nearly 0.2% data were with C3 below zero.
Study on C3 for FIT4 only, revealed less
scattered values than the other fits, though they still spread within a large
range. The medium value changed to about 0.22. Minimum value was still -0.5.
Table 1:
Statistics of the strain coefficient C3 for Flow
stress
|
Average
C3 |
Count |
Weight
%
(w/o 0) |
Weight
%
(w/ 0) |
|
< 0.0 |
60 |
0.2 |
0.1 |
|
0 |
18976 |
|
42.6 |
|
0 - 0.13 |
2255 |
8.8 |
5.1 |
|
0.14 |
881 |
3.4 |
2.0 |
|
0.16 |
826 |
3.2 |
1.9 |
|
0.18 |
1006 |
3.9 |
2.3 |
|
0.20 |
1478 |
5.8 |
3.3 |
|
0.22 |
1544 |
6.0 |
3.5 |
|
0.24 |
1763 |
6.9 |
4.0 |
|
0.26 |
1834 |
7.2 |
4.1 |
|
0.28 |
1735 |
6.8 |
3.9 |
|
0.30 |
2080 |
8.1 |
4.7 |
|
0.32 |
1807 |
7.1 |
4.1 |
|
0.34 |
1667 |
6.5 |
3.7 |
|
0.36 |
1585 |
6.2 |
3.6 |
|
0.38 |
1138 |
4.5 |
2.6 |
|
0.40 |
959 |
3.8 |
2.2 |
|
0.41 ?0.60 |
2728 |
10.7 |
6.1 |
|
> 0.60 |
219 |
0.9 |
0.5 |
The strain rate coefficient C4 for all fits was
also scattered within a large value range. The medium value, about 0.12,
however, was similar to the hot forming theoretical value 0.14 published
earlier. About 85% of C4 were zero, which means that in the most
learning fits, the strain rate factor was ignored. Among the non-zero values, 2%
were over 0.3, 7.5% were below zero. The minimum was -0.8 and the maximum 0.95.
Table 2: C4
values for the FIT4
|
C4 value |
Count |
Weight % |
|
≤0.02 |
885 |
21 |
|
0.04 |
218 |
5 |
|
0.06 |
156 |
4 |
|
0.08 |
369 |
9 |
|
0.1 |
342 |
8 |
|
0.12 |
709 |
17 |
|
0.14 |
516 |
12 |
|
0.16 |
454 |
11 |
|
0.18 |
151 |
4 |
|
0.2 |
139 |
3 |
|
≥ 0.22 |
248 |
6 |
The C4 values from the four-parameter learning
were mainly in the range of [0.02, 0.2] (Table 2). If the difference in
values of C4 in three temperature regions was also considered, a
rough range of the C4 boundaries for the three temperature regions
could be determined. About 1% of C4 were over 0.9 and around 5% below
zero, and 6% records were with C4=0. The influence of rolling stage
on C3 and C4 during a controlled rolling practice was also
studied. For single stage rolling, the average C3 value was 0.269,
and the average C4 was 0.115. For two stage rolling, the average C3
value was 0.268 and the average C4 0.119. This concluded that the
production stage had no influence on the C3 and C4.
<To
Be Continued>
<Previous>
|
<Next>
|
Part1
|
Part2
|
Part3
|
Part4
|
Part5
|
Part6
|
Part7
|
Part8
|
Part9
|
|
