Level 2 Model as a
An automation system in a steel mill consists of Level 1 (basic process control), Level 2 (production management and process modeling) and Level 3 (business and scheduling). Up to now, almost all the Level 2 models have treated hot rolling as a mechanical process and ignored metallurgical phenomena. Metallurgical models are not yet integrated into the reheating furnace Level 2 and controlled cooling Level 2 either. With recent technological development, widely applied metallurgical processes such as controlled rolling in steel mills demand that a Level 2 model should fully address metallurgical issues, among others (rolling process modeling and software engineering) , in order to maximize productivity and minimize costs.
The topics discussed in the following section
s demonstrate that the steel mill Level 2 should fully consider the metallurgical processes in order to achieve sufficient accuracy and to solve various quality problems:
Retained strain incurring due to incomplete recrystallization
Softening during the hold in the controlled rolling
Two-phase regions in the final passes
Dynamic changes of the grain size
Variable properties along the width, length and thickness
Incomplete recrystallization and retained strain
The level 2 models often use a simple algorithm of the flow stress which includes a dependence on temperature (T), strain (ε) and strain rate (u):
The constants for this algorithm were adjusted after the rolling of each piece as a means of adaptation (fitting) to the measured flow of the metal under particular conditions. Four fitting mechanisms are available: FIT2 (using C1 and C2), FIT3A (using C1, C2 and C3), FIT3B (with C1, C2 and C4) and FIT4 (with all coefficients). Coefficients not for learning are set to constants. To increase accuracy of the force prediction, the flow stress model maintains separate sets of flow stress coefficients for each model grade. A model grade is created based on the steel grade (chemical composition), the product (type and dimension) and the production practice (e.g., with or without hold). For each model grade, there are three sets of coefficients that are automatically adjusted by the long-term learning function to cover the three ranges (either thickness or temperature) expected during rolling.
To determine values of C3 and C4 for all model grades in all the three temperature regions, a study on the
C3 and C4 used in an existing Level 2 system was conducted. Adapted C3 from all fits reveals widely scattered values that were mostly much higher than the theoretical ones for hot forming. About 40% of the C3 were with value zero, including the C3 that was set to be zero in the FIT2 and FIT3B. The strain rate coefficient C4 from adaptation was also scattered. About 85% of values for C4 were zero; most of them were from FIT2 and FIT3A.
The larger values of the strain coefficient C3 than the theoretical ones, were mainly due to two factors: the retained strain and the improper learning fits, especially the FIT2,
FIT3A and FIT3B.
Modern rolling practice has significantly reduced the rolling temperature, in order to achieve better mechanical properties. Due to the low rolling temperature, the recrystallization often cannot be completed and some strain from the previous pass would be retained. Figure 1 shows the retained strain published by I. Tamura (IT), et al (left)  with Nb steel at an inter-pass time of 20 seconds.
Metal Pass also predicted the value of the retained strain for a plate mill assuming an inter-pass time 30-40 seconds. At an inter-pass time of 1 second and rolling temperature of 750°C (1380°F), almost the entire pass strain was retained.
Figure 1 Retained strain
<To be continued>
Part 1 |
Part 2 |
Part 3 |