Modelling Plate Mill Rolling An Expert Practical System Approach M. Rebellato* and R. Barbosa** *Companhia Brasileira de Mineração e Metalurgia, **Universidade Federal de Minas Gerais The Charles Hatchett Seminar, 16 th July 2014, London
Introduction • For structural steels, optimized mechanical properties are heavily dependent on how fine and homogeneous the cross section ferrite grains become • The key is to determine how to apply basic metallurgical fundamentals to the production line Charles Hatchett Seminar Plate mill: an expert practical system 2
In the field • Engineers face limitations such as plant layout design, customer and societal specifications, costly downgrades • There is little time to decide ; models must give outputs very rapidly Charles Hatchett Seminar Plate mill: an expert practical system 3
Motivation • An expert, metallurgically sound, practical system is necessary http://www.industry.siemens.com/datapool/industry/industrysolutions/metals/siroll/en/Dongkuk-Plate-Mill-No.2-en.pdf Charles Hatchett Seminar Plate mill: an expert practical system 4
This work • In what follows the first steps of a larger project aimed at building a practical expert system for industry rolling of microalloyed steels is presented Charles Hatchett Seminar Plate mill: an expert practical system 5
An industry case Charles Hatchett Seminar Plate mill: an expert practical system 6
Rolling 300 mm slab to 16 mm plate Microstructure at ¼ thickness Fine grains Coarse grain Mixed microstructure Charles Hatchett Seminar Plate mill: an expert practical system 7
Rolling 300 mm slab to 16 mm plate Microstructure at ¼ thickness Grain sizes Fine grains Average ~ 10 μ m However, grain sizes ranging from Coarse grain ~ 20-25 to ~ 5-6 μ m Very inhomogeneous Mixed microstructure structure Charles Hatchett Seminar Plate mill: an expert practical system 8
Plate: chemical composition Element % weight C 0.046 Mn 1.08 Nb 0.04 Ti 0.014 V 0 Cu 0 Cr, Ni, Mo < 0.35 N2 0.0051 Charles Hatchett Seminar Plate mill: an expert practical system 9
Plate: possible precipitation All Ti as TiN Ti = 0.014 3.4 : 1 0.0009 N2 available N2 = 0.0051 (almost no N2 left) • 0.046 C takes 0.006 N2 left ≈ 0 Nb 7.75 : 1 • Nb left in solid Nb = 0.040 solution = 0.040 – 0.006 = 0.034 Form TiN + NbCN (few) + NbC , ie, mixed particles + leaving Nb in solid solution during rolling Most Nb in solution Low RLT and RST Charles Hatchett Seminar Plate mill: an expert practical system 10
Hot rolling schedule Broadsizing Roughing @ 1150 o C R1 @ 1140 o C and R8 @ 1130 o C Temperature RLT @ 970 o C Holding period 240 s RST @ 890 o C Finishing Acc F1 @ 940 o C and Tstart @ 830 o C F6 @ 855 o C and Possible partial Tfinish @ 500 o C recrystallization case due to Time = 25 s low strain accumulation Time Charles Hatchett Seminar Plate mill: an expert practical system 11
Finishing stands: Partial recrystallization of austenite Pass Recrystallized Accumulated strain fraction after pass up to a given pass (%) F1 17 0.33 F2 73 0.61 F3 90 0.50 F4 22 0.29 F5 18 0.46 F6 5 0.50 Model indicates: possible partial recrystallization case low strain accumulation before transformation Charles Hatchett Seminar Plate mill: an expert practical system 12
Suggested alternative Start finishing below RST, stop above Ar3 Broadsizing Roughing @ 1150 o C R1 @ 1140 o C and R8 @ 1130 o C Temperature RLT @ 970 o C Holding period 300 s RST @ 890 o C Acc Tstart @ 830 o C Finishing and F1 to F6 below RST Tfinish @ 500 o C and above AR3 Time = 25 s Time Charles Hatchett Seminar Plate mill: an expert practical system 13
Suggested improvement in finishing Temperatures Temperature ( o C) Reference temperature RST 890 Finishing start temperature (F1) 910 (940 previous schedule) Finishing stop temperature (F6) 825 (855 previous schedule) AR3 810 Possible outcome No partial recrystallization and Increase in strain accumulation before transformation Charles Hatchett Seminar Plate mill: an expert practical system 14
Finishing Recrystallization and strain accumulation after changes Pass Recrystallized Accumulated strain fraction after pass up to a given pass (%) F1 Nil 0.38 F2 Nil 0.75 F3 Nil 1.08 F4 Nil 1.29 F5 Nil 1.51 F6 Nil 1.65 Model indicates: no recrystallization in all passes substantial increase in accumulated strain Charles Hatchett Seminar Plate mill: an expert practical system 15
Rolling 300 mm slab to 16 mm plate New trial Microstructure at ¼ thickness Grain sizes Average ~ 7 μ m Grain sizes ranging from ~ 10-12 to ~ 3-6 μ m More homogeneous structure Charles Hatchett Seminar Plate mill: an expert practical system 16
Rolling 300 mm slab to 16 mm plate Microstructure at ¼ thickness Before After Charles Hatchett Seminar Plate mill: an expert practical system 17
The model Charles Hatchett Seminar Plate mill: an expert practical system 18
Description • Two modules: reheating and hot rolling • Sellars ’ type model • Uses equations available in the literature Charles Hatchett Seminar Plate mill: an expert practical system 19
Reheating module Charles Hatchett Seminar Plate mill: an expert practical system 20
Reheating module Inputs Outputs Alloy design Nb content in solution Process parameters Time needed to dissolve • Slab thickness Nb • Furnace geometry • Furnace temperatures Charles Hatchett Seminar Plate mill: an expert practical system 21
Hot rolling module Charles Hatchett Seminar Plate mill: an expert practical system 22
Hot rolling module Nb in solution (from reheating module) Inputs Outputs Process parameters Recrystallization data • Pass temperatures Precipitation data • Strain Ferrite grain size • Strain rate Schedule optimization • Grain size • Delay times Charles Hatchett Seminar Plate mill: an expert practical system 23
Schedule optimization Suggested schedule Schedule thicknesses, mm 35 Stage at rolling Heaviest thickness reduction applied at 30 last roughing pass Slab 300 25 Reduction [%] End broadsizing 227 20 Hold 64 15 Final 16 10 Roughing Finishing Reduction % 5 BS Stage at Reduction Recommended 0 B1 B2 B3 R1 R2 R3 R4 R5 F1 F2 F3 F4 F5 F6 rolling (%) Pass Number Sizing + 24 broadsizing Obs.: ≥ 50 – 60% a) Heaviest thickness reduction at last roughing pass; Roughing 70 b) Pass reductions are progressive along metallurgical ≥ 20% Finishing 75 roughing phase. Charles Hatchett Seminar Plate mill: an expert practical system 24
Is this a reasonable type of model? Proposed and other types of models Charles Hatchett Seminar Plate mill: an expert practical system 25
Type of models (not exhaustive) Wide plate mill There are several types • Physically based • Physically based + Numerical • Numerical models • Empirically based + Numerical • Empirically based http://www.danieli.com/products/Flat-Products- Hot-Rolling-Mills/Plate-Mills/PLATE-MILLS/Wide- Plate-Mill on April, 12, 2014. Charles Hatchett Seminar Plate mill: an expert practical system 26
Physically based models Among strong points… • They render a better understanding of the physical variables and metallurgical phenomena behind Predicted results showing evolution of net drive force of recrystallization with time. the process C.L. MIAO, C.J. SHANG, H.S. ZUROB, G.D. ZHANG, and S.V. SUBRAMANIAN, Recrystallization, Precipitation Behaviors, and Refinement of Austenite Grains in High Mn, High Nb Steel, METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, FEBRUARY 2012 — 66 Charles Hatchett Seminar Plate mill: an expert practical system 27
Physically based models …however… • They rely on many variables that are difficult to obtain with a specific degree of accuracy S. Hore, S.K. Das, S. Banerjee, S. Mukherjee, A multiscale coupled Monte Carlo model to characterize microstructure evolution during hot rolling of Mo- TRIP steel, Acta Materialia 61 ( 2013 ) 7251 – 7259. Charles Hatchett Seminar Plate mill: an expert practical system 28
Physically based + Numerical models Multiscale model Possible strong and weak points • Still provides better understanding however, • Difficulty in obtaining key variables persists and Obs: Model uses a continuum dislocation • It is usually time density evolution model coupled with a heat consuming transfer model integrated with a mesoscale Monte Carlo (MC) simulation technique. S. Hore, S.K. Das, S. Banerjee, S. Mukherjee, A multiscale coupled Monte Carlo model to characterize microstructure evolution during hot rolling of Mo- TRIP steel, Acta Materialia 61 ( 2013 ) 7251 – 7259. Charles Hatchett Seminar Plate mill: an expert practical system 29
Empirical + Numerical models Possible strong and weak points • Usually very easy to use • Lack of generalization • Must be tuned by Average prediction errors for MFS values calculated using equations available in the literature numerical algorithm Antonella DIMATTEO, Marco VANNUCCI and Valentina COLLA , Prediction of Mean Flow Stress during Hot Strip Rolling Using Genetic Algorithms, ISIJ International, Vol. 54 (2014), No. 1, pp. 171 – 178 Charles Hatchett Seminar Plate mill: an expert practical system 30
Recommend
More recommend
