HOT STRIP MILL OPERATIONS
Volume V
SCALE
BY: DAVID T. BLAZEVIC
President
Hot Rolling Consultants
Sun Lakes, Arizona, USA
December 2005
right 2006
David T. Blazevie
Rights ReservedBOOKS on
HOT STRIP MILL OPERATIONS
by
David T. Blazevic
Volume I - COILING - 1992 (no longer in print)
Volume I - ROLL COOLING, ROLLING PROBLEMS, and SHAPE - 1995
ROLLS, and ROLL LIFE - 2002
Volume Il - ROLL WEAR, BRO!
! Volume IV — BUILD UP or RIDGES — 2003
also known as
Loma, Franja, Floppy Center, Quarter-line Buckles and Short Choppy Edge Waves
Volume V - SCALE ~ 2006
Volume VI - DESCALING ~ Planned 2007
| Volume VII — ROLLED in SCALE — Planned 2008HOT STRIP MILL OPERATIONS, Volume V
SCALE and ROLLED in SCALE
TABLE OF CONTENTS
- DEDICATION 1
- PREFACE 1
1 SCALE 7
2 SCALE GROWTH 6 8-13
3 DEFINITIONS of SCALE & SCALE GROWTH 6 14-19
4 TEMPERATURE & TIME VARIATIONS on HOT MILLS 19 20-38
5 SCALE GROWTH ELONGATION and COMPACTION 10 39-48
6 SCALE GROWTH ATMOSPHERES 4 49-52
7 SCALE GROWTH INHIBITORS and ACTIVATORS 20 53-72
8 _ FISSURING, BLISTING, VOIDS, and De-LAMINATION 9 73-81
9 SCALE GROWTH - “K”, “B” and “Fe” FACTORS 5 82-86
10 TRIAL CALCULATIONS ~ FURNACE SCALE 8 87-94
11 TRIAL CALCULATIONS ~ SECONDARY SCALE 9 95-103
12 TRIAL CALCULATIONS - TERTIARY SCALE 19 104-122
13. SCALE GROWTH VARIATIONS & OPTIONS 10° 123-132
14 SUMMARY and CONCLUSIONS 3 133-135
- ADDENDUM to Volume V~ ‘SCALE’ 3 136-138
References, Appendix I, II, and III, and Author
- REFERENCES 4 1-4
- Appendix I SCALE GROWTH FACTORS ~ “K”, “Fe”, “B” 5-9
and PERCENT OXYGEN in each LAYER 5
- Appendix I - FIGURES, PHOTOS, GRAPHS, ete. 4 10-13
= Appendix 111 - HOT STRIP MILLS. 7 14-20
- ABOUT THE AUTHOR:
21HOT STRIP MILL OPERATIONS, Volume V
SCALE
DEDICATION
This volume is dedicated to Messers MH Davies, Mt Simnad, and CE Birchenall
(1951), S. Garber and G.M. Sturgeon (1961), L. Luyckx and F. Lorang (1965 — only
found by the author is 1986), H Hojas and R Rheindt (1965), and T. Sheppard and
W.M. Steen (1970) who wrote the best technical papers on scale and descaling.
These helped this author to learn more about scale, descaling and rolled in scale than
any others. It was through their efforts that I began to learn, and they put me on the
right path of basic science, to solve this complex problem. They gave me very
practical views into different aspects on this subject that I use to this day, and have
extended the knowledge in this book and the next two on scale. It seems so many
years ago these people were like Professors in college helping me on my way. I wish
to thank them for their help in making this volume and succeeding volumes possible.
COPYRIGHT RESTRICTIONS
Copying of all, or parts, of this publication is prohibited except with the written permission of the
author. Authors who use short phrases, may do so if these phrases are identified, and reference is
made to this publication. Authors who wish to use extended quotations, drawings, or pictures can
only do so with the express written permission of the author.HOT STRIP MILL OPERATIONS, Volume V
SCALE and SCALE GROWTH
PREFACE
This is the fifth volume on hot strip mill operations. ‘The first book was on Coilers and Coiling,
the second one on Roll Cooling, Shape, and Rolling Problems, the third on Roll Wear, Broken
Rolls, and Roll Life, the fourth on Build Up (Ridges) and other “Narrow Shape Distortions”,
and this fifth volume begins a series of three books on Scale, Descaling, and Rolled in Scale. The
first of these books will look at scale to understand the nature of the quality problems caused by
scale for the operators to concentrate their efforts in the right direction,
As with the other volumes, the references and acknowledgments give credit to the most important
contributors to this work. The author has read so many articles on scale and descaling, some
credits may be overlooked; but blame this on my old age at 79 and my memory is failing a little.
It is again very important to‘ recognize the help of many hot strip and plate mill operators,
engineers, metallurgists, the Association of Iron & Steel Engineers, the Iron & Steel Society, the
Rolling Mill Committee of the American Iron & Steel Institute, and the many hourly workers and
mill rollers, which enhanced the author’s view of this area of hot strip mill rolling over all these 47
years.
This volume has two primary objectives [1] UNDERSTANDING of SCALE FORMATION
and SCALE GROWTH, and [2] DEVELOPING EMPIRICAL SCALE GROWTH
FORMULAS. Volumes VI and VII will concentrate on Desealing and Rolled in Scale on hot
strip and plate mills needed to eliminate rolled in scale. If these two objectives can be
accomplished, the hot mill operators can concentrate their efforts in relation to scale growth
control; and then on the descaling designs and temperature control in the mills to eliminate rolled
in scale to further enhance the surface quality off of the hot strip and plate mills
The success of the first four volumes on Hot Strip Mill Operations allows the continuation of this
series with its many new and different views on scale, while introducing the physical causes of
scale growth and rolled in scale. The steel industry of America has produced some of the best
technology in the modern world of steel rolling and processing. This fifth volume hopes to
contribute more back to the hot strip and plate mill operators some of the knowledge their
predecessors imparted to the author in his many travels through hot strip and plate mills around
the world.
David T. Blazevic
December 2005
Sun Lakes, Arizona USACHAPTER 1 - SCALE
Every hot mill or hot processing operation involving steel has a problem with sea/e, but few of the
people really know what it is, how it forms, and what are its characteristics. It took the author
many years to find these things out, and yet they are basic to eliminating surface defects due to
rolled in scale.
Scale, in relation to hot rolling or forming, is an iron oxide that forms at high temperatures. The
scale layer consists of three layers:
FeO —called Wustite
4 FesO, — ~ called Magnetite
‘ FeO; — —called Hematite
All scale layers grow in three layers according to the principles of LeChatelier. The inner layer is
FeO, the middle layer is Fes, and the outer layer is FezOs (2)(5)(9). In many papers on rolled in
scale and scale formation there is a reference to these three layers, but not many photos or
descriptions. In addition, each technical paper describes how thick the layers are in their experi-
ments or studies. Each paper’s thickness data is different, but all are 100% correct because scale
formation is a function of time and temperature, to some degree on the chemistry and surface
condition of the sample, and atmosphere in which the seale is formed. So each investigation
will create its own distinct scale formation depending on the time and temperature used in the
experiments. This same characteristic of different scale layer formation can be seen by picking up
pieces of scale at the furnace descaler.
In this book, we are primarily concerned with scale formation and conditions as follows:
> Carbon steel from Ultra Low Carbon to Medium Carbon (0.50%), HSLA, Silicon, and
some Stainless steel grades
Surface conditions of slabs and the rolled surface in the hot strip or plate mills
> Time range for scale formation is from hours in a reheat furnace to minutes in a tunnel
furnace, and to seconds in the rolling process, but the time ranges will be specified for
various situations of scale growth
Atmosphere conditions inside the furnaces are usually 0% to 2% excess Oxygen, and the
rolled slab is exposed to atmospheric air with 20% Oxygen at 100% humidity,
Ae
v
The best way to start is looking at a piece of scale is shown in Photo | (4). This is a large piece of
scale shows the top surface at the lower right is a relatively smooth surface. This FeyOz scale
layer is indicated by the black arrow, which has been “melted” in the heating process. The outer
surface is always a silver-gray color. The other part (light colored arrow) of the photo shows the
1:04 scale layer at the Fe,03/FeyO, interface. This interface shows some signs of “melting” the
porous, crystalline structure of the FeO, scale layer, and this middle layer has shiny areas like
“stars” in a crystalline structure. The color will appear to be black or refract a blue color at
times, This large piece of scale illustrates how the scale fractures through all three layers in a
general polygon shape following the scale layer’s grain boundaries, and perpendicular to the
layers. Notice there are no signs of a chemical bond between these two layers of scale because of
the “lack of sheared crystals” of the interface.Photo 1 TOP (FeO3) & MIDDLE (Fe30,) LAYERS of REHEAT FURNACE SCALE (4)
Photo 2 shows the interface of the FeO layer with the steel surface. In this case the surface texture
is relative uniform, but it will follow the oscillation marks on a cast slab, grinding marks on a
ground stainless steel slab, and the scarfing marks on a conditioned carbon slab. The color ranges
from black to dark gray, and has a texture ranging from “fine” to “coarse” emery cloth. It is the
difference in color and texture, which allows you to easily recognize the outer (Fe,0s) and inner
(FeO) parts of the scale layer. Again there is no sign of a chemical bond between the lower scale
layer and the steel slab because of the “lack of signs of a shearing action” on the FeO crystals.
Photo 2 - FeO INTERFACE with STEEL (4)Another view of the interface structure is seen in Photo 3 below. This is the Fe,Qs layer’s inter-
face with the FesOs layer. It shows a very solid crystalline structure with peaks and valleys like
the Himalayan Mountains in Asia. It creates a mechanical interface bond that would be difficult
for the descaling sprays to separate from the middle layer of this multi-layer scale particle. This
inner surface of the outer layer of Fe2O3 has as silver-gray color, and the crystals do not show any
signs of being porous
Photo 4 is a view of the fractured face of the large scale particle shown in Photo | to show the
structure of all three layers of scale. At the top is an extremely thin layer (red arrow) of the
Photo 4 — SIDE VIEW of SCALE LAYERS (4)mind when later discussions will be made on descaling affects and certain rolled in scale defects.
Photo 7— PRIMARY PURPLE & RED OXIDE SCALE PARTICLES (4)
In addition to the physical characteristics of the various scale layers as shown in Photo 4 are [a]
the density and the amount of Oxygen (O;) contained in each layer, and [b] the melting points of
the scale layers.
LE CULE %OXYGEN %IRON DENSITY —
Wastite FeO 22.3% 71.7% — 5.10g/em* |
Magnetite FesO. 27.6% TA% — 5.18g/em*
Hematite Fe:O3 30.1% 69.9% — 5.24g/em*
Table 1— PERCENT and DENS
Table 1 is standard data on the percent of each type of scale and its density. The density is a
combination of the amount of Oxygen and the crystalline structure of the scale molecules. In the
study (2) on scale formation there was a table of the scale thickness formed from 750°C to 950°C
of FeO, which was converted to density for comparison purposes and shown in Graph | in the
DENSITY of FeO vs TEMPERATURE.
7.0
60
5.0
4.0
3.0
20
1.0
0.0
Density (g/Sq.cm.
750°C 800°C 850°C 925°C 950°C FeO Fe304 Fe203
Temperature FeO Created Standard Data
Graph 1 — SCALE DENSITY VARIATIONS (4)
6.
five black columns to the left. The standard accepted density is shown in the three columns to the
right. The temperature at which the standard accepted density of FeO was created is not known,
but it is relatively close to the first three columns from 750°C to 850°C where the scale growth is
almost 100% FeO. At about 900°C the Magnetite begins to form and at 950°C Hematite begins to
form. The density in study (2) decreases at 925°C, but is close to the standard data. At 950°C
when both Magnetite and Hematite forms, the density does not match the standard data. This
illustrates how time, temperature, and the atmosphere used to create the scale can “skew” the
results.
Another factor in scale is the melting points. Almost all the particles out of a reheat furnace show
the outer surface is melted. After examining hundreds of scale particles at the furnace descaler on
many mills showed not only the outer surface of FexOs was melted, but also the interfaces
between the upper, middle and lower layers were often melted, The accepted standard melting
points and typical furnace flame temperatures are shown in Graph 2 below. The flame temper-
ature is shown using cold air and hot air coming from a recuperator to show the variations that
occur in different furnaces.
MELTING POINTS & FLAME TEMPERATURES. |
1900
1700
other!
Steel Flame- Flame -
Cold Air Hot Air
TEMPERATURE (°C)
Graph 2 MELTING pba and FLAME TEMPERATURES (4)
The black dotted line is a typical maximum temperature in a reheat furnace for thick slabs, but
some mills run some of the zones as high as the blue dotted line temperature. The scale melting
points are higher than the furnace operating temperatures, but far below the flame temperatures.
The affect on the scale formation and these higher flame temperatures will be reviewed when scale
growth in a reheat furnace is discussed.
With this initial look at “Seale” it is time to look at some classical studies, and what they tell us,
and what they don’t tell us to guide us down a path to greater understanding.CHAPTER I~ SCALE GROWTH
There have been many studies on scale growth, but this author always went back to the basic study
in 1951 (2) because in many years of observation it has proven to be a good guide. LeChatelier
defined the principles of multiple scale layer growth, but sadly to say is most of the studies
avoided this complication by limiting their investigations to below 1000°C (2)(4)(8). It becomes
necessary to look at the data and comments from Reference (2) in detail
This study (2) was made using a polished iron (Fe) sample heated in an inert atmosphere to the
test temperature. Then pure Oxygen was introduced into the test chamber, and oxidation was
allowed to continue for various lengths of time. The sample was continuously weighed to
measure the amount of Oxygen added to the sample, as the scale grew thicker. At various time
intervals the Oxygen gained was recorded. The result was scale growth followed a time!”
parabolic curve for each temperature. The results are reproduced in Graph 3 for test temperatures
from 650°C to 975°C. The vertical axis is Grams of Oxygen gained per Square Centimeter of
exposed surface, and the horizontal axis is the Square Root of Time in minutes. So, 10 is a 100
minutes, 20 is 400 minutes and 30 is 900 minutes of exposure time to oxidation of the sample at
cach temperature.
SCALE GROWTH - Oxygen Gain
0 5 10 15 20 25 30
TIME (Min40.5)
Graph 3 — SCALE GROWTH versus TIME (2)
This exponent “1/2” of time for scale growth has been questioned in references (5) & (19) because
this study assumes only a FeO scale layer is created. This assumption is true in the areas where
study (2) was primarily made in the 600°C to 950°C temperature range.
Another part of the experiments determined the percent of each scale layer growth rate by
temperature as shown in Table 2. This data was re-arranged into groups in relation to the exposure
time for oxidation at various temperatures. The data shows the percent of each scale layer stays
about the same regardless of the exposure time, but there is a distinct difference in the test data
above 900°C. At 975°C the percent of FesO4 more than doubled which indicates it seems the
conversion of FeO to FesO, requires more time. The other facet is most of the test data above
975°C was done for very short time periods from 25 to 64 minutes. This is in contrast to slabs
heated in a reheat furnace formlng significant amounts of Fe2Os and Fes at heating times of 150
to 180 minutes (4). Of significant importance is the conclusion on page 893 that “The results
indicate that in no case above a temperature about 625°C did the scale contain greater than 10% by
8.
Volume of higher FexOs and Fe20* (2). 10% can be considered a significant volume up to
1080°C.
{Temp | Min. [ % of Scale % of Scale] Min. | % of Scale
(°C) |110&<| FeO Fe,0, Fe,0,| 225 | FeO Fe,0, Fe,0,| 900 | FeO FeO. Fe,0.
| 600 100 50 50 O} 225 50 50 900 55 45
625 110 70 30 O} 225 7 25
700 225 98 2 900 97 3 oO
750 100 96 4 oj 900 94 6
800 100 96 2 2) 225 92 4 4| 900 98 2 <1
850 225 96 4 <1} 900 96 4 <4
900 100 30 9 <1) 225 90 9 <1] 900
975 100 96 3 <1] 225 96 3 <1} 900 92 7 <1
1000, 25 95 5 <1) 225 96 2 <1] 900
1035 | 49 9 4 <1
64 95 4 <1
1090 | 25 4 865 (Cf
49 94 5 <4
Table 2— PERCENT of SCALE LAYERS versus TEMPERATURE (2)
may be the difference is in the “steady state” of these experiments of using small samples and
the scale layer at a constant temperature in contrast to a furnace operation where the flame and
fumace temperature must be hotter than the steel to transfer heat. This means the le layer is
hotter than the steel in the reheating furnaces. This will be discussed when scale formation in
reheat furnaces is covered.
Figure 1 - CHEMISTRY of SCALE FORMATION (2)
9The chemical formation of scale is shown in Figure 1 (2), but translating it may provide a clue as.
to some of the problems of using “Oxygen weight gained” as a prediction of iron loss or scale
thickness. Scale is formed by negative Oxygen ions contacting positive Iron ions to form FeO
first. As many authors have stated: “that all scale formed is FeO” is 100% correct, but is all scale
formed only This quotation would violate the law of multiple layer oxidation principles of
LeChatelier (8), and to say the scale particle shown in Photos | to 4 do not exist. The answer is
“that all scale formed is 100% FeO is correct” because it is the basic form of iron oxide, but it
converts into three layers containing more Oxygen per unit of iron. The conversion of FeO to
FeO, reduces the FeO layer as it grows, and the same thing happens as the conversion of FesOs
to Fe:0s (26). Over a period of time all the FeO will be converted to FeyOs and disappes
leaving the remaining outer two layers to become “iron ore”. The phenomenon of the scale layer
growth becomes of importance to the hot mill operators only in relation to the “scale loss” or
rather the “iron loss” in reheating operations, and in rolled in scale problems.
The data shown in Table 2 above was expanded and added to from a variety of sources
(2)(4)(5)(11) to make an approximate graph of the scale layers if they were created at a specific
temperature. This data was converted into Graph 4 (4) is not 100% correct, but in practice it has
proven to be close and effective enough to predict the scale structure under varying conditions.
SCALE LAYERS by TYPE vs TEMPERATURE
100
80
70
5 eeu
& 60
& 50
cr eet sa
a
40 ‘
30
20
: ~~
300 400 500 600 700 800 900 1000 1100 1200 1300
TEMPERATURE (°C)
Graph 4 — PERCENT of SCALE LAYERS versus TEMPERATURE (2)(4)
This graph shows many things. First, most of the experimental data was conducted from 650°C to
950°C where the vast majority of scale formed is FeO, with minor percents of the other layers that
confirms the statements in reference (2), but this cannot be applied to all temperatures. Second,
almost all reheating and rolling occur above 800°C, and most of the heavy reduction during rolling
occurs above 950°C where the percent of each scale layer’s thickness varies considerably. Third,
is the temperature of the scale is higher than the slab temperature in a reheating process, and lower
than the slab or strip temperature in the rolling process, which requires some adjustments in
calculating the scale growth,
10.
‘The scale growth factor is called the “K” factor, which is the Oxygen Weight Gained by the
sample as it oxidizes as Oxygen molecules are added. The basic data (2) has been rearranged by
the test temperatures in the first column, and into scale groups as shown in the second group of
columns in Table 3. The “K” factor for FeO scale approximately doubles from 850°C to 950°C.
The “K” factor for FesO4 scale increases six times from 800°C to 986°C. The “K” factor for
multi-layer scale growth is shown from 400°C to 1200°C, but to compare growth rate changes to
the first two types of scale, the growth rate at 800°C triples at 965°C. This table includes the only
data on temperatures and scale growth at higher temperatures. The “K” factor increases six times
(600%) from 800°C compared to the growth rate at 1200°C.
Temp. K Factor|Temp. K Factor
ec__Feo | °c | FeO Fes0s Multi.
850 0.000230] 400 0.000003
910 0.000330 450 0.000011
930 0. cosa 500 0.000081
950 0.000490, 600 0.00009
625 0.000196
Fess | 650 0.000306
00 0.000045] 700 0.00060¢
900 0.000150 750 0.001106
950 0.000130 800 0.000045 0.001851
986 0.000300 850 | 0.000230 0.003001
900 0.000150 0.00390
Multi. | 910 |0.000330
400 0.000003 930 |0.000390 0.004601
450 0.000010 950 |0.000490 0.000130,
500 0.000050| 965 0.005501
600 0.000090|_986 0.000300
625 0.000190] 1035 0.00900
650 0.000300] 1090 0.01100
700 0.000600] 1130 0.012401
750 0.001100] 1200 0.018701
800 0.001851
850 0.003001
900 0.003901
930 0.004601
965 0.005501
1035 0.00900
1090 0.01100
1430 0.01240
1200 _ 0.01870
Table 3- OXYGEN WEIGHT GAIN (“K”) FACTORS (2)
The growth rates for three different types of scale are in the three columns to the right. Here the
values for each test temperature try to compare the different growth rates. The growth rates in the
800°C to 986°C range are quite different for the FeO and Fs, scale layers compared to the mu
seale layer’s growth. These figures are shown in Graph 5 plotted on a logarithmic scale. The data
in the last column was combined with other data on scale layers to develop the Oxygen Weight
Gain (“K” Factor) for temperatures from 300°C to 1300°C as shown in this graph.
iatOXYGEN WEIGHT GAIN
on
E
5 oot we
& oo?
E
2 ooo
z
z oer
8 soon a a
a a
g —F e304 Layer
one —& FeO Layer
—Mult.Layer
0.000001
400 60070050930 086130
TEMPERATURE (oC)
Graph 5 - OXYGEN WEIGHT GAIN (K) FACTORS (2)
Graph 6 is a composite of this data (2) that was extrapolated from 300°C to 1350°C to be used at
the rolling mill temperatures occurring from 500°C to 1300°C. The graph is on a logarithmic scale
OXYGEN WEIGHT GAIN ("K" Factor) - 2004
0.100000
0.010000
&
.
0.001000
0.000100
OXYGEN WEIGTH GAIN
0.000010
300 400 500 600 700 800 900 1000 1100 1200 1300
TEMPERATURE (°C)
Graph 6 - OXYGEN WEIGHT GAIN (“K” FACTOR) (2)(4)
to put in perspective. Notice there curves or changes in the Oxygen weight gain factor near 600°C
* 12.
and 1100°C. It is in the middle area between these two temperatures that mostly FeO occurs, and
toward cach end more multi-layered scale layers are being created.
From this excellent study (2) many years ago we have a good idea of the scale growth at various
temperatures, and the ability to translate this data into reasonable scale growth on hot rolling mills.
13CHAPTER 3 — DEFINITIONS of SCALE and ROLLED in SCALE,
Before proceeding it is necessary to introduce some definitions and terms used in the steel industry
in relation to scale and rolled in scale, On different mill configurations there are different terms
requiring some definitions. These are definitions made by the author or names (22), which are
those used mostly in the United States. Each country and mill will have its own definitions and
names, and later in the book Rolled in Scale, sketches and photographs will be provided to
identify any individual mill’s rolled in scale problems.
SCALE is a term used for the scale formed in the furnaces or during the rolling process
OXIDE is a term applied to the scale on a plate or coil surface when it leaves the rolling
mill
PRIMARY SCALE is the first scale (hence the term primary”) formed in the reheating
furnaces, In actual fact there is a thin layer of scale formed on the cast or rolled slab prior
to coming to a rolling mill, but this thin layer flakes off or combines with the new scale
growth in the furnace.
SECONDARY SCALE is the second layer (hence the term “secondary”) of scale formed
during the rolling process after the primary scale has been removed by a hydraulic
descaling system
TERTIARY SCALE is a more modem term applied to the last layer of scale (hence the
term “tertiary”) formed after the secondary seale has been removed by the descaling
system ahead of the finishing mills
OXIDE is a common term applied to the scale on the finished coil, but this is actually the
tertiary scale formed after the finishing mill descaler, and should be translated as such for
any explanation in this book
COMPACT MILLS can have these three types of scale if they have a roughing mill
ahead of the finishing or Steckel mill, but many roll directly from a tunnel furnace where
the first scale growth occurs. Then the slab is descaled and is rolled in the finishing mill.
For purposes of comparison for the compact mills without roughing mills the scale out of
the tunnel furnace is similar to heavy secondary scale, and the scale out of the finishing
mill descaler is similar to tertiary scale, When discussing scale or rolled in scale in this
book, the compact mill operators will have to interpret as to which scale formation is
similar to theirs.
PICKLING or SHOT BLASTING are processes to remove the tertiary seale from the
strip or plate for further processing. ‘These processing units need a uniform tertiary scale
(oxide) thickness on the strip or plate to maintain a high surface quality and production
rates, and will complain if they must process the coils at a slower rate.
ROLLED in SCALE is a general term for a piece or particle of scale embedded in the
steel surface creating a surface defect or condition, which is not a smooth and uniform
thickness tertiary scale laver after the final descaling is made ahead of the finishing mill.
The rolled in scale can be unacceptable because the particle is too deep and reduces the
14.
strip thickness, cannot be pickled out, requires slow or multiple pickling, and/or results in
an unsatisfactory surface appearance when painted.
+ ROLLED in SCALE DEFECTS get their names from the source, shape, color, when
related to the mill operation, or what happens on the next processing unit. This is the list
of specific rolled in scale defects as of June 2006, but knowing the mill operators they will
find a way to get into rolled in scale problems in some new way.
* DE-LAMINATION describes the separation of the scale layers either in the reheat
fumace, or due to mechanical fracturing the upper layer(s) during the width reduction by a
vertical edger (12i), and this definition does _not_agree with the conclusion of de-
lamination occurring as the scale layer cools down (24).
IARY ROLLED in SCALE
. Primary Rolled in Scale consists of all three layers of the scale formed in a reheat furnace
not removed by descaling ahead of the 2"! reduction pass.
. Primary Purple Oxide Scale has a purple color to give it its name, and consists of the
middle (FesO,) and inner (FeO) layers of the scale formed in the furnace overlaid with a
very thin layer of FeO formed after the original layer of FezOs formed in the furnace was
removed by descaling.
. Primary Red Oxide Seale has a reddish or rusty color to give it its name, and consists of
the inner (FeO) layer of the scale formed in the furnace overlaid with a very thin layer of
Fe,Os formed after the original layers of Fe:0s and Fes, formed in the furnace has been
removed by descaling,
|. Primary Head End Seale is primary seale, which consistently appears on the head end of
the plate or coil
Primary Tail End Scale is primary scale, which consistently appears on the last 1/4th of
the plate or coil.
Primary Scale Streaks is primary scale, which is a band of scale along the rolling
direction caused by mis-aligned or plugged descaling nozzles.
Primary Plugged Nozzle Scale is primary scale, which was not removed by the descalers
due to a plugged, partially plugged, or damaged nozzles.
Primary Rebound Scale consists of primary seale particles, which are removed by the
descaler, and then falls back on the slab surface to be carried into the roughing mill
Vertical Edger Collar Scale is primary purple or red oxide seale,which was on the slab
near the edges and is rolled into the slab when the slab “dogbones” (bulges upward and
downward) by the edger’s reducing the slab width forcing the scale down between the slab
and the edger collar.
0. Scale Pits (Stainless) are small particles of the original primary seale out of the furnace
15where the outer (Fe;O;) layer was removed by the descaler, but much of the two lower
layers remained on the slab. In the reduction process with a reversing rougher and
Steckel mill, more and more of these particles flake off the surface, but many small
particles were spread out to create scale pits which required slow or extra pickling to
remove the scale,
11. Primary “Knife Like” Gouges is the result of rough table rolls gouging the bottom of the
slab in the roughing mill due to a speed difference in the slab and the table roll’s speed.
12. “Hang Nail” slivers are gouges with entrapped scale cause by “ragged”, “fluted”, or
“grooves” cut into the first roughing mill rolls, which “slices” the steel surface in the
reduction process to make a steel sliver in which scale is entrapped.
13. Primary Rolled in Refractory occurs when refractory material falls on a slab in the
furnace, and gets combined with the scale and cannot be removed by the descaler.
14, Rolled in Kerf is cither slag from scarfing particles or the cutting of the slabs at the caster
with an Oxygen torch leaving a semi-steel/seale particle or bead on the slab that cannot be
removed by descaling
15. Gaul is created on Stainless Steel slabs where particles from carbon steel scale falls off the
top of the edger rolls and/or housing, and is rolled into the stainless steel slabs.
16. Furnace or “Sticky” Scale is a term that is no longer used. This type of scale has been
traced to Purple & Red Oxide Scale. (See 3.2 and 3.3 above)
SECONDARY ROLLED in SCALE
17, Secondary Scale consists of scale particles formed in the roughing mill (and scale formed
in a tunnel furnace on a compact mill without a rougher), which is not removed by the
finishing mill descaler.
18. Secondary Purple Oxide consists of the two lower layers of the scale formed in the
roughing mill or tunnel furnace on a compact mill that is not removed by the finishing
mill descaler.
19, Secondary Red Oxide consists of the an area or patch where stray water impacting the
slab between F1 & F2 mill cools to “freeze” a thin layer of Fe203 on the outer surface of
the tertiary scale layer created after the finishing mill descaler.
20. Secondary Head End Scale is created by the descaler effectiveness being decreased by
threading at a high speed, or an undulating strip on a coil box mill.
21. Secondary Seale Streaks are created by a gap between the descaling sprays due to poor
fabrication of the headers, or an “undulating” strip moving out of the pass line.
22. Secondary Plugged Nozzle Scale is created by a plugged, partially plugged, or damaged
nozzle in the finishing mill descaler.
16.
23. Secondary Rebound Scale results from removing the secondary scale in the finishing mill,
descaler, but allowing the tiny particles to “rebound” backwards into FI mill to be rolled
into the strip.
24, Scratches & Embedded Scale results from scratching the bottom surface and folding the
edges over to entrapped some secondary scale particles.
25. Gouges & Embedded Scale results from gouging the bottom surface and folding the
edges over to entrapped secondary scale.
26. Deflector Plate Scratches & Scale occurs in the finishing mill descaler hood when the
“deflectors” are allowed to wear too much and begin to scratch and entrapped secondary or
tertiary scale in these tiny scratches.
27. Drag Seale is a term used in the old days, but now these types of scale are classified as
Scratches, Gouges, or Looper Roll Scale.
TERTIARY ROLLED in SCALE
28, Heat Pattern (Wave) Scale got its name in the 1950's because it only occurred on thicker
strip that was hotter when rolling in the mill. The term “Waves” comes from the pattern
of the scale particles seen on the strip. This scale defect occurs due to excessive tertiary
scale growth caused by high strip temperatures after descaling ahead of the finishing mill
and the distance from the last descaler to F1 mill.
29, Salt & Pepper Scale is a finer and smaller relative of Heat Pattern Seale, which looks
like tiny specks of pepper on the strip surface.
30. Roll Wear Scale is another variation of Heat Pattern Seale, which occurs when a work
roll surface begins “peal” in bands, and is traced to the same pattern of work roll wear
bands usually in FI or F2 mill.
31. Spiral Roll Wear Scale (Stainless) is a variation of Roll Wear Scale on certain stainless
steel grades where the side flow during the reduction process produces a spiral pattern on
the edges of the work roll. These spiral ridges cause the tertiary scale to fracture and be
rolled into the strip, which requires a second pickling to remove the scale. This is similar
to Intermittent Scale Streaks (See definition 32 below), but is in curved lines instead of
being parallel to the rolling direction.
32. Intermittent Scale Streaks is another variation of Roll Wear Scale where the tertiary
scale layer is fractured in lines parallel to the rolling direction by the protrusions (ridges)
occurring on chrome and high speed steel work rolls as they become worn.
33. Black Digs is a variation of secondary rebound scale, which occurs when a descaler is
installed close to F1 mill (with the typical 15° lead angle) to remove some of the tertiary
seale formed after the finishing mill descaler and rebounds many tiny particles into F1 mill
to be rolled into the strip surface.
1734, Looper Roll Scale occurs when a looper roll is not turning freely, and it scrapes the
tertiary scale off the strip into a pile on the looper roll; and when the roll does begin to
turn, this pile of scale is picked up by the strip and rolled into the surface in the next mill
stand. This is similar to drag scale, which occurs in the roughing mills (see 26 above).
35. Furnace Coiler Drum Scale (Steckel mills) occurs when particles of the tertiary scale
are pulled off the strip onto the furnace drum, and slowly more scale layers are added until
it becomes a thick pile of many thin scale layers. This multiple layers of scale are pulled
off the drum and sticks to the strip to cause this defect.
36. Tunnel Furnace Seale (Compact mills) is similar to Furnace Coiler Drum Scale that
forms on the metal “collars” on the tunnel furnace rolls until a pile eventually sticks to the
strip to become so embedded the finishing mill descaler cannot remove these particles
37. Deflector Roll Scale (Steckel mills) also similar to Furnace Coiler Drum Scale where
the scale sticks to the defector roll, and when it is pulled off the roll, it is rolled into the
strip on the next rolling pass.
38. Ghost Spots (Steckel Mills) are light colored spots on the strip where the scale layer was
removed at the deflector roll, and indicates the strip may have Deflector Roll Scale on the
strip.
39. Plateaus & Valley Seale were found on a plate mill, which follows a “controlled
temperature” rolling practice where the plate was held to decrease the temperature below
the molecular transformation temperature. A heavy scale layer would be formed because it
took some time to reduce the temperature of the plate. The strip was sent back into the
mill without descaling, and this heavy scale is fractured into particles 90° to the rolling
direction, and some larger patches just break free of the surface. The valleys were created
by the 90° scale particles being embedded into the plate and separated by the narrow ridges
of clean “extruded steel” between these particles. The plateaus are where patches of the
scale broke free of the plate leaving a wide area of scale-free surface.
40. Stripe Scale is heavy tertiary scale formed in narrow bands where the strip is much
hotter. These narrow hot bands can be because the descaling sprays are not overlapped
allowing the gap between the sprays to remain hotter, or because the descaling sprays are
almost 100% overlapped with a band between the two adjacent sprays that is hotter.
41, Flaky Seale is a term used for scale that flakes off a finished coil or plate after it is cooled,
and is created by finishing at high temperatures allowing the scale to continue to grow as it
cools down. Flaky scale is often in bands because the strip temperature out of the mill is,
not uniform across the width, or the runout table sprays result in uneven cooling in bands.
42. Contraction Gouges are scale defects, which occurs inside a coil while it is cooling. A
combination of high coiling temperatures and a tightly wrapped coil will make intimate
contact between the adjacent wraps of the coil, and the two outer layers of Fe,O3 combine
to form a bond between them (10)).
PROCESSING of STEEL
18.
a plate or coil is finished off the hot strip mill it either processed inside the steel plant, or
to a customer. Here the processing equipment will bend, shape, or distort the steel to the
ited shape needed in the process. When this occurs, the rolled in scale problems may first
‘ome visible. For example, Flaky Seale, Contraction Gouges, Plateaus and Valleys, Black
and Secondary Rebound Scale defects are not seen until they have been processed. The
on processing units are listed below:
+ Pickle Lines unwind the coils and removes the scale with acid to remove all the tertiary
seale on the surface.
‘Temper Mills unwind the coils and temper rolls the strip to improve the shape and/or
hardness of the steel
Heat Tempering Lines take plates and put them into furnaces to temper them to get a
specific hardness.
Slitting & Shearing Lines take plates or coils and side trim, slit, or cut to length as
needed, and often two processes can occur the same line.
Stretcher Leveling Lines unwind a coil and flatten the strip by stretching it beyond its
elastic limit to remove minor shape distortions.
19CHAPTER 4- TEMPERATURE & TIME PARAMETERS
on HOT ROLLING MILLS for SCALE GROWTH
The previous chapters looked at various aspects of scale formation, growth, scale layers, etc., but
now it is necessary to look at the specific temperatures occurring on hot rolling mills in the various
stages of the process in relation to the mill configuration. It is necessary to distinguish between
the average slab or strip temperature, and the temperature of the scale layer. For example, a
modem furnace computer will calculate the “average” slab temperature at 1250°C coming out of a
furnace with the soak zone set point at 1300°C, but the outer layer of scale is melted smooth,
which can only occur above 1500°C, The same thing holds true in the rolling process in the
roughing and finishing mills. The surface is chilled by the cold descaling water, contact with the
colder work rolls, and exposure to the lower air temperature between passes.
A. TEMPERATURES when REHEATING THICK SLABS
The temperature of a slab usually enters the furnace at the ambient air temperature and rises to the
rolling temperature in 2-1/2 to 3 hours. For purposes of this paper to calculate the scale growth, a
36-meter long furnace requiring 3 hours to heat 225mm thick slabs will be used. The furnace
configuration is of a six-zone (or 8 zones if the soak zone is split into two control areas for tapered
FURNACE & TOP SURFACE TEMPERATURES.
1400
‘200 Furnace
4000
g
600
TEMPERATURE (°C)
g
200
3 6 9 2 15 18 2 2 27 30 33 36
LENGTH (meters)
Figure 2— FURNACE CONFIGUATION and TEMPERATURES (3)(4)
heating) furnace as shown in Figure 2. The Preheat Zone uses side-fired burners in the top and
bottom zones when the mill is rolling fast, and are shut off “in pairs” when not required to
maintain the rolling temperature in the slabs. The Heat Zone uses “backfired” burners for a more
uniform heating of the slab from the head to tail end. The Soak Zone uses “forward-fired” burners
in the bottom zone, and roof bumers in the top zone to provide uniform heating. The slabs are
5 20.
own from the charging end on the left side to the discharge end on the right side in this sketch,
colors of the slabs show an increasing temperature during the heating process. The two
curves overlaid on the sketch show the furnace temperature in red, and the top scale surface
‘temperature in orange.
Notice the top surface temperature of the slab is higher than the top furnace temperature when it is
‘exposed to the high flame temperatures as discussed in Chapter | where the outer layer of Fe:O3
yrbs and transfers the heat quickly. This outer layer of FeyOs is caught between the high flame
tures and the two porous insulating layers of scale next to the slab, which results in a
igher temperature on the slab’s scale surface compared to the furnace temperature. The scale
yer directly “under” or “over” each burner’s flame will be even hotter than shown here.
iph 7 illustrates the slab’s temperature in the furnace as they are heated. The red curve is the
ace temperature, and the orange curve is the top surface temperature as shown in Figure 2. A
ind curve (blue) has been added to show the slab’s “center” temperature when being heated.
FURNACE, SLAB & SURFACE TEMPERATURE
1400
1200
1000
800
TEMPERATURE (°C)
8
&
8
0 5 10 15 20 25 30 35
DISTANCE (Meters)
Graph 7 — FURNACE and SLAB TEMPERATURES (3)(4)
There is another temperature called “slab temperature”, which is the calculated average
perature of the slab in a computer-controlled furnace. These definitions will help in
lerstanding the differences.
> Furnace Temperature is the thermocouple temperature in each zone in the furnace, and
measures the approximate zone temperature depending on where the thermocouple is
21located and the burner flame locations. ‘The furnace temperature is not [a] the slab
temperature, [b] the slab surface temperature, [e] the “average” calculated slab temper-
ature; but [d] just each zone’s “thermocouple” temperature.
> Average Slab Temperature is a computer calculated average slab temperature from the
surface to the center of the slab including the cooling affects of the skid pipes, and does not
represent the slab’s surface or center temperature,
.
Slab Surface Temperature is the temperature of the outer scale layer balanced with the
internal temperature of the slab. That is, in order for heat to be transferred to the slab, the
outer surface must be hotter than the steel inside the slab.
Slab Center Temperature is calculated using the furnace thermocouple temperature data
by the furnace computer temperature of the slab’s center temperature for theoretical
discussions, and computer adjustments during the heating process.
v
» Scale Formation Temperature is the temperature of the scale layer, and is normally [a]
higher than the furnace temperature, and [b] higher than the average slab temperature
This chapter is interested in the Seale Formation Temperature needed to develop the scale
growth, and because the exact temperature of the scale is not known, there will have to be some
assumptions and extrapolations made to approximate the Seale Formation Temperature in the
furnaces and rolling mill.
The Scale Formation Temperature of the top surface shown by the orange curve is
approximated based on heating principles and years of observations in the furnaces. The slab
comes in cold with a very thin layer of scale on the surface. The slab increases in temperature
rapidly and expands in all directions, and causes much of this thin scale layer to “flake” off as can
be seen on the slab’s top surface just inside the furnace, and as indicated by the huge piles of
loose, un-melted scale “flakes” in the bottom preheat zone at the entry end of the furnace. The
surface temperature increases rapidly, but the center temperature (blue curve) “lags” the surface
temperature according to the formula derived from reference (3).
2 (Thickness in Inches? / 10) x (4/ The number of sides heated) = Equalization Time in
Minutes or 97/10 * 4/2 for a 225mm (9") slab = 16.2 minutes for the center temperature to
reach the same temperature as the surface (3)
@ Translating this formula into Metric terms changes to (Thickness in mm’s / 25.4)? / 10 * 4
/ the number of sides heated = Equalization time in Minutes (4).
The slab’s center temperature is shown by the blue curve in Graph 7 has a consistent “time lag”
up until the slab’s surface temperature reaches about 800°C as indicated by the black dot. At this
point the iron (Fe) begins to change its molecular structure, and requires additional heat and time
to occur. The vertical dotted arrow points to the temperature in the center when this
transformation begins to occur at the surface. The surface temperature will then continue to
increase faster than before as this transformation continues. The center temperature can now only
increase depending on the temperature differential at the molecular change area of +/-800°C and
the center temperature starting at about 450°C. As long as the “heat” is needed for the molecular
change, the center temperature will “lag” further behind the top surface temperature as indicated
by the horizontal dotted arrow. After all the steel molecules have changed, the center
. 22