THICKNESS PROFILE GAUGE

[0001] This patent application claims priority to and benefit of U.S. Provisional

Application No. 62/741,711, filed October 5, 2018, which is herein incorporated by reference.

BACKGROUND

[0002] In continuous casting of thin steel strip, molten metal is cast directly by casting rolls into thin strip. The shape of the thin cast strip is determined by, among other things, the surface of the casting surfaces of the casting rolls.

[0003] In a twin roll caster, molten metal is introduced between a pair of counter-rotated laterally positioned casting rolls, which are internally cooled, so that metal shells solidify on the moving casting roll surfaces and are brought together at the nip between the casting rolls to produce a thin cast strip product. The term“nip” is used herein to refer to the general region at which the casting rolls are closest together. The molten metal may be poured from a ladle through a metal delivery system comprised of a moveable tundish and a core nozzle located above the nip, to form a casting pool of molten metal supported on the casting surfaces of the rolls above the nip and extending along the length of the nip. This casting pool is usually confined between refractory side plates or dams held in sliding engagement with the end surfaces of the casting rolls so as to restrain the two ends of the casting pool.

[0004] The cast strip passes downwardly through the nip between the casting rolls and then into a transient path across a guide table to a pinch roll stand. After exiting the pinch roll stand, the cast strip passes into and through a hot rolling mill where the cast strip may be modified in a controlled manner, typically by reducing the thickness of the cast strip.

[0005] When a cast strip is reduced in thickness via axial compression, the strip will also experience a transverse expansion. The direction and amount that it expands is defined by the Poisson’s ratio of the material and the tension applied. Based on tension typically applied and the geometry of cast strip in the rolling process, this results in the horizontal expansion being almost entirely in the direction of rolling (length direction). This expansion is referred to as elongation. The percentage that the material is elongated is proportional to the percentage that it is reduced in thickness. If cast strip is reduced in thickness by the different amounts across the width of the strip (perpendicular to the rolling direction), then this will result in differential elongation of the cast strip along the length of the strip.

[0006] The different elongations are still part of the same piece of sheet metal, however, and the portions that have been elongated more are constrained by the less elongated portions. This will create a stress in the material that will eventually produce a“buckle” in the material when tension is removed from the sheet metal.

[0007] Another flatness defect is known as bending marks, which cause short wave thickness variations in the cast strip in the 4-7Hz frequency range. The thickness variations increase in amplitude as the bending marks get stronger. As the non-uniform rolling increases, it will start to create a tight buckle coming out of the mill before pinchers start to appear and eventually the strip will tear apart and cobble. The term“pinchers” refers to a buckle on the entry side of the rolling mill that becomes so large that it folds over on itself while going through the work roll.

[0008] Various controls have been developed for shaping the work rolls of the hot rolling mill to reduce flatness defects. For example, work roll bending cylinders have been provided to affect symmetrical changes in the roll gap profile central region of the work rolls relative to regions adjacent the edges. The roll bending is capable of correcting symmetrical shape defects that are common to the central region and both edges of the strip. Also, force cylinders can affect asymmetrical changes in the roll gap profile on one side relative to the other side. The roll force cylinders are capable of skewing or tilting the roll gap profile to correct for shape defects in the strip that occur asymmetrically at either side of the strip, with one side being tighter and the other side being looser than average tension stress across the strip.

[0009] Another method of controlling a shape of a work roll (and thus the elongation of cast strip passing between the work rolls) is by localized, segmented cooling of the work rolls. See, for example, U.S. Pat. No. 7,181,822, which is incorporated by reference. By controlling the localized cooling of the work surface of the work roll, both the upper and lower work roll profiles can be controlled by thermal expansion or contraction of the work rolls to reduce shape defects and localized buckling. Specifically, the control of localized cooling can be

accomplished by increasing an amount of time a pulse width modulated valve is open, thereby effectively increasing the relative quantity of coolant sprayed through nozzles onto the work roll surfaces in the zone or zones of an observed strip shape buckle area, causing the work roll diameter of either or both of the work rolls in that area to contract, increasing the roll gap profile, and effectively reducing elongation in that zone. Conversely, by effectively decreasing the relative quantity of the coolant sprayed by the nozzles onto the work surfaces of the work rolls causes the work roll diameter in that area to expand, decreasing the roll gap profile, and effectively increasing elongation. Alternatively or in combination, the control of localized cooling can be accomplished by internally controlling cooling the work surface of the work roll in zones across the work roll by localized control of temperature or volume water circulated through the work rolls adjacent the work surfaces. While such controls were known, such controls were typically operated manually and without real-time feedback of the existence of flatness defects.

[0010] Attempts to directly measure strip flatness downstream of the hot rolling mill have been found to be unsatisfactory to achieve practical control of the hot rolling mill. The high temperature of the cast strip at the exit of the hot rolling mill makes measurement of the strip flatness by direct contact difficult.

[0011] For example, measurement of differential tension across the width of the strip has been attempted to provide a means of detecting flatness. Typically, a physical device (often referred to as a“shape meter” roll) is placed in line with the sheet. The sheet should, as part of the process, have some deflection (or wrap angle) around the roll, and be under tension. The device typically measures differential tension across the width of the roll via either displacement or force measurement. Areas of low tension indicate where the buckles are. However, the devices used to measure differential tension across the width of strip tend to be very expensive. Further, they typically do not last long in a hot rolling environment.

[0012] Non-contact optical methods for flatness measurement have been used. Some measurement devices use either optical or radiological detection methods in a stereoscopic manner to detect the height of buckles in a strip. However, optical devices rely on the buckle to be visible. Such non-contact flatness measurement results in partial flatness measurement, since at any given time only part of the strip exhibits measured flatness defects. When the material is under tension it will be elastically deformed. This will tend to hide a buckle, which prevents optical detection until the flatness defect becomes very large.

[0013] What is needed is a system and method for determining flatness of a cast strip metal product that is sufficiently robust to survive a hot mill environment and detect properties which may cause flatness defects once tension is relaxed, even if flatness defects of the sheet metal are not optically detectable upon exiting the hot mill. This measurement can then be used to automate certain aspects of the rolling process to produce a product free of flatness defects.

SUMMARY

[0014] A system for controlling a plant having a rolling mill for producing thin strip product is provided herein. The system includes a thickness gauge and a controller. The thickness gauge is positioned at the exit of the rolling mill to make thickness measurements of the thin strip product at a plurality of locations across a width of the thin strip product. The controller is coupled to the thickness gauge and configured to receive the thickness

measurements, to process the thickness measurements to detect oscillations in the thickness of the thin strip product corresponding to a plurality of control locations, and to detect flatness defects in the thin strip product based on the thickness oscillations.

[0015] The controller may be further configured to determine a phase of a thickness oscillation at the each of the plurality of control locations and to detect flatness defects based on differences in phases of the thickness oscillations. The controller may be further configured to identify control locations having leading phase values as indicative of flatness defects.

[0016] The controller may be further configured to determine an amplitude of the thickness oscillation within a given frequency range at the each of the plurality of control locations and to detect flatness defects based on a magnitude of the amplitudes of the thickness oscillations. The frequency range may be 4-7 Hz. The controller may be further configured to identify control locations having higher amplitude values as indicative of flatness defects.

[0017] The plant may comprise a continuous twin roll casting plant and the rolling mill may have work rolls and a plurality of valves and nozzles to provide segmented water spray cooling to the work rolls, each spray area for a nozzle comprising a control location, wherein the controller actuates the valves to differentially cool the work rolls in response to detected flatness defects. The controller may be further configured to determine a phase of a thickness oscillation at the each of the plurality of locations corresponding to the nozzles and to actuate the valves to differentially cool the work rolls in response to detected phases of thickness oscillations.

[0018] According to another aspect of the present invention, a method for controlling a plant having a rolling mill for producing thin strip product comprises making thickness measurements of the thin strip product at a plurality of locations across a width of the thin strip product using a thickness gauge positioned at the exit of the rolling mill, receiving the thickness measurements at a controller coupled to the thickness gauge, processing the thickness measurements to detect oscillations in the thickness of the thin strip product corresponding to a plurality of control locations; and detecting flatness defects in the thin strip product based on the thickness oscillations.

[0019] The method may further comprise the steps of determining a phase of a thickness oscillation at the each of the plurality of control locations and detecting flatness defects based on differences in phases of the thickness oscillations. In one example, the method comprises identifying control locations having leading phase values as indicative of flatness defects.

[0020] The method may also comprise determining an amplitude of the thickness oscillation within a given frequency range at the each of the plurality of control locations and to detect flatness defects based on a magnitude of the amplitudes of the thickness oscillations. Control locations having higher amplitude values may be used to identify flatness defects.

[0021] The rolling mill may comprise work rolls and a plurality of valves and nozzles to provide segmented water spray cooling to the work rolls, each spray area for a nozzle comprising one of the plurality of control locations and the controller may actuate the valves to differentially cool the work rolls in response to detected flatness defects.

BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The operation of an illustrative twin roll casting plant in accordance with the present invention is described with reference to the accompanying drawings, in which: [0023] FIG. l is a schematic illustrating a thin strip casting plant having a hot rolling mill for controlling the shape of cast strip according to one aspect of the present invention;

[0024] FIG. 2 is an enlarged cut-away side view of the caster of the thin strip casting plant of FIG. 1;

[0025] FIG. 3 is a partial side view of the hot rolling mill of the thin strip casting plant of FIG. 1 showing the arrangement of the localized cooling devices;

[0026] FIG. 4 is a partial plan view showing the cooling pattern from the localized cooling devices of the hot rolling mill in the thin strip casting plant of FIG. 1;

[0027] FIG. 5 is a partial plan view showing the cooling pattern from the localized cooling devices of the hot rolling mill in the thin strip casting plant of FIG. 1;

[0028] FIG. 6 is a block diagram view of a control system according to another aspect of the present invention.

[0029] FIG. 7 is a representation of in-phase thickness oscillations;

[0030] FIG. 8 is a representation of out-of-phase thickness oscillations;

[0031] Fig. 9 is a flow chart of a method according to another aspect of the present invention;

[0032] FIG. 10 is a histogram showing the distribution of bending mark intensity

(frequency = number of coils), which is split by coils that showed bending mark scale vs coils that did not show bending mark scale when inspected downstream;

[0033] Fig. 11 is a flow chart of another method according to another aspect of the present invention.

DETAILED DESCRIPTION

[0034] The exemplary casting and rolling installation illustrated in Figures 1 and 2 comprises a twin-roll caster denoted generally by 11 which produces thin cast steel strip 12 which passes into a transient path across a guide table 13 to a pinch roll stand 14. After exiting the pinch roll stand 14, thin cast strip 12 passes into and through hot rolling mill 15 comprised of back up rolls 16 and upper and lower work rolls 16A and 16B, where the thickness of the strip reduced. The strip 12, upon exiting the rolling mill 15, passes onto a run out table 17 where it may be forced cooled by water jets 18, and then through pinch roll stand 20 comprising a pair of pinch rolls 20A and to a coder 19. An exit thickness gauge 90 measures the thickness of the cast strip after exiting the rolling mill 15 and provides signals indicating the measurements to a controller 92.

[0035] Referring to Figure 2, twin-roll caster 11 comprises a main machine frame 21 which supports a pair of laterally positioned casting rolls 22 having casting surfaces 22A and forming a nip 27 between them. Molten metal is supplied during a casting campaign from a ladle (not shown) to a tundish 23, through a refractory shroud 24 to a removable tundish 25 (also called distributor vessel or transition piece), and then through a metal delivery nozzle 26 (also called a core nozzle) between the casting rolls 22 above the nip 27. Molten steel is introduced into removable tundish 25 from tundish 23 via an outlet of shroud 24. The tundish 23 is fitted with a slide gate valve (not shown) to selectively open and close the outlet 24 and effectively control the flow of molten metal from the tundish 23 to the caster. The molten metal flows from removable tundish 25 through an outlet and optionally to and through the core nozzle 26.

[0036] Molten metal thus delivered to the casting rolls 22 forms a casting pool 30 above nip 27 supported by casting roll surfaces 22 A. This casting pool is confined at the ends of the rolls by a pair of side dams or plates 28, which are applied to the ends of the rolls by a pair of thrusters (not shown) comprising hydraulic cylinder units connected to the side dams. The upper surface of the casting pool 30 (generally referred to as the“meniscus” level) may rise above the lower end of the delivery nozzle 26 so that the lower end of the deliver nozzle 26 is immersed within the casting pool.

[0037] Casting rolls 22 are internally water cooled by coolant supply (not shown) and driven in counter rotational direction by drives (not shown) so that shells solidify on the moving casting roll surfaces and are brought together at the nip 27 to produce the thin cast strip 12, which is delivered downwardly from the nip between the casting rolls. [0038] Referring to Figure 1, below the twin roll caster 11, the cast steel strip 12 passes within a sealed enclosure 10 to the guide table 13, which guides the strip to a pinch roll, stand 14 through which it exits sealed enclosure 10. The seal of the enclosure 10 may not be complete, but is appropriate to allow control of the atmosphere within the enclosure and access of oxygen to the cast strip within the enclosure. After exiting the sealed enclosure 10, the strip may pass through further sealed enclosures (not shown ) after the pinch roll stand 14.

[0039] From the pinch roll stand 14, the thin cast strip 12 is delivered to the hot roll mill 15 comprised of upper work roll 16A and lower roll 16B. Referring to Figures 3, 4 and 5, adjacent the upper work roll 16A is header 70A supplying coolant to three rows of nozzles 71 A and 72 A. The row of nozzles 71 A closest to the strip contains 24 nozzles capable of delivering, for example, 470 gpm of coolant at 100 psi from the header 70A. Nozzles 71 A are not individually regulated during the casting campaign but cool the upper work roll 16A throughout the casting campaign. The remaining two rows of nozzles 72 A have a row of 12 nozzles capable of delivering, for example, 235 gpm of coolant at 100 psi and a further row of 13 nozzles interleaved with the previous row capable of delivering, for example, 400 gpm at 100 psi from the header 70A. The nozzles 72A in the two rows are spaced so that the sprays from the nozzles do not interfere with each other so as to reduce the cooling efficiency of the sprays. Control of the coolant sprays 75 from nozzles 71 A and the coolant sprays 76 from nozzles 72A may be manually controlled by upper header valve 73 A or by a flow meter 73 A that is preset by an operator to a desire flow rate.

[0040] In addition, sprays 76 from nozzles 72A are individually controllable by individual control valves 74A. The individual control valves 74A may be actuated by a controller 92 (Fig. 1, Fig. 6) or manually adjusted. The individual control valves 74 A may be pulse width modulated valves, and the controller may adjust a duty cycle of the pulses. It is understood that the individual control valves 74A can control more than one nozzle 72A if zoned cooling is desired depending on the particular embodiment of the hot rolling mill. However, typically an individual control valve 74A is provided for each nozzle 72A to impart more flexibility and effectiveness in operation of the hot mill to control the shape of the work roll 16A and in turn the shape of the cast strip. The nozzles 72A may be positioned typically about 50mm apart. The sprays from the nozzles 72A are set so that the spray spread substantially overlap between zones across the work surface 77A of the work roll 16A. In this way, the controllable nozzles 72A are able to respond to and effectively control shape defects anywhere across the entire strip 12. In particular, the control valves 74A may be controlled to increase or decrease roll gap profile to reduce or eliminate differential elongation. A swiper bar 81 is also provided to drain away the sprayed coolant from sprays 75 and 76 of nozzles 71 A and 72A after the coolant impacts onto the work surface 77A, so that the coolant is inhibited from contacting the strip 12 where it could cause defects from localized cooling.

[0041] The controlled cooling of adjacent the lower work roll 16B is header 70B supplies coolant to three rows of nozzles 71B and 72B. The row of nozzles 71B closest to the strip contains 24 nozzles capable of delivering, for example, 470 gpm of coolant at 100 psi from the header 70B. Nozzles 71B are not individually regulated during the casting campaign, but provide coolant to cool the lower work roll 16B throughout the casting campaign. The remaining two rows of nozzles 72B have a row of 12 nozzles capable of delivering, for example, 235 gpm of coolant at 100 psi and a further row of 13 nozzles interleaved with the previous row capable of delivering, for example, 400 gpm of coolant at 100 psi from the header 70B. Here again, the nozzles 72B in the two rows are spaced so that the sprays from the nozzles do not interfere with each other so as to reduce the cooling efficiency of the sprays. Manual control of the coolant sprays 75 from nozzles 71B and the coolant sprays 76 from nozzles 72B are manually controlled by lower header valve 73B.

[0042] In addition, sprays 76 from nozzles 72B are individually controlled by individual control valves 74B. Individual control valves 74B may be actuated by the controller 92 or manually adjusted. The individual control valves 74B may be pulse width modulated valves, and the controller may adjust a duty cycle of the pulses. It is understood that the individual control valves 74B can control more than one nozzle 72B if zoned cooling is desired depending on the particular embodiment of the hot rolling mill. However, typically an individual control valve 74B is provided for each nozzle 72B to impart more flexibility and effectiveness in operation of the hot rolling mill to control strip shape. The nozzles 72B may be positioned typically about 50mm apart. Nozzles 72B are set with spray spreads from the nozzle do substantially overlap between zones across the work surface 77B of the work roll 16B. In this way, the controllable nozzles 72B are able to respond to and control the shape of the work surface of the lower work roll 16B anywhere and in turn for shape defects anywhere in the strip 12. In particular, the control valves 74B may be controlled to increase or decrease roll gap profile to reduce or eliminate differential elongation.

[0043] Cyclic variations are often introduced into a material in a process with rotating equipment, or other factors that will oscillate at a periodic frequency. Being produced with rollers, cast metal strips will often have some level of periodic gauge (thickness) variation. This is typically minimized as much as possible, but it is not considered a defect unless it is outside of customer requirements. For example, in cast metal strips, a periodic oscillation in thickness can be detected by the exit profile thickness gauge 90.

[0044] For example, a metal strip may have a thickness which tapers from a center to an edge of the strip. The thickness at the center may oscillate, for example, between 1450 micron and 1470 micron, and the thickness at the edge may oscillate between 1410 micron and 1430 micron. As long as the peaks and troughs of these oscillations are aligned across a width of the metal strip, and a line defining a peak and a line defining a trough are perpendicular to the direction in which the metal strip is being rolled, then the strip is not experiencing differential elongation and low or no flatness defects would be expected. This may be referred to as in-phase oscillations. Figure 7 provides an illustration of in-phase oscillations. Figure 7 is not to scale and the oscillations are exaggerated to make the effect more visible. In Figure 7, the x axis represents the lengthwise direction of the cast strip, the y axis represents a widthwise direction of the cast strip, and the z axis represents the thickness direction of the cast strip. Lines 94 represent thickness measurements made at intervals across a width of the cast strip. The peaks 94A and throughs 94B of the measurements are generally in a straight line across the width of the strip.

[0045] It has been found that if the peaks 94A and troughs 94B of thickness measured at different points along the width of the strip do not generally follow a generally straight line perpendicular to the longitudinal direction, the metal strip has been differentially elongated. Figure 8 is an illustration of the peaks 94A and troughs 94B of thickness measurements across the width of a strip which do not follow a straight line. This may be referred to as out of phase oscillations. From the degree of differential elongation detected, a degree of flatness defect may be inferred. For example, if the peak of 1430 micron at the edge were detected to be ahead of the peak of 1470 micron at the center in the direction of travel of the strip, the strip has been differentially elongated more at the edge of the metal strip than in the center. When tension is removed from the strip, the elongated portion can be expected to buckle, causing a flatness defect. The amount of distance of elongation is indicative of degree of differential elongation and the degree of unflatness. In this way, flatness detection may be accomplished with a single thickness gauge, without directly measuring the actual flatness of the metal strip.

[0046] Because the peaks and troughs oscillate periodically at about 4-5 Hz, the relative phases of the oscillations at different measurement points across the width of the metal strip is indicative of whether the peaks and troughs are aligned and perpendicular to the direction of milling. When the oscillations are in phase, the peaks and troughs are aligned and no differential elongation or flatness defect is indicated. When the oscillations are out of phase, differential elongation and flatness defects are indicated.

[0047] Referring to Figure 9, a method 100 for detecting flatness defects is provided. Thickness measurements of the thin strip are made at a plurality of intervals across the strip in step 102. In one example, a cast strip is two meters wide. The thickness gauge 90 may make 400 measurements over the two meter width of the cast strip. This results in measurement being spaced at 5mm intervals. Different widths of cast strip may be measured and larger or smaller measurement intervals may be used, resulting in differing numbers of measurements. The measurements may be made at time intervals of 0.02 seconds. Shorter or longer time intervals may be employed.

[0048] The thickness measurements are received at controller 92 in step 104 and processed to detect oscillation in step 106. In one example, the thickness measurements are converted to a two-dimensional image, where thickness of the cast strip may be displayed as a color. The oscillations in thickness may be viewed by an operator of the casting plant to determine relative process of thickness oscillations in strip 108 and detect flatness defects based on phase differences in step 110. If the oscillations appear generally straight across the width of the cast strip, no corrective action may be required. However, if the oscillations appear to curve from the middle of the cast strip to the edge, the curved oscillations would be indicative that too much elongation may be occurring on parts of the cast strip. The operator may adjust the segmented cooling (or heating) of the casting rolls to reduce differential elongation in step 112. For example, the operator may manually adjust control valves 74A, 74B, to increase the flow of coolant to the work rolls in areas of the strip where oscillations are leading the center of the strip to reduce elongation.

[0049] In another example, the measurements may be analyzed to detect differential elongation (steps 108, 110) and valve 74 A, 74B automatically controlled to reduce or eliminate differential elongation (step 112). One illustrative example is provided herein in Figures 1 and 6, but the invention is not limited to this example. In one example, the thickness measurements may be received and stored for processing, for example, in a multi-dimensional array in step 104. One dimension of the array may be the distance of the measurement across the width of the cast strip (e.g., each measurement location across the width in 5mm segments). Another dimension may be the time of measurement (e.g., the 0.02 second interval).

[0050] If the measurements are to be used to control work roll shape, there may be higher resolution for the measurements than can be applied to control the work rolls. In this case, measurements made across the cast strip may be averaged together for a given work roll control. In the example of the adjustable work rolls described above, the nozzles 72A, 72B may be spaced 50mm apart, establishing control locations that are 50mm wide, whereas the

measurements may be made at 5 mm intervals. Accordingly, ten measurement locations may be averaged together in one averaged measurement for each nozzle.

[0051] In another example of step 108, the ten measurements corresponding to the center of the width of the cast strip are averaged for each of the 0.02 second interval measurements.

The frequency and phase of the oscillations in thickness may then be determined for the averaged measurements over time. For example, a Fast Fourier Transform (FFT) analysis on the resultant vector of averaged measurements along the time axis may be performed to identify the frequency of thickness oscillation.

[0052] This analysis may be carried out for each work roll control position. In the example of the segmented cooling work roll, described above, the spray area for each nozzle 72A, 72B comprises a control area. Because the nozzles are spaced at 50mm and the

measurements are at 5mm., averages may be calculated in segments of ten measurements across the width of the cast strip. Each segment would represent a 50 mm wide location corresponding to a nozzle, and provide an averaged thickness for that location. In one example, a 1.68 meter wide strip may have 33 resultant segments (the very edges may be ignored) with measurements every 0.02 seconds. The first maximum value within one sample period (as determined by the result of the FFT analysis) is identified in order to determine the phase of the oscillation of each 50mm segment across the width of the strip. Sample periods may be, for example, five seconds.

[0053] In another example of step 110 differential phase shifts may be measured by starting from the center and working out towards the two edges. The center measurement segment may be assigned a phase of zero, and the phases of the other measurement segment may be determined relative to the center. For example, leading phases will be positive and lagging phases will be negative relative to the center. Phase may be expressed in either angles of phase shift (by multiplying by frequency) or time delay or advance. In the current example, for a l.68m wide strip there should be 33 averaged measurements which indicate the phase shift from the waveform at the center of the strip. Measurements may be normalized by taking an average of all of the phase shift measurements, then subtracting that average from each measurement (so the average of all measurements becomes zero).

[0054] The relative phases of the oscillations at different segments may be used to control casting or rolling operations in another example of step 112, such as the segmented cooling for the work rolls. For each 50mm section across the width of the strip the resultant vector should have a numeric value indicating the phase shift from an“average” thickness oscillation phase. Zero phase shift is indicative of no flatness defect, so that is the target. Each of the values in the resultant vector, may be multiplied by a gain constant and integrated with respect to time. The resultant integrated values may be used as an offset to the associated segmented cooling sprays. For positive phase values, the control valve 74 A, 74B corresponding a differentially elongated area would be opened proportionally to the magnitude of the phase value, increasing the flow of cooling water. The increased cooling at the location causes the diameter of the work roll to contract, thereby decreasing elongation. For negative phase valves, the control valve 74A, 74B corresponding to a non-elongated portion would be closed proportionally to the magnitude of the phase value, reducing the flow of cooling water. As that location heats up the portion of the work roll will expand in diameter, increasing elongation. [0055] Measurements and control adjustments may be performed, iteratively to drive phase differences to approach zero, which is indicative of no flatness defect. At this point, measurements would continue, but work roll diameter controls would not require further adjustment until a flatness defect was detected.

[0056] To control bending, a quadratic curve fit of the resultant measurements may be performed. The quadratic term would indicate the symmetrical unflatness of the strip. The target for this term is zero. The quadratic term may be multiplied by a gain and integrated with respect to time and apply that as a bending offset.

[0057] Referring to Figure 11, another method of detecting flatness defects 120 involves ascertaining amplitudes of periodic oscillations. An increase in the amplitude of the oscillation in one part of the width of the strip width is indicative of bending marks. For example, referring to Figure 10, in strips where the most frequent amplitudes were in the 0-20 micron range, no bending mark defects were identified. However, when oscillations frequently included amplitudes in the 25-60 micron range, bending mark defects were found.

[0058] Thickness measurements are made at a plurality of intervals across the width of the strip as before in step 122 and received at the controller 92 in step 124. In one example, a one-dimensional array of data is created for each measurement point. The one dimensional array comprises measurement data at some point across the width from a given sensor over time. In one example, 500 measurements representing 20 seconds of data are stored in the one dimensional array. The measurement intervals may be in the 126. In one example, range of 0.02-0.04 seconds. The data is processed to detect oscillations in step 126. In one example, the data is filtered to remove variations outside of the 4Hz to 7Hz frequency of periodic oscillations. The filter may comprise an Infinite Impulse Response (IIR) bandpass filter with a third order Butterworth filter having a 3.75 Hz to 7.7 Hz passband.

[0059] A Blackman window is applied before applying a Fourier Transform to obtain an amplitude spectrum in step 128. Since non-bending mark frequencies were already filtered out, the remaining amplitudes are summed for each point across the width of the strip to give the total average thickness variation at each point. The minimum value of this new one dimensional array (representing all points across the width at one point it time) may be subtracted from all data points in the array to remove uniform thickness variation, which would be from other sources besides bending marks. The resulting data is a measure of the amplitude of the oscillations at each point across the width over the 20 second sampling period. The maximum value at each time step within a coil can then be averaged over the entirety of the coil to give a bending mark intensity for the coil. Flatness defects are identified by identifying oscillation amplitudes, for example, in the 25-60 micron range in step 130. This may be used to adjust mill operation in step 132.

[0060] The above phase and amplitude methods may be used in combination. Bending marks appear as localized "bending" of the thickness of the periodic oscillations. The phase detection technique disclosed above may be helpful in identifying bending marks with lower amplitude oscillations or screening out false positives.

[0061] The present invention is not limited to controlling segmented cooling nozzles. The detection of flatness defects may also be used to work roll bending and control of work roll force cylinders.