Dynamic roll eccentricity identification and control upgrade using extended Kalman filter state estimation for cold rolling mills

Abstract

To provide a controller for cold rolling mills to improve sheet metal thickness uniformity, and a method of programming the controller.SOLUTION: Sensors monitor the state of the cold rolling mill by measuring (i) roll eccentricity, (ii) roll slips during mill operation, (iii) mill disturbances from roll speed or roll force manifestations, and (iv) unknown disturbances referred to as process noise. The controller analyzes data from sensors to compensate. Data collected during the mill operation by the sensors are delayed in reaching the controller. This communication delay is takin into account by using a filter. Since an objective of controller software is dynamic identification of eccentricity of the back-up rolls, which is non-linear by nature, an extended Kalman filter may be used.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates generally to reducing sheet material thickness in cold rolling mill systems, and more particularly to techniques for improving cold rolling mill performance through dynamic identification of roll eccentricity and compensation. [Background technology]

[0002] In cold rolling, a sheet of metallic material is reduced in gauge or thickness by passing the metal strip between rolling cylinder surfaces under pressure. Typically, the rolling mill produces coils of sheet at a thinner, consistent gauge.

[0003] A single-stand cold rolling mill feeds material from an unwinding reel to a rewinding reel. The metal strip is threaded between work rolls acted upon by backup rolls. A force is applied to at least one of the backup rolls.

[0004] The cross sections of work rolls, backup rolls, supply reels, and rewind reels may not be perfectly circular in circumference for a variety of reasons: Grinding inaccuracies can occur due to axial deviations between the roll barrel and roll neck; Uneven thermal expansion can occur; Asymmetric adjustment of the bearing roll shell via the drive key can occur; Thermally induced wear and mechanically induced wear can occur with misalignment or aging.

[0005] Structural inconsistencies such as roll eccentricity can result. Each eccentricity may have a fundamental frequency, which is the rotational frequency of the roll, and several harmonic frequencies. The eccentricity frequency varies with the roll speed. The eccentricity results in periodic disturbance signals that manifest as thickness and tension errors.

[0006] Most cold rolling mills include two backup rolls that engage the outer surface of the work roll, so eccentricity of both backup rolls will cause variations in the exit strip gauge thickness. The variations may be in phase or out of phase. Filtering out a single eccentricity can be difficult because a pair of work mills may have similar discontinuities and frequencies.

[0007] In cold rolling mills, backup roll eccentricity results in gauge variations in the strip being rolled. This is caused by changes in the opening between the work rolls during processing of the work strip. This problem is becoming more pronounced as strip thickness specifications from cold rolling mills become tighter.

[0008] Over time, the eccentricity profile changes. Since eccentricity cannot be measured directly, it is determined by indirect measurement of the thickness profile for different mill speeds.

[0009] Because of the variables caused by backup roll eccentricity and other surface variations, cold rolling mills can use position control or automatic gauge control in conventional systems for controlling cold rolling mills, which can compensate for variations in feed gauge caused by rotational variations of the backup rolls.

[0010] Passive compensation involves avoiding the gain effect of roll eccentricity in the rolling mill stretch compensation loop (or gage meter loop).

[0011] Active compensation involves using a supplemental signal for the position control to calculate the roll eccentricity that is offset in the HGC output.

[0012] The eccentricity signal may be identified through indirect measurement using a learning algorithm.

[0013] The eccentricity information can be discerned from the roll forces and used for high speed compensation.

[0014] The eccentricity information can be inferred from the exit thickness measurement.

[0015] However, at higher speeds, it can become difficult to compensate for roll eccentricity with closed loop control of mass flow rate and strip tension, especially with time delay. Summary of the Invention

[0016] The present invention is based on the development of a control system for a cold rolling mill or a hot rolling mill to improve sheet metal thickness uniformity to meet or exceed specifications. Sheet metal thickness deviations from standard requirements can be significantly reduced.

[0017] Sensors and gauges can be positioned to obtain dynamic feedback of a cold or hot rolling mill by measuring (i) characteristics of the mill rollers, such as roll eccentricity, (ii) work roll slip during mill operation, (iii) mill disturbances due to manifestations of roll speed or roll force, and (iv) unknown disturbances, which can be referred to as process noise, which can be expressed as Gaussian white noise.

[0018] The roll eccentricity of the backup rolls is an important factor that is not easy to measure. For example, a cold rolling mill can be equipped with various encoders or proximity sensors configured to measure the angular velocity or position of each of the backup rolls. The frequency of the data generated by the various sensors should correspond at least to the execution frequency of the controller software. This should ensure a certain degree of functional accuracy for the control system of the present invention. In addition, the higher the software execution frequency per roller revolution, the more effective the eccentricity compensation will be.

[0019] The controller continuously analyzes data from various sensors to estimate the state of the cold rolling mill and, if necessary, initiates corrective or compensatory action through roll gap control. Feedback data collected during cold rolling mill operation by sensors or observers may be delayed in reaching the controller software. This may be referred to as a communication delay. This delay is taken into account by the software through the use of a filter such as a Kalman filter. Since one of the objectives of the controller software is the identification of eccentricity, which is inherently nonlinear, an extended Kalman filter (EKF) is preferably used.

[0020] In a preferred embodiment, the control system features dynamic identification of roll eccentricity based on sensor-based data and compensation using an extended Kalman filter (EKF) based on state estimation using dynamic partial state feedback. This is called partial because unknown mill defects or conditions can be corrected for as white noise or process noise. There is no direct feedback regarding the state or condition of the mill defects. They can be identified and estimated for the next operating state through the next execution frame of the controller software. Eccentricity identification is performed via a coarse measurement of the effective difference in thickness achieved relative to the roll gap setpoint relative to the desired sheet thickness. This may be referred to as an "initial estimate of thickness inaccuracy." With regard to roll force, feedback is supplemented with the setpoint / applied roll force through hydraulic screw control, which may be referred to as an "effective roll force signal," which is assumed to indirectly reflect mill eccentricity, roll slip, and other defects. This effective roll force is aligned across the above-referenced "initial estimate of thickness inaccuracy" to generate an initial estimate of the eccentricity signal across the circumference of the backup roller. This is refined through EKF-based state estimation of the mill state to generate an eccentricity compensation signal. Such dynamic generation of an eccentricity signal across the circumference of the backup roller adapts to the actual eccentricity changes in the mill over time during mill operation. Thickness deviations observed in current production in cold rolling mills range from 0.5 to 1.5%.

[0021] By implementing embodiments of the present invention, cold rolling mills can achieve improvements of at least 2-5%. Existing cold rolling mill control systems can be upgraded using the present invention. In addition to calculating compensation in response to determined mill conditions, sensors can provide adjustable process parameters that can be adjusted within a nominal range. Parameters may be scaled up or down to adjust the level of compensation to achieve the desired sheet metal results based on the mill operator's discretion. The identified eccentricity signal may also be scaled to different amplitudes by the operator for effective compensation, which is affected by the frequency of the rollers / mill speed.

[0022] The controller of the present invention may be tested as a function of the operation of a rolling sheet metal mill. The operation of the mill may be simulated, such as offline, to test the operation and then deploy the present invention to a particular mill of interest. The extended Kalman filter parameters may be adjusted accordingly at the discretion of the operator of the mill conditions. The extended Kalman filter parameters may be operator-adjustable by the mill operator according to guidance provided during mill commissioning. The operator may adjust the filter values, but typically should not exceed the guidance limits. These limits are designed or otherwise left to be specified by the mill commissioning engineer based on the state or condition of the mill. [Brief explanation of the drawings]

[0023] [Figure 1] 1 illustrates a sheet metal cold rolling mill system. [Figure 2] FIG. 1 is a schematic diagram of an ideal roll and a real roll superimposed to illustrate eccentricity. [Figure 3] FIG. 1 is a flowchart showing the operation of a control method for a cold rolling mill (plant model). [Figure 4]FIG. 10 is a flow chart illustrating the state of a method for identifying roll speed and its synchronization, and roll eccentricity. [Figure 5] FIG. 2 is a flow diagram illustrating a method for using an extended Kalman filter for operation of the cold rolling mill (plant model) shown in FIG. 1 to estimate the state or condition of the rolling mill using sensor feedback. [Figure 6] 1 illustrates a rolled sheet metal system for sensing and controlling sheet metal thickness. [Figure 7] 1 illustrates an example computer system for implementing the controller. DETAILED DESCRIPTION OF THE INVENTION

[0024] The methods of the present invention involve various inventive combinations of the following components: (1) Dynamic identification of roll eccentricity and compensation of roll eccentricity, (2) identification and compensation of roll slip, (3) Extended Kalman Filter (EKF) based state estimation using dynamic partial state feedback, and (4) operator adjustable parameters to improve estimation of rolling mill state.

[0025] If desired, mill operator adjustable parameters may be implemented. The control system of the present invention will be described in the context of controlling the thickness of sheet metal produced on a four-high stand cold rolling mill (four-high mill), which is taken as the baseline. However, it will be understood that the present invention is applicable to a variety of other cold rolling mill configurations. The description herein assumes that the frequency of controller calculations and the frequency of captured sensor or observer feedback data can be designed and implemented to occur at suitable periodic time intervals deemed valid for mill operation.

[0026] definition As used herein, the following terms have the following meanings:

[0027] The term "Work Roll" (WR) refers to a set of rollers that contact the surface of the sheet metal being produced.

[0028] The term "Back Up Rolls" (BUR) refers to a pair of rollers used to apply higher pressure to the WR. The two pairs of rollers, WR and BUR, are part of a four-high stand rolling mill.

[0029] The term "strip / workpiece" refers to the metal sheet being produced.

[0030] The term "roll gap" refers to the gap between a pair of WRs where the sheet is produced.

[0031] The term "material properties" (c m ) refers to the properties of the strip.

[0032] The term "Hydraulic Gap Control (HGC) time constant" (τ hyd ) refers to the time constant of the controller for the roll gap. For standard 4 high rolling mills, hydraulic actuation applies force to the rollers.

[0033] The term "WR main driving constant" (τ roll ) refers to the driving constant of the speed controller for the WR.

[0034] The term "reference angular velocity" (ω WR ) refers to the reference angular velocity calculated according to the setpoint WR speed.

[0035] "BUR diameter" (D BUR The term ) refers to the diameter of the BUR.

[0036] The following parameters describe sensors that monitor the state or condition of the rolling mill.

[0037] The term "entrance thickness" (h_1) refers to the thickness / height profile of the strip during entry into the roll gap.

[0038] The term "exit thickness" (h_2) refers to the thickness / height profile of the strip exiting the roll gap.

[0039] The term "WR angular velocity" (ω WR ) refers to the angular velocity of the WR (upper and lower).

[0040] The term "BUR angular velocity" (ω BUR ) refers to the angular velocity of the BUR, the upper and lower velocities are measured as N1 and N2, respectively, and (ω BUR ) is used to determine

[0041] The term "roll force" (F_roll) refers to the force applied to the roll.

[0042] The term "Avg_gap" refers to the average gap between WRs over that length.

[0043] The following parameters describe the actively controlled effects on the rolling mill:

[0044] The term "screw position" (S) refers to the screw position that is actively controlled to adjust the roll gap.

[0045] The following parameters are preset / predetermined before the rolling mill operation:

[0046] The term "strip velocity" (v strip ) refers to the speed of the strip exiting the roll gap.

[0047] The term "mill speed" refers to the mill speed set by the mill operator and converted to WR angular velocity.

[0048] The term "rolling mill stretching" g " refers to the stretch that a 4-high rolling mill stand undergoes during mill operation.

[0049] The term "material modulus" (c m ) refers to the bulk modulus of the metal strip being produced.

[0050] The term "reference screw position" S ref refers to the screw position set at the start of the mill operation.

[0051] The term "reference WR angular velocity" (ω WRref ) refers to the angular velocity of the WR at the start of the rolling mill operation.

[0052] The following parameters are system uncertainties that are estimated and compensated for:

[0053] The term "eccentricity" (e) refers to large BUR circumference eccentricity that leads to strip thickness error / deviation.

[0054] The term "roll eccentricity amplitude" (a ecc =[a t a b ]) refers to the eccentricity across the BUR.

[0055] The term "roll slip" refers to the difference in speed between the upper and lower WR.

[0056] The term "roll gap" refers to the final effective roll gap achieved as a result of controlled screw position changes and is indeterminate due to variations in strip properties, mill stretch, roll forces, and eccentricity.

[0057] 1 illustrates a metal rolling mill system 2 with thickness control. One example is the METALSMASTER process from Honeywell International, Inc., modified to include an embodiment of the present invention. The rolling mill stand may include four or more vertically mounted rolls, for example, two smaller diameter work rolls each in contact with a separately mounted larger diameter backup roll.

[0058] Incoming metal strip 18 of thickness H is provided by roll 16, which is reduced to thickness h through a plurality of rolls 4, 12, 14, 8 consisting of upper backup roll 4, upper work roll 12, lower work roll 14, and lower backup roll 8 (collectively referred to as the four-high stand 30). Metal strip 18 may be ferrous and / or non-ferrous. Strip 20 exiting stand 30 is collected by roll 22. Proximity sensor 6 measures the rotational position of upper BUR 4, and proximity sensor 10 measures the rotational position of lower BUR 8. For load signal sampling, the rotation around the roll circumference may be divided into several segments, such as 32 segments. These rotational positions are used to calculate the angular velocities of the two BURs 4 and 8, which are part of the rolling mill condition parameters. A feature of the present invention is measuring the angular velocity of the backup rolls. Various conventional devices can be used. Proximity sensors can be used. Representative proximity sensors include the Series IGMF / INFW inductive proximity switches manufactured by EGE-Electronik Spezial Sensoren GmbH (Gettorf, Delaware). The BUR 4,8 angular velocity can be assumed to be the same as the angular velocity of the work rolls 12, 14. Alternatively, the upper and lower work rolls 12, 14 may be equipped with proximity sensors (not shown) to measure their actual angular velocities.

[0059] The stand 30 is equipped with a gap positioning system 26, which may be mechanical, hydraulic, or a combination of both, and is controlled by a feedback device 28. The gap positioning system 26 can vary the work roll gap or opening and includes a pair of screw motors, each equipped with a screw or screws that clamp against the opposing ends of the two BURs 4, 8, thereby applying pressure (roll force) to the work rolls 12, 14. The strip 20 leaves the stand 30 at a thickness h, which is measured by a sensor 24, which may be an X-ray gauge. "L" is the centerline of the stand 30. The control objective is to adjust this exit thickness h as close as possible to a specified target thickness, e.g., at the centerline.

[0060] FIG. 2 is a schematic diagram of eccentricity on a roll, with the actual roll shape shown by solid line 42 and the ideal roll shape shown by substantially coaxial dashed line 40. The amount of eccentricity is shown as the difference between the ideal and actual radius at maximum eccentricity e at 46. Proximity sensor 44 is positioned in the lower BUR 14 of FIG. 1. The effect of eccentricity on thickness in rolling applications can usually be explained as increased force meaning increased exiting thickness (thus pushing the rolls apart). However, if the roll is eccentric, when the maximum radius passes through the roll gap, the force increases but the exiting thickness actually decreases. Therefore, when an eccentric component is present, the force change is misinterpreted.

[0061] Thickness variations not directly related to roll eccentricity may be compensated for separately. The pressure distribution from the upper work roll 12 and the lower work roll 14 may vary from the nominal neutral position for other reasons, such as roll bending due to roll torque.

[0062] The mill stretch factor can be compensated in the mill stand model. The HGC (Hydraulic Gap Control) model can account for mill stretch, which is a non-linear function of the rolling force.

[0063] A prime drive model comprising a set of nonlinear ordinary differential equations can be used to model the prime drive dynamics.

[0064] The deformation resistance may be based on the Bland, Ford, and Ellis model for rolling forces in cold strip rolling with tension.

[0065] The damping coefficient and spring constant can be adjusted.

[0066] Compensation for roll slip, including forward slip, can be included, such as by using roll gap control. The upper and lower work rolls may experience roll slip due to friction variations, etc., and may therefore rotate at different speeds. This complicates the task of reducing gauge variations in cold rolled strip. Slip ratio variables can include changes in speed between the work rolls and backup rolls. The slip factor may be exposed as a user-adjustable option.

[0067] The Roll Eccentricity Compensation (REC) model of the present invention uses real-time / dynamic feedback to dynamically generate an eccentricity signal using the EKF to account for communication delays in the feedback and process noise of the rolling mill operation. This involves correlation and learning of existing thickness deviations, effective forces, and emerging roll gaps into a model that generates an initial estimate of the eccentricity signal across the circumference of the backup roll. This is refined for real-time identification of an eccentricity compensation signal that is appropriate for the current state of rolling mill operation.

[0068] In the present invention, proximity sensors 6, 10 are used to track the angular position of the BURs 4, 8. Direct Control Concept - Method Eccentricity information is identified from the identification and estimation of roll forces and dynamic states. The theoretical model is described in sections 1-9 below.

[0069] Section 1. Definitions Work roll (WR): Reference angular velocity {ω WR} Backup roll (BUR): Diameter of upper and lower BUR {Diameter BUR} Strip / Workpiece: Material properties {c m} Roll gap: Hydraulic Gap Control (HGC) time constant {τ hyd} WR main driving constant {τ roll}

[0070] Section 2. Measurements: State Observer Strip thickness / height profile Entry Thickness {h1} Measures the thickness of the strip before it enters the work roll gap. An X-ray gauge can be used.

[0071] Exit Thickness {h2} Measures the thickness of the strip after it leaves the work roll gap.

[0072] Angular velocity / rotational speed WR angular velocity {ω WR} This velocity can be determined using a proximity sensor that measures rotational position and therefore derives angular velocity.

[0073] BUR angular velocity {ω BUR}This speed can be determined with a proximity sensor.

[0074] Roll Force {F roll} Pressure is typically applied by varying the screw position of a hydraulic press relative to the work rolls in response to feedback.

[0075] 3. Active Control Variables Screw position for adjusting the roll gap {S}

[0076] 4. System preset / constant parameters Strip velocity {νstrip} Rolling mill speed (or) WR angular velocity {ω WR} Rolling mill extension {c g} Material elastic modulus {c m} Reference screw position {S ref} Reference WR angular velocity

[0077]

number

[0078] Section 6. Mathematical Model The following assumptions can be made where appropriate:

[0079] 6.1 HGC Dynamics Assume the first order system

[0080]

number

[0081]

number

[0082] 6.2 Main Drive Dynamics Assume the first order system

[0083]

number

[0084]

number

[0085] 6.3 Estimation of Roll Force

[0086]

number

[0087] 6.4 Estimation of the thickness

[0088]

number

[0089]

number

[0090]

number

[0091]

number

[0092] 6.5 Estimation of screw compensation signal From the following,

[0093]

number

[0094]

number

[0095] The screw compensation is as follows:

[0096]

number

[0097] 6.6 Estimation of eccentricity signals Roll eccentricity is a periodic function that captures deformations on the BUR circumference. This periodic signal changes its characteristics with new work pieces, WR angular velocity, roll forces, and mill stretch.

[0098] Assuming,

[0099]

number

[0100]

number

[0101] where the single sine component is

[0102]

number

[0103]

number

[0104]

number

[0105] Let there be two components of eccentricity, one for the upper bur and one for the lower bur. next,

[0106]

number

[0107]

number

[0108]

number

[0109] Then the derivative is:

[0110]

number

[0111] Given the nonlinear profile, the Extended Kalman Filter (EKF) is chosen for state estimation.

[0112] The extended Kalman filter can be based on statistical inference.

[0113] Section 7. State-Space Representation of Mathematical Models 7.1 State Space Representation

[0114]

number

[0115] From the above, we can build:

[0116] State Transition Model

[0117]

number

[0118] X state

[0119] [Table 2]

[0120] The eccentricity profile is filtered by a complementary method using feedback from roll gap and roll force. The amplitude is scaled to millimeters to meet the compensation value for thickness correction. This calculated correction is added as a sum component to the average gap output.

[0121] Figure 3 is a control schematic diagram illustrating the process of identifying roll eccentricity during the running time of a cold rolling mill. In steps 100, 102, and 110, the "roll force," "average gap," and angular velocity of two BURs (N1 and N2) are measured. These contribute to the dynamic partial state feedback of the entire system. Step 124 determines whether there is a difference between the measured average roll gap and the set roll gap. If there is a difference, a calculation is used to estimate the noisy effective gap, referred to as "gap noise" 130. This is used to identify the effective "roll force" in step 100 and filtered using the "gap noise" in step 122. Meanwhile, steps 126 and 108 extract the instantaneous amplitude of the eccentricity from the waveform of the eccentricity signal. Optionally, instantaneous data indicating the waveform of the eccentricity signal may be extracted by the operator at any time for further investigation of the rolling mill condition, such as for maintenance and efficiency. The estimated amplitude of the eccentricity corresponds to the deviation in the "roll gap" that needs to be corrected. The amplitude of this eccentricity can be scaled up or down by the operator to accommodate other uncertainties in the rolling mill, as defined by a gain constant, which may be equal to 1 by default in step 128 .

[0122] Section 8. Operator Adjustable Parameters The method of the present invention also allows for the adjustability of various parameters appropriate to the rolling mill site. These parameters can be accessed by the operator at any time and used to scale the eccentricity amplitude for further improvement. Adjustments may only be required to solve operational problems. These parameters include, for example: Upper and lower radii, WR and BUR.

[0123] Rolling mill extension {c g}.

[0124] Material elastic modulus {c m}.

[0125] Eccentricity amplitude scaling.

[0126] Number of harmonics.

[0127] Execution speed.

[0128] Power Spectral Density (PSD) constants for process noise (q) and measurement noise (r): q_s, q_v, q_h, q_a, q_f, r_s, r_v, r_h, r_f PSD constants are general mathematical concepts that support noise functions. They can be adjusted by the mill operator.

[0129] Section 9. Methodology Description FIG. 4 lists the roll synchronization and eccentricity identification steps for the sensor and EKF based roll eccentricity and compensation technique of the present invention.

[0130] Roll synchronization begins with process step 140, which involves marker pulse tracking.

[0131] After process step 140, the next process step 142 involves roll circumference segmentation.

[0132] After process step 142, the next process step 144 involves synchronizing the BUR rotation with the state / reference variables.

[0133] After process step 144, the next process step 146 involves tracking the delta rotation position on a segment-by-segment basis.

[0134] Process step 146 completes roll synchronization, followed by eccentricity identification, which begins with process step 150 which involves tracking roll force sensor data on a segment-by-segment basis for the upper and lower BURs.

[0135] After process step 150, the next process step 152 involves building a delta roll force distribution between the upper and lower BURs on a segment-by-segment basis.

[0136] After process step 152, the next process step 154 ​​involves calculating an eccentric roll force signal.

[0137] Process step 154 ​​completes the eccentricity identification.

[0138] Process steps 144, 146, and 152 also use roll forces to identify roll slip / eccentricity profiling across each BUR surface.

[0139] FIG. 5 shows the overall method architecture in MATLAB / software / firmware implementation etc.

[0140] Process step 160 begins by initializing state variables.

[0141] After process step 160, the next process step 162 involves declaring a sampling time and an initial rotational position.

[0142] After process step 162, the next process step 163 involves eccentricity identification of the upper and lower BURs.

[0143] After process step 163, the next process step 164 involves constructing a state transition matrix [11x11].

[0144] After process step 164, the next process step 166 involves assuming a measurement noise contribution matrix [11×11].

[0145] After process step 166, the next process step 168 involves assuming a white noise PSD constant for the measurement noise using the measured state variables.

[0146] After process state 168, the next process state 170 involves discretizing the LTI ODE with Gaussian noise.

[0147] After process step 170, the next process state 172 involves building an EKF model that time-steps the state and covariance estimates.

[0148] After process step 172, the next process step 174 involves estimating the eccentricity setpoints that are calculated at each time step of the state and covariance estimation.

[0149] The process of the present invention may be incorporated into plants such as cold rolling mill systems 200, including those with automatic gauge control (AGC) and hydraulic gap control (HGC) controllers, to improve the thickness accuracy of the produced sheet. As shown in FIG. 6, the rolling mill 202 may include an output 210 from a controller 20. The output 210 may include roll eccentricity compensation, such as that described in U.S. Patent Application Publication No. 2018 / 0161839 to McGahan, which is incorporated herein by reference. The controller 206 may include inputs 208 from sensors 204, such as upper and lower burrs. The controller 206 may include a filter 220. The filter 220 may include various models 222, 224 by integrating the method of the present invention into a controller environment, such as in a MATLAB script or another suitable program. The model 222 may include a state estimate and a covariance (noise effect) estimate. The model 224 may include a state transition model with dynamic partial state feedback.

[0150] The method of the present invention can be implemented and validated in an HTS lab simulation environment for a variety of pre-defined conditions. In one embodiment of the present invention, it has demonstrated up to a 10% improvement over the existing 0.5-1.5% thickness deviation range. In another embodiment of the present invention, thickness accuracy has been improved by an additional 10% over the performance of an AGC controller. 1. The eccentricity estimation adapts to the actual eccentricity changes over time in the rolling mill. 2. Allows the operator to scale the eccentric amplitude for further refinement. 3. The noise model assumed in the plant process and measurements is operator adjustable for power spectral density (PSD) constants for further refinement.

[0151] FIG. 7 is a block schematic diagram of a computer system 240 for implementing a controller and method according to an exemplary embodiment. Not all components need be used in various embodiments. One exemplary computing device in the form of a computer system 240 may include a processing unit 242, memory 244, removable storage 252, and non-removable storage 256. While an exemplary computing device is illustrated and described as a computer system 240, the computing device may take different forms in different embodiments. For example, the computing device may instead be a smartphone, tablet, smartwatch, or other computing device including the same or similar elements as illustrated and described with respect to FIG. 7. Devices such as smartphones, tablets, and smartwatches are generally collectively referred to as mobile devices. Additionally, while various data storage elements are illustrated as part of the computer system 240, storage may also or alternatively include cloud-based storage accessible via a network such as the Internet.

[0152] Memory 244 may include volatile memory 248 and nonvolatile memory 259. Computer system 240 may include, or have access to a computing environment that includes, a variety of computer-readable media, such as volatile memory 248 and nonvolatile memory 259, removable storage 252 and non-removable storage 256. Computer storage may include random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) and electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD ROM), Digital Versatile Disk (DVD) or other optical disk storage, magnetic cassette, magnetic tape, magnetic disk storage, or other magnetic storage device capable of storing computer-readable instructions for performing the functions described herein.

[0153] Computer system 240 includes or has access to a computing environment that includes input 258, output 260, and communication interface 254. Output 260 may include a display device such as a touchscreen that can also function as an input device. Input 258 may include one or more of a touchscreen, touchpad, mouse, keyboard, camera, one or more device-specific buttons, one or more sensors integrated within computer system 240 or coupled via a wired or wireless data connection, and other input devices. The computer may operate in a networked environment using communication connections to connect to one or more remote computers, such as database servers, including cloud-based servers and storage. Remote computers may include personal computers (PCs), servers, routers, network PCs, peer devices or other common network nodes, etc. Communication connections may include a local area network (LAN), a wide area network (WAN), cellular, WiFi, Bluetooth, or other networks.

[0154] Computer-readable instructions stored on a computer-readable storage device are executable by the processing unit 242 of the computer system 240. Hard drives, CD-ROMs, and RAM are some examples of articles that include non-transitory computer-readable media, such as storage devices. The terms computer-readable media and storage devices do not include carrier waves. For example, the computer program 246 may be used to cause the processing unit 242 to perform one or more of the methods or algorithms described herein.

[0155] The foregoing has described the principles, preferred embodiments, and modes of operation of the present invention. However, the invention should not be construed as limited to the particular embodiments discussed. Accordingly, the above-described embodiments are to be considered illustrative rather than restrictive, and it will be understood that modifications can be made therein by those skilled in the art without departing from the scope of the invention as defined by the following claims.

Claims

1. 1. A controller for controlling thickness of sheet metal in a rolling mill exiting a rolling mill stand having first and second work rolls positioned between first and second backup rolls, respectively, the controller comprising: a processor; and code stored on a medium readable by the processor for controlling thickness of the sheet metal, the processor comprising: receiving input including at least a plurality of parameters related to the thickness of the sheet metal and the rolling mill stand; generating a model of the rolling mill stand based on the plurality of parameters; using the model to estimate eccentricity in the rolling mill stand based on feedback inputs from the rolling mill stand during operation, the feedback inputs including at least angular velocities of the first and second backup rolls, roll forces, and an average gap between the first and second work rolls; estimating a state of the rolling mill stand based at least on the estimated eccentricity and the model of the rolling mill stand using a filter; generating an output including a roll eccentricity compensation signal for controlling a gap between the first work roll and the second work roll based at least on the estimated condition of the rolling mill stand to control the thickness of the sheet metal produced by the rolling mill stand. The controller is configured as follows:

2. A method of programming a controller for controlling the thickness of sheet metal produced by a rolling mill stand, the rolling mill stand comprising a first work roll and a second work roll positioned between a first backup roll and a second backup roll, respectively, the method comprising: receiving input including at least a plurality of parameters related to the thickness of the sheet metal and the rolling mill stand; generating a model of the rolling mill stand based on the plurality of parameters; using the model to estimate eccentricity in the rolling mill stand based on feedback inputs from the rolling mill stand during operation, the feedback inputs including at least angular velocities of the first and second backup rolls, roll force, and an average gap between the first and second work rolls; estimating, by a filter, a state of the rolling mill stand based at least on the estimated eccentricity and the model of the rolling mill stand; generating an output comprising a roll eccentricity compensation signal for controlling a gap between the first work roll and the second work roll based at least on the estimated condition of the rolling mill stand to control the thickness of the sheet metal produced by the rolling mill stand.

3. A metal rolling mill system for controlling sheet metal thickness, comprising: a rolling mill stand including a first work roll and a second work roll positioned between a first backup roll and a second backup roll, respectively; a processor; one or more sensors for measuring a plurality of parameters related to the thickness of the sheet metal and the rolling mill stand; a memory device coupled to the processor; The memory device is receiving input from the one or more sensors including the plurality of parameters; generating a model of the rolling mill stand based on the plurality of parameters; using the model to estimate eccentricity in the rolling mill stand based on feedback inputs measured from the rolling mill stand during operation, the feedback inputs including at least angular velocities of the first and second backup rolls, roll force, and an average gap between the first and second work rolls; estimating a state of the rolling mill stand based at least on the estimated eccentricity and the model of the rolling mill stand using a filter; generating an output including a roll eccentricity compensation signal for controlling a roll gap between the first work roll and the second work roll based at least on the estimated condition of the rolling mill stand to control the thickness of the sheet metal produced by the rolling mill stand; and a program readable by the processor to execute the program.

Images