Slot die positioning with ringing constraints

Abstract

To provide methods of adjusting a slot die.SOLUTION: The slot die includes: an applicator slot in fluid communication with a fluid flow path through the slot die; and a plurality of actuators spaced along the width of the slot die. Each actuator is operatively coupled to the choker bar or flexible die lip to adjust a slot or choker bar height profile at its respective location, and comprises means of local adjustment of a fluid flow through the applicator slot. The method includes: measuring thickness values of an extrudate provided from the applicator slot; and adjusting a position of the actuators based on a residual vector representing deviation of the measured thickness values from desired thickness values so as to obtain a corrected shape for the choker bar or flexible die lip. The adjustment includes using the residual vector to predict a degree of ringing for actuator positions and obtaining a revised set of actuator positions that reduces the degree of ringing.SELECTED DRAWING: Figure 6

Description

[Technical Field]

[0001] The present invention relates to slot dies and related assemblies, systems, and methods. [Background technology]

[0002] Generally, a slot die includes a die lip that forms an applicator slot. The width of the applicator slot can extend along the width of a moving web or the width of a roller that receives an extrudate, such as a film. As used herein, with respect to slot dies and slot die components, "width" refers to the cross-web (or cross-roller) dimension of the slot die and its components. In this regard, the applicator slot of a slot die extends along the width of the slot die.

[0003] Slot dies are commonly used to form extrudates, coatings, and other extruded articles. For example, slot dies can be used in slot die coating to apply liquid materials to a moving flexible substrate or "web." There are many variations in slot die coating techniques. For example, the coating material may be at room temperature or at a controlled temperature. When the coating material temperature is elevated to ensure the coating material is melted or liquefied for processing, this is often referred to as "hot melt" coating. In other examples, the coating material may include a solvent diluent. The solvent may be water, an organic solvent, or any suitable fluid that dissolves or disperses the coating components. The solvent is typically removed in subsequent processing, such as by drying. Coatings can include single or multiple layers, and several slot dies can be used to simultaneously apply multiple layers. The coating may be a continuous coating across the width of the die, or alternatively, it may be composed of strips, each extending across only a portion of the die width and separated from adjacent strips.

[0004] Slot dies are also used to form extrudates, including thin film extrudates or other extrudates. In some instances, the extrudate may be an extrusion coating and may be applied to a web substrate, a process sometimes referred to as extrusion coating. In other instances, the extruded material directly forms a film or web. The extruded film may then be processed by length drawing or tentering operations. Like coatings, the extrudate may include a single layer or multiple layers.

[0005] The thickness of an extrudate, such as a film or coating, depends on the flow rate of the extrudate through the slot die, among other factors. In one example, a slot die can include an adjustable choker bar in the flow path that can be used to locally adjust the flow rate of the extrudate through the slot die to provide a desired thickness profile. The slot die can also include a flexible die lip that can be used to locally adjust the cross-sectional height of the applicator slot itself (i.e., the slot height) and control the flow rate of the extrudate from the applicator slot to provide a desired thickness profile.

[0006] The slot die may include multiple actuators spaced along the width of the applicator slot to provide a desired thickness profile for the extrudate. For example, each actuator can be configured to provide local position adjustment of a choker bar or flexible die lip. During the extrusion process using the slot die, the cross-web profile of the extrudate can be measured. Each actuator can then be individually adjusted to provide a desired thickness profile, such as a uniform thickness, for the extrudate across the width of the applicator slot. Summary of the Invention

[0007] Described herein is an improved method that uses a stiffness matrix and its inverse to efficiently and accurately convert actuator settings to slot die profiles, and vice versa. When using an inverse stiffness matrix to make adjustments to slot height at only one location, the calculated solution tends to exhibit a phenomenon known as "ringing." Ringing is an undesirable condition that occurs when adjacent actuators move in opposite directions on a large scale next to the location of a slot height adjustment. This occurs because adjacent actuators are physically connected to the same choker bar or flexible die lip and therefore must move in concert. Severe ringing can cause actuators to unnecessarily reach their force and bending limits, thereby limiting the degree of coating uniformity that can be achieved using this automated system.

[0008] A solution is provided in which the controller uses a ringing constraint to mathematically "penalize" actuator movements that cause ringing. Under this configuration, the degree of ringing is quantified, and solutions for actuator settings that reduce ringing below a certain value can be accepted at the expense of accepting some small deviation in slot gap accuracy. The ringing constraint can further aid in the efficiency, robustness, and coating uniformity achievable using the provided control system.

[0009] In a first aspect, a method of adjusting a slot die is provided, the slot die comprising: an applicator slot extending along a width of the slot die in fluid communication with a fluid flow path through the slot die, at least one of the group consisting of a choker bar and a flexible die lip, and a plurality of actuators spaced along the width of the slot die, each actuator operatively coupled to a choker bar or a flexible die lip to adjust a height profile of the slot or choker bar at its respective location to provide localized adjustment of fluid flow through the applicator slot. The method includes measuring thickness values ​​of the extrudate provided from the applicator slot, mapping the measured thickness values ​​to corresponding positions along the width of the slot die, performing data compression of the measured thickness values ​​to obtain a residual vector representing deviations of the measured thickness values ​​from a desired thickness value, and adjusting positions of one or more of a plurality of actuators based on the residual vector to obtain a corrected shape for the choker bar or flexible die lip, wherein adjusting includes using the residual vector to determine a predicted degree of ringing for a set of proposed actuator positions, and then applying a ringing constraint to obtain a modified set of actuator positions that reduces the degree of ringing relative to the predicted degree of ringing.

[0010] In a second aspect, there is provided a method of adjusting a slot die comprising: an applicator slot extending along a width of the slot die, the applicator slot being in fluid communication with a fluid flow path through the slot die; at least one of the group consisting of a choker bar and a flexible die lip; and a plurality of actuators spaced along the width of the slot die, each actuator operatively coupled to a choker bar or a flexible die lip to adjust a height profile of the slot or choker bar at its respective location to provide localized adjustment of fluid flow through the applicator slot, the method comprising: adjusting a thickness value of an extrudate provided from the applicator slot; A method is provided that includes measuring; mapping the measured thickness values ​​to corresponding positions along a width of the slot die; performing data compression of the measured thickness values ​​to obtain a residual vector representing deviations of the measured thickness values ​​from desired thickness values; and adjusting positions of one or more of a plurality of actuators based on the residual vector to obtain a corrected shape of the choker bar or flexible die lip, wherein the adjusting includes applying a slot or choker bar height constraint whereby the one or more adjusted actuator positions are predicted to maintain an average current slot or choker bar height along the corresponding portion of the choker bar or flexible die lip.

[0011] In a third aspect, there is provided a method of adjusting a slot die comprising: an applicator slot extending along the width of the slot die, the applicator slot being in fluid communication with a fluid flow path through the slot die; at least one of the group consisting of a choker bar and a flexible die lip; and a plurality of actuators spaced along the width of the slot die, each actuator operably coupled to the choker bar or the flexible die lip to adjust the height profile of the slot or choker bar at its respective location to provide localized adjustment of fluid flow through the applicator slot. The method includes measuring thickness values ​​of the extrudate provided from the applicator slot, mapping the measured thickness values ​​to corresponding positions along the width of the slot die, performing data compression of the measured thickness values ​​to obtain a residual vector representing deviations of the measured thickness values ​​from a desired thickness value, and adjusting the position of one or more of a plurality of actuators based on the residual vector to obtain a corrected shape of the choker bar or flexible die lip, wherein mapping the thickness values ​​to corresponding positions along the width of the slot die includes expanding and / or adjusting the measured thickness values ​​to account for positional variations at one or more transverse edges of the extrudate.

[0012] In a fourth aspect, there is provided a method for producing an extruded article, the method comprising: adjusting a slot die according to one of the methods described above; and extruding an extrudate based on the adjusted slot die through an applicator slot of the slot die to obtain the extruded article.

[0013] In a fifth aspect, a system is provided, the system comprising: a slot die including an applicator slot extending along a width of the slot die, the applicator slot being in fluid communication with a fluid flow path through the slot die, at least one of the group consisting of a choker bar and a flexible die lip, and a plurality of actuators spaced along the width of the slot die, each actuator operatively coupled to a choker bar or a flexible die lip to adjust a height profile of the slot or choker bar at its respective location to provide localized adjustment of fluid flow through the applicator slot; and a system for receiving a thickness value of the extrudate provided from the applicator slot. a controller configured to: map the received thickness values ​​to corresponding positions along a width of the slot die; perform data compression of the received thickness values ​​to obtain a residual vector representing a deviation of the received thickness values ​​from a desired thickness value; and adjust the positions of one or more of the plurality of actuators based on the residual vector to obtain a corrected shape of the choker bar or flexible die lip, wherein the adjustment includes using the residual vector to determine a predicted degree of ringing for a set of proposed actuator positions; and then applying a ringing constraint to obtain a modified set of actuator positions that reduces the degree of ringing relative to the predicted degree of ringing. [Brief explanation of the drawings]

[0014] [Figure 1] FIG. 10 illustrates an exemplary slot die in operation, including a choker bar with multiple actuators, each operable to adjust the height profile of the slot or choker bar at its respective location to provide localized adjustment of fluid flow at that location. [Figure 2] FIG. 2 is a top view of the slot die of FIG. 1. [Figure 3]FIG. 1 is a schematic diagram illustrating an actuator assembly including a position sensor and a controller for selecting a position of the actuator assembly based on the output of the position sensor. [Figure 4A] FIG. 1 is a schematic diagram illustrating an exemplary process for conditioning a slot die. [Figure 4B] FIG. 1 is a schematic diagram illustrating an exemplary process for conditioning a slot die. [Figure 5] 10 is a chart illustrating the ringing phenomenon observed for actuator positions resulting from a given target slot height profile. [Figure 6] 10 is a chart illustrating the effect of applying a ringing constraint to an actuator setting profile. [Figure 7] 1 is a chart illustrating the deviation of actuator position (units) relative to a reference point (i.e., nominal) actuator position in an exemplary slot die. [Figure 8] 8 is a chart showing the calculated flexible die lip profile of the slot die of FIG. [Figure 9] 9 is a chart showing the neutral (ie, resting) slot height profile of the slot die of FIGS. 7-8. [Figure 10] 10 is a chart showing thickness profiles of coatings produced using an exemplary die slot conditioning process based on the slot die configuration of FIGS. 7-9. [Figure 11] 10 is a chart showing the calculated die slot height profile calculated from the current actuator position along with the reference point slot and reference point actuator positions. [Figure 12] 10 is a chart showing how ringing can be affected by different methods of fitting the stiffness matrix used to determine actuator position. [Figure 13] 1 is a chart showing normalized stiffness parameters for various extrusion dies.

[0015] Repeat use of reference characters in the specification and drawings is intended to represent the same or analogous features or elements of the present disclosure. It is to be understood that those skilled in the art may devise numerous other modifications and embodiments that fall within the scope and spirit of the principles of the present disclosure. The figures may not be drawn to scale. DETAILED DESCRIPTION OF THE INVENTION

[0016] As used herein, the terms "preferred" and "preferably" refer to embodiments described herein that may offer certain advantages, under certain circumstances, although other embodiments may also be preferred, under the same or other circumstances. Furthermore, the recitation of one or more preferred embodiments does not imply that other embodiments are not useful, or is intended to exclude other embodiments from the scope of the invention.

[0017] As used in this specification and the appended claims, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. Thus, for example, reference to an element preceded by "a" or "the" may include one or more of the element and equivalents thereof known to those skilled in the art. Furthermore, the term "and / or" means one or all of the listed elements or a combination of any two or more of the listed elements.

[0018] It should be noted that the term "comprises" and variations thereof do not have a limiting meaning when these terms appear in the accompanying description. Furthermore, "a," "an," "the," "at least one," and "one or more" are used interchangeably herein. Relative terms such as left, right, front, rear, top, bottom, side, above, below, horizontal, vertical, etc. may be used herein, when taken from the perspective seen in a particular drawing. However, these terms are used merely for ease of description and in no way limit the scope of the present invention.

[0019] Throughout this specification, references to "one embodiment," "a particular embodiment," "one or more embodiments," or "an embodiment" mean that a particular feature, structure, material, or characteristic described with respect to that embodiment is included in at least one embodiment of the invention. Thus, the appearances of phrases such as "in one or more embodiments," "in a particular embodiment," "in one embodiment," or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment of the invention. Where applicable, product names are written in all capital letters.

[0020] 1 and 2 show a slot die 10. The slot die 10 includes an upper die block 2 and a lower die block 3. The upper die block 2 combines with the lower die block 3 to form a fluid flow path through the slot die 10. The fluid flow path includes an inlet 5, a die cavity 4, and an applicator slot 6. The applicator slot 6 is between a rotating rod 12 attached to the upper die block 2 and a die lip 13 of the lower die block 3. Because the slot die 10 includes a rotating rod 12 in its applicator slot, the slot die 10 is sometimes referred to as a rotating rod die.

[0021] The slot die 10 includes a choker bar 11 that spans the width of the fluid flow path within the slot die 10. As an example, the width of the fluid flow path within the slot die 10 at the choker bar 11 may be approximately the same as the width of the applicator slot 6, such that the choker bar 11 extends along the width of the applicator slot 6. The actuator assemblies 200 are mounted to a common mounting bracket 9 and are spaced apart along the width of the slot die 10. In some embodiments, the mounting bracket 9 may be segmented, e.g., the mounting bracket 9 may include separate structures for each actuator assembly 200. Each actuator assembly 200 is operable to adjust the height profile of the slot or choker bar at its respective location along the width of the slot die 10 by changing the position of the choker bar 11 within the extrudate fluid flow path within the slot die 10, thereby providing localized adjustment of fluid flow through the applicator slot 6. It should be understood that references herein to "slot height" in the examples and descriptions may be equated with "choker bar height" in similar slot die configurations.

[0022] During operation of slot die 10, extrudate enters slot die 10 at fluid flow path inlet 5 and continues through the fluid flow path of slot die 10, including die cavity 4, until the extrudate exits through applicator slot 6 and is applied to moving roller 7. In some embodiments, the extrudate can be applied to a moving web (not shown). In other embodiments, the extrudate can be applied directly to roller 7. Optionally, the extrudate and web can be conveyed through a series of rollers to cool the extrudate.

[0023] One or more additional processes may be performed on the extrudate downstream of roller 7. Such processes include, but are not limited to, stretching, coating, texturing, printing, cutting, rolling, and laminating. In some processes, the production release liner may be removed and a release liner added, or one or more additional layers (such as a laminated transfer tape) may be added. A curing step may also be performed and may be carried out by exposure to actinic radiation such as e-beam, heat, or ultraviolet (UV) light.

[0024] As shown in FIG. 2, the slot die includes a set of actuator assemblies 200 mounted on a common mounting bracket 9. Five actuator assemblies 200 are shown, although a different number of actuator assemblies are possible. Each actuator assembly 200 is attached to or otherwise engaged with a choker bar 11, and the actuator assemblies 200 are spaced apart along the width of the choker bar 11. Each of the actuators is operable to adjust the height profile of the choker bar at its respective location to provide local adjustment of the position of the choker bar 11 within the fluid flow path within the slot die.

[0025] As described below with respect to FIG. 3 , each of the actuator assemblies 200 can include a motor that drives a linear actuator. Each of the actuator assemblies 200 can also include a precision sensor, such as a linear variable differential transformer (LVDT) or linear encoder, that detects positional movement of the output shaft of the linear actuator. The output shafts of the linear actuator assemblies 200 are spaced along the width of the choker bar 11 such that each linear actuator assembly 200 is operable to adjust the local position of the choker bar. As described in further detail below, the position of each linear actuator can be individually selected to provide a desired cross-web profile of the extrudate. Additionally, the position of the linear actuator assemblies 200 can be precisely adjusted to provide a desired die cavity pressure within the die cavity 4 during operation of the slot die 10 by adjusting the total cross-sectional area of ​​the fluid flow path adjacent to the choker bar 11 within the slot die 10. In other embodiments, the position of each actuator assembly 200 can be actively controlled to produce extrudates having patterned features, such as repeating or random patterned features. When referred to herein, references to the position of an actuator or actuator assembly are intended to refer more specifically to the relative placement of the actuator output shaft.

[0026] Variations of the slot die 10 are also possible in which the actuator assembly is engaged with a flexible die lip or a rotating rod instead of a choker bar. Actuator assemblies may also be present to adjust the cross-sectional area of ​​the fluid flow path from both sides of the applicator slot. In these embodiments, various combinations are possible; for example, a rotating rod die can be configured with both a choker bar and a flexible die lip, each of which can be precisely adjusted via multiple actuators spaced along the width of the rotating rod die.

[0027] 3 shows an assembly including individual actuator assemblies 200, zero-backlash couplers 240, and a controller 300. As shown in FIGS. 1 and 2, actuator assemblies 200 may be used in slot dies to provide localized adjustment of the fluid flow paths of the slot die, for example, by adjusting the height of the fluid flow paths within the slot die, similar to slot die 10.

[0028] The actuator assembly 200 includes a motor 210, a linear actuator 220 coupled to the motor 210, and a position sensor 230. As an example, the motor 210 may be a stepper motor. An output shaft (not shown) of the motor 210 is mechanically coupled to the linear actuator 220. The sensor 230 senses the position of the linear actuator 220. For example, the sensor 230 may be an LVDT sensor or a linear encoder. The sensor 230 is secured to the output shaft 222 of the linear actuator 220 by a clamp 232 and accurately measures the relative position of the output shaft 222 of the linear actuator 220. In other embodiments, the sensor 230 may measure the zero backlash of the coupler 240, the die actuator linkage 252, the flexible die lip 32, the rotating rod 22, or the choker bar 11. An actuator assembly suitable for use as the actuator assembly 200 is available from Honeywell International Incorporated (Morristown, New Jersey).

[0029] The controller 300 is typically a dedicated computer that receives position inputs from both the motor 210 and the sensor 230. For example, the motor 210 may be a stepper motor and may provide an indication of the number of “steps” the stepper motor has taken from a known reference position. The sensor 230 may provide the controller 300 with more accurate position information than provided by the motor 210. The controller 300 commands the motor 210 to drive the output shaft 222 of the actuator 220 to a preselected position. For example, the controller 300 may monitor the position of the output shaft 222 of the actuator 220 using the sensor 230 while operating the motor 210 to position the output shaft 222 of the actuator 220 according to the preselected position. In some embodiments, the controller 300 can control a set of actuator assemblies 200 either simultaneously or sequentially. For example, the controller 300 can control each of the actuator assemblies 200 of the slot die 10 shown in FIG. 2 .

[0030] In the slot die 10, the output shaft 222 of the actuator 220 can be advantageously connected to the die actuator linkage 252 by a zero-backlash coupler 240. The zero-backlash coupler 240, in this embodiment, includes two halves, a lower half 242 and an upper half 244, that are screwed together. The lower half 242 is directly attached to the die actuator linkage 252 with screws. In addition, the zero-backlash coupler 240 includes a stacked ejection assembly that is bolted to the end of the output shaft 222 of the actuator 220. The stacked ejection assembly includes two metal discs 246 that surround an insulating disc 248. By way of example, the insulating disc 248 can include a ceramic material. The lower half 242 and the upper half 244 combine to surround the stacked ejection assembly that includes the metal disc 246 and the insulating disc 248 that are bolted to the end of the output shaft 222 of the actuator 220. When the top half 244 is securely screwed onto the bottom half 242 , the output shaft 222 of the actuator 220 is effectively connected to the zero-backlash coupler 240 and the die actuator linkage 252 .

[0031] The zero-backlash coupler 240 functions to thermally isolate the actuator assembly 200 from the slot die. Specifically, the insulating disk 248 significantly limits the metal-to-metal contact path between the output shaft 222 of the actuator 220 and the die actuator linkage 252. This helps protect the actuator assembly 200 from the damaging heat of the slot die. For example, slot dies typically operate at temperatures exceeding 300°F (149°C). In contrast, components of the actuator assembly 200, including the motor 210 and the sensor 230, may experience limited functionality or even permanent damage when exposed to temperatures exceeding 130°F (54°C). Thus, the zero-backlash coupler 240 can function to maintain the temperature of the actuator assembly 200 at or below 130°F (54°C). In some embodiments, the metal disk 246 can also be formed from a non-metallic material to prevent metal-to-metal contact between the output shaft 222 of the actuator 220 and the die actuator linkage 252. Such a feature can further thermally isolate the actuator assembly 200 from the slot die housing.

[0032] Although not described in full herein, the actuator assembly 200 may include further options and advantages as described elsewhere in U.S. Pat. Nos. 9,044,894 (Loukusa et al.), 9,216,535 (Trice et al.), 9,579,684 (Yapel et al.), and 9,744,708 (Loukusa et al.).

[0033] Alternative actuator mechanisms are possible. In other embodiments, the actuator mechanism includes one or more of a thermally adjustable bolt, a differential bolt, a piezoelectric actuator, a pneumatic actuator, and / or a hydraulic actuator. In one embodiment, the actuator mechanism includes a thermally adjustable bolt, and a technique for retrofitting a slot die with the thermally adjustable bolt includes evaluating a cross-web profile of the extrudate after it exits the applicator slot and adjusting the relative position of one or more of the actuator mechanisms with its respective thermally adjustable bolt so that the cross-web profile of the extrudate after it exits the applicator slot more closely matches a preselected cross-web profile.

[0034] In an exemplary actuator mechanism, the applicator slot can be adjusted by applying a compressive or tensile load to the flexible die lip using a lever supported by a rotating shaft as a fulcrum, with an actuation rod that is axially displaced by the body of the slot die. The rotational force of the lever can be converted into a force along the axial direction of the actuation rod, which becomes a compressive or tensile load acting on the flexible die lip. The lever can apply a force directly to the actuation rod at the point of application of the lever.

[0035] In another embodiment, a thermally adjustable bolt automatically adjusts the applicator slot using multiple adjustment pins coupled to respective thermoelectric elements disposed on the flexible die lip. The thermoelectric elements may be controllable by a controller to adjust the applicator slot through the action of mechanical forces applied to the flexible die lip by the corresponding adjustment pins through expansion or contraction of the thermoelectric elements. As a further option, the actuator mechanism can include providing at least two adjustment pins and / or thermoelectric elements that are adjusted simultaneously.

[0036] Further aspects of the above, along with other variations, are described in U.S. Pat. No. 9,700,911 (Nakano) and PCT Patent Publication No. WO2019 / 219724 (Colell et al.).

[0037] 4A is a flowchart illustrating an exemplary workflow 400 for incrementally adjusting the position of each of a plurality of actuators of a slot die during an extrusion process to achieve a target extrudate thickness profile. While not intended to be limiting, workflow 400 is described with respect to slot die 10 (FIGS. 1 and 2), actuator assembly 200 (FIG. 3), and controller 300 (FIG. 3) during an extrusion operation. In various embodiments, the described techniques can also be utilized with strip coating, film slot dies, multilayer slot dies, hot melt extrusion coating dies, drop dies, rotating rod dies, adhesive slot dies, solvent coating slot dies, aqueous coating dies, slot-fed knife dies, extrusion replication dies, vacuum contact dies, or other slot dies.

[0038] The first step in workflow 400 is to provide a profile measurement (block 402) on the extrudate obtained from the applicator slot of the slot die. Known measurement techniques include the use of beta gauges. Beta gauges measure the normalized mass of a web by penetrating the web with radiation emitted from a radioisotope source (e.g., based on promethium, krypton, or strontium) and detecting the resulting attenuation of beta particles. Because the measurement footprint of the beta gauge is small, profile measurements are taken at various locations across the width of the web and tracked over time to account for time differences between measurements taken along the web. A complete profile measurement may require averaging of several scans (potentially six or more scans) to have statistical significance.

[0039] Alternative metrology methods are also available, including those based on direct thickness measurement such as UV fluorescence, laser triangulation, reflectance methods such as gamma backscattering, and capacitance-based caliper (i.e., web thickness) sensing. Some of these methods can capture the web thickness profile across the entire width of the web at once. It has been determined that applying the workflow 400 described herein while measuring the coating thickness along its entire width at once can provide a 4x to 12x improvement in cycle time compared to workflows using traditional scan-based metrology methods.

[0040] Data from profile measurements typically includes many individual measurements for each cross-web scan. Measurement systems can process the raw individual measurements into fewer data values, essentially utilizing a form of cross-web and down-web averaging as defined by Equation 1 below.

[0041]

number

[0042] The provided technique can be used for both continuous full-web coated extrudates and stripe-coated extrudates. In full-web coatings, the multiple measurements represent the entire coated width. In stripe coatings, the measured thickness values ​​correspond to multiple coated areas spaced apart from one another along the cross-web direction, so that the measurements cover only the coated stripes and not the uncoated lanes between them.

[0043] Referring back to FIG. 4A, the next step is to perform a web-to-die mapping of these profile measurements (block 404). When mapping the gauge measurements to the die, certain assumptions can be made. First, since modern metrology systems have edge detection, it can be assumed that the data received from the profile measurements is taken edge-to-edge. Second, it is assumed that the gauge data corresponds directly to the actual extrudate weight / thickness. If the web is placed on a release liner, the basis weight of the release liner is subtracted and the data is assumed to represent the coat weight only. Finally, for simplicity, it is assumed that there is a symmetric neck-in from both edges of the extruded web.

[0044] The number and locations of coarse-grain measurements may differ from the number and locations of die lip actuators, and therefore a general method of mapping measurements to actuator locations is desirable. The film neck-in calculation can account for the effects of different flow rates, draw distances, and web speeds. In one embodiment, coated web coordinates in physical units (e.g., millimeters) are mapped to coordinates in cross-web units (e.g., in simple actuator coordinates, actuator #1 has a coordinate of 1, actuator #2 has a coordinate of 2, etc.). This process allows for mapping from coating measurement array numbers to physical coated web locations, and then, via a web-to-die mapping function, to physical die locations. Here, the coating measurement array can represent, for example, the latest array of thickness measurements generated each time the scanner traverses the moving web.

[0045] It is also possible to reverse the process (i.e., die-to-web mapping) to include similar steps in the reverse order. Instead of direct mapping, the inverse procedure can involve interpolation of the coating measurement array: direct mapping of die actuator positions to corresponding locations on the coated web (via a die-to-web mapping function, i.e., the inverse of the web-to-die mapping function), followed by interpolation of this direct mapping to the actual measurement array values ​​(low-order polynomial interpolation, such as a Hermite cubic polynomial characterized by a continuous function and first derivative).

[0046] When mapping function evaluation from either the web to the die or from the die to the web is computationally inefficient or impractical, the reverse procedure may be preferable, as it tends to allow for significantly faster response times. Additionally, it may be useful to simplify the mapping by avoiding the use of physical coordinates altogether. The simplest mapping may be directly from the die actuator coordinates to the coating measurement array coordinates (coated web coordinates based on their relative position compared to the measurement array number). This logical extension is likely to be the most computationally efficient (e.g., using third-order polynomial interpolation), but may be less transparent to understand because no physical coordinate system is involved.

[0047] With an appropriate die-to-web or web-to-die mapping function, it may be advantageous to utilize this function elsewhere in the profile control strategy. For example, the mapping can incorporate cross-web variations in control parameters such as controller gains or actuator interactions through the coating bead flow field. This can be done in an ad hoc manner based on manufacturing experience, or in a more scientific manner using a fluid-mechanical model of the coating bead flow field.

[0048] In some embodiments, this mapping incorporates any scan position to die bolt position mapping already built into the measurement system. Various commercially available systems, such as Profile Control Solutions from NDC Infrared Engineering Inc. (Irwindale, CA), can perform web-to-die mapping.

[0049] Following the web-to-die mapping step, the profile control strategy proceeds to a data compression step (block 406), sometimes referred to as a coarse-graining step. Typically, coating measurement arrays contain many more elements than lip actuators (500-1000 measurement elements compared to 20-40 actuators). Advantageously, data compression (block 406) can improve computational efficiency and provide faster controller response when processing multidimensional profile data, especially when complex mechanical and fluid dynamic interactions are involved. Data compression is sometimes referred to as "coarse-graining" because it can translate small-scale variations into the scale of variations that the actuators can affect.

[0050] Once the actuator zone and measurement zone are defined along a common coordinate system (which generally corresponds to the die position), the transformation between the two can be performed using a variety of methods.

[0051] One method is based on the use of an overlap matrix OL defined according to Equation 2.

[0052]

number

[0053] where S(a,x,l) is a square wave function that depends on the cross-web coordinate a, the zone position x, and the zone width l. In the above notation, A i is the cross-web coordinate of the ith actuator zone, and G j is the cross-web coordinate of the jth measurement zone, and wi a and w j s are the actuator zone width and scan-based zone width, respectively, defined along the cross-web coordinate.

[0054] This overlap matrix represents the convolution of the actuator zone of influence with the measurement zone. This rectangular matrix contains one row for each actuator and one column for each measurement. The coordinate system used above (i.e., coordinate a) can be based on the actuator number, where actuators are numbered from 0 to N. act -1, where N act is the total number of actuators. The spacing between actuators is unity (i.e., 1) in this coordinate system, and the zone of influence of each actuator extends from its position to 1 / 2 the actuator spacing in both cross-web directions. That is, the zone of actuator j is from j-1 / 2 to j+1 / 2. This is represented by the square function S(a,j,1) defined above for actuator j. Similarly, each measurement has a zone associated with it, and this zone is typically expressed as S(a,M i ,D x ) the cross-web zone D of the data points that make up that value x is.

[0055] When coating stripes, D x can represent the width of the stripe associated with the measurement. For full width coatings, D x can represent the overall cross-web width of the coating. Applying Equation 2 above to the exemplary stripe coating, the actuator zone width

[0056]

number

[0057] By calculating the overlap matrix above for a stripe or full width coating, a general method for converting measurements into values ​​associated with individual actuators can be realized. For example, the measurement M s A given array of M can be simply multiplied by the overlap matrix to obtain the array M associated with the actuator according to Equation 3 below: a can be created.

[0058]

number

[0059] In the above equation, the superscript a denotes an actuator-based value, and the superscript s denotes a scan-based value.

[0060] Alternative methods are possible, for example, interpolation and extrapolation methods can be used to take into account the influence of actuator zones on measurement zones, for example, spline interpolation can be used to map measurements to individual actuators.

[0061] Table 1 (next page) shows how stripe position and coating thickness can be converted between stripe position and actuator position using two different methods. Here, the third column represents the measured coating thickness, and the last column shows the overlap matrix mapping data. This chart shows the overlap matrix mapping of coating thickness to each actuator position. The fifth column shows spline-interpolated values. These do not exactly match the overlap matrix values, reflecting the fact that these values ​​were mapped using two different methods.

[0062] In some embodiments, mapping the thickness values ​​to corresponding positions along the width of the slot die includes expanding and / or adjusting the measured thickness values ​​to account for positional variations at one or more transverse edges of the extrudate, which can include not only the global leftmost and rightmost edges of the extrudate, but also local edges in a stripe coating extrusion process.

[0063] One implementation applies when a metrology system uses automatic edge detection to record non-physical input data as a result of the absence of coating during at least some portion of the extrusion process. This detectable absence of coating tends to occur when the coating edge moves left or right over time, causing one or more "0" thickness values ​​to be recorded at some locations. To avoid averaging these zeros with actual thickness values, any zero values ​​can be automatically replaced with the last known thickness value reported by the nearest adjacent metrology sensor.

[0064] Another implementation applies when significant die neck-in is present, a phenomenon that tends to occur when there is a long stretching zone after the extrudate exits the slot die. In these instances, web-to-die mapping can be performed according to a mapping function that not only redistributes thickness values ​​to reflect the reduction in lateral width, but also supplements the measured thickness values ​​with interpolated values. Expanding to include these interpolated thickness values ​​can ensure sufficient resolution of thickness measurements along the transverse direction, thereby allowing meaningful extrudate thickness data to be provided for each actuator zone. Such expansion can take many forms, including linear interpolation, trapezoidal mapping routines, spline interpolation, or even uniform replication of thickness measurements across the width of the measurement zone.

[0065] [Table 1]

[0066] The degree of expansion can be determined based on a calculation of the minimum gauge data resolution for a given extrusion process. This effective gauge resolution can be determined based on, for example, the degree of neck-in, the mapping function, and the width of the actuator zones used. Generally, it is preferred to have at least two thickness measurements per actuator zone, and more preferably three or more.

[0067] The workflow 400 then proceeds to obtain the profile deviation (block 408). One simple technique uses a square wave convolution with the difference between the measured and desired coating weights. The square wave has a value of 1 within 1 / 2 actuator spacing on either side of the actuator position and a value of 0 everywhere else. The width between actuators is defined as unity (i.e., 1) in this coordinate system. This difference is then converted into a residual value R, which can be calculated for each actuator i according to Equation 4 below: i can be represented as a set of R i =∫Sq[A i ](t measured -t desired )dX (Equation 4) where S is the cross-web coordinate A i is a generalized square wave function centered at t measured and t desired represent the measured coating weight and the desired coating weight, respectively.

[0068] According to Equation 4, data compression combines knowledge of the extrudate thickness measured at the die with input for the desired cross-web profile of extrudate thickness that defines the target profile (block 407) to obtain the profile deviation (block 408). Thus, the residual vector R represents a measure of how the current coat weight measurement deviates from the target profile (block 407) at each actuator position.

[0069] The target profile (block 407) can be set manually by an operator or automatically by a controller according to a control algorithm for the application at hand. In some applications, a flat profile is desirable, but there are situations in which other profiles are preferred. For example, a "dog-bone" shaped target profile may be desirable to control edge instabilities in the extrusion process, such as those associated with draw resonance or edge scalloping.

[0070] While the aforementioned square wave convolution is simple to implement, it does not consider some important aspects of flow field physics. For example, there is no interaction between the residual vector elements and neighboring actuators, but from the perspective of the flow field, there may be a wide zone of influence. In general, the movement of a single actuator affects the flow field globally through the relative flow rates, and in other instances locally in the vicinity of each actuator.

[0071] The strength and extent of the actuator interaction can be affected by a variety of factors, including the slot die geometry and structure, the materials of construction, the coating fluid, the coating method, and the coating bead flow. One way to capture the actuator interaction in the data compression step is to construct a residual vector using a flow sensitivity convolution, ∂t, as shown in Equation 5. measured / ∂DH i indicates the flow rate sensitivity, i.e., the sensitivity of the coating weight t (extrudate thickness) to changes in slot height (DH).

[0072]

number

[0073] In this configuration, each residual vector element represents the sensitivity of that portion of the coating weight profile to slot height at a particular actuator cross-web position.

[0074] Alternatively, the residual vector can be provided in terms of filtering, such as based on a Fourier transform or a wavelet transform, in which the residual vector contains specific modes that, when removed from the coating profile, result in the desired thickness profile. For example, a wavelet transform with two parameters, wavelet position a and scale σ, is given by Equation 6 below: R i =Ψ i (a,σ)=∫Ψ i (a,σ)(t measured -t desired )dX (Equation 6)

[0075] Although flow rate sensitivity and wavelet transform convolution can be quite different, it was surprisingly discovered that thickness sensitivity to actuator movement can be closely represented by the wavelet function of the extrusion die. In particular, the Ricker Wavelet (sometimes called the "Mexican Hat Wavelet" because of its shape) provides a very close approximation to thickness sensitivity to actuator movement for actuators not adjacent to the side edges of the slot die.

[0076] The Ricker Wavelet shape offers a unique technical advantage because it extends both above and below the x-axis so that the entire wavelet has a total integrated area of ​​0. For a particular slot die, this shape can accurately simulate the decrease in fluid flow that occurs at adjacent actuator positions when fluid flow increases at a given actuator position.

[0077] Wavelet transforms can potentially be more useful than Fourier transforms because they provide more localized information. Fourier transforms typically extract only frequency information, whereas wavelet transforms extract both wavelet scale (analogous to wavelength) and position. However, when using actuators to control choker bars or flexible die lips, certain wavelet transforms are particularly useful for guiding adjustments near the actuator and at the same general scale as the magnitude of the actuator adjustment. Advantageously, wavelets are an efficient representation because they capture only the portion of the thickness profile that the actuator can mediate.

[0078] In an exemplary workflow, the wavelet position for a given actuator is centered on the actuator and the wavelet scale is based on the calculated flow rate / actuator sensitivity, both of which are used in transforming the slot height profile data.

[0079] Referring back to FIG. 4A , once the residual vector R is determined, an appropriate slot height adjustment can then be determined (block 410). When predicting the adjustments to the actuators engaged with the flexible die lip or choker bar appropriate to achieve the target profile (block 407), it is advantageous to apply a pressure die deflection model that allows the effects of fluid flow through the die to be separated from the mechanical effects associated with a slot die in the absence of fluid flow. A suitable model can predict the pressure profile across the applicator slot and the corresponding amount of pressure deflection in the flexible die lip or choker bar. In some cases, it may be possible to empirically measure this pressure-based deflection (i.e., pressure deflection D).

[0080] The pressure deflection can be accounted for by adding it to the calculated slot height, and the sum is compared to the pressure-compensated slot height profile from the current slot height (block 409), which includes the pressure deflection due to fluid flowing through the applicator slot. Then, assuming any change in the amount of deflection D (i.e., D), is small, the deflection D due to fluid alone can be subtracted from the pressure-compensated slot height profile (H+D) to obtain the zero-pressure slot height profile (H), thereby providing a nexus for calculating the target actuator position.

[0081] In an exemplary method, the pressure die deflection model is based on the local dependence of coating thickness on slot height. This can be modeled by deriving a Taylor series around the current slot height. Assuming there are only small perturbations from the set point, only the linear terms need to be retained, as shown in Equation 7.

[0082]

number

[0083] This linearized form may be sufficient if many small profile corrections are made. However, for rapid control, large profile corrections are required. Unfortunately, using the linearized form in these cases can result in unstable control.

[0084] In one embodiment, the nonlinear control scheme can be set in a non-interacting situation: For each actuator, an adjustment to the slot height is made according to Equation 8.

[0085]

number

[0086]

number

[0087]

number

[0088] Equation 8 still requires a prediction of D, or die deflection, across the coating die width, which can be obtained using Equation 8A. D=K0(1+K1x 2 +K2x 4 +K3x 6 )P (Formula 8A) where K1, K2, and K3 are polynomial coefficients, P is the internal die cavity pressure, and x is the cross-web position along the applicator slot as measured from the die center.

[0089] Through Equation 8A, the pressure die deflection model considers the effects of die cavity pressure, but Equation 8A is based on certain assumptions. For example, lip deflection is assumed to be approximately symmetric about the die centerline and is adequately approximated by a low-order polynomial. The polynomial coefficients can be provided by calculation and / or measurement. The deflection of the flexible die lip or choker bar can also be assumed to be proportional to the die cavity pressure. This assumption is generally well justified when P is the actual die cavity pressure, but may require modification if P is measured significantly upstream of the die, for example, at the feed hose or compounder exit. Furthermore, it is recognized that not all dies necessarily deflect symmetrically (e.g., end-fed dies may not), and therefore a different appropriate correlation model of die deflection is used in those cases.

[0090] In one embodiment of a flexible die lip configuration, the die deflection is characterized using an array of measurement indicators that detect the deflection of both the fixed die lip side and the flexible die lip side. Alternatively, a taper gauge can be used to directly measure the die slot height under several conditions: (1) with flow rate and pressure, and (2) without flow rate and pressure, and D is calculated by subtracting slot height (2) from slot height (1).

[0091] Advantageously, the nonlinear nature of this scheme allows for rapid control using relatively large corrections without encountering unstable control, but the number of control steps can be further reduced by also incorporating interactions between the actuators that are transmitted through the flow field.

[0092] In a preferred embodiment, the thickness sensitivity represented by the wavelet function in Equation 6 above can be used and combined with the previous nonlinear functional form to provide a nonlinear interaction form of the coating thickness as shown in Equation 9.

[0093]

number

[0094] In the typical Newton-Raphson situation, the actuator position update proceeds using the solution of the following matrix equation, which incorporates the flow field-actuator interaction, as reflected in Equation 10:

[0095]

number

[0096] This is essentially equivalent to the nonlinear non-interacting form above, except that the wavelet transform convolution is preconditioned by the fluid flow interaction matrix as shown in Equation 11.

[0097]

number

[0098] In its final form, the actuator update can proceed to obtain the slot height adjustment according to Equation 12.

[0099]

number

[0100] To fully account for the dependence of this profile control algorithm on the hydrodynamics of the actual fluid, the exponent E should accurately reflect the thickness sensitivity to slot height changes. Manifold flow calculations provided by commercially available calculation software can provide this information. Calculations of flow sensitivity to gap changes of different magnitudes can be used to determine the most physically relevant exponent E. This can be obtained by measuring coating thickness over a range of different slot heights and applying the general power-law hydrodynamic relationship in Equation 12A.

[0101]

number

[0102] The exponent E can be estimated using a curve fitting method based on the relationship in Equation 8, which can be re-expressed as shown below in Equation 12B:

[0103]

number

[0104] The curve fitting examples show that different slot dies can behave differently, resulting in different indices. In this way, the profile control algorithm can be customized to perform optimally depending on the die design and mode of use.

[0105] Next, given knowledge of the profile deviation (block 408) obtained above and the current slot height (block 409), an appropriate slot height adjustment (block 410) is determined, where the current slot height (block 409) incorporates both the original slot height profile (without pressure correction) and the pressure correction to create a pressure corrected slot profile. The slot height adjustment (block 410) involves application of Equations 8 and 12 and subtraction of the pressure correction to determine a predicted zero-pressure slot height profile corresponding to the current slot height (block 409).

[0106] Using the zero-pressure slot height profile obtained above, controller 300 can then predict a suitable set of individual actuator settings to achieve the desired slot height profile. In a preferred embodiment, this set of individual actuator settings is based on a plurality of actuator settings corresponding to a preselected cross-web profile. As described below, a plurality of individual actuator settings can be determined for a series of preselected cross-web profiles, thereby providing a predictive model that is later used to determine actuator adjustments (block 412).

[0107] In various embodiments, the controller 300 may retrieve the pre-selected cross-web profile from a non-transitory computer-readable medium or may receive the pre-selected cross-web profile from user input.

[0108] In a preferred embodiment, the prediction of a set of individual settings from multiple individual settings and / or the zero pressure current slot height (block 409) can be obtained through a mathematical construct called a stiffness matrix.

[0109] The stiffness matrix converts the actuator position profile into a slot height profile, which excludes the effects of deflections due to fluid flow through the applicator slot. Whether the adjustable gap is a final die slot or choker bar slot, knife coater, or other, the slot height profile can be determined by empirical measurements at numerous cross-web locations as expressed by Equation 13. h i ≡h(x i ), where i is from 1 to N m (Equation 13) In the formula, h i is the slot height at the position of actuator i, and x i is the cross-web coordinate, and N m is the number of measurement points.

[0110] The slot height values ​​at these measurement points depend on the position of the die lip actuator. In general, the functional form of this dependence can be expressed by the Taylor series in Equation 14:

[0111]

number

[0112] The stiffness matrix K is given by

[0113]

number

[0114]

number

[0115]

number

[0116] The stiffness matrix K is N m ×N a and the jth actuator position A j Measurement point x due to changes in i If the die lip were infinitely flexible, almost all of the stiffness matrix elements would be zero only if the actuator position coincided with the measurement point, and the stiffness matrix elements would be non-zero. If the measurement point were co-located with the actuator, in this scenario the stiffness matrix would be diagonal, as represented by Equation 16.

[0117]

number

[0118] In reality, the die lip and choker bar are not infinitely flexible, and changes in individual actuators change the slot height in a region around the actuator location. How far this region of influence extends is reflected in how far the non-zero components of the diagonal appear in the stiffness matrix. In this way, the lip stiffness matrix incorporates in its components all interactions between actuator movement and slot height changes across the entire width of the die.

[0119] Empirically measuring the stiffness matrix of the die lip can be accomplished by finite differencing. The difference between slot height measurements across the die slot width as each actuator is displaced independently provides the rows of the stiffness matrix. For example, central differencing proceeds as shown in Equation 17:

[0120]

number

[0121] In Equation 17, Δ j means that all actuators are in their initial positions except for the jth actuator, which is displaced by a distance Δ, first in the positive direction and then by a similar displacement in the negative direction. The difference between the slot height measurements at these two settings relative to the magnitude of the displacement provides a central difference approximation to the stiffness matrix elements. Central differencing in this method provides greater accuracy than forward or backward differencing, but requires twice as many measurements. Additional measurements at different displacements allow the use of higher-order differencing techniques, but central differencing provides reasonable accuracy with limited effort.

[0122] Numerical difference theory suggests that there is an optimal displacement scale that should be used in central differences to produce the most accurate stiffness matrix entries (Richard L. Burden et al., Numerical Analysis 131-133 (2 nd(ed.) 1981). The error inherent in central differencing is quadratic in displacement size, i.e., proportional to the square of the displacement. However, slot height measurements have uncertainties associated with them, and therefore, when both of these errors are considered, the optimized actuator unit displacement Δ opt is given by Equation 18.

[0123]

number

[0124] where e is the slot height measurement error and M is the third derivative of the gap-actuator relationship

[0125]

number

[0126] A typical bump test used actuator displacements of approximately 1 mil (25 micrometers) of change in slot height. Generally, significant interaction occurred between the actuators, with a change in actuator position resulting in a change in slot height that was two or three actuators away. In some embodiments, this is represented by a band stiffness matrix. Often, the band stiffness matrix can be assumed to be symmetric. For example, the matrix may have significant non-zero elements only along the five diagonals of the matrix (the main diagonal plus the two diagonals on either side). International Patent Publication No. WO 2012 / 170713 (Secor et al.) provides a more detailed example of how "bump test" data can be used to derive a stiffness matrix.

[0127] Advantageously, an inverse stiffness matrix can also be determined that converts the actuator position profile to a slot height profile, excluding the effects of deflection due to fluid flow through the applicator slot. Conveniently, the inverse stiffness matrix fit can be obtained using the same bump test measurements used to fit the stiffness matrix. Ringing can be further reduced by performing a dual fit method that incorporates a direct transformation using both the stiffness matrix and the inverse stiffness matrix. This is shown in FIG. 12, which illustrates the reduced ringing achieved using the dual fit method relative to a single fit method that includes only the stiffness matrix. Also shown in FIG. 12 is the ringing reduction achieved using a direct raw differencing technique based on Equation 17.

[0128] The main advantages of performing fits to empirical bump test data to obtain a banded matrix structure include computational efficiency, reduced computer memory storage requirements, improved accuracy by mitigating the effects of outliers and errors in data input, and improved robustness in the operation of the PLC code. In addition, a parameterized model representation of the stiffness matrix is ​​highly desirable for control system safety benefits. Verifying that a small number of parameters have the correct values ​​and are not corrupted is a much easier task than verifying the values ​​of the entire stiffness matrix, which can contain 400 to over 1000 elements. Further details regarding these fits are provided below.

[0129] Once the stiffness matrix is ​​determined, the slot height measurements can be used in two ways. First, for any set of actuator settings, the resulting slot height profile can be calculated by multiplying the stiffness matrix by the actuator displacement. Second, the actuator settings that result in a specified slot height profile are calculated by multiplying the inverse of the stiffness matrix by the desired change in slot height. These calculations are represented by Equations 19 and 20, respectively. (h-h0)=K·(A-A0) (Equation 19) (A-A0)=K-1 ·(h-h0)(Equation 20)

[0130] Now, referring back to the actuator adjustment (block 412) of FIG. 4A, if the desired zero pressure slot height adjustment (block 410) is known, then the inverse stiffness matrix K -1 can finally be used according to Equation 20 to obtain the actuator adjustments (block 412) to achieve the target profile (block 407).

[0131] 3, where the controller 300 predicts settings for at least one actuator assembly 200 based on measurements from the sensors 230 and the operating motors 210, and places the output shaft 222 in a position corresponding to the predicted actuator adjustments (block 412). Once the controller 300 has thus determined the settings for the actuator assemblies 200 of the slot die 10 corresponding to the preselected cross-web profile, the slot die 10 is operated by passing the extrudate through the fluid flow path and out of the applicator slot 6, with the actuator assemblies 200 positioned according to the actuator adjustments (block 412).

[0132] A ringing constraint (block 411) can be applied to arrive at a different solution if the degree of ringing is considered important. This is based on a discovery made when calculating the actuator adjustments (block 412) without any such constraint. This somewhat surprising discovery arose when comparing different model representations of the stiffness matrix K. The exact representation of the stiffness matrix is ​​given by the inverse stiffness matrix K. -1 It was observed that the results do not necessarily translate into a good representation of the

[0133] Physically, the inverse stiffness matrix represents the combination of actuator movements that must occur to change the slot gap at only one location. To change the gap at only one location, multiple actuators must be adjusted to counteract undesirable actuator interactions. Figure 5 also illustrates what is known as "ringing." To achieve a particular gap profile, adjacent actuators tend to move in opposite directions by a large magnitude. That is, adjacent actuators tend to work in large opposition to each other. Excessive ringing, as shown in the target slot height and actuator setting (the latter in arbitrary units) profiles in Figure 5, is undesirable because it can unnecessarily push the actuators to their force and bending limits, thereby limiting the degree of coating uniformity that can be achieved with the control system.

[0134] If the stiffness matrix and inverse system are linear systems, then there is one unique set of actuator settings corresponding to a particular set of slot gap positions. A set of actuator settings with substantial ringing is not flawed in any way; it is the only set of settings that will result in the desired slot gap profile given a particular stiffness matrix. However, as shown below, some additional uncertainty in the slot gap profile can be accepted in exchange for reducing the degree of ringing.

[0135] First, the degree of ringing can be quantified. A convenient ringing measure is the root mean square of the actuator curvature (Equation 21), hereafter denoted R value Or abbreviated as ring value.

[0136]

number

[0137] Implicit in the ring value is the fact that the curvature is calculated at the actuator scale (i.e., a coordinate system where the distance between actuators is 1). In some embodiments, Equation 20 calculates each instance of actuator setting A as a function of its deviation from the reference point value (AA o ) Further, each A can be defined as an actuator position, corresponding to an actuator setting, and having distance units (e.g., mm).

[0138] A suitable penalty function may be based on the ring value above, but may also consider other metrics such as the mean absolute ringing sawtooth amplitude as provided in Equation 22 below.

[0139]

number

[0140] Following this quantification step, it is desirable to penalize the actuator movement so that ringing is below some value, instead of accepting some small deviation in slot gap accuracy. While the backlash of the actuator assembly is generally small and the movement is highly repeatable, the uncertainty in the actuator position is on the order of 2-3 micrometers. Given this variability, small slot gap deviations (e.g., at least 2-3 micrometers) can be accepted if they provide benefits such as reduced ringing and enhanced control schemes.

[0141] A common method for enforcing such constraints is the Lagrangian multiplier method. Essentially, the original system of equations for the motion of a set of actuators is expanded with a penalty parameter called the Lagrangian multiplier and multiplied by the constraints. Physically, the Lagrangian multiplier represents the uncertainty in the solution of the system of equations that is accepted in exchange for obeying the constraints. In the provided control scheme, the constraints are applied by the ringing measure mentioned above, referred to herein as the ring constant ("R"). max The goal is to keep the force constant below a certain value, which can be in either distance units or actuator units, called the force constant . Mathematically, this can be expressed as maximizing a function f(x,y) subject to the condition g(x,y)=c, where both f and g are assumed to have continuous partial derivatives.

[0142] To solve this, a Lagrangian function L can be defined and solved according to Equation 23. L(x,y,l)=f(x,y)-l(g(x,y)-c)(Equation 23) where l is the Lagrange multiplier, which is then solved based on Equation 24.

number

[0143] Next, a solution is found for the expanded set of equations, where the values ​​of the Lagrange multipliers are the additional unknowns to solve for. Various solution techniques are possible for the expanded system, and the Lagrange bound algorithm is a convenient choice. This algorithm allows the solution of the equation system to proceed in the same way as if there were no ringing constraint, but then follows it with several matrix multiplication operations to generate the constrained solution. A basic application of the Lagrange bound algorithm can be derived, for example, from Equation 24A below.

[0144]

number

[0145] In generalizing the application of the ring constant, two observations were made. First, with respect to the curvature / bending of the die lip or choker bar, one would expect the ring constant to increase if adjacent actuators were spaced further apart. This may also be a function of the leverage provided to the actuators in manipulating the flexible die lip or choker bar. This leverage or mechanical advantage is calculated as the ratio of actuator travel to die lip travel and is referred to herein as its Motion Amplification. See Equation 25 below.

[0146]

number

[0147] Second, the ring constant can be increased if the motion amplification is increased with respect to the curvature or bending of the die lip or choker bar, i.e., moving each actuator more for a given die slot change.

[0148] This allows for the creation of a unitless metric called the Ring Factor, defined by Equation 26 below.

[0149]

number

[0150] The ring factor is a generalized form of the ring constant that takes into account the actuator spacing and motion amplification due to actuator movement relative to changes in the die slot via flexible die lip (flexure) or choker bar movement. To illustrate the application of this relationship, let us double the actuator spacing while keeping the die lip curvature deflection constant, and also double the ring constant (R maxSimilarly, if the motion amplification is doubled, for example, by doubling the moment arm length of the flexible die lip, the ring constant (R max ) can also be doubled to provide the same degree of die lip deflection.

[0151] 7-11 show exemplary progressions of slot profiles calculated using the ringing constraint as described herein for a slot die configured with flexible die lips. The Y-axis values ​​in these bar graphs represent either the slot height or coating thickness dimension, as shown below, and have been automatically scaled to show cross-web variation along the applicator slot.

[0152] Figure 7 shows the difference between the current actuator position and the reference point actuator position in actuator units. The deviation from the reference point represents the actuator unit displacement from the "reference point" or "flat slot" position, which is used along with the stiffness matrix to calculate the slot position (without pressure deflection effects) and ultimately the actuator position corresponding to this slot position. In the inset of Figure 7, "SP" refers to the ring constant, "Before" refers to the ring value before application of the Lagrangian boundary algorithm as described above, and "After" refers to the ring value after application of the Lagrangian boundary algorithm.

[0153] Figure 8 shows a calculated flexible die lip profile representing the position of one side of the applicator slot relative to the neutral slot height profile. This profile shows the degree to which the flexible die lip is deflected away from its rest state at various positions along the width of the applicator slot.

[0154] FIG. 9 shows the neutral (i.e., resting) slot height profile, which represents the design die slot obtained when there is no flexing of the flexible die lip.

[0155] Figure 10 shows the coating thickness profile resulting from this die slot adjustment process.

[0156] Figure 11 shows the calculated die slot obtained by applying the stiffness matrix method to the current actuator position combined with the reference point slot position and reference point actuator position. Note that the reference point slot position and reference point actuator position can be determined from Equations 19 and 20 above, respectively.

[0157] Exemplary values ​​of the ring factor calculated for various slot dies are shown below in Table 2. As shown in the table, the dimensionless ring factor ranges from 0.0018 (Die N) to 0.0096 (Die H), averaging 0.0041 across all dies considered.

[0158] [Table 2]

[0159] Below are two simulation examples showing how this technique can reduce ringing. Table 3 below shows the results for various R max This figure also shows the reduction in ringing achieved in the rotating rod die embodiment with choker bars for an R of 1000. This is also shown in Figure 6, which shows a comparison between the actuator setting profile plotted against the reference point profile. max This demonstrates the benefit of imposing a ringing constraint using , which reduces the ring value from 2800 (without ringing constraint) to 852 (with ringing constraint).

[0160] [Table 3]

[0161] Thus, at a high level, the residual vector can be used to determine the predicted degree of ringing for a set of proposed actuator positions (or actuator settings), and then a ringing constraint is applied to obtain a modified set of actuator positions that reduces the degree of ringing relative to the predicted degree of ringing.

[0162] Collectively, these techniques can provide significantly greater efficiency in die control methods. Increased efficiency can be derived, for example, from reducing the number of iteration cycles in workflow 400 of FIG. 4A to achieve the best achievable coating thickness profile. Advantageously, this can result in less waste in large-scale manufacturing processes. More accurate and faster convergence can also help eliminate higher-frequency disturbances to coating thickness in manufacturing operations than would otherwise be possible.

[0163] Following the performance of the actuator adjustments (block 412), the workflow 400 resumes with an updated profile measurement (block 402), and the process can begin again. This profile measurement (block 402) occurs after a period of time, allowing for iterative adjustments of the actuators. Over the course of many iterations, successive refinements can be made to the individual actuator settings so that the extrudate profile converges to and closely matches the target profile (block 407).

[0164] The workflow 400 can be repeated until the controller 300 determines that the predicted set of settings cannot be improved and / or at periodic intervals to maintain the desired cross-web profile. The set of individual settings can be saved as a recipe for future take-outs where similar materials, extrusion or coating properties, and processing conditions are required, and for future use minutes, hours, or years later.

[0165] As suggested earlier, the inverse stiffness matrix K-1 Obtaining a good representation of K can be difficult. Physically, the inverse stiffness matrix represents the combination of actuator movements that must be made to change the slot gap at only one position. Therefore, the selection of K can reveal tendencies to produce ringing (i.e., larger ring values) in the actuator position. The elements of the stiffness matrix K can be determined by bump testing and a difference equation such as Equation 17. While this "raw" stiffness matrix from difference (i.e., the raw difference matrix) can be used for die control, it has been discovered that a statistical fitting of the raw matrix to a model representation can mitigate the effects of experimental error on K as well as account for the banded matrices that are expected due to the limited range of influence of neighboring actuators.

[0166] Stiffness matrix K and / or inverse stiffness matrix K -1 It has been discovered that an advantageous method for fitting (Eq. 19) is to use Bayesian statistical methods such as nested sampling and / or the Monte Carlo Markov Chain algorithm. It has been discovered that small changes in the stiffness matrix result in unusually large changes in its inverse. A preferred procedure involves fitting model representations of both the stiffness matrix and its inverse to ensure that the operations of Eqs. 19 and 20 work well. A Bayesian nested sampling procedure for model comparison can be found in John Skilling's "Nested Sampling for General Bayesian Computation," Bayesian Analysis, Vol. 1, Number 4, pp. 833-860 (2006)

[0167] In a Bayesian nested sampling framework, the likelihood of a slot gap measurement given an actuator setting is calculated, as well as the likelihood of an actuator setting given a slot gap measurement. In this way, both the stiffness matrix and the inverse matrix multiplication operation are sampled. In addition, bump test data includes measurements from coordinated motion, where all actuators move simultaneously with a single actuator motion. In this way, the behavior of row or column sums of the stiffness matrix and its inverse matrix is ​​queried in addition to values ​​near the stiffness matrix diagonal. As an additional advantage, these nested sampling methods can determine parameter probability distributions across parameter space, which can incorporate measurement uncertainty and prior knowledge. Different model representations of the raw stiffness matrix can be compared and ranked based on model data likelihood or model evidence. In some embodiments, the highest evidence model with the most likely parameter values ​​(mode or mean) is used to calculate the actuator settings. Parameter value estimates for the highest evidence stiffness matrix model can be obtained using, for example, a Monte Carlo Markov chain algorithm.

[0168] Another useful parameter is the band-wise motion amplification and stiffness matrix K, as shown in Equation 27. ij is the dimensionless stiffness movement number, which is the product of the sum of

[0169]

number

[0170] In this case, it is useful to account for the motion amplification using an adjusted flexible die lip length, represented by the flexible die lip length reduced by the measured distance into the die slot.

[0171] The stiffness movement number is calculated by combining the knowledge of the die type movement amplification to obtain the stiffness matrix

[0172]

number

[0173]

number

[0174] [Table 4]

[0175] In some embodiments, K ij The relative values ​​of can be predicted from the normalized stiffness parameters for similar die configurations.

[0176]

number

[0177]

number

[0178] Exemplary stiffness parameters, both normalized and unnormalized, are listed below in Table 5 for fluid bearing coating die "A" (from Figure 13).

[0179] This chart illustrates and demonstrates the characteristic relative influence of actuator position on its neighbors. Depending on the mechanical details of the die lip or choker bar in these particular examples, K ii It is observed that accounts for approximately 39% to 80% of the influence on the slot position. In an exemplary control method, the stiffness matrix and / or inverse stiffness matrix are calculated from a predetermined normalized stiffness.

[0180] [Table 5]

[0181] K i,i is in the range of 0.35 to 0.8, and K i,i±1 is in the range of 0.1 to 0.3, and K i,i±2 is in the range of 0.0 to 0.1, and K i,i±3 is in the range of 0.0 to 0.03 or 0.0 to 0.05, and K i,i±4 is in the range of 0.0 to 0.02.

[0182] In addition, the stiffness matrix K and the inverse stiffness matrix K -1 must accurately convert the physical actuator movement into slot height change and vice versa according to Equations 19 and 20. This overall power is given by K over the band of the stiffness matrix. ij The sum of

[0183]

number

[0184] In some embodiments, the techniques described herein, including wavelet-based techniques, can be used in combination with other methods to control coating thickness. For example, actuator adjustments based on these techniques can provide a more optimized or refined extrudate profile using cross-web heat distribution systems such as those described in International Patent Publications WO 2006 / 130141 (Karg et al.), WO 2006 / 130142 (Karg et al.), and WO 2006 / 130143 (Karg et al.). The wavelet scale and position of each transform can be tailored to the sensitivity and type of actuator being used, resulting in a more efficient control system.

[0185] An alternative workflow 500 is depicted in Figure 4B. Workflow 500 includes many steps essentially similar to those of workflow 400 of Figure 4A, including profile measurement (block 502), web-to-die mapping (block 504), data compression (block 506), obtaining profile deviation (block 508) based on a target profile (block 507), slot height adjustment (block 510), and actuator adjustment (block 512). Additionally, similar to workflow 400, slot height adjustment (block 510) optionally incorporates data representing a ringing constraint (block 511) and a current slot height (block 509). These operations have already been described above and will not be described again here.

[0186] Workflow 500 differs from previous workflows in that a slot height constraint (block 513) is applied over at least some portion of the applicator slot after determination of the current slot height (block 509). In this mode of operation, this constraint dictates that the calculated slot height profile should essentially be "frozen," regardless of any calculated adjustments that would normally be required based on Equations 8 and 12. While the slot height may be "frozen" over some portion of the applicator slot, this does not necessarily mean that the corresponding actuator settings (i.e., A i ) will be frozen. Variations in fluid dynamic effects may require repeated actuator adjustments to maintain a constant slot height.

[0187] In some embodiments, workflow 500 first determines an average actuator setting and ensures that this average actuator setting is held constant for each subsequent iteration of workflow 500. This can be achieved by offsetting the actuator setting profile consistently upward or downward, as necessary, to prevent the average actuator setting from changing over time. In practice, this operation can be achieved using the following subroutine: 1) Calculate the NON-FROZEN AVERAGE NEW SETPOINT ACTUATOR SETTINGS {which represents the actuator settings without the slot height constraint (block 513)}. 2) Define (AVERAGE_DIFFERENCE) = (NON-FROZEN AVERAGE NEW SETPOINT ACTUATOR SETTINGS) - (NON-FROZEN AVERAGE ACTUAL ACTUATOR SETTINGS). 3) Run the following algorithm: FOR i=1 to (number of actuators) IF actuator [i] is not frozen Define ACTUATOR_SP[i] = ACTUATOR_SP_NEW[i] - AVERAGE_DIFFERENCE END_IF END_FOR {thereby restoring the NON-FROZEN AVERAGE NEW SETPOINT ACTUATOR SETTINGS}.

[0188] The modified workflow 500 can be highly advantageous in avoiding undesirable drift, also known as creep, in the average slot height over the course of repeated iterations. Such drift can occur as a result of circumstances not accounted for by the mathematical construction of typical workflows. For example, situations can arise where actuators are unable to reach their desired setting / position due to physical limitations. Such limitations can be based on predicted force values ​​that exceed the allowable amount of force that can be safely applied without damaging the die. Actuators can also reach bending limits due to the continuity of flexible die lips or choker bars, which prevent adjacent actuators from having set points that are too far apart from each other. Yet another limitation can arise from discontinuities in the slot height profile created when one or more deckles are inserted into the applicator slot, particularly when the blocked and unblocked zones of the applicator slot come together. Advantageously, workflow 500 can help stabilize the slot height profile, along with the corresponding extrudate thickness profile, in each of these situations.

[0189] The techniques described in this disclosure, such as those described with respect to controller 300, may be implemented at least in part in hardware, software, firmware, or any combination thereof. For example, various embodiments of the techniques may be implemented within one or more microprocessors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), or any other equivalent integrated or discrete logic circuitry, as well as any combination of such components, embodied in a controller, user interface, or other device. The term "controller" may generally refer to any of the foregoing logic circuitry alone or in combination with other logic circuitry, or any other equivalent circuitry.

[0190] When implemented in software, the functions attributed to the systems and controllers described in this disclosure may be embodied as instructions on a computer-readable storage medium, such as random access memory (RAM), read-only memory (ROM), non-volatile random access memory (NVRAM), electrically erasable programmable read-only memory (EEPROM), flash memory, magnetic media, optical media, etc. The instructions may be executed by one or more processors to support one or more embodiments of the functions described in this disclosure.

[0191] Various embodiments have been described in the preceding sections. These and other embodiments are within the scope of the following claims.

[0192] All documents, patent documents, or patent applications cited in the above patent application are incorporated herein by reference in their entirety for consistency. In the event of any inconsistency or contradiction between any of the incorporated references and this application, the information in the foregoing description shall prevail. The foregoing description is intended to enable a person skilled in the art to practice the disclosure set forth in the claims, and should not be construed as limiting the scope of the present disclosure, which is defined by the claims and all equivalents thereof. In addition to the above-described embodiments, the following aspects will be noted. (Appendix 1) 1. A method of adjusting a slot die comprising: an applicator slot extending along a width of the slot die, the applicator slot being in fluid communication with a fluid flow path through the slot die; at least one of the group consisting of a choker bar and a flexible die lip; and a plurality of actuators spaced along the width of the slot die, each actuator operatively coupled to the choker bar or the flexible die lip to adjust a height profile of the slot or choker bar at a respective location of each actuator to provide localized adjustment of fluid flow through the applicator slot, the method comprising: measuring a thickness value of the extrudate provided from said applicator slot; mapping the measured thickness values ​​to corresponding positions along the width of the slot die; performing data compression of the measured thickness values ​​to obtain a residual vector representing deviations of the measured thickness values ​​from a desired thickness value; adjusting a position of one or more of the plurality of actuators based on the residual vector to obtain a corrected shape for the choker bar or the flexible die lip, wherein the adjusting includes: using the residual vector to determine a predicted degree of ringing for a proposed set of actuator positions; then applying a ringing constraint to obtain a set of modified actuator positions that reduces the extent of the ringing relative to the predicted extent of the ringing; A method comprising: (Appendix 2) 2. The method of claim 1, wherein the degree of ringing is predicted according to a calculated curvature of a height profile of the slot or choker bar. (Appendix 3) 3. The method of claim 1 or 2, wherein the degree of ringing is reduced by tolerating some imprecision in the set of corrected actuator positions in exchange for reducing the degree of ringing below a threshold. (Appendix 4) 4. The method of claim 3, wherein the threshold is correlated with a curvature in an actuator position profile. (Appendix 5) 5. The method of claim 4, wherein the curvature corresponds to a ring coefficient of 0 to 0.01. (Appendix 6) 6. The method of claim 5, wherein the curvature corresponds to a ring coefficient of 0 to 0.006. (Appendix 7) 7. The method of any one of claims 3 to 6, wherein the condition that tolerates some inaccuracy in the corrected shape in exchange for reducing the degree of ringing below the threshold is applied via a Lagrange multiplier method. (Appendix 8) 8. The method of claim 7, wherein the Lagrangian multiplier method includes solving for a Lagrangian multiplier based on the threshold. (Appendix 9) 9. The method of claim 8, wherein solving for the Lagrange multipliers includes using a bound algorithm for extended systems. (Appendix 10) 10. The method of any one of claims 1 to 9, wherein one or more actuators are adjusted based on a set of individual actuator settings based on a stiffness matrix that converts an actuator position profile into a slot or choker bar height profile, the slot or choker bar height profile excluding the effects of deflection due to fluid flow through the applicator slot. (Appendix 11) 11. The method of claim 10, further comprising fitting a model representation of the stiffness matrix using bump test data, wherein the set of individual settings is determined from both the stiffness matrix and an inverse stiffness matrix that converts an actuator position profile into a slot or choker bar height profile, the slot or choker bar height profile excluding the effects of deflection due to fluid flow through the applicator slot. (Appendix 12) 12. The method of claim 11, further comprising fitting an inverse stiffness matrix using the same bump test data used to fit the model representation of the stiffness matrix. (Appendix 13) 13. The method of claim 11 or 12, wherein parameters of the stiffness matrix model representation and / or the inverse of the stiffness matrix model representation are estimated using a nested sampling algorithm, a Monte Carlo Markov Chain algorithm, or both. (Appendix 14) When the stiffness matrix is ​​normalized, K ranges from 0.35 to 0.8. i,i value and K in the range of 0.1 to 0.3 i,i±1 value and K in the range 0 to 0.1 i,i±2 value and K in the range of 0.0 to 0.03, or optionally in the range of 0.0 to 0.05. i,i±3 value and K in the range of 0.0 to 0.02 i,i±4 14. The method of any one of appendices 10 to 13, comprising: (Appendix 15) 1. A method of adjusting a slot die comprising: an applicator slot extending along a width of the slot die, the applicator slot being in fluid communication with a fluid flow path through the slot die; at least one of the group consisting of a choker bar and a flexible die lip; and a plurality of actuators spaced along the width of the slot die, each actuator operatively coupled to the choker bar or the flexible die lip to adjust a height profile of the slot or choker bar at a respective location of each actuator to provide localized adjustment of fluid flow through the applicator slot, the method comprising: measuring a thickness value of the extrudate provided from said applicator slot; mapping the measured thickness values ​​to corresponding positions along the width of the slot die; performing data compression of the measured thickness values ​​to obtain a residual vector representing deviations of the measured thickness values ​​from a desired thickness value; adjusting positions of one or more of the plurality of actuators based on the residual vector to obtain a corrected shape of the choker bar or the flexible die lip, wherein the adjusting includes applying a slot or choker bar height constraint whereby one or more adjusted actuator positions are predicted to maintain an average current slot or choker bar height along a corresponding portion of the choker bar or the flexible die lip. (Appendix 16) 16. The method of claim 15, wherein a height constraint of the slot or choker bar is applied, whereby one or more adjusted actuator positions are predicted to hold one or more current slot or choker bar heights constant along the choker bar or the corresponding portion of the flexible die lip. (Appendix 17) 17. A method of adjusting a slot die, comprising repeatedly performing the method of claim 15 or 16 to reduce drift over time of average slot or choker bar height along the choker bar or the respective portion of the flexible die lip. (Appendix 18) 1. A method of adjusting a slot die comprising: an applicator slot extending along a width of the slot die, the applicator slot being in fluid communication with a fluid flow path through the slot die; at least one of the group consisting of a choker bar and a flexible die lip; and a plurality of actuators spaced along the width of the slot die, each actuator operatively coupled to the choker bar or the flexible die lip to adjust a height profile of the slot or choker bar at a respective location of each actuator to provide localized adjustment of fluid flow through the applicator slot, the method comprising: measuring a thickness value of the extrudate provided from said applicator slot; mapping the measured thickness values ​​to corresponding positions along the width of the slot die; performing data compression of the measured thickness values ​​to obtain a residual vector representing deviations of the measured thickness values ​​from a desired thickness value; adjusting a position of one or more of the plurality of actuators based on the residual vector to obtain a corrected shape of the choker bar or the flexible die lip; mapping the thickness values ​​to corresponding positions along the width of the slot die includes expanding and / or adjusting measured thickness values ​​to account for positional variations at one or more transverse edges of the extrudate. method. (Appendix 19) 1. A method for producing an extruded article, comprising: Preparing a slot die according to the method of any one of appendices 1 to 18; extruding an extrudate based on the adjusted slot die through the applicator slot of the slot die to obtain the extruded article; A manufacturing method comprising: (Appendix 20) A slot die, an applicator slot extending along the width of the slot die, the applicator slot being in fluid communication with a fluid flow path through the slot die; at least one of the group consisting of a choker bar and a flexible die lip; a plurality of actuators spaced along the width of the slot die, each actuator operatively coupled to the choker bar or the flexible die lip to adjust the height profile of the slot or choker bar at each actuator's respective location to provide localized adjustment of fluid flow through the applicator slot; a slot die including a controller configured to receive thickness values ​​of the extrudate provided from the applicator slot, map the received thickness values ​​to corresponding positions along the width of the slot die, perform data compression of the received thickness values ​​to obtain a residual vector representing deviation of the received thickness values ​​from a desired thickness value, and adjust positions of one or more of the plurality of actuators based on the residual vector to obtain a corrected shape of the choker bar or the flexible die lip, the adjustment including using the residual vector to determine a predicted degree of ringing for a set of proposed actuator positions, and then applying a ringing constraint to obtain a set of modified actuator positions that reduces the degree of ringing relative to the predicted degree of ringing; A system comprising:

Claims

1. 1. A method of adjusting a slot die comprising: an applicator slot extending along a width of the slot die, the applicator slot being in fluid communication with a fluid flow path through the slot die; at least one of the group consisting of a choker bar and a flexible die lip; and a plurality of actuators spaced along the width of the slot die, each actuator operatively coupled to the choker bar or the flexible die lip to adjust a height profile of the slot or choker bar at a respective location of the actuator to provide localized adjustment of fluid flow through the applicator slot, the method comprising: measuring a thickness value of the extrudate provided from said applicator slot; mapping the measured thickness values ​​to corresponding positions along the width of the slot die; performing data compression of the measured thickness values ​​to obtain a residual vector representing deviations of the measured thickness values ​​from a desired thickness value; adjusting positions of one or more of the plurality of actuators based on the residual vector to obtain a corrected shape of the choker bar or the flexible die lip, wherein the adjusting includes applying a slot or choker bar height constraint whereby one or more adjusted actuator positions are predicted such that an average slot height or choker bar height is maintained before and after the adjustment along a corresponding portion of the choker bar or the flexible die lip.

2. 2. The method of claim 1, wherein a height constraint of the slot or choker bar is applied whereby one or more adjusted actuator positions are predicted to hold a slot height or choker bar height constant before and after adjustment along the corresponding portion of the choker bar or flexible die lip.

3. 3. A method of adjusting a slot die, comprising repeatedly performing the method of claim 1 or 2 to reduce drift over time of the slot height or choker bar height along the respective portion of the choker bar or flexible die lip.

4. 1. A method of adjusting a slot die comprising: an applicator slot extending along a width of the slot die, the applicator slot being in fluid communication with a fluid flow path through the slot die; at least one of the group consisting of a choker bar and a flexible die lip; and a plurality of actuators spaced along the width of the slot die, each actuator operatively coupled to the choker bar or the flexible die lip to adjust a height profile of the slot or choker bar at a respective location of the actuator to provide localized adjustment of fluid flow through the applicator slot, the method comprising: measuring a thickness value of the extrudate provided from said applicator slot; mapping the measured thickness values ​​to corresponding positions along the width of the slot die; performing data compression of the measured thickness values ​​to obtain a residual vector representing deviations of the measured thickness values ​​from a desired thickness value; adjusting a position of one or more of the plurality of actuators based on the residual vector to obtain a corrected shape of the choker bar or the flexible die lip; mapping the thickness values ​​to corresponding positions along the width of the slot die includes expanding and / or adjusting measured thickness values ​​to account for positional variations at one or more transverse edges of the extrudate. method.

5. 1. A method for producing an extruded article, comprising: Adjusting the slot die according to the method of any one of claims 1 to 4; extruding an extrudate based on the adjusted slot die through the applicator slot of the slot die to obtain the extruded article; A manufacturing method comprising:

Images