Parameterization of traction controllers

Abstract

The traction force controller (9) controls the rotational speed (v9) of the controlled roller (2) in the web processing machine (1) in order to transport the material (3) from the controlled roller (2) to the further roller (6) or from the further roller (6) to the controlled roller (2) at a line speed (v) and with an applied traction force (F) on the web processing machine (1). For parameterizing the traction force controller (9) of the controlled roller (2). During a stationary test (TO) at a line speed (v) of zero, the traction force (F) is increased to 90% of an identified traction force (F w2 ), preferably a pre-given stationary traction force operating point (Fop) in order to determine stationary control section parameters of a traction force control section (G F,0 ), and from the stationary control section parameters of the traction force control section (G F,0 ), preferably by means of a frequency characteristic curve method, stationary controller parameters (R F,o ) of the traction force controller (9) are determined. The traction force controller (9) is parameterized with the stationary controller parameters (R F,o ).

Description

Technical Field

[0001] This invention relates to a method for parameterizing a traction force controller for a controlled roller of a web processing machine, the traction force controller controlling the traction force via the rotational speed of the controlled roller to transport material at a linear velocity from one controlled roller to another or from another roller to another on the web processing machine when traction force is applied; and to the use of a traction force controller parameterized according to the invention for controlling the rotational speed of a controlled roller in a web processing machine to transport material at a linear velocity from one controlled roller to another or from another roller to another on the web processing machine when traction force is applied. Furthermore, the invention relates to a parameterization unit for parameterizing a traction force controller for a controlled roller of a web processing machine, transporting material at a linear velocity from one controlled roller to another or from another roller to another on the web processing machine when traction force is applied, wherein the traction force controller is designed to control the traction force via the rotational speed of the controlled roller. Background Technology

[0002] In a web processing machine, materials in the form of webs, foils, tubes, wires, or strips are transported along a conveyor path at a linear speed and processed into finished or intermediate products during the process. Materials such as metals, plastics, carbon fibers, textiles, paper, and composite materials can be used. Here, the material is wound as a winding item on a winder (roller, roller, cylinder, etc.), unwound from the winder, and subsequently processed. Uncontrolled stretching or compression of the material negatively impacts both the material's inherent properties and the processing quality within the web processing machine. It can be specified that the finished or intermediate product is rewound onto the winding device at the end of the processing step or fed into another (processing) step.

[0003] During the processing, the material is subjected to traction. This traction is generated by the non-slip transport of the material between rollers (e.g., the winder and the traction roller). The traction roller is equipped with a clamping roller to transport the material without slippage along the transport track between the traction roller and the clamping roller. Therefore, the material is transported between the winder and the traction roller at a linear velocity and is tensioned by traction during this time. In principle, both rollers (i.e., the winder and the traction roller) are driven. Therefore, a set rotational speed is pre-given to both axes, preferably generated at the center. It is stipulated that only one roller (preferably the winder) is controlled by adding a corrective rotational speed to the set rotational speed, because controlling both rollers can lead to instability in the traction force.

[0004] Even under varying profile conditions (e.g., geometric irregularities), uniform removal (i.e., uniform unfolding) of material from the winder is a crucial prerequisite for the quality of the product or intermediate product. For example, uneven winder movement can so strongly affect traction that it is nearly impossible for a traction controller to correct this effect. Fluctuating traction can also result from irregularly wound material or lateral offsets during winding. The traction controller contributes particularly when the linear velocity of the material accelerates to the operating speed. Therefore, traction controllers are especially useful when it is necessary to ensure precise traction stress in the material at all stages of the production process. Unlike traction control, traction adjustment incorporates a measuring unit that measures the actual traction force in the material, which is then fed back to the traction controller.

[0005] Traction control involves a controlled slave roll (typically a winder) and a master roll, with a closed control loop. The slave roll, like the master roll, has a rotational speed, and the traction controller applies a correction speed to the slave roll's speed, thereby adjusting the traction force. The correction speed is determined by the traction controller based on the actual traction force and a pre-defined set traction force.

[0006] Preferably, different influencing variables (e.g., the diameter of the material and / or the current linear velocity) are considered during traction control to ensure optimal material processing. When suitable means are provided, the diameter of the material can be precisely measured, for example. If the diameter of the material is not measured due to its unavailability or impossibility, an estimate of the diameter can be utilized.

[0007] The determination of the controller parameters for the traction controller is typically only performed when the web processing machine is in production operation. This process is very time-consuming, and the user must have extensive process knowledge to obtain useful results.

[0008] Controller parameters can also be determined automatically. DE 11 2014 005 964 T5 illustrates a method in which the traction force is slowly increased to the traction force operating point and parameterization of the traction force controller begins when the traction force operating point is reached. Summary of the Invention

[0009] The objective of this invention is to ensure the simplicity and automatic parameterization of the traction controller for a web processing machine.

[0010] According to the invention, this task is solved by increasing the traction force to 90% of the identified traction force, preferably a pre-defined static traction force operating point, during a stationary test at zero online speed, to determine the stationary control segment parameters of the traction force control segment, and determining the stationary controller parameters of the traction force controller from the stationary traction force control segment, preferably using a frequency response curve method, wherein the traction force controller is parameterized using the stationary controller parameters. Furthermore, according to the invention, this task is solved by a parameterization unit designed to increase the traction force to 90% of the identified traction force, preferably a pre-defined static traction force operating point, during a stationary test at zero online speed, to determine the stationary control segment parameters of the traction force control segment, and determining the stationary controller parameters of the traction force controller from the stationary control segment parameters, preferably using a frequency response curve method, and parameterizing the traction force controller using the stationary controller parameters. The identified traction force is selected so that no components of the material or web processing machine are damaged, and therefore also depends on the selection of the static traction force operating point. In principle, it is also possible to have a marked traction force exceeding 100% of the pre-defined static traction force operating point. The controller parameters can be automatically determined using the parameterization unit according to the invention. The described method, along with static tests, creep tests, speed tests, etc., can be automatically executed by the parameterization unit. The (static) traction force control segment is preferably modeled as an integrator. The integrator amplification largely depends on the material parameters. Materials with high stiffness have high amplification. Therefore, a length change in a material with higher stiffness results in a higher traction force in the material compared to the same length change in a material with relatively low stiffness.

[0011] Therefore, as disclosed, for example, in publication DE 11 2014 005 964 T5, instead of applying oscillations to the traction force, the traction controller is driven at rest (zero linear velocity) and with a traction force of magnitude equal to the identified traction force. The elastic modulus of the material can be directly determined by the control parameters of the traction control segment. Since the controller parameters, preferably the PI controller parameters of the PI traction controller, are determined rapidly and automatically, the user requires little to no knowledge of control technology and only minimal process knowledge. Thus, the control parameters of the traction control segment and the controller parameters can be determined automatically. The controller parameters are optimally matched to a given requirement (e.g., rise time or overshoot), which would likely be very costly to determine manually.

[0012] The stationary controller parameters of the traction controller can be determined from the stationary control segment parameters of the stationary traction control segment using the frequency response curve method. The frequency response curve method is used in the frequency domain. The requirements for the start-up characteristics of the closed-loop response to certain selected test functions are considered and transferred to the requirements of the Bode plot of the open-loop. The start-up characteristics of the closed-loop are evaluated based on the parameters rise time (a measure of velocity), overshoot (a measure of damping), and maintained control deviation (a measure of steady-state accuracy). These parameters of the time characteristics of the step response of the closed-loop are related to the frequency characteristics of the open loop. The rise time is related to the crossover frequency by an approximation. The crossover frequency separates those frequencies amplified by the open-loop from those attenuated by the open-loop, thus the crossover frequency is a measure of the bandwidth of the open-loop, where the dynamics of the closed-loop become faster with increasing crossover frequency. The percentage overshoot is related to the phase margin by an approximation. The phase margin is a measure of the distance to the stability limit, thus a decrease in the phase margin increases the oscillation tilt (i.e., overshoot). Conversely, the maintained control deviation is directly related to the amplification factor of the open-loop transfer function. The frequency response curve method is largely known and will not be elaborated upon here. See, for example, Chapter 5 of the lecture notes from the Automation Lectures and Exercises given by Univ.-Prof. Dr. Techn. Andreas Kugi at TU Wien during the Winter Semester 2019 / 2020.

[0013] Preferably, the traction force is increased to 10% of the tensile traction force, preferably the static traction force, before the marked traction force is increased. This ensures that the material is under mechanical tension at the start of the static test.

[0014] The traction controller can be parameterized using stationary controller parameters, and the traction force can be increased to the stationary operating traction force. After reaching the traction operating point, a traction step is applied to the traction force to determine the quality of the stationary controller parameters based on the first stationary quality step response, preferably using a recursive best-fit method (i.e., in the case of using the residual sum of squares). The stationary controller parameters are calculated based on the stationary control segment parameters. Furthermore, the step response of the closed control loop is measured, compared with the expected step response of the identified closed control segment, and the residual sum of squares is calculated.

[0015] The sum of squared residuals is a quality criterion for the accuracy of the identified model. If the results meet expectations, the static test is complete. Otherwise, the test can be repeated based on new requirements for the parameterization of the traction controller.

[0016] Preferably, a creep test is performed after the static test, wherein the traction controller is parameterized using the static controller parameters, and wherein a first operating linear velocity and a traction force of magnitude equal to a first traction force operating point are provided. A traction force step is applied to the traction force, the creep step response is determined, and preferably, the fine control segment parameters of the traction control segment are identified from the creep step response using a (recursive) least squares method. The fine controller parameters are determined from the creep step response and the fine control segment parameters, preferably using a frequency response curve method, and the traction controller is parameterized using the fine controller parameters. Therefore, the creep test is used to optimize the controller parameters determined in the static test. The result of this optimization is the fine controller parameters.

[0017] A step in traction force also causes a change in the circumferential velocity of the shaft. Therefore, the response to a step in traction force and the resulting change in the circumferential velocity of the winder is called the creep step response.

[0018] (Recursive) least squares (i.e., the method of minimizing the mean square of the estimation error in the recursive variant (RLS)) uses a parameterized model in the form of a classification model structure, preferably an ARX (autoregressive with exogenous input) model. This algorithm is based on least squares and is used to estimate model parameters in the identification of a linear system. Here, the optimization problem is chosen such that the square of the difference between the measured data and the model data is minimized. Therefore, the solution that minimizes the mean square of the error is sought. The desired parameter vector (optimal solution) can be computed by setting the derivative of the optimization problem to zero. The recursive variant allows for minimal computational overhead when adding new data because previous results are used as a starting point, and the estimated parameter vector improves with each new measurement. The RLS algorithm requires at most as many recursive steps as the parameters to be identified to obtain good results. The starting values ​​must be chosen meaningfully. Only recursion enables online use of system identification. (Recursive) least squares is generally known, so it will not be described in more detail here. For example, see Chapter 1.3.4 of the lecture notes for the course "Regelungssysteme (Control Systems)" written by Assoc.-Prof.Dr.-Ing.Wolfgang Kemmetmüller and Univ.-Prof.Dr.techn.Andreas Kugi in the winter semester of TU Wien 2018 / 2019.

[0019] In principle, if coarse controller parameters have been predetermined and the traction controller is parameterized using these parameters, a creep test can be performed without prior static testing. With a first operating linear velocity and a traction force equal to the magnitude of the first traction force operating point, a traction force step can be applied to the traction force, and the creep step response can be determined. This creep step response can then be used to identify the fine control segment parameters of the traction control segment, and the fine controller parameters can be determined based on the creep step response and the fine control segment parameters.

[0020] Coarse controller parameters can be determined manually, for example, by manually analyzing the step response of the closed control loop. Therefore, a traction controller (e.g., in the form of a PI controller, PID controller, state controller, etc.) can be designed with zero linear velocity. A suitable value for proportional amplification is found by first setting the integral time to zero for all tests. Preferably, a very conservative initial value is chosen for the amplification until the correct range of values ​​has been found. The amplification is increased until slight oscillations are observed and the step response corresponds as expected. With zero linear velocity, usable results can be obtained using a pure P controller because the system has integral characteristics at zero linear velocity. However, for operational conditions (i.e., linear velocity greater than zero), the integral component of the traction controller is mandatory to maintain no residual control deviation. Therefore, once a suitable proportional amplification has been found, the integral component of the traction controller must be changed. The controller's integral time is selected to satisfy the specified requirements for rise time and overshoot. A shorter integral time results in a faster attainment of the setpoint, but overshoot is more likely. A longer integral time has the opposite effect. The adjustment of the integral component is determined by the user.

[0021] The elastic modulus and / or length of the material for the first working linear velocity can be determined based on the fine control section parameters. Preferably, the fine control section parameters are determined by the product of the material's elastic modulus and its cross-sectional area, wherein the elastic modulus can be directly calculated when the cross-sectional area is known.

[0022] A traction force step can be applied to the traction force so that the quality of the fine controller parameters for the first operating linear velocity can be determined, preferably using a best-fit method, based on the creep quality step response. High-quality controller parameters ensure the mechanical stability of the material, and thus the quality of production results, especially during acceleration, braking, and high linear speeds. This reduces defects and waste.

[0023] Preferably, a speed test is performed after the creep test, wherein the traction controller is parameterized using fine controller parameters, and a second operating linear velocity and a traction force of magnitude equal to the second traction force operating point are provided. A traction force step is applied to the traction force, the speed test step response is determined, and additional fine control parameters of the traction force control segment are thus identified. These additional fine controller parameters can be determined based on the speed test step response and the additional control segment parameters, thereby allowing the traction controller to be parameterized using fine controller parameters. The response to the traction force step and the associated step of the circumferential speed of the winder 2 is called the speed test step response.

[0024] Therefore, the speed test is used to optimize the fine controller parameters determined for the first operating speed in the creep test for the second operating speed. The result of this optimization is additional fine controller parameters. Thus, in addition to the control parameters for the first operating linear speed, there are also control parameters for the second operating linear speed. Preferably, the second operating linear speed is selected to its maximum to obtain control parameters for the maximum operating linear speed.

[0025] Furthermore, for example, by interpolation, additional fine controller parameters for an additional operating line speed can be determined based on the fine controller parameters and other fine controller parameters. This determination of additional controller parameters for an additional operating line speed is particularly efficient if the first operating line speed has been selected as low and the second operating line speed has been selected as maximum.

[0026] Additional fine controller parameters for the additional operating line speed can also be determined based solely on the fine control segment parameters and the fine controller parameters, for example, by means of coefficient comparison.

[0027] The stationary controller parameters can be stored, from which extrapolation speed controller parameters for several extrapolation linear velocities can be derived. During extrapolation speed operation within one of the extrapolation linear velocities of the web processing machine, the relevant extrapolation speed controller parameters can be recalled to parameterize the traction controller. For this purpose, the stationary control segment parameters identified in stationary mode and the desired linear velocity can be used in the general transfer function G(s) of the traction control segment, and controller design for the system can be performed.

[0028] Preferably, additional fine controller parameters for the additional working line speed are stored, and during the operation of the web processing machine at a line speed within the corresponding additional working line speed range with the relevant fine controller parameters, the relevant additional fine controller parameters are invoked to parameterize the traction controller, and the traction controller is parameterized using the relevant additional fine controller parameters.

[0029] Similarly, fine controller parameters for the first working line speed can also be stored, and during the operation of the web processing machine at a line speed within the first working line speed range, the fine controller parameters are invoked to parameterize the traction controller, and the traction controller is parameterized using the fine controller parameters.

[0030] Similarly, additional fine controller parameters for the second working line speed can also be stored, and during web processing machine operation at a line speed within the range of the second working line speed, these additional fine controller parameters are invoked to parameterize the traction controller, and the traction controller is parameterized using these additional fine controller parameters.

[0031] This allows for the creation of various additional fine controller parameter sets for different working line speeds, which can be accessed as needed during the operation of the web processing machine to parameterize the traction controller.

[0032] The parameterized traction controller according to the method of the present invention can be used to control the traction force of material in a web processing machine, wherein the material is transported at a linear speed from one controlled roll to another or from another roll to a controlled roll when traction force is applied. Attached Figure Description

[0033] The following will refer to Figures 1 to 5 To explain the invention in more detail, Figures 1 to 5 The advantageous design features of the invention are illustrated, illustrative, and non-limiting. The accompanying drawings show:

[0034] Figure 1 The overall web processing machine is shown.

[0035] Figure 2 The various areas of the web processing machine are shown.

[0036] Figure 3 The traction adjustment when stationary is shown.

[0037] Figure 4 A static test was shown.

[0038] Figure 5 The creep test or speed test is shown. Detailed Implementation

[0039] exist Figure 1The diagram illustrates a web processing machine 1 for a continuous process. A winder 2 is configured as a controlled roller, designed to wind material 3 onto or unwind material 3 from the winding core 20, depending on whether the winder is at the beginning or end of the web processing machine 1. Therefore, it is always assumed that material 3 is unwinded from the winding core 20, but it is also always possible to wind material 3 onto the winding core 20 in a similar manner. The wound material 3 is pre-tensioned on the winder 2 and therefore has a fundamental elongation ∈ 0.

[0040] Furthermore, a traction roller 6 with a pressure roller 60 is provided to transport material between the traction roller 6 and the pressure roller 60 without slippage. The pressure roller 60 is not actively driven and presses against the traction roller 6. Changes in the rotational speed of the winder 2, coupled via the material 3, also affect the traction roller 6. The traction roller 6 itself is driven at a traction roller speed v6, does not have a superimposed traction force controller, and therefore represents the main roller. The linear velocity v of the material 3 is adjusted by the traction roller speed v6. Therefore, the linear velocity v is controlled by the circumferential speed of the traction roller 6, where linear velocities exceeding 1000 m / min are possible. The linear velocity v preferably has a trapezoidal profile, i.e., a linear increase from zero to the working linear velocities v1 and v2 at the start of the production process. The linear velocity v remains constant for the desired working linear velocities v1 and v2 during the production process and linearly decreases back to zero at the end of the production process.

[0041] The winder 2 has a winder speed v9′, which consists of a set winder speed v9 and a correction speed Δv9. The circumferential speed of the winder 2 remains constant, thus the set winder speed v9 varies depending on the diameter of the winder 2. The correction speed Δv9 is preset by the traction controller 9 to control the winder speed v9′. Therefore, the winder 2 is an adjusting element for controlling the traction force F in the material 3. When using the measuring unit 5 (e.g., a force sensor), the actual traction force F... ist It is measured as a process variable and fed back to the traction controller 9. The traction controller 9 measures the actual traction force F. ist and setting the traction force F soll To determine the correction speed Δv9. The traction roller speed v6 and the winder set speed v9 are given in advance by the traction controller 9 as examples only, and can also be given in advance by other components.

[0042] Because the winder 2 and the traction roller 6 are connected in a force-locked and non-slip manner in the contact area with the material 3, the linear rotational speed v can be approximately equal to the circumferential speed of the traction roller 6 and the winder 2. However, depending on the traction force F, the deviation between the circumferential speed of the winder 2 and the linear rotational speed v is minimal. Since the material 3 unwinds from the winder 2, it is advantageous to consider the variation in the winder diameter when determining the relationship between the circumferential speed of the winder 2 and the linear rotational speed v. For this purpose, the winder diameter can be measured or estimated.

[0043] Conversely, if the web processing machine 1 has a runout control, then the actual traction force F is used instead. ist The position of the runout component is set as a process variable for feedback. If the web processing machine 1 has traction control instead of traction controller 9, then there is no feedback for the process variable at all.

[0044] Figure 1 An optional deflection roller 4 is also provided for guiding the material 3 but is not driven itself. The mass moment of inertia of the deflection roller 4 is small and can usually be ignored. However, during acceleration and braking, it may be necessary to take into account the mass moment of inertia of the deflection roller 4 and produce a smooth linear velocity curve to minimize negative inertial effects.

[0045] Web processing machine 1 typically consists of multiple sections (also called zones). The term "zone" refers to the area in web processing machine 1 between two drive rollers, with material 3 held between them without slippage. The state of material 3 within a zone is influenced by the two drive rollers, which define the corresponding zone. Within a zone, one roller serves as the master roller, and the other as the slave roller. Typically, web processing machine 1 has at least three zones: an inlet zone A, a processing zone B, and an outlet zone C, as shown below. Figure 2 As shown. In the inlet area A, the traction force F is controlled by the traction force controller 9, which pre-sets the winding speed v9′ to the winding machine 2. ist Material 3 is unwound from winder 2. Preferably, a web motion control to correct lateral deviation of material 3 is provided in the entry area A. Additionally, a material buffer may be present to store material. This refers to a structure with deflecting rollers that increase the distance between each other and thus can accommodate more material 3. Performing roller changes without stopping the machine is particularly meaningful in the winding and unwinding areas. During roller changes, the material is removed from the buffer; the web processing machine does not need to be stopped during this period. Processing procedures (e.g., pressing, packaging, coating, stamping, etc.) occur in processing area B, therefore the highest requirements are placed on the accuracy of the traction force F in processing area B. Figure 2As shown, in exit region C, material 3 is removed and / or wound onto winding device 7. Similar to in entry region A, web movement control and / or material buffers may be provided in exit region C. After removal, material 3 can be transferred to further (e.g., discontinuous) processes.

[0046] Because all rollers that are in force-locked contact with material 3 (i.e., Figure 2 The winder 2, traction roller 6, another traction roller 6′ and winding device 7 are coupled by material 3, so the material properties of material 3 can have a significant impact on this coupling and therefore on the design of traction controller 9.

[0047] Therefore, the linear velocity v in a region (entry region A, processing region B, exit region C) is determined by the main roller (e.g., by the traction roller 6 in entry region A). Entry region A is then considered. However, for the traction controller 9 in processing region B or exit region C, the determination of the controller parameters can also be done in a essentially similar manner—assuming there are a main roller and a slave roller.

[0048] When stationary (i.e., when the online speed v is zero), the elongation of material 3 can be determined by the position of the winder 2. Material 3, located between the winder 2 and the traction roller 6, has a basic length L0. Therefore, the traction force F in the inlet region A corresponds to the basic traction force F0 used to wind material 3 onto the winder 2. Figure 3 As shown, if the position of the winder 2 changes, the adjustment angle will be adjusted. The change in the length of material 3 is represented by a length difference ΔL, which results in a traction force difference ΔF. Therefore, for the traction force F, we obtain the sum of the basic traction force F0 and the traction force difference ΔF: F = F0 + ΔF.

[0049] As mentioned, in order to change the traction force F during operation (i.e., when the linear velocity v is greater than zero), a corresponding correction speed Δv9 is applied to the set speed v9 of the winder, thereby obtaining the winder speed v9′ of the winder 2. With a constant correction speed Δv9, a constant change ΔF in the traction stress occurs after a certain period of time, the magnitude of which largely depends on the occurring linear velocity v. Therefore, the traction roller 6, as the main roller, is given a predetermined linear velocity v, and the winder speed v9′ and thus the angular speed ω of the winder 2 are thus changed via the correction speed Δv9 to achieve the desired traction force F. ist This is generated in material 3. Therefore, the winder 2 operates to a certain extent relative to the traction roller 6, and thus generates a traction force F in material 3.

[0050] The angular rotational speed ω of the winder 2 also changes with ω = v / r depending on the changing radius r of the winder 2, while maintaining a constant linear velocity v. Therefore, in order to ensure that the circumferential velocity of the winder 2 corresponds to the linear velocity v of the system, the angular rotational speed ω or the correction speed Δv9 must always adapt to the current radius r.

[0051] The traction controller 9 may correspond to a PI controller, for example, while other types of controllers (e.g., PID controllers, state controllers, etc.) are also possible.

[0052] It can be used for various working linear speeds v1, v2, v x To determine the controller parameters R of traction controller 9 F，v1 R F，v2 R F，vx .

[0053] Material 3 can take different forms (webs, threads, etc.) and can be composed of, for example, paper, fabric, plastic, metal, etc. Material 3 can be considered a three-dimensional volume. Material 3 has a length L, which is initially at least roughly known and can subsequently be determined precisely. Furthermore, material 3 has an elastic modulus E that is generally unknown. In addition, the material has a cross-section A, which is preferably known as precisely as possible so that the elastic modulus E can be calculated based on control segment parameters (representing the product of the cross-section and the elastic modulus E) (e.g., based on fine control segment parameters – see below).

[0054] If material 3 is subjected to a traction force F in the longitudinal direction, the traction stress σ = F / A will be generated depending on the cross-sectional area A of the material. Assuming that the cross-sectional area A does not change significantly due to the externally applied traction force F, the traction stress σ is directly proportional to the traction force F. Furthermore, the externally applied traction force F also generates the elongation ε of material 3. For the design of the traction controller 9, only a region with a linear elastic relationship between the traction stress σ and the elongation ε is considered. This means that in this region, the elongation ε increases linearly with the traction stress σ, where the slope is described by the elastic modulus E. If the traction stress σ decreases again, material 3 will again exhibit its original length L. The traction stress in material 3 can be described by Hooke's Law σ = E*ε. Because it is assumed that the traction stress σ is directly proportional to the traction force F, it can also be assumed that the traction force F is directly proportional to the elongation ε. The elongation ε describes the relationship between the length change ΔL caused by the applied traction force F and the initial length L0. The relationship between the elongation ε and the traction force F yields F = E*A*ε.

[0055] Now, in order to determine the static controller parameters R F，o Perform a stationary test T0 with a linear velocity v of zero, where in Figure 4The diagram illustrates an exemplary path for the traction force F and the set rotational speed v9 of the winder. A stationary test T0 can be performed on the parameterization unit 90 to determine the stationary controller parameters R. F，o The parameterization unit 90 can be a component of the traction controller 9.

[0056] During the static test T0, the traction roller speed v6 is zero. During the static test T0, material 3 is stretched through a negative winder set speed v9 = v0. This means that the winder set speed v9 = v0 acts in the opposite direction to the rotation of winder 2 during production operation. Because the traction roller 6 does not move or moves only negligibly, a counter-torque is established; however, the linear velocity v remains zero during the static test T0. In the first part T of the static test T0... 01 In the middle, the winder 2 is driven at a negative winder set speed v9 = v0 until the traction force F reaches the stretching traction force F. w1 Preferably, the traction force operating point F op 10%. If the traction force F reaches the stretching traction force F w1 Then during the initialization phase T 02 In the middle, the winding speed v9 is again preset to zero in order to keep the traction force F constant at the tension traction force F. w1 superior.

[0057] Subsequently, in the identification phase T 03 During this period, the traction force F is increased to the indicated traction force F by pre-setting a negative winding speed v9 = v0 to the winder 2. w2 Preferably, the static traction force operating point F op 90%. During the identification phase T 03 In the middle, the static controller parameter R F，o According to traction control section G F，0 The parameters of the stationary control segment are determined, which is preferably done using the frequency response curve method.

[0058] Traction controller 9 can utilize the stationary controller parameter R F，o To parameterize, the traction force F increases to the stationary traction force operating point F. op A step ΔF can then be applied to the traction force F in order to determine the stationary controller parameters R based on the first stationary mass step response g0. F，o The quality (e.g., by using recursive best-fit methods).

[0059] If the static controller parameter R F，o With sufficient quality, a creep test T1 can be performed to determine the fine controller parameters for the first operating linear velocity v1, where... Figure 5The diagram shows the traction force F and the constant first linear velocity v1. This is used to determine the fine controller parameters R. F，v1 The creep test T1 can be performed on parameterized unit 90, or on a separately implemented fine parameterized unit.

[0060] During the creep test T1, material 3 moves at a first working linear velocity v1. The traction controller 9 utilizes the static controller parameters R determined during the static test T0. F，o To parameterize. During the creep test T1, a first traction force operating point F is pre-defined for the traction force F. op1 It can correspond to the static traction force operating point F of the static test T0. op The traction controller 9 controls the winding set speed v9 of the winder 2 to adjust the traction force F to the first traction force operating point F. op1 .

[0061] The advantage of using a closed control loop is that unknown disturbances can be compensated for by the traction controller 9. However, it must also be considered that the controlled system is fully excited by the corrected speed Δv9 provided by the traction controller 9. Therefore, the traction force F is applied at the traction force operating point F. op1 The step change in traction force corresponds to a step change ΔF to ensure sufficient excitation of the controlled system. As the traction force step ΔF increases, the marked stage T... 11 Let's begin again. Fine-tuning parameter G for the traction control segment at the first working linear velocity v1. F，v1 Identify the creep step response h1 using (recursive) least squares. Since the control loop is closed, the traction force F in the ideal case has a rise time t. r After reaching the traction working point F opl Because the rise time t is given in advance for the stationary controller. r Therefore, according to the static controller parameter R F，o The parameterized traction controller 9 does not necessarily have to meet the rise time t in the creep test. r .

[0062] Fine controller parameters R for the first working linear velocity v1 F，v1 Based on the first step response h1 obtained for the traction control segment at the first working linear velocity v1 and the determined fine control segment parameters G F，vl The elastic modulus E and / or length L of material 3 with respect to the first working linear velocity v1 can be determined based on the fine control segment parameter G. F，v1 To determine.

[0063] A second traction force step ΔF can be applied to the traction force F in order to determine the fine controller parameter R for the first operating linear velocity v1 based on the creep quality step response h2. F，v1 The quality can be assessed using a recursive best-fit method.

[0064] Additionally, similar to creep test T1, a speed test T2 can also be performed, which corresponds to creep test T1 at a higher linear velocity v2, preferably at the maximum linear velocity. Here, the traction controller 9 utilizes fine controller parameters R determined within the range of creep test T1. F，v1 To parameterize, in order to determine additional fine-grained controller parameters G for the second operating linear velocity v2. F，v2 The speed test T2 can be performed during the creep test T1 by setting a second working linear speed v2 and, for example, the working point F of the first traction force. op1 The corresponding second traction force working point F op2 The magnitude of the traction force F is used. A traction force step ΔF is applied to the traction force F, the speed test step response h2 is determined, and additional fine control parameters G of the traction control segment are identified from the speed test step response h2. F，v2 Other fine controller parameters R F，v2 Based on the speed test step response h2 and other fine control parameters G F，v2 This can be determined by applying a step ΔF to the traction force F. Furthermore, the quality of the fine controller parameters can be checked by applying a step ΔF to the traction force F.

[0065] Used to determine additional fine controller parameters R F，v2 The speed test T2 can also be performed on parameterized unit 90, or it can be performed on a separate fine parameterized unit.

[0066] If no controller parameters are determined between the first operating linear speed v1 and the second (preferably maximum) operating linear speed v2, then the additional operating linear speed v can also be determined offline. x Additional controller parameters R F，vx This means no additional test procedures are required. This can be done by performing coefficient comparisons within the range of the frequency response curve method, or by tracing the function along the fine controller parameters R. F，v1 With other fine controller parameters R F，v2 Interpolation is performed between them.

[0067] For the additional linear velocity v x The determined additional controller parameters R F，vx (e.g., the fine controller parameter R for the first working linear velocity v1) F，v1And / or additional fine-tuning controller parameters R for the second operating linear velocity v2 F，v2 These parameters can be stored as a set for the traction controller 9 and called up as needed during operation.

[0068] The parameterization unit 90 and / or the fine parameterization unit and / or additional fine parameterization units may include microprocessor-based hardware, such as a computer or digital signal processor (DSP) on which corresponding software for performing corresponding functions is implemented. The parameterization unit 90 and / or the fine parameterization unit and / or additional fine parameterization units may also include integrated circuits that also have microprocessors, such as application-specific integrated circuits (ASICs) or field-programmable gate arrays (FPGAs). The parameterization unit 90 and / or the fine parameterization unit and / or additional fine parameterization units may also include analog circuits or analog computers. Hybrid forms are also conceivable. Furthermore, different functions can be implemented on the same hardware.

[0069] The identifiers of the control segment parameters of the traction control segment G(s) are then shown as an example. Traction control segment G F，0 The stationary control segment parameters of (s) are determined and used to determine the stationary controller parameters R. F，0 (Static test T0). Additionally, traction control section G... F，v1 The fine control segment parameters of (s) are determined and used to determine the fine control parameters R of the traction controller 9. F，v1 (Creep Test T1). The identification of control segment parameters is first performed using the static test T0 (i.e., when the online velocity v is zero) and then using the creep test T1 (i.e., when the online velocity v is not equal to 0).

[0070] Material 3 has a fundamental elongation ∈ 0, wherein in the case of slightly wound material 3, the fundamental elongation ∈ 0 is also zero or at least negligible.

[0071] Assume the elastic modulus is The length is L = 4.5 m, and the cross-section is A = 2.8 × 10⁻⁶ m. -5 m and the basic elongation is ò0 = 0.1786. Traction control section G F，v0 The identification of the static control segment parameters of (s) is first executed uncontrolled in the open control loop and then in the closed control loop.

[0072] The general transfer function of the traction control segment G(s) is derived from... or To describe, where the coefficient is and

[0073] Therefore, when the online speed v is greater than 0, the traction control segment G can be estimated.F，v The two static control segment parameters of (s) can be used to determine the length L and the elastic modulus E.

[0074] For a stationary state (i.e., linear velocity v = 0), the traction control segment is applicable. in

[0075] Therefore, there is only one coefficient K when at rest. S Therefore, only one control segment parameter can be estimated here. The elastic modulus E can be determined based on the static control segment parameter only when the length L of material 3 is known.

[0076] Now, a static test T0 is performed with an open control loop, assuming that the material 3 in the web processing machine 1 has a linear velocity v of 0 m / min. Figure 4 As shown, a step is applied to the set winder speed v9 of winder 2. Furthermore, the step response is determined by observing how the traction force F behaves.

[0077] Subsequently, with coefficient K S Traction control section G in the form of F，0 The stationary control segment parameters of (s) are determined from the step response using a (recursive) least squares method. Therefore, in this example, the coefficient is obtained as K. S =144.66. Given a length L = 4.5m, the elastic modulus is:

[0078]

[0079] Because of traction control segment G F，0 All the necessary static control segment parameters of (s) are now known, so the static controller parameters R can be determined. F，v0 And therefore, a controller can be designed.

[0080] The controller ultimately has the form And according to the regulations ω c t r ≈1,5 and To design, where ω c Describes the open-loop crossover frequency, t r Describe the rise time of the closed-loop step response. Describes the phase margin and ü describes the overshoot of the closed-loop step response.

[0081] Now we use the frequency response curve method. For this, the desired rise time t is given in advance. r For example, 0.17s. Therefore, this yields the crossover frequency. When the overshoot ü is ü = 10%, the phase margin is The transfer function at the crossover frequency ω c The independent variable at the location is used To calculate.

[0082] Time constant T R use To further calculate, the time constant T... R As the first controller parameter R F，v0 It has been confirmed.

[0083] The control segment at the crossover frequency is:

[0084]

[0085] Therefore, the amplification rate is obtained for the traction controller 9.

[0086] Therefore, for the controller Time constant T R and magnification V R As a static controller parameter R F，0 It was determined, and the traction controller 9 used the stationary controller parameter R F，v0 Parameterization was performed. This completed the controller design for the static test T0.

[0087] The traction controller 9, parameterized by means of a static test T0, is now used to perform a creep test T1 with a closed control loop, where, for example, it is assumed that the material 3 has a linear velocity v of 15 m / min in the web processing machine 1. A traction force step ΔF is then applied to the traction force F, as follows: Figure 5 As shown in the image.

[0088] Subsequently, the traction control segment G is in the form of coefficients a1 and b0. F，v1 The fine control segment parameters (s) (which are obtained, for example, with a1 = 18,095 and b0 = 2691,1) are determined using (recursive) least squares based on the traction force step ΔF. Therefore, the length L = a1v = 4.52 m, and the elastic modulus is equal to... This corresponds to the result E = 1.97.10 for the elastic modulus determined above within the range of the static test T0. 7 The comparison of N / m shows that the results of the static test are accurate enough.

[0089] Fine controller parameters R of traction controller 6 used for creep test T1 F，v1 The determination is essentially similar to that used for the static controller parameter R during the static test T0. F，0 The determination will be carried out accordingly.

[0090] However, in order to determine the stationary control segment parameter GF，0 The traction force F increases to the marked traction force F. w2 In order to determine the fine control segment parameter G F，v1 Apply a step traction force ΔF.

[0091] The transfer function at the crossover frequency ω c The independent variable at the location is used To calculate. Time constant T R use To further calculate, the time constant T... R As a controller parameter R F，v1 It has been confirmed.

[0092] The control segment at the crossover frequency is: Therefore, the amplification rate is obtained for the traction controller 9.

[0093] Therefore, for the controller Time constant T R and magnification V R As a parameter of the fine controller R F，v1 The design of the controller for the creep test was thus completed, allowing the traction controller 9 to utilize the fine controller parameters R. F，v1 To parameterize.

Claims

1. A method for parameterizing a traction controller (9) for a controlled roller (2) of a web processing machine (1), wherein the traction controller (9) controls the rotational speed (v9) of the controlled roller (2) to transfer material (3) from the controlled roller (2) to another roller (6) or from another roller (6) to the controlled roller (2) at a linear velocity (v) and when a traction force (F) is applied on the web processing machine (1), characterized in that, During the stationary test (T0) with an online speed (v) of zero, the traction force (F) is increased to the indicated traction force (F). w2 In order to determine the traction control section (G) F,0 The stationary control segment parameters, and the parameters based on the traction control segment (G) F,0 The stationary control parameters (R) of the traction controller (9) are determined by the stationary control segment parameters of the traction controller (9). F,o ), and the traction controller (9) utilizes the stationary controller parameters (R) F,o ) to parameterize.

2. The method as described in claim 1, characterized in that, During the stationary test (T0) at which the online speed (v) is zero, the traction force (F) is increased to a pre-defined stationary traction force operating point (F). op 90% of ).

3. The method as described in claim 1, characterized in that, The traction control section (G) F,0 The parameters of the stationary control segment are determined using the least squares method.

4. The method according to any one of claims 1 to 3, characterized in that, The traction control section (G) F,0 The parameters of the stationary control segment are determined using the recursive least squares method.

5. The method according to any one of claims 1 to 3, characterized in that, The static controller parameters (R) F,o It can be determined using the frequency response curve method.

6. The method according to any one of claims 1 to 3, characterized in that, The elastic modulus (E) of the material (3) is determined according to the traction control section (G). F,0 The parameters of the stationary control segment are determined by the static control segment.

7. The method according to any one of claims 1 to 3, characterized in that, The traction force (F) increases to the marked traction force (F) w2 Before increasing to tensile traction force (F) w1 ).

8. The method as described in claim 1, characterized in that, The traction force (F) increases to the marked traction force (F) w2 Before that, the static traction force operating point (F) was increased. op 10% of ).

9. The method as described in claim 1, characterized in that, The traction force (F) is increased to the stationary traction force operating point (F). op ), and when the static traction force operating point (F) is reached op After that, a traction step (ΔF) is applied to the traction force (F) to determine the stationary controller parameters (R) based on the first stationary quality step response (g0). F,o (quality) 10. The method according to any one of claims 1 to 3, characterized in that, After the static test (T0), a creep test (T1) is performed, setting a first working linear velocity (v1) and a magnitude of a first traction force working point (F). op1 A traction force (F) is applied, a traction force step (ΔF) is applied to the traction force (F), the creep step response (h1) is determined, and the fine control segment parameters (G) of the traction force control segment are determined. F,v1 The parameters are identified from the creep step response (h1) and the fine controller parameters (R) F,v1 Based on the creep step response (h1) and the fine control segment parameters (G) F,v1 The traction controller (9) determines this using the fine controller parameters (R). F,v1 ) to parameterize.

11. The method as described in claim 10, characterized in that, The fine control parameters (G) of the traction control section F,v1 The least squares method is used to identify them.

12. The method as described in claim 10, characterized in that, The fine control parameters (G) of the traction control section F,v1 The recursive least squares method is used for identification.

13. The method as described in claim 10, characterized in that, The fine controller parameters (R) F,v1 It can be determined using the frequency response curve method.

14. The method as described in claim 10, characterized in that, The elastic modulus (E) and / or length (L) of the material (3) at the first working linear velocity (v1) are determined according to the fine control segment parameter (G). F,v1 To determine.

15. The method as described in claim 10, characterized in that, A traction step (ΔF) is applied to the traction force (F) to determine the fine controller parameter (R) for the first operating linear velocity (v1) based on the creep quality step response. F,v1 (quality) 16. The method as described in claim 15, characterized in that, The fine controller parameters (R) for the first operating linear velocity (v1) F,v1 The quality of () is determined using the best-fit method.

17. The method according to any one of claims 1 to 3, characterized in that, The static controller parameters (R) f,v0 The parameters are stored, and extrapolation speed controller parameters for several extrapolation linear velocities (v) are extrapolated from the static controller parameters. During the operation of the web processing machine (1) at an extrapolation linear velocity in one of the ranges of the extrapolation linear velocities (v), the relevant extrapolation speed controller parameters are invoked to parameterize the traction controller (9).

18. The method as described in claim 10, characterized in that, Fine controller parameters (R) for the first operating linear velocity (v1) f,v1 The fine controller parameters (R) are stored and, during operation of the web processing machine (1) at a linear speed (v) within the range of the first working linear speed (v1), are... f,v1 The traction controller (9) is invoked to parameterize the traction controller (9).

19. The method as described in claim 10, characterized in that, After the creep test (T1), a speed test (T2) is performed, in which a second working linear velocity (v2) and a second traction force working point (F) are set. op2 The traction force (F) is applied, a traction force step (ΔF) is applied to the traction force (F), the speed test step response (h2) is determined, and additional fine control parameters (G) of the traction force control segment are determined. F,v2 The parameters are identified from the speed test step response (h2) and the additional fine controller parameters (R) F,v2 The speed test step response (h2) and the additional fine control segment parameters (G) are used as a basis. F,v2 The traction controller (9) determines this using the additional fine controller parameters (R). F,v2 ) to parameterize.

20. The method as described in claim 19, characterized in that, Additional fine control parameters (R) for the second operating linear velocity (v2) f,v2 The additional fine controller parameters (R) are stored and, during operation of the web processing machine (1) at a linear speed (v) within the range of the second working linear speed (v2), are also stored. f,v2 The traction controller (9) is invoked to parameterize the traction controller (9), and the traction controller (9) utilizes the additional fine controller parameters (R). F,v2 ) to parameterize.

21. The method as described in claim 19, characterized in that, For additional working linear speed (v) x Additional fine controller parameters (R) f,vx According to the fine controller parameters (R) f,v1 ) and the additional fine controller parameters (R) f,v2 The traction controller (9) determines this using the additional fine controller parameters (R). F,vx ) to parameterize.

22. The method as described in claim 21, characterized in that, For the additional working linear velocity (v) x Additional fine controller parameters (R) f,vx ) is stored, and in the web processing machine (1) with relevant fine controller parameters (R f,vx At the corresponding additional working linear speed (v) x During operation at linear velocities (v) within the range of ), the associated additional fine controller parameters (R) f,vx The traction controller (9) is invoked to parameterize the traction controller (9), and the traction controller (9) utilizes the associated additional fine controller parameters (R). f,vx ) to parameterize.

23. A traction controller (9) for controlling the traction force (F) of a material (3) in a web processing machine (1), wherein the traction controller (9) is parameterized according to any one of claims 1 to 22, wherein the material (3) is transported at a linear velocity (v) and when the traction force (F) is applied from a controlled roller (2) to another roller (6) or from another roller (6) to a controlled roller (2).

24. A parameterization unit (90) for parameterizing a traction controller (9) of a controlled roller (2) of a web processing machine (1), on which material (3) is transported at a linear velocity (v) and when a traction force (F) is applied from the controlled roller (2) to another roller (6) or from another roller (6) to the controlled roller (2), wherein the traction force (F) can be controlled by means of the traction controller (9) via the rotational speed (v9) of the controlled roller (2), characterized in that, The parameterization unit (90) increases the traction force (F) to the marked traction force (F) during a stationary test (T0) when the online speed (v) is zero. w2 ), determine the traction control section (G F,0 The stationary control segment parameters, and the parameters based on the traction control segment (G) F,0 The stationary control parameters (R) of the traction controller (9) are determined by the stationary control segment parameters of the traction controller (9). F,o ), and the traction controller (9) utilizes the stationary controller parameters (R) F,o ) to parameterize.

25. The parameterization unit (90) as described in claim 24, characterized in that, During the stationary test (T0) at which the online speed (v) is zero, the traction force (F) is increased to a pre-defined stationary traction force operating point (F). op 90% of ).

26. The parameterization unit (90) as described in claim 24, characterized in that, The static controller parameters (R) F,o It can be determined using the frequency response curve method.

Images