JP2003048226A - Method for detecting shift of corresponding relation between actuator and thickness sensor in film manufacturing apparatus, film thickness control method and apparatus and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a film thickness control unit and method capable of controlling the discharge amount of a resin from an actuator so as not to generate the fluctuation of the thickness in a width direction of a film even if shift is generated between the actuator and a film thickness measuring position. SOLUTION: The film thickness control method includes a step for measuring the thickness of the film, a step (90) for calculating the control input to the corresponding actuator so that the thickness of the film becomes predetermined target quantity over the width direction of the film corresponding to the measured thickness, steps (92 and 96) for performing wavelet analysis with respect to the waveform of the distribution in the width direction of the film of the control input to a plurality of actuators to extract a wavelet component and steps (98, 100 and 102) for correcting the control input to the actuator using the wavelet component to control the casting quantity of the resin.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a film for uniform distribution of film thickness in a film manufacturing process in which a resin material is cast from a die having an automatic film thickness controller to form a thin sheet. In particular, a plurality of actuators arranged in a casting step in order to control a casting amount of a resin as a raw material of the film, and a thickness sensor for detecting the thickness of the film. The present invention relates to a method for detecting a shift in correspondence between the above, a device and a method for controlling film thickness capable of preventing performance deterioration due to the method, and a program therefor. The present invention can be applied to general films regardless of the material, including films made of polyimide.

[0002]

2. Description of the Related Art Generally, the film manufacturing process is as shown in FIG.
As shown in FIG. 4, it can be divided into five processes: casting, drying, firing, observation, and winding. First, in the casting process,
The resin solution is cast from the die 150. Next, in the drying process, the cast liquid film 152 is dried to form the film 40 having self-supporting property. In the baking step, the film 40 is heat-treated. In the observation step, the quality of the film 40 is inspected by measuring the thickness of the film 40 with a non-contact type thickness sensor 42, and the film is wound into a roll 154 in the winding step.

Referring to FIG. 25, the width of the film is controlled in the width direction by using a die having several to several hundreds of actuators 48A to 48N capable of minutely changing the ejection amount in the width direction. According to the thickness measurement values measured by the plurality of non-contact type thickness sensors 42A to 42N, a plurality of actuator operation amounts are calculated so that the thickness of the film becomes a predetermined target value. PID (Proportional plus Integral
plus Derivative action) Controller 44A-44
A film having a uniform thickness is produced by feedback-controlling the actuators 48A to 48N by controlling N.

[0004]

However, the thickness of the produced film is measured after the resin material solution is discharged from the die and subjected to post-treatments such as drying and heat treatment. Therefore, the correspondence between the measurement position and the operation position often occurs. When deviation occurs, there is a problem that strict thickness control cannot be performed due to deterioration of control performance due to deviation.

As shown in FIG. 25, the film thickness control is performed using the same number of thickness sensors 42A to 42N and single loop PID controllers 44A to 44N as the number of actuators 48A to 48N. If the film portion whose thickness is detected by the thickness sensors 42A to 42N corresponds to the portion ejected from the previously associated one of the actuators 48A to 48N, the thickness of the film 40 is not so problematic. Absent.

However, as shown in FIG. 26, the film portion where the thickness sensors 42A to 42N detect the thickness is the thickness sensor 42A to 42N among the actuators 48A to 48N.
In the case of the object other than the one previously associated with 42N, vibration occurs in the operation amount and the controlled amount in the lateral direction,
A phenomenon in which this vibration is amplified occurs.

For example, in FIG. 26, the actuator 4
The thickness of the film portion formed of the resin discharged from 8A and 48C is in the plus direction (increase direction), and the thickness of the film portion formed of the resin discharged from the actuator 48B is in the minus direction (decrease direction). And Then, as shown in FIG. 26, the thickness sensor 42A that should originally be paired with the actuator 48A.
However, it is assumed that the position is set so as to increase the thickness of the film portion discharged from the actuator 48B during film production. In this case, the sensor 42A outputs a signal "the thickness is decreasing". According to this signal, the PID controller 44A causes the actuator 48A
On the other hand, control is performed such that the discharge amount is increased to increase the thickness. As a result, although the actuator 48A should originally reduce the discharge amount of the resin, it actually increases the discharge amount of the resin.

On the other hand, the sensor 42B is an actuator 48.
The thickness of the film portion formed of the resin discharged from C is detected. Therefore, in this case, the thickness of the film portion detected by the sensor 42B becomes larger than the target value. Therefore, the PID controller 44B controls to reduce the amount of resin discharged from the actuator 48B according to the output of the sensor 42B. As a result, although the amount of resin discharged from the actuator 48B was originally too small, it is controlled to be smaller.

As described above, when the positions of the actuator and the thickness sensor that are adjacent to each other are deviated from each other, the PID control does not work correctly for both the manipulated variable and the controlled variable, and the phenomenon of lateral vibration amplification occurs. This vibration tends to increase with time. Such a problem inevitably occurs in a film manufacturing process in which the position of the casting process and the observation process are separated as shown in FIG. 24, and the film position formed by the resin discharged from a certain actuator is not always constant. It's a problem. This is an important issue especially in the process of manufacturing a film, which requires surface smoothness and thickness uniformity.

Therefore, it is preferable that the misalignment of the pairing between the actuator and the thickness sensor, which causes vibration in the thickness of the film in the width direction, can be accurately detected. Further, it is more preferable that the thickness of the film can be kept uniform for a long time by controlling each actuator so as to cancel such vibration.

An object of the present invention is to provide a method for surely detecting a deviation between an actuator and a film thickness measuring position.

Another object of the present invention is to control the amount of resin discharged from the actuator so that vibration of the thickness in the width direction does not occur even if a deviation occurs between the actuator and the film thickness measuring position, thereby achieving uniform thickness. An object of the present invention is to provide a film thickness control device and method capable of continuously producing an excellent film for a long period of time.

[0013]

Means for Solving the Problems As a result of intensive studies on the above problems, the present inventors have found that the displacement of the actuator can be detected by performing a wavelet analysis of the operation amount of the actuator. We found a method to prevent the deterioration of control performance due to deviation by applying saturation after applying saturation. In addition, we proposed an algorithm that can secure high control performance in the non-shifted part by relaxing the saturation constraint in the non-shifted part. As a result, they have found that it is possible to continuously produce a film having excellent thickness uniformity for a long period of time, thus completing the present invention.

That is, the method according to claim 1 is controlled by a plurality of actuators arranged in the casting step and a plurality of actuators in order to control the casting amount of the resin as the raw material of the film. For detecting the thickness of the film produced by the resin cast at the casting amount,
A method for detecting a shift in a correspondence relationship between an actuator and a thickness sensor in a film manufacturing apparatus including a plurality of thickness sensors arranged corresponding to a plurality of actuators in a width direction of a film. , The step of measuring the thickness of the film using each of the plurality of thickness sensors, and the adjustment of the thickness of the film to a predetermined target value according to the thickness measured by each of the plurality of thickness sensors. The step of calculating the operation amount of each of the actuators and the wavelet analysis of the distribution of the operation amounts of the plurality of actuators in the width direction of the film are performed to determine the correspondence between the plurality of actuators and the plurality of thickness sensors. Detecting a deviation.

According to another aspect of the present invention, a plurality of actuators arranged in the casting process and a plurality of actuators are used to control the casting amount of the resin as the raw material of the film. In a film manufacturing apparatus including a plurality of thickness sensors arranged corresponding to a plurality of actuators in the width direction of the film for detecting the thickness of the film produced by the resin cast in a casting amount , Controlling a plurality of actuators, the film thickness control method for making the thickness of the film in the width direction of the film uniform, a step of measuring the thickness of the film, respectively, using a plurality of thickness sensors, According to the thickness measured by each of the multiple thickness sensors, the corresponding actuation is performed so that the film thickness reaches the predetermined target value. The step of calculating the operation amount of each actuator and the waveform of the distribution of the operation amount of the actuator in the width direction of the film by performing wavelet analysis on the waveform of the distribution of the operation amount of the plurality of actuators in the width direction of the film. Of the wavelet component, and the step of correcting the operation amount of the plurality of actuators using the wavelet component, and controlling the casting amount of the resin by the plurality of actuators by the corrected operation amount.

According to another aspect of the present invention, in order to control the casting amount of the resin as the raw material of the film, a plurality of actuators arranged in the casting step and a plurality of actuators are used. In a film manufacturing apparatus including a plurality of thickness sensors arranged corresponding to a plurality of actuators in the width direction of the film for detecting the thickness of the film produced by the resin cast in a casting amount , A film thickness control program that controls a plurality of actuators to operate a computer so as to make the film thickness uniform in the width direction of the film is a value measured by a plurality of thickness sensors for each film thickness. According to the routines that control the computer to obtain the A routine that controls the computer to calculate the operation amount of each corresponding actuator so that the film thickness becomes a predetermined target value, and a distribution of the operation amounts of the plurality of actuators in the width direction of the film. A routine that performs wavelet analysis on the waveform to extract the wavelet component of the waveform of the actuator operation amount distribution in the width direction of the film, and the operation amount of multiple actuators that was corrected using the wavelet component and the corrected operation And a routine for controlling the amount of resin cast by a plurality of actuators depending on the amount.

A film thickness control apparatus according to still another aspect of the present invention comprises a plurality of actuators arranged in a casting process and a resin cast by a casting amount controlled by the plurality of actuators. In order to detect the thickness of the produced film, a plurality of thickness sensors arranged corresponding to the plurality of actuators in the width direction of the film, and the thickness of each film measured by the plurality of thickness sensors Operation amount calculation means for calculating the operation amounts of the corresponding actuators so that the thickness of the film becomes a predetermined target value, and a waveform of the distribution of the operation amounts of the plurality of actuators in the width direction of the film. Wavelet analysis is performed for the waveform of the waveform of the actuator manipulated variable distribution in the width direction of the film. And an operation amount correction means for correcting the operation amount of a plurality of actuators using the wavelet component and controlling the casting amount of the resin by the plurality of actuators by the corrected operation amount. Including and

Preferably, the wavelet analysis is Ha
Let the ar function be the wavelet generating function.

Preferably, the wavelet analysis is the first
It is a wavelet analysis for a stepped wavelet component.

Alternatively, the wavelet analysis is a wavelet multiple analysis. The wavelet analysis may be a wavelet analysis for the second stage wavelet component.

Prior to controlling the casting amount of the resin, the wavelet component may be corrected so that the value of the wavelet component falls within a predetermined upper limit and lower limit. It is more preferable to set the upper limit, the lower limit, or both to a predetermined initial value, and adjust the upper limit, the lower limit, or both values in response to the outputs of the plurality of thickness sensors.

[0022]

BEST MODE FOR CARRYING OUT THE INVENTION An example of an embodiment of the film thickness control method according to the present invention will be described below.

[First Embodiment] The vibration of the film thickness described with reference to FIG. 26 is an amplification in the width direction of the film, and is not a change in the flow direction. If this deviation occurs in the entire film, the vibration will continue to be amplified and diverge, but there is also a part where the sensor and actuator are correctly combined, and if the controller of that part shows good control performance , This widthwise vibration is absorbed at that part and may not diverge as a whole. When there is a partial deviation, vibration may be amplified in the part where the deviation occurs, but stable control may occur in the part that is successfully paired. By detecting the part where the pairing is inappropriate and preventing the vibration from increasing, the control performance as good as possible can be achieved.

With respect to the vibration of the thickness in the width direction due to the displacement shown in FIG. 26, the directions of increase of the amplitude alternate, that is, the thickness becomes excessive in the film portion formed by the resin discharged from a certain actuator. On the other hand, there is a feature that the film portion formed of the resin discharged from the actuator adjacent thereto has an excessively small thickness. Therefore, the inventors of the present invention
With respect to the operation amount of the actuator, it was considered to detect the vibration of the thickness by detecting whether or not the difference between the actuator and the adjacent actuator is amplified. As a result, the inventors of the present invention have conducted a wavelet analysis capable of analyzing the frequency component of each curve in a curve formed by plotting the film thickness in the width direction (hereinafter referred to as “thickness curve”). I thought about using it.

FIG. 1 shows a schematic block diagram of a film thickness control apparatus according to the first embodiment of the present invention. Figure 1
Referring to FIG. 25, the device 30 includes PID controllers 44A to 44A in addition to the components of the conventional device shown in FIG.
Connected between N and the actuators 48A-48N, P
Wavelet analysis, which will be described later, is performed on the manipulated variables output from the ID controllers 44A to 44N, and the manipulated variables are corrected based on the results of the analysis to correct the actuators 48A to 48N.
It includes a control computer 46 for feeding N.

Typical actuators include, but are not limited to, lip heaters and heat bolts, and any actuator can be used as long as it can control the flow rate of the resin as the raw material of the film. Good. Although a non-contact thickness sensor is used in the present embodiment, a contact type thickness sensor may be used.

Referring to FIG. 2, the control computer 46.
Includes a computer 50, a console 72, a mouse 74 and a monitor 76 connected to the computer 50.

The computer 50 has a CPU (Central Pr
cessing unit) 60, a bus 62 connected to the CPU 60, and a ROM (Read-O) connected to the bus 62
nlyMemory) 64, RAM (Random Access Memory) 6
6, a CD (Compact Disc) -ROM drive 68,
An input / output unit 70 for inputting / outputting data to / from PID controllers 44A to 44N and actuators 48A to 48N is included.

The ROM 64 stores a program for realizing the device of this embodiment. Alternatively, the CD-ROM 78 storing the program is stored in the CD-ROM 6
It may be mounted on the RAM 8, loaded into the RAM 66, and executed by the CPU 60.

The configuration and operation itself of the control computer 46 are well known. Therefore, detailed description thereof will not be repeated here.

The outline of the wavelet analysis process executed by the CPU 60 will be described below. The following wavelet analysis process is particularly suitable for analyzing input waveforms,
The Haar function, which facilitates the analysis process, is used as the generating function of the wavelet, but other generating functions, for example, generating functions of Mexican hat, French hat, Shannon, Gabor, etc. may be used. In general, the wavelet generating function has a value of 0 when integrated in the interval of (−∞, ∞), and the region where the value is other than 0 is limited to the finite interval. Therefore, the average value of the wavelet generating function in the finite section is 0.

As shown in FIG. 4, the Haar function is a wavelet function which is stepwise, takes a value other than 0 only in a finite section, and has an average of zero. Using this Haar function,
As shown in FIG. 4A, from the data string in which the output from the PID controller is arranged in the width direction of the film,
As shown in (B) and FIG. 4 (C), the difference between adjacent data is cut out.

In the case of FIG. 4B, the difference is cut out as a Haar function from the combination of (odd data, even data) among the combinations of two adjacent data strings. Figure 4
In the case of (C), the difference is cut out as a Haar function from the combination of (even data, odd data). Figure 4 (B)
In (C), the dotted line indicates the average of two adjacent data. A waveform that matches the Haar wavelet waveform centered on the average represents the difference between the two data. In the case of FIG. 4C, since there is no paired data for the first data and the last data, the difference is calculated assuming that the same data as itself forms a pair.

As a result of the cutting process shown in FIG. 4 (B), the data string shown in FIG. 4 (A) is the difference data string shown in FIG. 4 (D1) and the difference data string shown in FIG. 4 (D2). It is decomposed into a residual data string with the difference removed. Similarly, FIG.
As a result of the cutting process shown in (C), the data string shown in FIG. 4 (A) is the difference data string shown in FIG. 4 (E1).
It is decomposed into a residual data string obtained by removing the difference shown in FIG. 4 (E2). In the remaining data string, the unit of fluctuation width is twice as wide as the unit of fluctuation width of the original data string shown in FIG. 4 (A). Wavelet analysis is to analyze the strength of vibration at each place by this procedure.

As will be described later, the residual is cut out by a Haar function having a width twice the width, and the residual having a width double is obtained to obtain a difference data string at the second stage. Is a Haar function that is 4 times the original, and has a width of 4
The difference data string at the third stage can be found by finding the double residue, and the difference data string at the nth stage having the original width of 2 ⁿ can be found in the same manner. Wavelet multiple analysis is used to analyze the strength of vibrations with different fluctuation widths at each place by this procedure. In the present embodiment, only the difference data string of the first stage is used.

In this embodiment, the wavelet components shown in FIGS. 4 (D1) and 4 (E1) are converted to those shown in FIG. 4 (F).
4G, and take their average as shown in FIG. 4 (G). At a place where the operation position of the actuator and the film thickness measurement position deviate from each other, amplification of the variation in the thickness in the width direction occurs. Therefore, by applying this wavelet analysis to the output of the actuator, it is possible to detect the deviation between the operation position of the actuator and the film thickness measurement position.

The increase in the width direction unevenness of the actuator operation amount creates the film thickness unevenness in the width direction. Therefore, it is effective to prevent the unevenness of the actuator operation amount in the width direction from becoming large in order to prevent the unevenness of the film thickness in the width direction from occurring. Therefore, in the present embodiment, the wavelet component of the first stage as a result of the wavelet decomposition of the actuator operation amount output from the single loop PID controller (FIG. 4 (G)).
However, the upper limit 80 and the lower limit 82 are set, and those exceeding the upper and lower limits are corrected to the limit values and then returned to the actuator operation amount shown in FIG. 4 (H).

By doing so, it is possible to prevent the difference in the operation amount between the adjacent actuators from increasing, and it is possible to prevent an increase in the width-direction unevenness of the film thickness.
In addition, since the residual components of the wavelet remain unchanged, large movements of the whole are not interrupted and the average thickness can be brought close to the target value even if there is a deviation in pairing.

It should be noted that if the upper and lower limits of the wavelet are made smaller, the stability is increased, but since the difference between the manipulated values next to each other is made small, it becomes impossible to eliminate the unevenness of the film thickness due to disturbance. Therefore, in the present embodiment, as will be described later, when there is no deviation, an operation for canceling the disturbance is performed by increasing the upper and lower limit values, and when there is a deviation, the upper and lower limit values are reduced to stabilize. Try to increase your sex.

The control error moves in the direction where it is attenuated by the PID controller in the part where there is no deviation, and in the direction where it diverges in the part where there is no deviation. Therefore, the behavior of the controlled variable is monitored while loosening the limit value of the wavelet component of the manipulated variable, and the limit value of the part that is not displaced is large, and the limit value of the part that is displaced is small, and the limit values are adjusted respectively.

FIG. 5 shows a flowchart of the processing performed by the CPU 60 for controlling the thickness. First, step 90
Then, the data string output by each PID controller 44A to 44N at that time is acquired. This data string uses a plurality of thickness sensors to measure the thickness of the film, respectively, so that the thickness of the film 40 has a predetermined value according to the thickness measured by each of the thickness sensors. , Corresponding to the operation amount of each actuator calculated by the PID controllers 44A to 44N. Therefore, the PID controllers 44A to 44N correspond to operation amount calculation means.

Then, for each odd-numbered data, the difference from the right adjacent is calculated (92). This process corresponds to the process shown in FIG. Further, the difference between the odd-numbered data and the data on the left is taken (94). This process corresponds to the process shown in FIG. Then in step 96, step 9
Calculate the average of the difference data sequence calculated at 2,94. This process corresponds to the process shown in FIGS. 4 (F) to 4 (G). Thus, the first stage wavelet component is calculated. That is, in steps 92 and 94, the wavelet analysis is performed on the waveform of the distribution of the operation amounts of the plurality of actuators in the width direction of the film, and the wavelet component of the waveform of the distribution of the operation amount of the actuator in the width direction of the film is obtained. To be extracted.

Among the calculated first-stage wavelet components, it is judged whether or not there is a value exceeding the upper limit value or a value lower than the lower limit value (98), and if there is such a component, the upper limit value and the lower limit value, respectively. Replace with the value (100). The corrected first-stage wavelet component is returned to the original data string and output as the operation amount for controlling the actuators 48A to 48N, and this processing ends. That is, in steps 98 and 100, the operation amounts of the plurality of actuators are corrected using the wavelet component, and the resin casting amounts by the plurality of actuators are controlled by the corrected operation amounts.

By repeating the above-described steps 90 to 102 in units of a predetermined time, the film thickness can be uniformly controlled even if the actuator is displaced from the thickness measuring position.

Next, when there is no deviation, an operation for canceling the disturbance is performed by increasing the upper and lower limit values, and when there is a deviation, the upper and lower limit values are decreased to increase stability. The control method will be described with reference to FIG. The process of FIG. 6 is for controlling the upper limit value and the addition / subtraction value referred to in step 98 of FIG. 5, as a process different from the process shown in FIG.

When the process is started, first, an appropriate predetermined initial value is set as a saturation value (upper limit value or lower limit value) (112). This initial value can be determined in advance by various experiments or simulations as a value capable of stably controlling the film thickness.

Subsequently, it is determined whether a fixed time has elapsed after setting this saturation value (114). This time is a design matter and can be set in advance by experiments or simulations according to system requirements.

When it is determined in step 114 that the fixed time has elapsed, the saturation value is once relaxed and the absolute value is increased (116). Further, the absolute value of the control response PV at this time is stored in the variable PVstock.

Furthermore, it is determined whether a fixed time has passed (118). After a certain time has elapsed, it is determined whether the absolute value of the control response PV at that time is larger than the value of PVstock (120). The absolute value of the control response PV is P
If it is larger than the value of Vstock, the saturation value is returned to the initial value (122), and the control returns to step 114. If the absolute value of the control response PV is equal to or smaller than the value of PVstock, the control returns to step 118 while keeping the saturated value.

By this processing, when there is no deviation between the actuator and the thickness measurement position, the response to the fluctuation of the control amount can be speeded up to perform the operation of canceling the disturbance. The stability of thickness control can be increased by slowing the response. It is not always necessary to perform such upper and lower limit control as will be described later. However, the thickness control can be performed more stably by performing this control.

Specific simulation results will be described below. In the following simulation, a model having 13 lip heaters in the width direction of the die was created,
Under the condition that a constant disturbance is present in the shape shown in FIG. 7, control is performed so as to reduce the thickness deviation in the width direction to zero.

(Simulation 1) As shown in Table 1, a simulation was performed under the condition that the correspondence between the measurement end and the operation end was partially deviated. The results are as shown in FIGS. 8 to 10 show the controlled variables at t = 0, 500 and 1000, respectively. 11 to 13 show t =
The operation amount at 0, 500, and 1000 is shown.

In this example, since the central PID controllers Nos. 6 to 8 worked normally, the central part could be controlled, but the controlled amount and the manipulated amount were large at the ends.

[0054]

[Table 1]

Here, the amount of operation of the 13 lip heaters at each time is subjected to wavelet analysis, and t = 0, 5
Values at 00 and 1000 are shown in FIGS.
14 to 16, it can be seen that the wavelet components of Nos. 1 to 4 and 10 to 13 greatly increase with time, and the change in the width direction of the film thickness due to the deviation is amplified. That is, by examining the wavelet component, it can be seen that there is a shift in the correspondence between the actuator and the thickness measurement position.

When the wavelet analysis is not performed in this simulation, the deviation must be detected from the time response of the controlled variable and the manipulated variable shown in FIGS. 8 to 10 and 11 to 13, respectively. However,
From FIG. 10 and FIG. 11 to FIG. 13, it is difficult to understand the difference from the adjacent one, and it can be seen that it takes time to detect the deviation.

(Simulation 2) As shown in Table 1, under the condition that the correspondence between the measurement end and the operation end is partially deviated, the process shown in FIG. The results are as shown in FIGS. Figure 1
7 to 19 show the controlled variables at t = 0, 500, and 1000, respectively. 20 to 22 show t =
The operation amount at 0, 500, and 1000 is shown.

As shown in FIG. 22, in the portion where the pairing of the measuring end and the operating end is appropriate (6 to 8 in the center), the difference in the operating amount between the adjacent ones is relatively large, and the operating amount is relatively large. Is well controlled. Further, in the displaced portion, the difference in the operation amount from the adjacent portion is suppressed, and the deterioration of the thickness due to the displacement can be suppressed.

Simulation 1 does not use an algorithm such as saturation or relaxation of wavelet components. Therefore, when the result of the simulation 1 is compared with the result of the simulation 2, the following can be seen. That is, in FIGS. 8 to 10, as a result of being continuously affected by the shift, the thickness of the end portion is deteriorated.
17 to 19, it can be seen that the influence of the displacement is small and the thickness of the end portion is relatively stable.

As described above, when the deviation detecting method and the film thickness controlling method according to the present embodiment are used,
The corresponding shift between the actuator and the measurement position can be detected. Further, it is possible to prevent the control performance from deteriorating due to the shift, and it is possible to secure high control performance in the non-shift portion. As a result, it becomes possible to continuously produce a film having excellent thickness uniformity for a long period of time.

[Second Embodiment] In the first embodiment, only the first-stage wavelet component is used to detect the corresponding shift between the actuator and the measurement position, and the film thickness is controlled. . However, as the wavelet components, not only the first stage but also higher-order wavelet components can be used. The apparatus of the second embodiment described below uses the wavelet component of the second stage.

When the second-stage wavelet component is used, as shown in FIGS. 4 (D2) and 4 (E2),
The residual data string from which the first-stage wavelet component is extracted is used to extract the wavelet component in the same manner as the first-stage wavelet component is extracted. The result is the second stage wavelet component.

The remaining data string shown in FIG. 4D2 is
As described above, it consists of the average of the odd-numbered data and the data on the right of the original data sequence. Its width is twice the width of the original data string. The remaining data string shown in FIG. 4 (E2) is an average of the odd-numbered data and the data adjacent to the left of the original data string. Its width is twice that of the original data string.

For the remaining data string thus obtained,
The wavelet component can be extracted in the same manner as in the first embodiment. The wavelet component thus obtained is the second-stage wavelet component, and even if the second-stage wavelet component is used, the corresponding deviation between the actuator and the thickness measurement position is detected, and the film thickness is uniformly controlled. be able to.

FIG. 23 shows a schematic flow chart of a program for realizing the extraction of the wavelet component at this time. Referring to FIG. 23, first of all, similarly to the case of the first embodiment, PID controllers 44A to 44A at that time point.
The manipulated variable output from N is acquired as a data string.

Subsequently, the average of the odd-numbered data in the data string and the data on the right is calculated (13).
4). With respect to each odd-numbered data of the obtained average data sequence, two kinds of wavelet components are extracted by taking the difference between the right neighbor and the left neighbor, respectively (136).
The width of the wavelet generating function at this time is twice as wide as that used in the first embodiment.

Further, the odd-numbered data in the data string is averaged with the data on the left (138). With respect to each odd-numbered data of the obtained average data string, two kinds of wavelet components are extracted by taking the difference between the right neighbor and the left neighbor, respectively (140). The width of the wavelet generating function at this time is also twice as wide as that used in the first embodiment.

The average of the two wavelet components obtained in step 136 and the two wavelet components obtained in step 140 is calculated (142). As a result, the second stage wavelet component is obtained.

With respect to the wavelet component thus obtained, it is judged whether or not there is a component exceeding a predetermined upper and lower limit value (144), and if there is, that value is replaced with the upper and lower limits (146).

In this way, the second-stage wavelet components stored within the upper and lower limits are returned to the original data string and output as the manipulated variables for the actuators 48A to 48N (148).

Even when the second-stage wavelet is used in this manner, the magnitude of fluctuation in the distribution waveform of the manipulated variable in the width direction is detected and the actuator and the thickness are measured in the same manner as in the first embodiment. Detecting the shift in correspondence with the position,
And the thickness of the film can be controlled uniformly.

However, since the width of the wavelet generating function used in the second embodiment is twice as wide as that in the first embodiment, the response to the disturbance becomes small. On the contrary, the stability of the film thickness is higher in the second embodiment.

Similar to the second embodiment, it is possible to detect the corresponding shift between the actuator and the thickness measurement position and control the film thickness by using the nth-order wavelet component. The method will be apparent to those skilled in the art from the above description of the first and second embodiments. The wavelet analysis using the nth-order wavelet component is generally called a wavelet multiple analysis.

The embodiments disclosed this time are to be considered as illustrative in all points and not restrictive. The scope of the present invention is shown not by the above description but by the claims, and is intended to include meanings equivalent to the claims and all modifications within the scope.

[0075]

With the thickness control method according to the present invention, it is possible to detect the corresponding shift between the actuator and the thickness measurement position. Further, it is possible to prevent the control performance from deteriorating due to the displacement, and it is possible to secure high control performance in the portion where the correspondence between the actuator and the thickness measurement position is not displaced. As a result, it becomes possible to continuously produce a film having excellent thickness uniformity for a long period of time.

[Brief description of drawings]

FIG. 1 is a schematic block diagram of a film thickness control device according to an embodiment of the present invention.

FIG. 2 is a block diagram of the control computer shown in FIG.

FIG. 3 is a schematic diagram for explaining a Haar function.

FIG. 4 is a schematic diagram of wavelet decomposition using a Haar function as a wavelet generating function.

FIG. 5 is a flowchart of a program that realizes the processing shown in FIG.

FIG. 6 is a flowchart of a program for adjusting the limit value of the wavelet component of the manipulated variable.

FIG. 7 is a diagram showing a constant disturbance in a simulation.

FIG. 8 is a diagram showing a controlled variable at time t = 0 in the first embodiment.

FIG. 9 is a diagram showing a controlled variable at time t = 500 in the first embodiment.

FIG. 10 is a diagram showing a controlled variable at time t = 1000 in the first embodiment.

FIG. 11 is a diagram showing an operation amount at time t = 0 in the first embodiment.

FIG. 12 is a diagram showing a manipulated variable at time t = 500 in the first embodiment.

FIG. 13 is a diagram showing an operation amount at time t = 1000 in the first embodiment.

FIG. 14 is a diagram showing a first-stage wavelet component at time t = 0 in the first embodiment.

FIG. 15 is a diagram showing a first-stage wavelet component at time t = 500 in the first embodiment.

FIG. 16 is a diagram showing a first-stage wavelet component at time t = 1000 in the first embodiment.

FIG. 17 is a diagram showing a controlled variable at time t = 0 when the algorithm of the second embodiment is applied.

FIG. 18 is a diagram showing a controlled variable at time t = 500 when the algorithm of the second embodiment is applied.

FIG. 19 is a diagram showing an operation amount at t = 1000 when the algorithm of the second embodiment is applied.

FIG. 20 is a diagram showing a manipulated variable at time t = 0 when the algorithm of the second embodiment is applied.

FIG. 21 is a diagram showing an operation amount at time t = 500 when the algorithm of the second embodiment is applied.

FIG. 22 is a diagram showing a controlled variable at time t = 1000 when the algorithm of the second embodiment is applied.

FIG. 23 is a flowchart of a program that realizes processing for obtaining a second-stage wavelet component.

FIG. 24 is a process drawing of a general film manufacturing process.

FIG. 25 is a diagram for explaining a mutual relationship assumed for control between an operation amount and a measurement end.

FIG. 26 is a schematic diagram for explaining a problem when the operation amount and the position of the measurement end are deviated.

[Explanation of symbols]

40 film, 42, 42A to 42N thickness sensor,
44A to 44N PID controller, 46 control computer, 48A to 48N actuator, 50
Computer, 60 CPU, 62 bus, 64 RO
M, 66 RAM, 68 CD-ROM drive, 70
Input / output unit, 78 CD-ROM, 150 dies, 15
2 resin.

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Haruhiko Maki             16 Tanaka Higashikogen Town, Sakyo Ward, Kyoto City, Kyoto Prefecture F-term (reference) 4F205 AA40 AG01 AM22 AP11 AR12                       GA07 GB02 GC02 GC07 GN24                       GN25

Claims (20)

[Claims] 1. A plurality of actuators arranged in a casting process and a casting amount controlled by the plurality of actuators for controlling a casting amount of a resin as a raw material of a film. And a thickness sensor in a film manufacturing apparatus including a plurality of thickness sensors arranged corresponding to the plurality of actuators in a width direction of the film for detecting the thickness of the film generated by the resin. Is a method for detecting the deviation of the correspondence relationship between and, using the plurality of thickness sensors, respectively measuring the thickness of the film, the thickness measured respectively by the plurality of thickness sensors In accordance with the above, the step of calculating the operation amount of the corresponding actuator is performed so that the film thickness reaches a predetermined target value. And performing a wavelet analysis on the distribution of the operation amounts of the plurality of actuators in the width direction of the film, and detecting a corresponding shift between the plurality of actuators and the plurality of thickness sensors. A method for detecting a shift in correspondence between an actuator and a thickness sensor in a film manufacturing apparatus including. 2. The method according to claim 1, wherein the wavelet analysis uses a Haar function as a wavelet generating function. 3. The wavelet analysis is a wavelet analysis for a first stage wavelet component,
The method according to claim 1 or 2. 4. The method according to claim 1, wherein the wavelet analysis is a wavelet multiplex analysis. 5. The wavelet analysis is a wavelet analysis for a second stage wavelet component,
The method of claim 4. 6. A plurality of actuators arranged in a casting process and a casting amount controlled by the plurality of actuators for controlling a casting amount of a resin as a raw material of a film. A plurality of thickness sensors arranged corresponding to the plurality of actuators in the width direction of the film for detecting the thickness of the film generated by the resin, Controlling the actuator, a film thickness control method for making the film thickness uniform in the width direction of the film, the step of measuring the thickness of the film, respectively, using the plurality of thickness sensors, Before taking measures so that the thickness of the film reaches a predetermined target value according to the thickness measured by each of the multiple thickness sensors. A step of calculating the operation amount of each of the actuators, of the distribution of the operation amount of the actuator in the width direction of the film by performing wavelet analysis on the waveform of the distribution of the operation amount of the plurality of actuators in the width direction of the film. A step of extracting a wavelet component of a waveform, a step of correcting the operation amount of the plurality of actuators using the wavelet component, and a step of controlling the casting amount of the resin by the plurality of actuators by the corrected operation amount. A method for controlling the thickness of a film, comprising: 7. The method according to claim 6, wherein the wavelet analysis uses a Haar function as a wavelet generating function. 8. The wavelet analysis is a wavelet analysis for a first stage wavelet component,
The method according to claim 6 or 7. 9. The method according to claim 6, further comprising the step of correcting the wavelet component so that the value of the wavelet component falls within a predetermined upper limit and lower limit prior to the step of controlling the casting amount of the resin. 9. The method according to any one of to 8. 10. A step of setting the upper limit or the lower limit or both of them to a predetermined initial value, and adjusting the value of the upper limit or the lower limit or both of them in response to outputs of the plurality of thickness sensors. The method of claim 9, further comprising: 11. A plurality of actuators arranged in a casting process and a casting amount controlled by the plurality of actuators for controlling a casting amount of a resin as a raw material of a film. A plurality of thickness sensors arranged corresponding to the plurality of actuators in the width direction of the film for detecting the thickness of the film generated by the resin, A film thickness control program for controlling an actuator to operate a computer so as to make the film thickness uniform in the width direction of the film, wherein the plurality of thickness sensors measure the thickness of each film. And a routine for controlling the computer to obtain the thickness measured by each of the plurality of thickness sensors. Therefore, so that the thickness of the film becomes a predetermined target value, a routine for controlling the computer to calculate the operation amount of the corresponding actuator, respectively, the operation amount of the plurality of actuators, the film A routine for performing a wavelet analysis on the waveform of the distribution in the width direction of the film to extract the wavelet component of the waveform of the distribution of the actuator manipulated variable in the width direction of the film, and using the wavelet component of the plurality of actuators. A film thickness control program, comprising a routine for correcting an operation amount and controlling a resin casting amount by the plurality of actuators according to the corrected operation amount. 12. The wavelet analysis is Haar
The program according to claim 11, wherein the function is a wavelet generating function. 13. The program according to claim 11, wherein the wavelet analysis is a wavelet analysis for a first stage wavelet component. 14. The computer is controlled so as to correct the wavelet component so that the value of the wavelet component is within a predetermined upper limit and lower limit prior to the operation of a routine for controlling the casting amount of the resin. The program according to any one of claims 11 to 13, further comprising a routine that: 15. A routine for setting the upper limit, the lower limit, or both to a predetermined initial value, and adjusting the upper limit, the lower limit, or both of them in response to the outputs of the plurality of thickness sensors. 15. The program according to claim 14, further comprising a routine. 16. A plurality of actuators arranged in a casting process and a casting amount controlled by the plurality of actuators for controlling a casting amount of a resin as a raw material of a film. A plurality of thickness sensors arranged corresponding to the plurality of actuators in the width direction of the film for detecting the thickness of the film generated by the resin, and the thickness sensors respectively measure the thickness. The operation amount calculation means for calculating the operation amounts of the corresponding actuators so that the thickness of the film becomes a predetermined target value according to the thickness of the film, and the operation amounts of the plurality of actuators. Wavelet analysis is performed on the waveform of the distribution in the width direction of the film to calculate the amount of actuator operation in the width direction of the film. Wavelet analysis means for extracting the wavelet component of the waveform of, the operation amount of the plurality of actuators is corrected using the wavelet component, the casting amount of the resin by the plurality of actuators by the corrected operation amount A film thickness control device including a manipulated variable correcting means for controlling. 17. The wavelet analysis is Haar
The apparatus according to claim 16, wherein the function is a wavelet generating function. 18. The apparatus according to claim 16, wherein the wavelet analysis is a wavelet analysis for a first stage wavelet component. 19. The method according to claim 16, further comprising means for correcting the wavelet component so that the value of the wavelet component input to the manipulated variable correcting means is within a predetermined upper limit and lower limit. The device according to any one of 1. 20. Means for setting the upper limit, the lower limit, or both to a predetermined initial value; and the upper limit, the lower limit, or both values in response to the outputs of the plurality of thickness sensors. 20. The device of claim 19, further comprising means for adjusting.