JPH0637759B2 - Profile control method for thickness and absolute dry basis weight - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a method for controlling a thickness profile, and particularly to a thickness profile control method suitable for controlling a thickness profile of paper. It is about.

<Prior Art> In a paper machine, the breadth-wise profile of the paper, which is a product, in the width direction is measured using a B / M meter or the like, and the manipulated variable is calculated from the deviation between the measured value and the target value. The profile of the opening of the slice lip, which is the pulp outlet, is controlled. That is, the absolute dry basis weight vector , The operation amount vector of the opening of the slice lip , And the interference matrix is A, Is established. Here, if there is a slice lip and the opening degree is changed, there is an interfering property that the absolute dry basis weight around the slice lip is changed. Therefore, the interference matrix A is usually a band diagonal matrix as shown below.

From this relationship, the fluctuation amount of the operation amount of the slice lip opening is calculated. , The difference between the target value of the absolute dry basis weight and the measured value Then, G _C (S): Transfer function of PI controller M: Partition coefficient matrix for decoupling is obtained and the manipulated variable is calculated from this equation. For M, the inverse matrix of the interference matrix A is usually used.

FIG. 6 shows the configuration of a control device used for profile control of such an absolutely dry basis weight. In FIG. 6, the deviation calculation block 2 receives the output of the process 1, that is, the measured value of the absolute dry basis weight, averages the basis weights near the measurement points corresponding to the respective operation points, and calculates the basis weights of all the measurement points. The difference from the average value of is calculated, and this value is output to the PI controller 3 as a deviation. PI
The controller 3 calculates the manipulated variable from this deviation and outputs it to the decoupling block 4. The decoupling block 4 calculates the distribution coefficient matrix M of the equation (2), and outputs the final manipulated variable to the process 1.

<Problems to be Solved by the Invention> However, in the profile control method of such a thickness, the interference matrix A is treated as a deterministic one, but in reality, the process varies and the operating point of the slice lip and the measurement of the basis weight are measured. The positional correspondence with points may shift. When such a deviation occurs, there is a problem that control becomes unstable.

<Object of the Invention> An object of the present invention is to provide a thickness profile control method capable of realizing stable control by applying fuzzy control.

<Means for Solving the Problems> The present invention that achieves such an object is to measure the thickness of the sheet in the width direction and the profile of the absolute dry basis weight, and to determine the opening of the slice lip based on the measured profile. When controlling the profile and operating the i-th slice lip, the i-th actual measurement point (M _i ) in the width direction defined corresponding to the i-th slice lip
And the position correspondence correction value (Δgap _i ) based on the distance (gap _i ) from the reference measurement point (u _i ) at which the influence of the operation of the i-th slice lip appears to the maximum. In the profile control method of thickness and absolute dry basis weight, which is calculated by calculating a position-corresponding correction signal, and correcting the position-correspondence relationship between the operating point and the measurement point by the position-corresponding correction signal to perform control,
The following steps are provided.

In the first step, the deviation of the profile measurement value of the thickness or the absolute dry basis weight before and after the operation is corresponding to the operation amount difference (Δu _{i, k} ) of the opening degree of the i-th slice lip. The minute (Δe _{i, k} ) is calculated.

The second step is the so-called fuzzy operation,
Sign of change in deviation of the th measurement value, sign of operation amount difference of opening of the i th or i + 1 th or i−1 th slice lip adjacent thereto, the actual measurement point and reference A plurality of membership functions defined in accordance with the distance to the measurement point and the position-corresponding correction value, and the change amount of the difference between the manipulated variable and the profile measurement value obtained in the first step and the reality. Calculate the truth value of the membership value by calculating the membership value according to the distance between the measurement point of and the reference measurement point, apply each membership function to the truth value of this membership value, and add the fuzzy set Then, the position correction signal is calculated by obtaining the center of gravity.

<Embodiment> FIG. 1 shows the configuration of an apparatus for carrying out an embodiment of a thickness profile control method according to the present invention. In FIG. 1, 10 is a position correspondence correction block, and process 1
The input value of the operation amount and the output value of the absolute dry basis weight are input, and the fuzzy rule is used to infer a corrected value of the positional correspondence between the absolute dry basis weight measuring point and the slice lip opening operating point. And outputs it to the deviation calculation block 12. A fuzzy controller 11 receives the deviation calculated by the deviation calculation block 12, calculates a manipulated variable using an appropriate fuzzy rule, and outputs it to the process 1. The measured value of the absolute dry basis weight from the process 1 is also input to the deviation calculation block 12.

Next, the operation of the position correspondence correction block 10 will be described. The subscript i is the position where the opening of the slice lip is operated,
k represents a time point. First, the variation Δe _{i, k} of the deviation from the input measurement value of the absolute dry basis weight is obtained from the following equations (3) and (4), and the difference Δu _{i, k} of the operation amount is obtained from the equation (5).

Δe _{i, k} = e _{i, k} −e _{i, k-1} (3) e _{i, k} = av (e _k ) −x _{i, k} (4) av (e _k ): at time k Average value of measured values of absolute dry basis weight x _{i, k} : Measured value of absolute dry basis weight at position i at time k Δu _{i, k} = u _{i, k} −u _i , k−1 (5) u _{i, k} : manipulated variable of slice lip opening at position i at time _k Using these values, fuzzy inference is executed by the following six rules and the membership function shown in FIG. Calculate the output value. FIG. 2 (A) shows Δe _i , Δ
e _i-1 , Δe _{i + 1} , (B) is Δu _i , (C) is gap _i ,
(D) represents the membership function for Δ gap _i . It should be noted that Δe _i , Δe _i−1 , and Δe _{i + 1} represent the difference between the manipulated variables at the position i obtained by the above equation (3) and on both sides thereof. Since the inference is performed at each time point, the subscript k is omitted. Further, gap _i is the current position shift, and Δ gap _i is the position correspondence correction value. Furthermore, P, N, and Z0 are positive, negative, and negative among the membership functions of FIGS. 2 (A) to (D), respectively.
Represents choosing a function of zeros.

if Δe _i is P and Δu _i is P and gap _i is Z0 th
en Δ gap _i is Z0 if Δe _{i + 1} is P and Δu _i is P and gap _i is Z0
then Δ gap _i is P if Δe _i-1 is P and Δu _i is P and gap _i is Z0
then Δ gap _i is N if Δe _i is N and Δu _i is N and gap _i is Z0 th
en Δ gap _i is Z0 if Δe _{i + 1} is N and Δu _i is N and gap _i is Z0
then Δ gap _i is P if Δe _{i + 1} is N and Δu _i is N and gap _i is Z0
then Δ gap _i is N Next, this rule will be described based on FIG. In FIG. 3, S _i is the operating point of the opening of the slice lip, u _i
Is the reference measurement point for this operating point S _i , M _i is the operating point S _i
Corresponding to _i : The actual measurement points, M _i−1 and M _{i + 1} are the measurement points on both sides thereof. The distance between M _i and u _i corresponds to gap. If the Δu _i is positive (opening the opening of the operating point S _i ) and Δe _{i + 1} is positive (the measured value at M _{j + 1} increases), the above rule applies to the measurement corresponding to S _i. This means that the points are closer to the M _{i + 1} side. The rule similarly indicates that as Δe _i-1 increases, the corresponding measurement point is closer to the M _i-1 side. The rule is a rule of the shift of the corresponding measurement point when the operation amount decreases. Further, as can be seen from FIG. 2 (C), “gap _i is Z0” plays a role of a limiter for gap _i whose absolute value is too large.

Next, the procedure of fuzzy inference is shown. The formula is Δe _i , Δ
u _i is a positive membership function, gap _i is a membership function of zero, and each membership value is calculated, and the truth value which is the minimum value of them is calculated as the zero membership function of FIG. 2 (D). Represents the application of. That is, the membership value a ₁ is obtained for the value e _{1 of} Δe _i as shown in FIG. 4 (A), and similarly, the values u ₁ of Δu _i and gap _i are obtained as in (B) and (C). , G _{1 the} membership values a ₂ , a ₃ are obtained. Next, as shown in (D), the minimum membership value a ₁ (truth value) is applied to the zero membership function of Δ gap _i to obtain the hatched area S ₁ . The formula applies Δe _{i + 1} , Δu _i is a positive membership function, gap _i is a membership function of zero, and the positive membership function is applied to the truth value of the calculated membership value. Δe _i−1 and Δu _i represent applying a positive membership function, and gap _i represents applying a negative membership function to the truth value of the membership value calculated by selecting a membership function of zero. Also, the formula is
Δe _i , Δu _i are negative membership functions, gap _i is a membership function of zero, and the membership function of zero is applied to the truth value of the calculated membership value.
The formula is Δe _{i + 1} , Δu _i is a negative membership function,
gap _i represents that the positive membership function is applied to the truth value of the membership value calculated by selecting the membership function of zero. Further, the equations are Δe _i−1 , Δ
u _i represents a negative membership function, and gap _i represents applying a negative membership function to the truth value of the membership value calculated by selecting a zero membership function. The calculation is the same as that described with reference to FIG. Next, in order to defuzzify, the six trapezoids thus obtained are combined and the fuzzy set is summed,
The center of gravity is obtained and this is designated as Δ gap _{i, k} . This Δ gap
_The position correction signal gap _{i, k} is calculated from _{i, k} by the following equation (6) and is output to the deviation calculation block 12.

gap _{i, k} = (1-α) (gap _{i, k-1} + Δ gap _{i, k} ) + α · gap _{i, k-1} ……
(6) α: parameter (0 <α <1) The deviation calculation block 12 calculates the average of the measured values of the peripheral measurement points around the position _{where the} gap _{i, k} is added to the reference measurement point u _i at the position i. As a measured value, the deviation from the average of all measured points is calculated and output to the fuzzy controller 11. The fuzzy controller 11 calculates an operation amount using an appropriate fuzzy rule and outputs it to the process 1. In addition to the above-mentioned methods, some methods have been proposed for the fuzzy inference method. The present invention is not limited to the inference method described above.

FIG. 5 shows a flowchart summarizing the operation of the position correspondence correction block 12. First, Δe _{i-1, k} , Δe _{i, k} , Δe _i + 1, k, and Δu _{i, k} are calculated by the equations (3) and (5) _{, and} are input to the position correspondence correction block 12 together with gap _1. . Next, the membership value is calculated from the rules of the above formulas and the membership function of FIG. Next, the minimum value of the membership value of the antecedent part of each rule is set as a truth value, and the fuzzy sets are summed by the method described above.
Next, defuzzify this fuzzy set to obtain Δ gap _{i, k}
And the position correction signal gap _{i, k} is calculated from the equation (6). Next, the deviation of the average value of the measurement values of the peripheral measurement points centered on the point obtained by adding the reference measurement point to this position correction signal is output. This operation is performed for operation amounts corresponding to all operation points.

Although the fuzzy controller 11 is used to calculate the operation amount in FIG. 1, another controller, for example, a PI controller may be used.

Further, the membership function is not limited to the one shown in FIG. 2 and can be appropriately selected according to the conditions. It may be non-linear.

Further, in this embodiment, the profile control of the absolute dry basis weight of the paper is described, but it can be applied to the thickness control of the film or iron.

<Effects of the Invention> As described above, according to the present invention, fuzzy calculation is performed using the change amount Δe of the deviation of the measured value of the operation amount Δu and the thickness to determine the positional correspondence between the operation point and the measurement point. Since the correction is performed, there is an effect that stable control can be realized for a long time even in a process in which the positional correspondence changes with time.

In addition, this fuzzy operation realizes adaptive control for the gap gap corresponding to the position by clarifying the adaptive law performed by the operator in the form of fuzzy rule. In practice, when performing adaptive control, what kind of variable should be focused on and what kind of adaptive law should be used for control is a problem. Here, since it is expressed in the form of a fuzzy rule by an adaptive law that is easy for an operator to recognize, there is also an effect that it is a human-friendly control method.

[Brief description of drawings]

FIG. 1 is a block diagram of an apparatus for realizing a thickness profile control method according to the present invention, FIG. 2 is a characteristic curve diagram showing an example of a membership function, and FIGS. 3 and 4 are fuzzy rules. FIG. 5 is a flow chart showing the operation of the position correspondence correction block, and FIG. 6 is a block diagram of a conventional thickness profile control device. 10 ... Position correction block, 11 ... Fuzzy controller, 12 ... Deviation calculation block.

Claims (1)

[Claims] 1. A profile of a thickness in a width direction of a sheet and an absolute dry basis weight are measured, a profile of an opening of a slice lip is controlled based on the measured profile, and an ith slice lip is operated. At this time, the width-direction i-th actual measurement point (M _i ) determined in correspondence with the i-th slice lip and the influence of the operation of the i-th slice lip are maximum with respect to the measurement value. The position correspondence correction value (Δgap _i ) is calculated based on the distance (gap _i ) from the reference measurement point (u _i ) appearing in (1) to obtain the position correspondence correction signal. In the profile control method of the thickness and the absolute dry basis weight for performing control by correcting the positional correspondence with the point, the operation amount difference (Δu of the opening degree of the i-th slice lip)
Corresponding to _{i, k} ), the amount of change (Δe) in the deviation of the profile measurement values of the thickness and the absolute dry basis weight before and after the operation.
_{i, k} ), the sign of the change in the deviation of the i-th measurement value, and the opening degree of the i-th or the (i + 1) th or (i-1) th slice lip adjacent to the i-th measurement value Of the operation amount difference, the distance between the actual measurement point and the reference measurement point, and a plurality of membership functions defined according to the position correspondence correction value, and the operation obtained in the first step The truth value of the membership value is calculated by calculating the truth value of the membership value by calculating the membership value according to the variation of the quantity difference and the deviation of the profile measurement value and the distance between the actual measurement point and the reference measurement point. The second step of calculating the position correction signal by calculating the center of gravity by applying each membership function to the value and calculating the sum of the fuzzy sets, and the thickness and the absolute dry basis weight are provided. Profile control method .