JPH042896A - Papermaking machine control unit and control method therefor - Google Patents

Abstract

PURPOSE:To provide the title control unit capable of ensuring sawtooth wave- shaped weighing profile to be flattened, made up of a position correspondence section, profile operation section, operation rule selection section, hypothetical slice operation section, smoothing slice determination section and output control section. CONSTITUTION:With the reference of a position correspondence section 34, based on the deviation profile from a profile operation section 32 to express the deviation between a measured profile and the target profile corresponding to slice, a hypothetical slice operation section 36 determines the distribution of the deviation profile by adding hypothetical slice to real slice. A nominee selection section 40 selects an operation rule which seems to be the most desirable from the slice and the deviation profile of the hypothetical slice close thereto through fuzzy inference, and a slice determination section 50 determines a slice for smoothing, and an output control section 60 controls the slice in accordance with the operation rule determined by the selection section.

Description

DETAILED DESCRIPTION OF THE INVENTION <Industrial Application Field> The present invention relates to a device for controlling a paper machine that produces paper, and particularly relates to an improvement suitable for producing paper of uniform quality.

<Prior Art> A control device for a paper machine is known, for example, from Japanese Patent Application Laid-Open No. 63-75197 proposed by the present applicant. FIG. 13 is a perspective view of the overall structure of the paper machine. In FIG. The paper is dried by passing through a drying part 15, the surface of the paper is glossed by passing through a calender 11, and the paper is wound onto a reel 12.

The head box 5 is provided with a slicing lip 6 for discharging the raw material liquid, and several dozen slicing bolts 7 for adjusting the opening degree of the slicing lip 6 at regular intervals in the width direction of the paper. The white water silo 3 receives and collects white water from the wire pat 3, which is returned to the head box 5 by a pump 4.

At the entrance of the pump 4, the raw material (seed) in the seed box 1 is injected by the seed port valve 2.

The sensor 14 that measures the basis weight and moisture content of paper is connected to the calendar 1.
It measures the paper width by reciprocating along a frame 13 installed between the reel 1 and the reel 12.
620,146 is used. This data in the paper width direction is called a profile.The controller 16 inputs the basis weight profile measured by the sensor 14 and adjusts the slice bolt 7 so that it has the target basis weight.
is operated to control the opening degree of the slice lip 6.

Regarding the moisture content, input the moisture content profile and control the steam amount in the dry part 15 so that the target moisture content is achieved. This is done in a controlled manner.

FIG. 14 is a functional block diagram of the paper machine control device. In FIG. The measurement profile signal is made to correspond to the slice to be actually operated by referring to the section 34, and the control deviation C(k) for each slice (k) is thereby determined. For the measurement profile, for example, the total number of measurement points should be 360 points in the width direction, and this should be averaged over 6 adjacent points to smooth it.The target profile may be, for example, to slightly increase the basis weight at the center, while increasing the basis weight at both ends. The amount is slightly reduced so that the reel 12 can perform stable winding. The position correspondence unit 34 describes the correspondence between the measurement points of the measurement profile signal and the slices, and the slice operation amount calculation unit 38 describes the correspondence between the slices and the measurement points of the measurement profile signal. , the amount of change in the manipulated variable △IJJ1k> is calculated, and the controller 1
This is the core part of 5. The output control unit 60 changes the manipulated variable Δ1J for each slice obtained by the slice manipulated variable calculation unit 38.
It adds I(k) to the previous manipulated variable "ul(k) and outputs it as the current manipulated variable 1 & (k). For example, it can be set to a natural number multiple of the sensor scanning period, or the time constant of the paper machine. For example, sensor 1 is determined based on the computing power of the controller.
If the scanning period of No. 4 is 30 seconds, the control period of the output control section 60 is determined to be every 2 minutes.

Incidentally, in many paper mills, the slicing operation amount calculation unit 38 is manually operated by a skilled operator.

FIG. 15 is a block diagram showing the configuration of the paper machine control system focusing on the controller 16, and explains the control state in detail. Here, in order to simplify the explanation, it is shown in a state corresponding to slices. In the controller 16, a target value (a set value, here expressed as y)
The manipulated variable is determined from the deviation e(to) from the fed-back basis weight profile y(to). The slice bolt 7 is adjusted by this operation amount, and the basis weight profile that changes as a result is fed back. And the deviation e(K) is 0
controlled so that it approaches

The manipulated variable in such a control system is the slice lip opening pattern (u, , , , unl), and the control amount is the basis weight profile (y, -0, yHl).

The plant can be approximated by first-order delay and interference in the width direction, and this can be expressed as the S domain in the Laplace transform as follows.

Y(S) = (k/ (1+Ts) AU(S)
(1) Here, Y(S) is the Laplace transform of vector V(t), and V(t) is the matrix [yl (t)+, , +
yH(t) ] of the transposed matrix ([3'1 (t),, 1., ys (t) ]
), is U(S) a vector? j/(t)
The Laplace transform of 1JI(t) is [u 1 (t)21
1. , u N(t) ], k is the gain, T
is a time constant, A is an interference matrix (NxN), and satisfies the following relationship.

And t is time.

The conventional control law combines computation that takes interference into account and PI (proportional/integral) control that takes first-order delay into consideration, and this can be expressed in discrete time and velocity form as follows.

ΔM(k) =K MΔe(k) +K Me (
k) (2)P 1 Here, k is discrete time, vl(k) is discrete time slice lip opening vector, △IJ/(k) is the difference, e (k
) is a discrete-time control deviation vector, e(k) is a control deviation, and Δt'(k) is a difference, each of which satisfies the following relationship.

Further, K is a proportional gain, K is an integral gain, and M is an NXN matrix considering interference.

Next, the operation will be explained. The arithmetic unit 22 performs calculations that take interference into account on the deviation (control deviation) e (k) between the target value and the controlled amount obtained at the addition point 21 .

This result is expressed by the following equation.

■ e'(k)−= [e −(k), ,,, e −
N (k) In this case, the following equation holds between e'(k) and e(k).

e'(k)=Me(k)
(3) The controller 25 has t”(k) and Δt”(
k) is input. The difference △t'' (k) is obtained from the addition point 24 by the difference between the output C (k-1') of the arithmetic unit 22 and the previous output value C'' (t1) of the arithmetic unit via the buffer 23. It is expressed by the following formula.

Δe'(k)=t''(k)-4'-(k-1) Then, the controller 25 performs a PI calculation as shown in the following equation.

△tJI(k)=K Δt”lk)+x, t”(k)
(4) This operation can also be expressed as a scalar operation for each element of the vector as follows.

Δu(k)−K Δe−(k)+K e
(k)l P I
1 1 (1=1~N) (
5) The adder 26 adds the output △ from the controller 25 to the previous slice lip opening vector held by itself.
v1(k) is added and output as the current slice lip opening vector vI(k). This slice lip opening vector 1#fk) is applied to the plant 28 via the zero-order hold circuit 27 as a manipulated variable 1j7(t).

The output (basis weight profile) y(t) of the plant 28 is taken out via a sampler 29 that performs sampling every time, and is fed back to the summing point 21.

<Problem to be solved by the invention> By the way, a plant is ideally expressed by equation (1), but in reality it contains complex nonlinear characteristics, so (2)
The problem was that linear control such as the one shown in the equation is insufficient.

(2) FIG. 16 is an explanatory diagram of the inter-slice sawtooth wave, with (A) showing before the operation and (B) after the operation.

In the figure, the horizontal line represents the width direction of the paper, the curve represents the basis weight profile at the relevant position, and the vertical line (+) represents the position of the measurement edge where the influence of the relevant slice is most pronounced (referred to as positional correspondence). , and the arrow (↓) represents the operation of the corresponding slice. In order to eliminate the sawtooth wave between slices, two slices are placed in between and the slices on both sides are manipulated (Fig. 16A). However, a valley in the basis weight occurs at the position of the manipulated slice, and the sawtooth shape is instead created. In some cases, the waves were increased (Figure 16B).

■ Due to the complexity of the process, the response of the basis weight profile to the slicing operation often does not appear as ideal.
In particular, if many slices are operated at once, the basis weight profile may be disturbed and the assistance of a skilled operator may be required.

■ When the process is highly disturbed, such as when a major brand change is made or when the paper machine starts operating, it takes a very long time to converge the profile to a quality that can be recognized as a product, resulting in a huge amount of waste paper. become.

The first object of the present invention to solve the above-mentioned problems is to provide a paper machine control system that can reliably flatten the basis weight profile of sawtooth waves by adopting a fuzzy control method that incorporates the manual operation method of a skilled operator. The second aspect of the present invention is to provide an apparatus.
The purpose of this is to provide an apparatus that can converge the profile to the product level in a short time even when the process is greatly disturbed.

Means for Solving the Problems> The present invention achieves the above objects by using a plurality of slices (slices) arranged in the width direction of the paper for controlling the supply amount of raw materials for the paper.
6, 7) and a sensor (14) that measures the quality of the paper produced in the width direction of the paper, the value detected by this sensor matches a predetermined target profile. A paper machine control device for operating the slice in the following direction has the following configuration.

That is, a position correspondence unit (34) that determines which of the slices the measurement point of the sensor corresponds to, and a deviation between the detected basis weight profile and the target profile, according to the correspondence of the position correspondence unit, A profile calculation unit (32) that calculates the slices and the position between the slices! a virtual slice calculation unit (36) that calculates a deviation between the detected basis weight profile and the target profile with respect to a virtual slice provided in the area; and at least one or more rules for manipulating slices, the profile calculation unit and the virtual slice calculation unit; Using the deviation of each slice calculated in and the virtual slice adjacent to it, calculate the degree of compatibility with the operation rule for each of these rules, and then calculate one of the operation rules based on the calculated degree of compatibility. A candidate selection unit (40) extracts one slice for each slice, and compares the degree of suitability for each slice obtained by this candidate selection unit with a predetermined threshold value, and selects the slice as an operation target according to the wrinkle smoothing standard. The slice determination unit (50) determines a slice, and the output control unit (60) sends a control amount to the slice determined by the slice determination unit in accordance with the operation rule selected by the candidate selection unit. It is characterized by:

Functions Each component of the present invention has the following functions. The profile calculation unit refers to the position correspondence unit and represents the deviation between the profile measured by the sensor and the target profile in correspondence with slices. The virtual slice calculation unit virtualizes a slice that is finer than the actual slice, and calculates the distribution state of the deviation profile together with the actual slice.

The candidate selection unit uses so-called fuzzy inference to propose the most desirable operation rule from the deviation profiles of the slice and neighboring virtual slices. The slice determining unit sequentially selects end slices from the starting slice using the wrinkle smoothing standard, and gradually reduces the size of the sawtooth wrinkles like smoothing the wrinkles of paper.

The output control section controls and outputs only the slices determined by the slice determination section according to the operation rules determined by the candidate selection section.

<Example> The present invention will be explained below using the drawings.

(Embodiment 1) FIG. 1 is a configuration block diagram showing an embodiment of the present invention. The paper machine has a plurality of slices arranged in the width direction of the paper for controlling the supply amount of paper raw materials, and a sensor 14 for measuring the quality of the manufactured paper in the width direction of the paper. A device such as a dryer (not shown) is also provided. The slice specifically refers to the slice bolt 7 that is the object of control, but may also include the slice lip 6 that actually discharges the raw material.The slices are provided in, for example, 37 m rows in the width direction. The sensor 14 scans in the width direction of the paper and has, for example, 360 measurement points.

The paper machine control device operates the slice in a direction in which the value detected by the sensor 14 matches the target profile, and includes the following components. The profile calculation section 32 and the position correspondence section 34 have been described above, and a description of the same functions will be omitted.

The profile calculation unit 32 further correlates the data of the measurement points for each slice with positional correspondence, and for example, simply averages the data of five consecutive measurement points to determine the deviation of the slice. The virtual slice calculation unit 36 calculates the slice 12
A virtual slice is provided at a position between the two, and the deviation between the detected basis weight and the target basis weight in this virtual slice is calculated. One virtual slice may be provided between each slice, or two or more virtual slices may be provided between each slice. This virtual slice is, for example, 1
In the case of points, it is called a half slice, and it obtains more detailed information than the original slice, for example, measurement points that were not used in the slice (in this example, there are 9 measurement points per slice on average) ) is obtained by simply averaging 3 to 5 pieces of data.

The candidate selection unit 40 provides guidelines on how to perform operations on each slice using fuzzy calculations. The operation rule section 41 stores rules for manipulating slices, and the number of rules may be one or more.The evaluation function section 42 stores rules for manipulating slices.
An evaluation function is stored that calculates the degree of fitness (B) using the values of each slice and the virtual slices adjacent thereto in accordance with the rule. Details of these operation rule section 41 and evaluation function section 42 will be explained later. Operation rule inference unit 43
extracts one of the operation rules held by the operation rule section 41 based on the fitness of the evaluation function section 42. For example, Mandani's method used in the field of fuzzy control is adopted, and the calculation method uses the centroid method. or height method is used. If there is a single operation rule, this rule is used.

The slice determining unit 50 compares the degree of suitability for each slice obtained by the candidate selecting unit 40 with a predetermined threshold, and
This determines the slice to be operated this time. The slice selection unit 51 performs actual determination work, with reference to the wrinkle smoothing standard 52. Wrinkle smoothing standard 5
2, in principle, selects only the slices in the unfinished area in a certain direction, and performs an operation similar to the work that T-reverse does to smooth out the wrinkles of paper, that is, for wrinkled cloth, the wrinkles are sequentially moved from the center to the edges. This enables the work of stretching and obtaining a flat cloth. The slice counter 53 distinguishes between an operated area and an unoperated area, and stores the boundary value thereof. This boundary value should be included in the unoperated area.The reason for this is that if it is an operated area, this slice will be excluded from the operation target in the next control cycle, but if the operation is performed only - times, the basis weight or the target value will not be included. This is because there may be instances where the desired value is not reached and the operation needs to be repeated. Details of the wrinkle smoothing standard 53 will be explained later. The output control unit 60 uses the operation rule determined by the operation rule inference unit 41 to determine the slice determination unit 50.
For the slice selected in , a control amount corresponding to the deviation between the detected basis weight and the target basis weight is output.

FIG. 2 is a flowchart explaining the operation flow of the device shown in FIG. 1. In FIG. With respect to the detected basis weight profile, it is determined which of the slices the position corresponds to with reference to the position correspondence unit 34, and the position in the virtual slice provided at the position corresponding to this position and the position between these slices is determined. The deviation between the detected basis weight and the target basis weight is calculated (S12).

Next, the evaluation function unit 42 calculates the degree of fitness (B) using the deviation determined in step 312 according to the slice operation rule predetermined by the operation rule unit 41 for each slice (S13), and If there are multiple operation rules in the rule unit 41, the operation rule inference unit 43 compares the fitness degree B for each rule, and selects the rule with the slice that gives the largest value as the rule for performing the current operation. (S14).

In the wrinkle smoothing norm 52, the value of the degree of conformity B and the differential threshold value V
1 is used to determine which of the first norm, which is used for a stable state, or the second norm, which is sufficient to maintain the current slice state, is to be adopted in the current control cycle. First, the operation rule inference unit 43 selects an operation rule that is suitable for the state of each slice.
Fine adjustment threshold value 71 of the degree of conformance to this operation rule
It is determined whether the above conditions exist (S50).

First, the first norm is used because there is a slice whose fitness is larger than the fine-tuning threshold in S50 (B>Vl) (
S51), and if it is determined that this slice is located closer to the unfinished operation area than the value indicated by the slice counter 58 (S52), the slice is selected (353). When the number of slices to be operated is limited, select the slices located on the unoperated side from the boundary value of the slice counter 58 up to the limit number, and do not select any more slices this time. A control amount corresponding to the deviation between the detected basis weight and the target basis weight is sent to the slice determined in S14 according to the rule determined in S14 (S54). Next, this manipulated slice is written into the slice counter 58 (S55), and the process moves to the next control cycle.

Next, the second norm is used when it is determined in S50 that the fitness B is smaller than the fine-tuning threshold for all slices (Vl≧B), or when there is a slice that is larger than the fine-tuning threshold ( B>Vl) (351), and if it is determined that this slice exists only on the operated area side of the value indicated by the slice counter 58 (S52), the slice to be selected is set to none and the slice counter 5 is
8 (552) and move on to the next control cycle.

Next, regarding the main components of the device shown in Figure 1,
Further details will be explained.

(1) Position correspondence and virtual slices FIG. 3 is a diagram clearly displaying the position correspondence between the basis weight profile detected by the sensor 14 and slices, and FIG. 4 shows the slices and virtual slices displayed by the virtual slice calculation unit 36. There is a deviation display diagram. In the figure, white circles O represent measurement points, and the basis weight profile of the sawtooth wave is a wave with a wavelength as wide as two slices. Then, according to signal processing theory, to remove this wavelength component, it is necessary to sample even darker blue data. Here, one virtual slice (half slice) is provided between each slice to cope with the problem. In the figure, measurement points 1 to 9 exist between slices A and B, measurement points 2 to 0 exist on the left side of slice A, and measurement points 10 to 18 exist on the right side of slice B. . For the deviation of slice A, use the simple average of the data of 5 measurement points -1 to 3, and for the deviation of slice B, use the data of measurement points 7 to 11.
A simple average of the data is used. For the deviation of the virtual slice H3 provided between the slices A and B, a simple average of data from five measurement points 3 to 7 is used, for example.

(11) Operation rules FIG. 5 is a diagram showing an example of the rules of the operation rule section 41.
(A) shows - slices, (B) shows two slices in the same direction, and (C) shows two slices in opposite directions. In the figure, Sl, 32 represents a slice, H3I, H32°H83 represents a virtual slice, and the arrow ↓ represents the operating direction of the slice. Rule 1
(FIG. 5A) is an effective operation method when there is a deviation only in the negative slice and there is no deviation in the virtual slices on both sides, and only that slice is operated in the direction of decreasing the deviation. Rule 2 (FIG. 5B) is that there is a deviation in the middle virtual slice H32, and the slices Sl and S2 on both sides and the virtual slice H
This is an effective method when there is no deviation in 3I and H33 (when there are peaks or valleys between slices).
are operated in the same direction to reduce the deviation of the intermediate virtual slice H82. Rule 3 (Fig. 5C) is an effective method when the sign of the deviation changes in the slices Sl and S2 on both sides (when the peaks and valleys are shifted from the slices). Manipulating S2 in the opposite direction reduces the deviation of each slice Sl, S2 and virtual slice H5I~3.

Note that although a rule for operating up to two slices at the same time is considered here, a rule for operating three or more slices at the same time may be provided.

In addition, although the number of operation rules is 3, a single
There is no problem even if two or four or more are provided.

(To) Calculation of goodness of fit FIG. 6 is a diagram showing an example of the evaluation function section 42, and here, the one for rule 1 is shown. In FIG. 6, (A) is the evaluation function 71 of the virtual slice H3I in FIG. 5A. (B)
is the evaluation function 72 of slice S1 in FIG. 5B. (C) shows the evaluation function/3 of the virtual slice H82 in FIG. 5C. To explain the evaluation function /1, the reason why it is set to zero when the deviation is less than 11 is to prevent the slice from being manipulated in this direction. ), the deviation is 1.
The reason why we set a constant value above 5 is because it is rare to take such a value under normal operating conditions (therefore, there is no problem in maintaining the upward trend), so we set the value to 1 when the deviation is 0. The reason for this is that multiplication has a neutral effect on calculation. If the deviation value is in the interval (-1, 0), set the value smaller than 1,
The slicing operation is performed by reducing the fitness level B.

On the other hand, when the value of the deviation is in the interval (0, 1, 5), the value is set larger than 1, and the degree of adaptation B is increased to perform the slicing operation. Evaluation function 72. The same applies to /3, and the value is determined empirically. The goodness of fit B is determined by the following formula.

B=/1x/2x/3 (7
) FIG. 7 is an explanatory diagram of an example of calculating the fitness degree B according to the rules of FIG. 5. In the figure, the deviation χ21-2 in the virtual slice H3I is 0.5, the deviation χ2I-1 in the slice S1 is 1.0, and the deviation χ21-2 in the virtual slice H3I is 0.5.
The deviation χ21 at 32 is -0.5.

Then, from FIG. 5, the evaluation function/1(χ2i-2) is 1.
5, the number of evaluation intervals/2 (χ ) is 4.7, and the evaluation function/3 (χ21) is 0.5, so the goodness of fit B is (7
) is 3.52. Similar calculations are made for other rules, and the operation rule inference unit 43 selects the rule with the highest degree of suitability.

(P/) Slice Counter First, a case where all slices are handled in one zone will be explained. The slice counter 53 is located at the center of the profile measured by the sensor 14 (position of 180 points if there are 360 points).
towards both ends (if there are 37 slices, 1
This indicates the slice position (there are two values, one near the main opening and one near the 37th paper). Normally, the slicing operation in a paper machine is performed in a batch manner, so if you perform slicing this time, the effect will be on the sensor 14 position. After stabilizing, perform the next slicing operation. Therefore, the value of the outermost slice among the slices operated last time is recorded in the slice counter 53 as a boundary value.

Next, a case will be described in which all slices are divided into several zones and handled. A slice counter 53 and wrinkle smoothing standard 52 are used for each zone. Then, slice positions are shown in the direction of both ends with the center of the profile measured by the sensor 14 of the corresponding zone as a reference. Here, the edge refers to the nearest slice of the slice in the adjacent zone. Division into zones in this manner has the effect of converging quickly to the target profile while stably controlling the profile.

M Output control unit The control amount output by the output control unit 60 is one of +A (up), O, and -A (down) in the case of a digital type, and the manipulated variable A is specified as a discrete value. do. When performing a simple operation, stable control can be achieved by setting the -time slice operation amount to a constant value (for example, 0.51). In the case of an analog type, it is preferable to change the manipulated variable in an analog manner according to the amount of deviation.

(Embodiment 2) Next, a control device that relaxes the restrictions on slice selection in the wrinkle smoothing standard 52 to reach a target profile more quickly will be described. FIG. 8 is a flowchart illustrating a second embodiment of the present invention. To explain the comparison with the embodiment shown in Fig. 2,
In addition to the fine translation threshold value v1, the rough threshold value vo is added to the threshold value.
has been added. Note that the configuration block diagram is the same as that in FIG. 1, so this is applied mutatis mutandis. In this wrinkle smoothing standard 52,
The following cases (1) to (to) are divided using the goodness of fit.

(1) When the operation rule inference unit 43 determines that there is a slice whose fitness degree B is larger than the rough threshold value Vo (
B>Vo) selects the relevant slice. The slice counter 53 is left unchanged.

(11) The operation rule inference unit 43 determines that there is a slice whose fitness level B is smaller than the rough threshold value Vo and larger than the fine threshold value v1 (V o ≧ B > V l), and this slice is detected by the slice counter 53 . When it is determined that the slice exists outside the indicated value, the slice is selected and the operated slice is written in the slice counter 53.

@ When the operation rule inference unit 43 determines that the fitness degree B is smaller than the fine adjustment threshold v1 in all slices (
Vl≧B), there is no slice to be selected and the value of the slice counter 53 is reset. In addition, there is a slice whose fitness B is smaller than the rough threshold ff1vO and larger than the fine threshold v1 (Vo≧B>Vl
), and also when it is determined that this slice exists only inside the value indicated by the slice counter, the slice to be selected is set to none and the slice counter 5 is
Reset the value of 3.

Next, explain the flowchart in Figure 8. 0, tzu, and bare hands execute Sll-S14 as explained in Figure 2. 6 Next, use the value of fitness B to determine the stable state. It is determined which of the first norm, the second norm, which is sufficient to maintain the current slice state, or the third norm, which is used when an abnormal profile occurs, is to be adopted in the current control cycle. First, the operation rule inference unit 43 selects an operation rule that is suitable for the state of each slice, and calculates the difference between the maximum value B max of the degree of conformance B for this operation rule and the rough threshold value ■0 or fine threshold value v1. It is determined which norm to apply based on the size relationship (S50).

First, the third norm is used when it is determined in S50 that there is a slice whose fitness is greater than the rough threshold (BIlax>Vo), and the slice is selected (S
41) When the number of slices to be operated is limited, select the slices from the boundary value of the slice counter 53 up to the limit number of slices located on the unoperated side, and do not select any more slices this time. A control amount corresponding to the deviation between the detected basis weight and the target basis weight is sent to the selected slice according to the rule determined in S15 (342
), the current control cycle is completed.

Next, the first norm is used because it is not adopted as the third norm in S50 and there is a slice whose fitness is smaller than the rough threshold and larger than the fine threshold (V o
≧Blax>vl) (S51), and if it is determined that this slice is located closer to the unfinished operation area than the value indicated by the slice counter 53 (S52), the same processing as in the case of FIG. 2 is performed.

Third, when using the second norm, the third norm or the first norm is used.
If none of the norms apply, the degree of conformance is B at S50.
is smaller than the fine threshold in all slices (V1≧B) IIa× (S50), or there is a slice that is smaller than the rough threshold and larger than the fine threshold (V).

≧B1. ax>vl) (S51) and it is determined that this slice exists only inside the value indicated by the slice counter (352), the slice to be selected is set to none and the value of the slice counter 53 is reset. 352), move on to the next control cycle.

(Third Embodiment) As in the second embodiment, restrictions on slice selection are relaxed to allow faster convergence to the target profile. Example 2
In the above, the relationship between the rough threshold value VO and the maximum fitness degree B 11ax was looked at, but the same function is realized using the deviation profile obtained by the profile calculation unit 32.

FIG. 9 is a block diagram showing a third embodiment of the present invention. In comparison with FIG. 1, in FIG. 9, the slice determination unit 5
Since only 0 is different, this part will be explained and the others will be omitted.

In the figure, the stability determination unit 55 is configured to include the profile calculation unit 32
The calculation result is used to judge whether the state of the process is stable or disordered. For example, the maximum value of deviation or the sum of squares of deviation is used as an index. Mode setting section 56
inputs the status signal of the stability determination unit 51, selects the coarse adjustment mode at the start of process operation, and automatically switches to the fine adjustment mode when the process becomes stable. 11j! The slice selection unit 57 is used in the coarse adjustment mode, and when it is recognized that there is a slice whose fitness degree B is larger than the coarse adjustment threshold value V2, it selects the slice. Fine adjustment slice selection section 51a
is used in the fine adjustment mode, and operates almost the same as the slice selection section 51 in FIG. 1. That is, when it is determined that there is a slice larger than a predetermined threshold value, the slice is selected according to the wrinkle smoothing rule 52. When the slice is divided into several regions by the zone dividing section 54, the wrinkle smoothing norm 52 is applied to each region.

The output control unit 60 controls the coarse slice selection unit 57 in the coarse adjustment mode, and the fine slice selection unit 51a in the fine adjustment mode.
Outputs the manipulated variable to the slice according to the decision. Preferably, the amount of operation in the coarse adjustment mode is set to be larger than that in the fine adjustment mode. In this case, with the digital method, the manipulated variable A is determined separately for coarse adjustment mode and fine adjustment mode, and with the analog method, the calculated manipulated variable in fine adjustment mode is multiplied by a coarse adjustment mode coefficient larger than 1 to perform coarse adjustment. For example, if the operation amount in the coarse adjustment mode is twice that in the fine adjustment mode, the profile converges quickly.

The operation of the device configured in this manner will be described next. 1st
FIG. 0 is a flowchart of the third embodiment. First, steps 811 to S14 are executed in the same way as in FIG. Next, the mode setting section 56
It is determined whether the current control mode is the coarse adjustment mode or the fine adjustment mode (S20). If the current control mode is the coarse adjustment mode, the coarse adjustment slice selection unit 57 selects a value larger than the coarse adjustment threshold value ■2. An unlimited number of slices with degree B are selected (S21
).

Then, the output control unit 60 operates the slice determined in S21 according to the operation rule determined by the operation rule inference unit 43 (S22).

If it is the fine adjustment mode, the fine adjustment slice selection unit 51a determines a slice to be a candidate for operation according to the wrinkle smoothing standard 52, and further confirms that the slice exists in an area where the operation is permitted using the slice counter 53 (331); Then, the output control unit 60 operates the slice determined in S31 according to the operation rule determined by the operation rule inference unit 43 (
S32). Next, the process moves to the next control cycle.

(Embodiment 4) FIG. 11 is a configuration block diagram of a fourth embodiment of the present invention. A comparison with the third embodiment (FIG. 9) is that the rough slice selection section 57 refers to the wrinkle smoothing standard 52 and the zone division section 54. Coarse slice selection section 57
In this case, the zone division unit 54 divides the slices into eight slices, for example, and increases the number of slices that can be operated in the negative times to speed up the convergence of the profile. The amount of slice operation is also larger compared to the fine adjustment mode. On the other hand, in the fine adjustment slice selection section 51a, the zone division section 54 divides into two slices, for example, to reduce the number of slices that can be operated in the negative times, and thereby maintain the profile stably.

FIG. 12 is a block diagram of the configuration of the present invention, focusing on control functions. In addition, in FIG. 12, the 15th
Components having the same functions as those in the drawings are designated by the same reference numerals, and explanations thereof will be omitted. The virtual slice calculation unit 36 calculates the control deviation of the virtual slice, and sends it to the candidate selection unit 40 together with the control deviation of the slice calculated by the calculation unit 22. The candidate selection unit 40 uses fuzzy calculations to select an operation rule according to the control deviation of each slice. The slice determining unit 50 is the candidate selecting unit 4
The suitability used in determining the operation rule with 0 is compared with a predetermined threshold value to determine the slice to be actually operated in the current control cycle. Since profiles with a low degree of fitness are considered to be in a stable state, we will ignore the slicing operation this time. The criterion for determining this slice is the wrinkle smoothing criterion 52, which sends the result to the adder 26 and the zero-order hold circuit 27.
The necessary slicing operations are accomplished via .

<Effects of the Invention> As explained above, the present invention has the following effects.

■ Since a virtual slice is provided to understand the operating state in more detail than the slice deviation, slicing operations that increase sawtooth waves can be avoided.

(2) If a plurality of operation rules are provided as in the embodiment and a slicing operation is selected according to the waveform of the sawtooth wave in the relevant slice, a slice operation that removes the sawtooth wave can be performed.

■ As in Examples 3 and 4, 'Ili adjustment mode and fine adjustment mode are provided, and when the process is unstable, the coarse adjustment mode quickly converges the profile, and when the process is stabilized, the fine adjustment mode allows the profile to be more precisely adjusted. By controlling this, you can reduce the amount of paper waste.

■ As in the embodiment, if a slice counter 53 is provided and the slice is gradually operated from the slice on the operated area side to the slice on the unoperated area side, a situation where the profile diverges during control can be prevented. .

(2) If the number of slices to be operated is limited to 1 or 2 in the first wrinkle smoothing criterion as in the embodiment, the controllability of the profile becomes stable.

[Brief explanation of drawings]

FIG. 1 is a configuration block diagram showing one embodiment of the present invention, and FIG.
The figure is a flowchart explaining the operation flow of the device in figure 1,
The figure shows the basis weight profile detected by the sensor 14 clearly showing the positional correspondence with the slice, FIG. 4 shows the slice displayed by the virtual slice calculation unit 36 and the deviation display of the virtual slice, and FIG. 5 shows the operation. FIG. 6 is a diagram showing an example of the rules of the rule section 41, FIG. 6 is a diagram showing an example of the evaluation function section 42, and FIG. 7 is an explanatory diagram of an example of calculating the fitness B according to the rules of FIG. FIG. 8 is a flowchart explaining the second embodiment of the present invention, FIGS. 9 to 10 are explanatory diagrams of the third embodiment of the present invention, and FIG. 11 is an explanatory diagram of the fourth embodiment of the present invention. FIG. 12 is a block diagram of the configuration of the present invention, focusing on control. Fig. 13 is a perspective view of the overall configuration of the paper machine, Fig. 14 is a block diagram of the configuration of the paper machine control device, and Fig. 15 is the controller 1.
6. Configuration block diagram of paper machine control device focusing on No. 6, No. 16
The figure is an explanatory diagram of inter-slice sawtooth waves. 6.7...Slice, 14...Sensor, 32...
Profile calculation section 32. 34... position correspondence section, 36.
... Virtual slice calculation section, 40 ... Candidate selection section, 50
...Slice determining section-152...Wrinkle smoothing standard, 5
3... Sura figure figure figure (Al (B) figure δ figure O

Claims (11)

[Claims] (1) For a paper machine having a plurality of slices arranged in the width direction of the paper for controlling the supply amount of paper raw materials, and a sensor for measuring the quality of the manufactured paper in the width direction of the paper,
In a paper machine control device that operates the slice in a direction in which a value detected by the sensor corresponds to a predetermined target profile, a position correspondence unit determines to which slice of the slice a measurement point of the sensor corresponds. and a profile calculation unit that calculates the deviation between the detected basis weight profile and the target profile for the slice according to the correspondence of the position correspondence unit, and the detected basis weight for the virtual slice provided at a position between the slices. It has a virtual slice calculation section that calculates the deviation between the profile and the target profile, and at least one rule for manipulating the slices, and each slice calculated by the profile calculation section and the virtual slice calculation section and the virtual slice adjacent thereto. a candidate selection unit that calculates the degree of compatibility of each of these rules with the operation rule using the deviation of the operation rule, and extracts one of the operation rules for each slice based on the calculated degree of compatibility; a slice determining unit that compares the suitability of each slice obtained by the candidate selecting unit with a predetermined threshold value and determines a slice to be operated according to the wrinkle smoothing standard; A paper machine control device comprising: an output control section that sends a control amount to a slice according to an operation rule selected by a candidate selection section for the slice. (2) The operation rules of the candidate selection section include a first rule that operates one slice, and a second rule that operates the slice and the slice on one side adjacent to it in the same direction.
2. The paper machine control device according to claim 1, further comprising a third rule for operating the slice and the adjacent slice on one side in different directions. (3) The operation rules of the candidate selection unit include a first rule that operates one slice, and a second rule that operates the slice and the slice on one side adjacent to it in the same direction.
and a third rule that operates the slice and the slice on one side adjacent to it in different directions, and the calculation of the fitness is performed for the first rule by operating the slice and the slice on one side adjacent to it in different directions. For the second and third rules, use the deviations in the virtual slices on both sides, and for the second and third rules, use the deviations in the slice and the adjacent slices, the virtual slices that exist between them, and the virtual slices that exist adjacent to these ends. 2. The basis weight paper machine control device according to claim 1. (4) The wrinkle smoothing criterion displays slices in a predetermined direction as those that can be adopted, and includes a slice counter indicating the boundary between the operated region and the unoperated region, and the degree of suitability of the slice counter. An operation inference unit that classifies each current slice operation by comparing it with a predetermined fine-tuning threshold; When it is determined that the slice is located closer to the unoperated area than the value indicated by the slice counter, the first step selects the slice and writes the position of the operated slice to the slice counter.
When the reference unit and the operational reasoning unit determine that the fitness is smaller than the fine threshold for all slices, or there is a slice that is larger than the fine threshold,
and a second reference unit that, when determining that this slice exists only on the operated area side of the value indicated by the slice counter, sets the slice to be selected as none and resets the value of the slice counter. The paper machine control device according to claim 1, characterized in that: (5) The wrinkle smoothing norm is for displaying slices in a predetermined direction as those that can be adopted, and includes a slice counter indicating the boundary between the operated region and the unoperated region, and the degree of suitability of the slice counter. Comparing a predetermined rough threshold value and a fine adjustment threshold value smaller than this rough threshold value,
an operational inference unit that classifies each slice; a third reference unit that selects the slice when the operational inference unit determines that there is a slice whose fitness is greater than a rough threshold; When the part determines that there is a slice whose fitness level is smaller than the rough threshold value and larger than the fine threshold value, and that this slice exists in the unfinished operation region than the value indicated by the slice counter, a first reference section that selects a slice and writes the manipulated slice to the slice counter; When it is determined that there is a slice that is smaller than the threshold value and larger than the fine adjustment threshold value, and that this slice exists only on the operated area side of the value indicated by the slice counter, the slice to be selected is determined to be none. 2. The paper machine control device according to claim 1, further comprising: a second reference section that also resets the value of the slice counter. (6) The slice counter stores the most recently operated slice located outward from the center of the profile measured by the sensor as a boundary value on the side of the uncompleted operation area. Claim 4 or 5
The paper machine control device described. (7) The first reference unit selects slices to be operated within a predetermined number of slices to be operated on the unoperated area side from the slices stored in the slice counter. Or the paper machine control device according to 5. (8) The slice determining unit includes a mode setting unit that determines whether the current control mode is a coarse adjustment mode or a fine adjustment mode; includes a coarse slice selection unit that selects the relevant slice, and a fine slice selection unit that selects a slice to be operated according to the wrinkle smoothing norm when it is determined that there is a slice whose fitness level is greater than the fine adjustment threshold. , The paper machine control device according to claim 1, characterized in that it has the following. (9) The slice determining section is characterized in that the mode setting section selects the coarse slice selection section when the profile state is disturbed, and selects the fine slice selection section when the profile state is stable. Item 8. The paper machine control device according to item 8. (10) The slice determining unit divides the slice into several zones in advance, and determines the slice to be operated in each zone according to a wrinkle smoothing standard for each zone. The paper machine control device according to item 1 or 8. (11) For a paper machine having a plurality of slices arranged in the width direction of the paper for controlling the supply amount of paper raw materials, and a sensor for measuring the quality of the manufactured paper in the width direction of the paper, In the paper machine control method of operating the slice in a direction in which the value detected by the sensor corresponds to a predetermined target profile, a position correspondence step of determining which of the slices the measurement point of the sensor corresponds to. and,
The deviation between the detected basis weight profile and the target profile is
a profile calculation step of calculating for the slice according to the correspondence of the position correspondence step; a virtual slice calculation step of calculating a deviation between the detected basis weight profile and the target profile for a virtual slice provided at a position between the slices; , has at least one or more rules for manipulating slices, and uses the deviation of each slice calculated in the profile calculation step and the virtual slice calculation step and the virtual slice adjacent thereto to apply the corresponding operation rule to each of these rules. a candidate selection step of calculating the degree of compatibility between the two and taking out one of the operation rules for each slice based on the calculated degree of compatibility; and the degree of compatibility for each slice obtained in this candidate selection step; a slice determination step in which a slice to be operated is determined in accordance with a wrinkle smoothing standard by comparing the slices with a predetermined threshold value; and a slice determination step in which a slice to be operated is determined in accordance with a wrinkle smoothing norm by comparing the slices with a predetermined threshold; A paper machine control method comprising: an output control step of sending a controlled amount to the slice.