JPH07334246A - Automatic control device - Google Patents

Abstract

(57) [Summary] (Correction) [Purpose] Especially for automatic control devices that control movements after a target according to a tendency based on the preceding control movement of the target, and feedback predictive control based on continuously changing measured values. It is an object of the present invention to propose an automatic control device that can be accurately performed. A feedback signal S output from a feedback signal correction means 9 indicating the processing results of the workpieces carried out from the processing means 1 for sequentially processing the workpieces successively carried in.
3, the non-correction processing result calculation means 5 calculates the processing result of the workpiece when processed by the feedback signal S3 which is not corrected by the feedback signal correction means 9, and the calculated processing of the workpiece is performed. The predicted tendency calculation means 6 predictively calculates the tendency prediction value when the processing is performed without correction based on the result, and the feedback signal correction means 9 corrects the feedback signal S3 based on the predicted tendency calculation value. Output to the processing means 1.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic controller for controlling various objects such as numerical control (NC) machine tools, and more particularly to an automatic controller for controlling the operation after the object according to the tendency based on the preceding control operation of the object. Regarding the control device.

[0002]

2. Description of the Related Art Conventionally, as an automatic control device of this type, FIG.
There was one shown in 0. FIG. 10 shows a block diagram of a conventional automatic control device and its operation characteristic diagram. In the figure, a conventional automatic control device includes a machining unit 1 for machining a workpiece based on a control operation signal S2 calculated from a machining target value S1 and a feedback signal S3, and machining of the machined workpiece. Dimension measuring unit 2 for measuring precision dimensions
And a suitability judging unit 40 for judging whether or not it is within a conforming range showing an allowable value of the deviation value from the processing target value S1, and a correcting unit 50 for correcting the value of the feedback signal S3 based on this judgment. Is. The suitability determination unit 40 is configured such that an allowable value preset based on the processing accuracy of the processing unit 1 or the like is input from the allowable value generation unit 41.

Next, the control operation of the conventional device based on the above configuration will be described. First, the work piece is processed by the processing section 1, and the work piece is measured by the dimension measuring section 3. The dimension (measurement dimension) of the measured processing accuracy is the suitability determination unit 4
According to the above, it is determined whether or not it falls within the applicable range. When it is determined that the deviation is out of the conforming range, the correcting unit 50 corrects and outputs the feedback signal S3 so as to be at the center of the conforming range. Based on the corrected feedback signal S3 and the processing target value S1, the processing unit 1 processes the next workpiece to be processed.

FIG. 1 shows another conventional automatic control device.
There was one shown in 1. FIG. 11 shows a block diagram of another conventional device and its operation characteristic diagram. In the same figure, another conventional automatic control device is provided with a processing section 1, a dimension measuring section 3 and a correcting section 50 in common as in the conventional apparatus shown in FIG. 3 and the correction unit 50, and includes a moving average unit 60 that calculates a moving average value from a plurality of measurement values measured by the dimension measuring unit 3.

Next, the operation of the other conventional device will be described. First, the workpiece processed by the processing unit 1 is measured by the dimension measuring unit 3, and the measured values are Δ1, Δ2, Δ3, ...
・ And The moving average unit 60 obtains moving average values q3, q4, ... Based on the measured values Δ1, Δ2, Δ3 ,. This moving average value q3 = (△ 1 + △ 2) / 2, q4 ＝ (△
1 + △ 2 + △ 3) / 3, and so on. This moving average value q
The correction unit 50 corrects the feedback signal S3 based on 3, q4, .... Based on the control operation signal produced from the corrected feedback signal S3 and the processing target value S1, the processing unit 1 processes the workpiece.

As described above, even when the processing target value S1 is set to the same value and processing such as processing of the workpiece is executed,
The measured value will change due to various factors such as wear of the cutting tool (blade) in the processing unit 1, temperature change, etc., but the moving average value can be calculated from the measured value of the preceding phenomenon of the workpiece processed in advance. Since the next process of feedback control of the workpiece is executed based on the obtained moving average value, the fluctuation of the measured dimension can be corrected.

Further, the moving average value will be described in detail. The measured value includes a tendency change and a processing variation (variation). If this processing variation is subjected to feedback control by dimensional correction, accuracy will be deteriorated. On the other hand, if feedback control is performed using the moving average value, it is possible to separate the tendency change and the processing variation and use only the tendency change as the correction value. When the measured value Δ containing the processing variation error is processed as several groups, the trend change is calculated as a loop gain, and the processing variation error can be eliminated.

In the first dimension measurement, k1 = Δ1
And _Letting k _{(i <limit)} be a correction value until the number of times of dimension measurement reaches i _(limit) , k _{(i <limit)} = (1 / i) .multidot..DELTA.i + {(i-1) / i } ・
When the correction value of greater stability region number of measurements than i _(limit) of the k _i-1 the dimension measurement and _{k (i> limi t),} k (i> limit) = g · △ i + (1-g) · k _i-1 where k _i is the correction value after the i-th measurement, k _i-1 is the correction value before the i-th measurement, Δi is the i-th measurement value, and g is the loop gain (0 ,. ., 1) and i _(limit) are transition points to the stable region at the end of the unstable state. i _(limit) ≈ (1 / g).

[0009]

Since the conventional automatic control device is constructed as described above, it is subject to a large variation in the area defined as the conforming range as shown in FIG. There is a problem that feedback control is difficult. Further, since another conventional automatic control device sequentially performs the processing of the workpiece to be the current target by the feedback control based on the moving average value q calculated from the measured value Δ of the preceding phenomenon, a plurality of processing steps are performed. In the case of the workpiece to be carried out through the above, the measured value of the workpiece immediately before cannot be obtained, and there is a problem that accurate feedback control processing cannot be performed on the current workpiece.

For example, when the work piece is continuously carried out successively and a large number of steps such as processing, cleaning and measurement are required for this work, wear of the cutting tool and temperature change Therefore, for a continuous change in the measured value, the moving average is obtained from the measured value of the work piece delayed corresponding to the number of processing steps. Thus, since the feedback predictive control is performed by the moving average value based on the measured value of the workpiece delayed by the delay amount corresponding to the number of processing steps, the trend of the lagging phenomenon cannot be accurately predicted. The present invention has been made to solve the above problems, and an object of the present invention is to propose an automatic control device that can accurately perform feedback prediction control based on continuously changing measurement values.

[0011]

FIG. 1 shows the principle of the present invention. In the figure, the automatic control device according to the present invention is
Processing means for sequentially executing various kinds of processing such as processing to a preset target value for the workpieces sequentially carried in;
A feedback signal correction means for correcting the feedback signal S3 based on the processing result of the work piece processed by the processing means and the target value, and for the work piece to be subsequently carried in by the corrected feedback signal. In the automatic control device for controlling the processing of the processing means by the processing means, when the processing is performed by the feedback signal which is not corrected by the feedback signal correction means, based on the processing result and the feedback signal of the workpiece carried out from the processing means. Non-correction processing result calculation means for calculating the processing result of the work piece, and prediction tendency calculation means for predicting calculation of a tendency prediction value when processing is performed without correction based on the calculated processing result of the work piece. And the feedback signal correction means corrects the feedback signal based on the trend prediction value. Is output to the processing means.

Further, according to the present invention, if necessary, a tendency correction for calculating a corrected tendency prediction value based on a tendency correction value determined by the characteristic of the processing means for the tendency prediction value calculated by the prediction tendency calculation means. Computation means is provided, and the processing means executes processing based on the corrected tendency prediction value.

[0013]

In the present invention, the non-correction processing result is obtained based on the processing result of the workpiece carried out from the processing means for sequentially processing the workpieces sequentially carried in and the feedback signal output from the feedback signal correcting means. The calculation means
The processing result of the work piece is calculated when the processing is performed by the feedback signal that is not corrected by the feedback signal correction means, and the trend prediction value when the processing is performed without correction based on the calculated processing result of the work piece Is calculated by the predictive tendency calculating means, and the feedback signal correcting means corrects the feedback signal based on the predicted tendency predicted value and outputs the corrected feedback signal to the processing means. By performing feedback prediction control based on the trend prediction value of the preceding phenomenon, it is possible to control fluctuations in the processing results of the processing means caused by various factors as much as possible, and accurate feedback prediction control can be performed.

Further, in the present invention, the tendency correction calculation means calculates the corrected tendency prediction value based on the tendency correction value determined by the characteristic of the processing means. Since the processing means executes the processing based on this corrected tendency prediction value, the tendency prediction value is converged and matched with the target value originally intended in consideration of the characteristics of the processing means, and the feedback with higher accuracy is provided. Predictive control is possible.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS. 2 is an overall block diagram of the present embodiment, FIG. 3 is an operation flowchart of the embodiment shown in FIG. 2, and FIGS. 4 and 6 are dimensional change characteristic diagrams of the embodiment shown in FIG. In each of the above drawings, the automatic control apparatus according to the present embodiment targets the processing unit 1 and the cleaning unit 2 that perform processing such as processing on a workpiece for feedback control, and measures the dimension measurement unit 3, the mode determination unit 4, and the non-operation unit 4. A correction processing dimension calculation unit 5, a predicted tendency calculation unit 6, a tendency correction determination unit 7,
The configuration is such that the control target is controlled based on the feedback signal S3 output via the tendency correction calculation unit 8 and the feedback signal correction unit 9.

In the dimension measuring unit 3, the workpiece is processed by the processing unit 1 based on the feedback signal S3, and the processed workpiece is cleaned by the cleaning unit 2, and then the processing accuracy of This is a configuration for measuring dimensions. The mode determination unit 4 determines whether the measured processing accuracy dimension is to be a target of the moving average mode or a non-moving average mode according to whether or not a preset number of times of measurement has been reached. Is. When the mode determination unit 4 determines that the moving average mode is to be the target, the uncorrected processing size calculation unit 5 corrects the feedback signal based on the processing accuracy size and the feedback signal S3 of the feedback signal correction unit 9. The processing accuracy dimension (hereinafter, uncorrected processing dimension) of the workpiece when it is processed by the feedback control that is not corrected by the unit 9 and the tendency correction calculation unit 8 is calculated.

The predictive tendency calculation unit 6 predicts a predicted tendency value for a workpiece processed by the processing unit 1 without being corrected based on the calculated non-corrected processing size and further cleaned by the cleaning unit 2. This is a configuration for calculating. The tendency correction judgment unit 7 is forced to converge the predicted tendency value calculated by prediction into a processing target value based on the characteristics of the system, for example, every time a predetermined number of processing dimensions are output (every predetermined number of measurements). It is configured to determine whether or not there is a correction. When the tendency correction judging section 7 judges that the correction is to be performed, the tendency correction calculating section 8 predicts a tendency prediction value based on a tendency correction value set in advance for a predetermined number of times of dimension measurement (for example, 4 times). Is corrected and the corrected tendency predicted value is output to the feedback signal correction unit 9 in the subsequent stage and the non-corrected machining dimension calculation unit 5 in the previous stage.

The feedback signal correction section 9 receives the feedback signal S based on the corrected correction tendency prediction value.
3 is corrected and the corrected feedback signal S3 is output to the processing unit 1 which is a control target. Next, the feedback prediction control operation of this embodiment based on the above configuration will be described. First, in the case of the present embodiment, the delay number α of the feedback signal S3 is "3" (processing unit 1, cleaning unit 2,
(3 delays by the dimension measuring unit 3), and the number of samples of the moving average value, etc. are set as initial conditions (step 1).
When this initial condition is set, the above-described processing such as correction value is performed (step 2).

The workpiece is loaded into the processing section 1 and processed by the processing section 1 so as to match the processing target value S1.
Further, it is cleaned by the cleaning unit 2, and the cleaned workpiece is measured by the dimension measuring unit 3 and the machining accuracy dimension is output. When a plurality (for example, two) of the measured processing accuracy dimensions are output, the tendency correction calculator 8 calculates a correction value based on the plurality of processing accuracy dimensions (step 3). In the calculation of this correction value, first, the non-corrected machining dimension calculation unit 5 calculates the non-corrected machining dimension (in FIG. 5) based on the machining precision dimension (shown as an X mark in FIG. 5) and the feedback signal S3 of the feedback signal correction unit 9. Is shown as a circle).

Next, the tendency calculator 6 calculates a predicted trend value (see FIG. 5) based on the calculated non-corrected machining dimension, and the tendency correction calculator 8 determines the tendency correction calculator 8 by the judgment of the tendency correction judger 7. In step 3, the correction tendency prediction value is calculated. That is, as shown in FIG. 4, the correction value ki is calculated from a plurality of correction dimensions to be processed (circle in the figure) to P11 to P12,
The correction values k11, k21, ... Are calculated on the basis of the predicted tendency values represented as the inclined line segments passing through P21 to P22. The mode determination unit 4 determines whether the moving average mode is set or not depending on whether or not the processing accuracy dimension is output a predetermined number of times (step 4). When the moving average mode is determined, the feedback signal correction unit 9 is based on this correction value.
Executes a calculation for correcting the feedback signal S3 (step 5). This corrected feedback signal S
3 is added to the processing target value S1 by the addition unit 101C and is output to the processing unit 1 as a control operation signal S2 (step 5) to control the processing operation of this processing unit 1.

Further, when it is determined that the loading of the workpieces sequentially loaded into the processing section 1 is not completed (step 6), the operations from step 2 to step 6 are repeated. When it is determined in step 6 that the loading of the workpiece has been completed, this control operation is also completed. In the above embodiment, the non-corrected machining dimension is calculated based on the machining accuracy dimension, the predicted tendency value is calculated from the non-corrected machining dimension, and the correction value is calculated from the predicted tendency value. Although the feedback signal is corrected by the value, the tendency correction calculation unit 8 calculates the correction tendency prediction value forcibly converging the absolute value of the processing precision dimension to the processing target value in step 3 described above. You can also A characteristic diagram of the amount of dimensional change in this case is shown in FIG.

In the figure, the processing unit 1 is operated at time T1.
When the tendency prediction value changes due to the temperature change or the like, the tendency prediction value is forcibly corrected every time a predetermined number (for example, 4) of machining accuracy dimensions are detected, and the tendency is converged to the machining target value. Let This forced correction value β is used when the correction is performed at time T2 after the four processing precision dimensions. further,
The theoretical calculation required for the feedback predictive control of the present embodiment based on each of the above configurations will be described. What is the main purpose of this correction method is if the processing accuracy dimension (measured value: detected as a deviation from the target value) measured by the dimension measuring unit 3 is not directly used for dimension correction, but is processed without correction. The amount of trend change (trend prediction value) indicating whether the trend change will occur is calculated and the correction value is obtained. every time·
Measurement-feedback system Dimensional measurement part 3 in operation
The measured value of does not directly represent the tendency change, but is output as a dimensional change amount that appears as a result of adding a certain correction amount each time. In the case of out-of-machine inspection, there is a delay, so this process is performed to correct the trend change. It should be noted that the correction function of the absolute value is weak only with this correction, and therefore the absolute value forced correction term β can be further added.

First, the symbols used in the calculation formula will be defined.
α indicates the number of delays, and in the case of the dimension correction in the in-machine measurement (feedback is one delay), α = 1. Also, in the case of dimension correction by out-of-machine measurement (feedback is 2
(After a delay), α = 2. Further, in the case of the dimension correction in the out-of-machine measurement (when the feedback is delayed by 3), α = 3. i = order number (1,2,3,4, ...) k _i = correction number (correction value when machining i-th workpiece) δ _i = measured value (measured value of i-th processed workpiece) ) N = number of moving averages (obtained from processing results ... Example n =
5) g = 1 / n loop gain u = assumed measurement value without correction u _i = δ _i −Σk _i (assumed measurement value when the _i -th workpiece is processed without correction) γ = uncorrected Assumed adjacent difference

Here, γ _i is the i-th workpiece and (i-
1) It is an assumed adjacent difference with the non-corrected work. γ _i = u _i −u _(i-1) = (δ _i −Σk _i ) − (δ _(i-1) −Σk _(i-1) ) = δ _i −Σk _i −δ _(i-1) + Σk _(i-1) = δ _i −δ _(i-1) + Σk _{(i-1) −} Σ k _i = δ _i −δ _(i-1) + Σ k _(i-1) − (Σ k _(i-1) + k _i ) = Δ _i −δ _(i-1) −k _i

Next, the formula for calculating the correction number ki is as follows. When (i-α-1) <0, a measured value has not been obtained, so k _i = 0. When (i-α-1) = 0, the measured value δ ₁ of the first processed workpiece is obtained, but γ _i
Cannot be calculated, so k = −δ _i . (I-α-1)
= 1, the measured value δ ₁ of the first machined workpiece,
Next, since the measured value of the machined work δ _{2 and} the measured value are obtained, k _i =
Let −δ _(i- α ₎ + δ _(i- α _-1) . The previous correction value k
_(i-1) does not consider the change in trend, so it is not reflected in this correction value. 1 <(i-α-1) <n
In the case of, g = 1 / (i-α-1).

When (i-α-1)> n, g =
1 / n. Therefore, the correction value k _i is as follows. k _i = −g (δ _(i- α ₎ −δ _(i- α ₋₁₎ −k _(i- α ₎ ) + (1
-G) k _(i-1) Further, the forced correction of the absolute dimension value is as follows. Since the above equation is corrected by the moving average value of the tendency change, the force for forcibly converging the absolute value of the dimension to the center is weak. Moreover, when the tendency change characteristic changes, it deviates largely from the center and exhibits hunting-like behavior. The following compulsory correction term βi is added for the purpose of correcting this.

However, this compulsory correction term β _i must be taken into consideration in the calculation of the adjacent difference of the tendency change, but the previous compulsory correction term β _(i-1) must be excluded from the calculation of the moving average. I won't. Considering this term, the above equation becomes: k _i = −g (δ _(i- α ₎ −δ _(i- α ₋₁₎ −k _(i- α ₎ ) + (1
_{-G) · (k (i-} 1) -β (i-1)) + and β _i.

What should be used as the value of β _m should be determined from the characteristics of the system. As an example, measured value δ
It is conceivable to use several averages of _i . However, in consideration of the delay of the measurement value, the correction of β _i is occasionally added. The measured value used for the average value must be a value processed without correction of β _i . As a specific example, two consecutive measurements are obtained for trend correction, so the average value is β
_i is corrected once every m. When correction is performed once every m times, this forced correction term β _i is β _i = − (δ _(i-
α ₎ + δ _(i- α _-1) ) / 2. Further, when the average value of the two is m ≧ α + 1. Further, when the average value of the three is m ≧ α + 2.

Further, the operation of the feedback predictive control by the calculation as in the above embodiment will be described with reference to FIGS. 7 to 9. 7 to 9 are detailed flowcharts of the steps shown in FIG. First, the initial conditions are set by setting the feedback delay number α to “3”, the moving average number n to “5”, and the counter reset setting number i to “0” (step 1). It is determined whether the integrated value i of the counter is "0" (step 2
1), when it is judged to be "0", the past correction values k _i , k _(i-1) , k _(i-2) , k _(i- α ₎ , δ _(i- α ₎ Are reset to "0" (step 22). After the reset value is reset to "0" in this step and when it is determined in step 21 that the integrated value is not 0, it is determined whether the feedback delay number α is "1" (step 2
3) Further, it is determined whether or not the feedback delay number α is “2” (step 24).

If α = 1 is determined in step 23, the correction value k _{(i-1 ) is set} to k _i (step 251). If it is determined that α is not equal to ₁ and α is equal to 2 in step 24, the correction value k _(i- α _{) is set} to k _(i-1) (step 252). If α ≠ 2 is determined in step 24, the correction values are set to k _(i- α ₎ and k _(i-2), and the correction value k _(i-2)
Be k _(i-1) (step 253). Step 2
In any of 51, 252, and 253, the correction value k _{(i-1 )} is further set to k _i (step 26), and the forced correction value (term) is set.
β _{(i-1) is set} to β _i (step 27), and the measured value (machining accuracy dimension) δ _(i- α _{-1) is set} to δ _(i-1), and the measured value δ _(i- α ₎ is set. Input (step 28). Further, the integrated value i of the counter is set to i + 1 (step 29).

The step 27 is not executed unless the absolute dimension value is forcibly corrected, and the process proceeds to the next step 28. Further, when the absolute value of the dimension is forcibly corrected, if the integrated value i is set to i + 1 in step 29,
The procedure moves from step 31 to the forced correction step of step 33. In the forced correction step, if it is determined that the forced correction is performed every predetermined number m (for example, “4”) at which the preset measurement value is output, whether or not the predetermined number of times m = 4 is determined. to decide. If it is determined that m = 4, the forced correction term β _i is calculated as follows (step 32). [beta] _i =-([delta] _{(i- [} alpha] _-1) + [delta] _{(i- [} alpha] ₎ ) / 2 After this forced correction term [beta] _i is calculated, the predetermined number m is set to "0" (step 33). In addition, the step 3
Predetermined number of times m ≠ 4 in 1 (when it has not reached 4)
The forced correction term β _i is output as “0” (step 34).

If the forced correction is not performed, after step 29, if the forced correction is performed, step 3 is performed.
After 3 (or 34), it is determined whether or not the feedback prediction control in the moving average mode is performed based on the integrated value of the counter (step 46). In this step 41, the integrated value is not (i-α-1) <0,
If the integrated value is not (i-α-1) = 0 in 3 and the integrated value is not (i-α-1) = 1 in step 45, the moving average mode is entered (steps 51 to 54).

In this moving average mode, the integrated value is (i
-[Alpha] -1)> n is determined (step 5
1) If it is determined that the integrated value is (i-α-1)> n, the loop gain g is set to g = (1 / n) (step 52). Further, the integrated value is (i-α-1)> n
If it is determined that the loop gain g is not g = {1
/ (I-α-1)}. Based on the loop gain g set in step 52 or 53, the correction value k when the forced correction is not performed is calculated by the following equation (step 54). k _i = −g (δ _(i- α ₎ −δ _(i- α ₋₁₎ −k _(i- α ₎ ) + (1
-G) k _(i-1)

The calculation of the correction value k _i will be described more concretely from the relationship with k ₁ and k ₂ . Now, the assumed adjacent difference γ without correction is γ ₁ , γ ₂ , γ ₃ , γ ₄ , γ ₅ , and γ ₂ , γ ₃ , γ ₄ , γ ₅ , γ ₆ are detected. And When the correction values k ₁ and k ₂ are obtained from the respective assumed adjacent differences γ, the following is obtained. k ₁ = (γ ₁ + γ ₂ + γ ₃ + γ ₄ + γ ₅ ) / 5 k ₂ = (γ ₂ + γ ₃ + γ ₄ + γ ₅ + γ ₆ ) / 5 Substituting equation k ₁ to the k ₂ formula, and _{k 2 = {(5k 1 -γ} 1) + γ 6} / 5. If k ₁ ≈ and γ ₁ are set here, k ₂ = {(5k ₁ −k ₁ ) + γ ₆ } / 5 = (4/5) k ₁ + (1/5) γ ₆

Here, when corresponding to the above k equation, k _i is k ₂
And k _(i-1) corresponds to k. Also, g
Corresponds to (1/5) and (1-g) corresponds to (4 /
Since they correspond respectively to 5), the equation of the correction value k is approximately established. A case where the calculation in step 54 is forcibly corrected will be described. The correction value k ′ _i for performing this forcible correction is calculated in step 52,
The loop gain g set by 53 is set to step 3
It is obtained by the forced correction β term calculated by 2 and 24. _{k 'i = -g (δ (} i- α) -δ (i- α -1) -k (i- α)) +
(1-g) · (k (i-1) -β (i-1)) + β i

It should be noted, terms _{-g (δ (i- α) -δ} (i- α -1) -k (i- α)) in the correction value k _i, the formula of k _'i is uncorrected The ratio of the assumed adjacent difference at the time is shown. Also, (1-g)
k _{(i- 1)} and (1-g) * (k _(i-1)-[ beta] _(i-1) ) + [beta] _i indicate the term of the previous correction value. As described above, the correction values k _i ,
The feedback signal correction unit 9 can execute correction of the feedback signal using k ′ _i . When it is determined in step 41 that the integrated value is (i-α-1) <0, the feedback signal correction unit 9 sets the correction value k _i to k _i = 1 and corrects the feedback signal.

When it is determined in step 43 that the integrated value is (i-α-1) = 0, the correction value k _{i is set} to k _i = -δ _(i- α ₎ to correct the feedback signal. The unit 9 performs the correction of the feedback signal. further,
In step 45, the integrated value is (i-α-1) = 1
If it is determined that the correction value k _i is k _i = −δ
_The feedback signal correction unit 9 corrects the feedback signal as the adjacent difference of _(i- α ₎ + δ _(i- α _-1) .

[0038]

As described above, in the present invention, the processing results of the workpieces carried out from the processing means for sequentially processing the workpieces sequentially carried in are used as feedback signals output from the feedback signal correcting means. Based on this, the non-correction processing result calculation means calculates the processing result of the workpiece when processed by the feedback signal that is not corrected by the feedback signal correction means, and the correction is performed based on the calculated processing result of the workpiece. The prediction tendency calculation means predictively calculates the trend prediction value in the case of processing without being performed, and the feedback signal correction means corrects the feedback signal based on the predicted prediction tendency value and outputs the feedback signal to the processing means. Therefore, the feedback prediction control is performed by the trend prediction value of the preceding phenomenon based on the original processing result in the processing. It will be as much as possible control of the processing result variation processing means caused by species cause an effect that it is accurate feedback predictive control. Further, the present invention provides a trend correction calculation means for calculating a corrected trend prediction value based on a trend correction value determined by the characteristics of the processing means, if necessary, for the trend prediction value calculated by the prediction tendency calculation means. It is possible to provide the effect that the processing means executes processing based on the corrected tendency prediction value.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of an automatic control device according to the present invention.

FIG. 2 is a block diagram of an automatic control device according to an embodiment of the present invention.

FIG. 3 is an operation flowchart of the automatic control device according to the embodiment described in FIG.

FIG. 4 is a dimensional change characteristic diagram of the automatic control device according to the embodiment shown in FIG.

FIG. 5 is a dimensional change amount characteristic diagram of the automatic control device according to the embodiment shown in FIG.

FIG. 6 is a dimensional change amount characteristic diagram of the automatic control device according to the embodiment shown in FIG.

7 is a detailed flowchart of steps 1 and 2 in the apparatus of the embodiment shown in FIG.

FIG. 8 is a detailed flowchart of step 3 in the embodiment apparatus shown in FIG.

9 is a detailed flowchart of steps 4 and 5 in the apparatus of the embodiment shown in FIG.

FIG. 10 is a block configuration diagram and a dimensional change amount characteristic diagram of a conventional automatic control device.

FIG. 11 is a block configuration diagram and a dimensional change amount characteristic diagram of a conventional automatic control device.

[Explanation of symbols]

 1 Processing Section 2 Cleaning Section 3 Dimension Measuring Section 4 Mode Judgment Section 5 Uncorrected Machining Dimension Calculation Section 6 Prediction Trend Calculation Section 7 Trend Correction Judgment Section 8 Trend Correction Calculation Section 9 Feedback Signal Correction Section 10 Addition Section

Claims (2)

[Claims] 1. A processing means for sequentially performing various kinds of processing such as processing to a preset target value for workpieces sequentially carried in, and processing results of the workpieces processed by the processing means. And an automatic control device that includes a feedback signal correction unit that corrects a feedback signal based on a target value, and controls the processing of the processing unit with respect to a workpiece that is subsequently carried in by the corrected feedback signal. Non-correction processing result calculation means for calculating the processing result of the workpiece when processed by the feedback signal not corrected by the feedback signal correction means, based on the processing result and the feedback signal of the processing object carried out from the processing means. And predicting a trend prediction value when processing is performed without correction based on the calculated processing result of the workpiece And a prediction trend calculation means calculated for said feedback signal correcting means automatic control apparatus characterized by outputs to the correction to the processing means a feedback signal based on the trend prediction value. 2. The automatic control device according to claim 1, wherein the tendency prediction value predicted and calculated by the prediction tendency calculation means is a corrected tendency prediction value based on a tendency correction value determined by a characteristic of the processing means. An automatic control device, comprising: a tendency correction calculation unit that calculates, and the processing unit executes a process based on a corrected tendency prediction value.