JP7275492B2 - Control device, control method and program - Google Patents

Description

The present invention relates to a control device, control method and program.

Control devices such as temperature control devices, PLCs (Programmable Logic Controllers), DCSs (Distributed Control Systems), and control devices mounted on personal computers and built-in control devices are widely used in industry.

In addition, PID (Proportional-Integral-Differential), model-predictive-control (MPC), internal model control, LQG ( Control methods such as Linear-Quadratic-Gaussian) control, H2 control, and H∞ control are known. Further, techniques related to these control methods are disclosed in Patent Documents 1 to 5, for example.

For example, in Patent Document 1, in model predictive control, after calculating an estimated disturbance value based on a prediction error, which is an error between a predicted value and an actual measured value of a control amount, and a terminal value of a step response, this disturbance A technique is disclosed for correcting the manipulated variable for the controlled object based on the estimated value.

Further, for example, in Patent Document 2, a new value is calculated based on a corrected target deviation, which is the difference between a predicted convergence value of the controlled variable according to changes in the manipulated variable in the past up to the present time and the target value. Techniques for determining manipulated variables are disclosed.

Further, for example, Patent Literature 3 discloses a technique for displaying on a graph a region in which a control logical expression representing the relationship between the current value of the target deviation and the amount of change in the manipulated variable is established.

In addition, for example, Patent Document 4 discloses a technique for feeding forward output to the operation of the master regulator by using the error in the manipulated variables of two controllers, the master regulator and the slave regulator, using model predictive control as an estimated disturbance. is disclosed.

Further, for example, in Patent Document 5, when a disturbance having the same magnitude and application timing is applied in a plurality of consecutive operation cycles, a measured value of a control amount in a certain operation cycle, A technique is disclosed for correcting a predicted value in an operation cycle after a certain operation cycle using a model prediction error, which is an error from the predicted value of the control amount in a certain operation cycle.

Japanese Patent No. 5396915 WO2016/092872 WO2015/060149 Japanese Patent No. 4177171 JP 2018-41150 A

By the way, in model predictive control, it is often important to suppress the influence of disturbances on the control amount. Therefore, for example, as disclosed in Patent Document 1, a method of correcting the manipulated variable based on the estimated disturbance value is often effective in making the controlled variable of the controlled object follow the target value. .

However, in the prior art, it was not possible to dynamically change the disturbance estimation model according to the time constant and duration of the disturbance that could actually occur, and it was necessary to assume the disturbance estimation model and the like in advance. On the other hand, disturbance models include various models such as impulse disturbance, step disturbance, and lamp disturbance. Therefore, depending on the disturbance that actually occurs, it may not be possible to effectively suppress the influence of the disturbance on the controlled variable.

An embodiment of the present invention has been made in view of the above points, and an object of the present invention is to effectively suppress the influence of disturbance on the controlled variable.

In order to achieve the above object, one embodiment of the present invention provides a control device that outputs a manipulated variable to a controlled object and causes the controlled variable of the controlled object to follow a target value, comprising: a plant response model of the controlled object; a predicted value calculation means for calculating a predicted value of the control amount based on the amount of change in the manipulated variable in the past up to the present time, and storing the calculated predicted value as a time series in a first storage means; Measured value acquisition means for acquiring measured values of the control amount and storing the acquired measured values as a time series in a second storage means; and past reading of the time series of the predicted values and the time series of the measured values When the past reading length indicating the time width for reading is input, the latest prediction value among the prediction values stored in the first storage means and the prediction value at the time point indicated by the past reading length; Measurement at the time indicated by the latest measured value and the past reading length using the latest measured value among the measured values stored in the second storage means and the measured value at the time indicated by the past reading length the target value and the control; second correction value calculating means for calculating a corrected target deviation by correcting either the difference between the target value in the future or the difference between the target value in the future and the controlled variable in the future using the past reading correction value; and the corrected target deviation and a manipulated variable calculation means for calculating the new manipulated variable based on.

It is possible to effectively suppress the influence of disturbance on the controlled variable.

It is a figure which shows an example of a structure of the control apparatus which concerns on 1st embodiment. It is a figure for demonstrating an example of operation|movement of a target value prefetch part. FIG. 11 is a diagram (part 1) for explaining an example of the operation of the response correction unit; FIG. 11 is a diagram (part 2) for explaining an example of the operation of the response correction unit; FIG. 12 is a diagram (part 3) for explaining an example of the operation of the response correction unit; 9 is a flowchart for explaining an example of update processing of a prediction time series storage unit; It is a figure for demonstrating an example of an update of a prediction time-series storage part. 9 is a flowchart for explaining an example of update processing of a past measured value time-series storage unit; It is a figure for demonstrating an example of an update of a past measured value time-series memory|storage part. 9 is a flowchart for explaining an example of update processing of a past prediction time series storage unit; It is a figure for demonstrating an example of an update of a past prediction time-series storage part. FIG. 10 is a diagram for explaining an example of the operation of an operation change amount calculation unit; It is a figure which shows an example at the time of comprising the control loop of a controlled object plant using the control apparatus which concerns on 1st embodiment. It is a figure which shows an example of the hardware constitutions of the control apparatus which concerns on 1st embodiment. It is a figure which shows an example of a structure of the control apparatus which concerns on 2nd embodiment. It is a figure which shows the plant response model in an Example. FIG. 4 is a diagram showing information used for designing control gains in the embodiment; It is a figure which shows the control gain in an Example. FIG. 11 is a diagram (part 1) showing control results in the embodiment; FIG. 11 is a diagram (part 2) showing control results in the embodiment; FIG. 11 is a diagram (part 3) showing a control result in the embodiment; FIG. 10 is a diagram (part 4) showing a control result in the embodiment;

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention (hereinafter also referred to as "present embodiments") will be described in detail with reference to the drawings. Hereinafter, the control device 10 that calculates the manipulated variable for following (or settling) the controlled variable to the target value when the target value time series from the present to the future is input will be described. The control device 10 according to the present embodiment is based on the target value time series {r(t)}, which is time series data of arbitrary target values, the control amount y indicating the state of the controlled plant 20, etc. A manipulated variable u for the plant 20 is calculated. Then, the control device 10 according to the present embodiment measures the controlled variable y of the controlled plant 20 according to the manipulated variable u, and based on the controlled variable y, the target value time series {r(t)}, etc. , the following manipulated variable u is calculated. The controlled variable y is, for example, the temperature of the controlled plant 20, and the target value r is, for example, the set temperature. However, the controlled variable y and the target value r are not limited to the temperature and the set temperature, and any controlled variable in the controlled plant 20 and any desired target value of the controlled variable can be used.

Assume that the control device 10 according to the present embodiment is, for example, a built-in computing device, a PLC, or the like. However, the control device 10 according to this embodiment is not limited to this, and may be any device or device.

[First embodiment]
<Configuration of control device 10>
First, the configuration of the control device 10 according to this embodiment will be described with reference to FIG. FIG. 1 is a diagram showing an example of the configuration of a control device 10 according to the first embodiment.

As shown in FIG. 1 , the control device 10 according to this embodiment includes a target value prefetching unit 101 , a measuring unit 102 , a differentiator 103 , an operation amount updating unit 104 and a timer 105 . Each of these functional units is implemented by, for example, processing that one or more programs installed in the control device 10 cause a CPU (Central Processing Unit) or the like to execute.

The target value look-ahead unit 101 inputs the target value time series {r(t)} and the look-ahead length _Tp at each predetermined control cycle _Tc , and calculates the target value at time t+ _Tp after the look-ahead length from the current time t. Output the value r(t+T _p ). The look-ahead length T _p is the time length for determining the look-ahead target value r(t+T _p ) in the target value time series {r(t)}. In addition, henceforth, target value r (t+ _Tp ) is also represented as "prefetch target value r (t+ _Tp )."

The measuring unit 102 measures the controlled variable y of the plant 20 to be controlled at each control cycle _Tc . Then, the measurement unit 102 outputs the latest value of the measured control amount y as the control amount current value _y0 . Since y ₀ is the control amount measured by the measuring unit 102, it is also expressed as "control amount measurement value y ₀ ". Therefore, the current control amount value is the latest control amount measurement value among the control amount measurement values.

The controlled variable y of the controlled plant 20 is determined according to the manipulated variable u and the disturbance v. The disturbance v includes, for example, a drop or rise in outside air temperature when the controlled variable y is temperature.

The measurement unit 102 also acquires the operation amount u output from the operation amount update unit 104, and outputs the latest value of the acquired operation amount u as the operation amount current value _u0 .

The differentiator 103 outputs the difference (deviation) between the prefetch target value r(t+T _p ) output from the target value prefetch unit 101 and the control amount current value y ₀ as a target deviation e ₀ (t+T _p |t). . The target deviation e ₀ (t+T _p |t) is calculated by e ₀ (t+T _p |t)=r(t+T _p )−y ₀ (t). In addition, henceforth, target deviation _e0 (t+ _Tp |t) is also represented as "prefetching target deviation _e0 (t+ _Tp |t)."

The manipulated variable update unit 104 outputs the manipulated variable u for the controlled plant 20 every control cycle _Tc . The manipulated variable updating unit 104 includes a response correcting unit 111 , a manipulated change amount calculating unit 112 , and an adder 113 .

The response correction unit 111 calculates the plant response model {S(t)} of the controlled plant 20, the look-ahead target deviation e ₀ (t+T _p |t), the look-ahead length T _p , the past look length T _q , and the past Based on the operation change amount time series {du(t)}, which is the time series data of the change amount du of the operation amount u, the corrected target deviation e ^* (t) used to calculate the operation change amount du is calculated. The past reading length T _q is the time-series data {y ⁰ (t)} of the control amount measurement value y ₀ measured in the past by the measurement unit 102 and _the response This is the time length for determining how far back the time-series data {y _b (t)} of the predicted value of the control amount y predicted by the correction unit 111 in the past. The details of the method of calculating the corrected target deviation e ^* (t) will be described later.

Hereinafter, the change amount du of the manipulated variable u will also be referred to as the “manipulation change amount du”. In addition, the time series data {y ₀ (t)} of the control amount measurement value y ₀ measured in the past is also expressed as “past measurement value time series {y ₀ (t)}”, and the control amount y predicted in the past The time-series data {y _b (t)} of predicted values is also expressed as “past prediction time-series {y _b (t)}”. Each y _b (t) is a predicted value of the control amount y at each time t (a first predicted response value to be described later) at that time t. The details of the calculation method of {y _b (t)} will be described later.

The operation change amount calculation unit 112 calculates the operation change amount du based on the corrected target deviation e ^* (t) calculated by the response correction unit 111 for each control cycle _Tc . The manipulation change amount calculation unit 112 calculates and outputs the manipulation change amount du(t) in the order of du(t−3T _c ), du(t−2T _c ), and du(t−T _c ), for example. The amount of change in operation du is the amount by which the amount of operation u changes in each control cycle _Tc .

The adder 113 adds the operation amount current value _u0 output from the measurement unit 102 and the operation change amount du output from the operation change amount calculation unit 112 to calculate a new operation amount u. The adder 113 then outputs this manipulated variable u to the controlled plant 20 . This manipulated variable u is calculated by u(t)=u ₀ +du(t)=u(t−T _c )+du(t).

The timer 105 operates the target value prefetching unit 101 and the measuring unit 102 every control cycle _Tc . Note that the target value prefetching unit 101 and the measuring unit 102 operate in each control period _Tc , so that the manipulated variable updating unit 104 also operates in each control period _Tc .

<Operation of Target Value Prefetch Unit 101>
Next, the operation of target value prefetching section 101 will be described with reference to FIG. FIG. 2 is a diagram for explaining an example of the operation of the target value prefetching unit 101. As shown in FIG.

As shown in FIG. 2 , when target value time series {r(t)} and look-ahead length T _p are input, target value look-ahead unit 101 calculates a look-ahead target at time t+T _p after look-ahead length from current time t. Output the value r(t+T _p ). Note that SV in FIG. 2 represents the target value of the control amount.

In this way, the target value look-ahead unit 101 outputs the target value r(t+T _p ) at the time t+T _p after the look-ahead length in the target value time series {r(t)}.

Although the example shown in FIG. 2 shows the case where the target value time series {r(t)} is represented by a straight line, the present invention is not limited to this. The target value time series {r(t)} may be represented by an arbitrary curve, rectangle, or the like. In particular, the target value time series {r(t)} may be represented by a curve that periodically changes according to time t. Also, the target value time series {r(t)} may be set in advance, or the future target value r(t) may be updated as needed.

<Operation of Response Corrector 111>
Next, the operation of response correction section 111 will be described with reference to FIG. FIG. 3 is a diagram (part 1) for explaining an example of the operation of the response correction unit 111. FIG.

As shown in FIG. 3, the response correction unit 111 includes a plant response model {S(t)}, a look-ahead target deviation e ₀ (t+T _p |t), a look-ahead length T _p , a past look-ahead length T _q , When the operation change amount time series {du(t)} and the control amount current value y ₀ (t) are input, the corrected target deviation e ^* (t) is calculated and output by the following steps S1 to S4. do. In FIG. 3, PV represents the controlled variable, MV represents the manipulated variable, and dMV represents the manipulated variable.

Step S1: The response correction unit 111 calculates, as a look-ahead response correction value y _n (t), a value that is predicted to change the control amount current value y ₀ after T _p based on the past operation change amount du. For example, when the current time is t, the past operation change amount du is expressed as du(t−T _c ), du(t−2T _c ), and the like.

Here, as an example, a method of calculating the look-ahead response correction value y _n (t) when the plant response model {S(t)} is a step response model will be described.

Using the predicted value of the control amount at time predicted at time t as the generalized predicted value y _n,C (s|t), the response correction unit 111 calculates the generalized predicted value y _n, Compute _C (s|t).

ここで、Ｍは、一般化予測値ｙ_ｎ，Ｃ（ｓ｜ｔ）の計算に使用するモデルの長さ（モデル区間）である。

Here, M is the length (model interval) of the model used to calculate the generalized prediction value y _n,C (s|t).

At this time, the first response prediction value y _n,B (t) corresponding to the current time t is set to y _n,B (t)=y _n,C (t|t), and the second response prediction value y n,B (t) corresponding to the future time t+T _p The response prediction value y _n,A (t) is set to y _n,A (t)=y _n,C (t+T _p |t), and the response correction unit 111 calculates y _n (t)=y _n,A (t )-y _n,B (t) to calculate the look-ahead response correction value y _n (t). Note that y _b (t) described above is the first response prediction value at time t, and y _b (t)=y _n,C (t|t)=y _n,B (t).

Step S2: Next, the response correction unit 111 corrects the look-ahead target deviation e ₀ (t+T _p |t) with the look-ahead response correction value y _n (t) as a look-ahead correction target deviation e _f ^* (t), and _ef ^* (t)=r(t+T _p )−(y ₀ (t)+y _n (t))=e ₀ (t+T _p |t)−y _n (t).

Step S3: Next, the response correction unit 111 sets the difference between the current controlled variable value y ₀ and the predicted controlled variable value y _p (t|t−T _q ) as the past reading correction value e _b ^* (t), It is calculated by e _b ^* (t)=y ₀ (t)−y _p (t|t−T _q ). Here, the controlled variable predicted value y _p (t|t−T _q ) is the predicted value of the controlled variable at the current time t predicted at the past time t−T _q .

Here, a method of calculating the control amount predicted value y _p (t|t−T _q ) will be described with reference to FIG. FIG. 4 is a diagram (part 2) for explaining an example of the operation of the response correction unit 111. As shown in FIG.

As shown in FIG. 4, the response correction unit 111 inputs the past reading length T _q , the past measured value time series {y ₀ (t)}, and the past prediction time series {y _b (t)}. , the control amount prediction value y _p (t|t−T _q ) is calculated by the following equation.

y _p (t|t−T _q )=y ₀ (t−T _q )+(y _b (t)−y _b (t−T _q ))
That is, the past prediction time series {y _b (t)} is calculated based on the plant response model {S (t)} and the operation change amount {du (t)} output by the control device 10, and the past Since the _controlled variable measured value y ₀ (t−T _q ) is an actually measured value, the controlled variable predicted value y _p (t|t−T _q ) is calculated under the premise that correct prediction can be made by the plant response model {S(t)} and the operation change amount {du(t)} from the past time t−T _q to the current time t. It can be said that it is the predicted value of the control amount. Therefore, if an error occurs between the control amount current value y ₀ (t) actually measured at the current time t and the control amount predicted value y _p (t|t−T _q ), this error is , represents the deviation between the plant response model {S(t)} and the actual plant response, the influence of some disturbance, and the like. Therefore, as described above, assuming that this error is the past reading correction value e _b ^* (t), e _b ^* (t)=y ₀ (t)−y _p (t|t−T _q )=(y ₀ (t)-y ₀ (tT _q ))-(y _b (t)-y _b (tT _q )).

Return to FIG. Step S4: Finally, the response correction unit 111 corrects the read-ahead target deviation e _f ^* (t) with the past-read correction value e _b ^* (t) to obtain the corrected target deviation e ^* (t) as e ^* (t) = e _f ^* (t) - e _b ^* (t).

Here, as an example, a prediction time series storage unit 121 that stores the time series data of the generalized prediction value y _n,C (hereinafter also referred to as a “prediction time series”), and a past that stores the past measurement value time series Regarding the case where the response correction unit 111 calculates the corrected target deviation e ^* using the measured value time series storage unit 122 and the past prediction time series storage unit 123 that stores the past prediction time series, refer to FIG. while explaining. FIG. 5 is a diagram (part 3) for explaining an example of the operation of the response correction unit 111. FIG. Note that the predicted time series storage unit 121, the past measured value time series storage unit 122, and the past predicted time series storage unit 123 can be realized using a storage device such as an auxiliary storage device or a random access memory (RAM), for example. .

As shown in FIG. 5, the prediction time series storage unit 121 stores the prediction value (generalized prediction value) y _{n ,C} ( t−Δt|t), y _n,C (t|t), y _n,C (t+Δt|t), . . . , y _n,C (t+T _b |t) are stored. Note that T _b is a constant that determines the length (time length) of the generalized prediction value y _n,C stored in the prediction time series storage unit 121, and T _b =N ₁ Δt (N ₁ is an arbitrary positive integer). Also, Δt is the prediction interval, and Δt= _Tc .

Further, the past measured value time-series storage unit 122 stores control amount measured values y ₀ (t ₎ , y ₀ (t−Δt), . , y ₀ (t−T _e ) are stored. Note that T _e is a constant that determines the length (time length) of the control amount measured value y ₀ stored in the past measured value time-series storage unit 122, and T _e =N ₂ Δt (N ₂ is an arbitrary positive integer). Here, N ₂ =N ₁ may be satisfied.

Furthermore, in the past prediction time series storage unit 123, where t is the current time, predicted values (first _response predicted values) y _b (t), y _b (t−Δt), . . . , y _b (t−T _e ) are stored. Here, since y _b (t)=y _n,B (t)=y _n,C (t|t), the first response prediction value y _b (t)=y _n,B (t) is the generalized predicted value y _n,C (t|t) at s=t among the generalized predicted values y n _,C (s|t) at time s predicted at the current time t .

At this time, as shown in FIG. 5, the look-ahead response correction value y _n (t) is y _n,A (t)=y _n,C (t+T _p |t) stored in the prediction time series storage unit 121 and y _n,B (t)=y _n,C (t|t), y _n (t)=y _n,A (t)−y _n,B (t). As a result, using this look-ahead response correction value y _n (t), the look-ahead correction target deviation e _f ^* (t) is: e _f ^* (t)=e ₀ (t+T _p |t)−y _n (t) Calculated by

Further, as shown in FIG. 5, y ₀ (t) and y ₀ (t−T _q ) stored in the past measured value time series storage unit 122, and stored in the past prediction time series storage unit 123 Using y _b (t) and y _b (t - T _q ), the past reading correction value e _b ^* (t) is e _b ^* (t) = (y ₀ (t) - y ₀ (t - T _q ))−(y _b (t)−y _b (t−T _q )).

As described above, the corrected target deviation e ^* (t) is calculated by e ^* (t)= _ef ^* (t) _-eb ^* (t). In this way, by using the prediction time series storage unit 121, the past measurement value time series storage unit 122, and the past prediction time series storage unit 123, the response correction unit 111 can perform correction with a small amount of calculation and a small amount of memory. A target deviation e ^* can be calculated.

Also, the predicted time series storage unit 121 and the past predicted time series storage unit 123 are updated by the time series updating unit 124 each time the operation change amount du is input to the response correcting unit 111 . Similarly, the past measured value time series storage unit 122 is updated by the time series update unit 124 each time the current control amount value _y0 is input to the response correction unit 111 .

Note that the response correction unit 111 may calculate the correction target deviation e ^* (t) by e ^* (t)= _ef ^* (t) _-eb ^* (t)/ _Tq . As a result, for example, when the past reading length T _q is a very large value, it is possible to mitigate the influence of e _b ^* (t) also becoming large.

<Update processing of prediction time series storage unit 121>
Here, processing for updating the prediction time series storage unit 121 by the time series update unit 124 will be described with reference to FIG. FIG. 6 is a flowchart for explaining an example of update processing of the prediction time series storage unit 121. As shown in FIG. Hereinafter, a case will be described in which the generalized prediction value y _n,C at time t stored in the prediction time series storage unit 121 is updated to the generalized prediction value y _n,C at time t+Δt.

Step S11: The time series update unit 124 shifts the generalized predicted value y _n,C at time t stored in the prediction time series storage unit 121 by time Δt.

For example, as shown in FIG. 7A, y _n,C (t−Δ|t), y _n,C (t|t), y _n,C (t+Δ| _t ), _{, C (} t+T _p |t), _. . . , N= _N1 . In this case, the time-series updating unit 124 decrements the relative position m of the generalized predicted value y _n,C for which the relative position m≧0 is -1. That is, the time-series updating unit 124 converts y _n,C (t|t) at the relative position m=0 to the relative position m=−1, and y _n,C (t+Δt|t) at the relative position m=1. to the relative position m=0 in order. Thereafter, similarly, the time-series updating unit 124 shifts y _n,C (t+T _b |t) at the relative position m=N to the relative position m=N−1 until the relative position of each y _n,C Shift m in order.

Step S12: The time-series updating unit 124 updates the generalized predicted value y _n,C at the final time (that is, the relative position m=N). The generalized predicted value _{y n,C} at the final time t + NΔt + Δt predicted at time t (t + NΔt + Δt | t) is, for example, using the generalized predicted value y _n,C before time t + NΔt, for example,

により推定する。これは、時刻ｔにおける時刻ｔ＋ＮΔｔ－（ｊ－１）Δｔでの一般化予測値ｙ_ｎ，Ｃに対して重みａ_ｊを掛けて和を取った式である。

estimated by This is an equation obtained by multiplying the generalized predicted value y _n,C at time t+NΔt−(j−1)Δt at time t by weight a _j and calculating the sum.

In addition to this, the generalized predicted value y _n,C (t + NΔt + Δt | t) is also used, for example, when the generalized predicted value _{y n,C at time t + NΔt predicted at time t} is used as it is, that is, y _n,C ( A case where t+NΔt+Δt|t)=y _n,C (t+NΔt|t) is also conceivable.

Further, for example, considering the speed change, the generalized predicted value y _n,C at time t + NΔt and the generalized predicted value y _n ,C at time t + NΔt-Δt are extrapolated to obtain y _n,C (t + NΔt + Δt | t)=(1+h)×y _n,C (t+NΔt|t)−h×y _n,C (t+NΔt−Δt|t), h≧0. In this case, if h = 1, it corresponds to estimating that the difference between the generalized predicted value y _n,C at time t + NΔt and the generalized predicted value _{y n,C} at time t + NΔt-Δt continues for Δt .

As a result, as shown in FIG. 7B, the relative position N in the prediction time series storage unit 121 is updated to y _n,C (t+T _b +Δt|t)=y _n,C (t+NΔt+Δt|t).

Step S13: The time-series updating unit 124 reflects the influence of the latest operation change amount du(t+Δt). That is, as shown in FIG. 7(c), the time-series updating unit 124, for each generalized prediction value y _n,C (t+(m+1)Δt|t) from relative position m=0 to N, Add S(mΔt)du(t+Δt).

Specifically, S(0)du(t+Δt) is added to the generalized predicted value y _n,C (t+Δt|t) at the relative position m=0. Also, S(Δt)du(t+Δt) is added to the generalized predicted value y _n,C (t+2Δt|t) at the relative position m=1. The same applies when m≧3.

Step S14: The time series updating unit 124 sets each generalized prediction value y _n,C stored in the prediction time series storage unit 121 as a prediction time series at time t+Δt. That is, by the above S13, each generalized prediction value y _n,C from m = 0 to N is y _n,C (t + Δt + mΔt|t + Δt) = S (mΔt) du (t + Δt) + y _n,C (t + ( m+1)Δt|t).

Therefore, if t'=t+Δt, as shown in FIG _. A generalized predicted value of 0 can be expressed as y _n,C (t′|t′), and a generalized predicted value of relative position m=1 can be expressed as y _n,C (t′+Δt|t′). Similarly, when m≧2, it can be expressed as y _n,C (t′+mΔt|t′). As a result, the generalized prediction value y _n,C at time t stored in the prediction time series storage unit 121 is updated to the generalized prediction value y _n,C at time t+Δt.

<Updating process of past measured value time-series storage unit 122>
Next, processing for updating the past measured value time series storage unit 122 by the time series update unit 124 will be described with reference to FIG. FIG. 8 is a flowchart for explaining an example of update processing of the past measured value time series storage unit 122 . Hereinafter, a case will be described in which the past measured value time series at time t stored in the past measured value time series storage unit 122 is updated to the past measured value time series at time t+Δt.

Step S21: The time-series update unit 124 shifts the control amount measurement value _y0 stored in the past measurement value time-series storage unit 122 by time Δt.

For example, as _shown in FIG. 9A, y ₀ (t ₎ , y ₀ (t− _{Δt )} , . +Δt), y ₀ (t−T _e ) are stored in the past measurement value time series storage unit 122, and the respective relative positions m are 0, 1, . . . , N−1, N=N ₂ shall be In this case, the time-series updating unit 124 increments the relative position m of the control amount measurement value y0 below the relative position N− ₁ by +1. That is, the time-series updating unit 124 changes the relative position of y ₀ with the relative position m=N−1 to the relative position m=N, and changes the relative position of y ₀ with the relative position m=N−2 to the relative position m= Shift sequentially to N−1. Likewise, the time-series updating unit 124 sequentially shifts the relative position m of each _y0 until the relative position m= ₀ of y0 is shifted to the relative position m=1.

Step S22: The time-series updating unit 124 reflects the latest control amount measurement value y ₀ (t+Δt). That is, for example, as shown in FIG. 9B, the time series update unit 124 stores the latest time (that is, relative position = 0) in each storage area of the past measurement value time series storage unit 122. In contrast, the latest controlled variable measurement value y ₀ (t+Δt) is set.

Step S23: The time-series updating unit 124 updates the past measured value time-series storage unit 122. FIG. That is, for example, as shown in FIG. 9C, the time-series updating unit 124 sets each control amount measurement value y ₀ stored in the past measurement value time-series storage unit 122 to the time t as t′=t+Δt. '=t+Δt is the time series of past measured values. As a result, the past measured value time series at time t stored in the past measured value time series storage unit 122 is updated to the past measured value time series at time t′=t+Δt.

<Update processing of the past prediction time series storage unit 123>
Next, processing for updating the past prediction time series storage unit 123 by the time series update unit 124 will be described with reference to FIG. 10 . FIG. 10 is a flowchart for explaining an example of update processing of the past prediction time series storage unit 123. As shown in FIG. Hereinafter, a case will be described in which the past prediction time series at time t stored in the past prediction time series storage unit 123 is updated to the past prediction time series at time t+Δt.

Step S31: The time series updating unit 124 shifts the prediction value (first response prediction value) _yb stored in the past prediction time series storage unit 123 by time Δt.

For example, as shown in FIG. 11(a), y _b (t ₎ , y _b (t− _{Δt )} , _. +Δt), y _b (t−T _e ) are stored in the past prediction time series storage unit 123, and the respective relative positions m are 0, 1, . . . , N−1, N=N ₂ and In this case, the time-series updating unit 124 increments the relative position m of the predicted value _yb below the relative position N−1 by +1. That is, the time-series updating unit 124 changes the relative position of _yb with the relative position m=N−1 to the relative position m=N, and changes the relative position of _yb with the relative position m=N−2 to the relative position m= Shift sequentially to N−1. Likewise, the time-series updating unit 124 sequentially shifts the relative position m of each _yb until _{the yb} with the relative position m=0 is shifted to the relative position m=1.

Step S32: The time-series updating unit 124 reflects the latest control amount predicted value y _n,C (t+Δt|t+Δt)=y _b (t+Δt). That is, for example, as shown in FIG. 11B, the time series update unit 124 updates the storage area of the latest time (that is, relative position=0) among the storage areas of the past prediction time series storage unit 123. to set the latest control amount prediction value y _n,C (t+Δt|t+Δt)=y _b (t+Δt).

Step S<b>33 : The time series updating unit 124 updates the past prediction time series storage unit 123 . That is, for example, as shown in FIG. 11(c), the time series updating unit 124 sets each prediction value _yb stored in the past prediction time series storage unit 123 to time t'=t+Δt as t'=t+Δt. is the past forecast time series at As a result, the past prediction time series at time t stored in the past prediction time series storage unit 123 is updated to the past prediction time series at time t′=t+Δt.

<Operation of Operation Change Amount Calculation Unit 112>
Next, the operation of the operation change amount calculation unit 112 will be described with reference to FIG. 12 . 12A and 12B are diagrams for explaining an example of the operation of the operation change amount calculation unit 112. FIG.

As shown in FIG. 12 , when the correction target deviation e ^* (t) is input, the operation change amount calculation unit 112 multiplies the correction target deviation e ^* (t) by a predetermined control gain. A change amount du(t) is calculated, and the calculated du(t) is output. For example, when the integral gain _kI is used as the predetermined control gain, the operation change amount du(t) is calculated as du(t)= _kI ×e ^* (t).

However, if the result of multiplying the correction target deviation e ^* (t) by a predetermined control gain exceeds the upper limit value du _{max of} the amount of change in operation du (t). Similarly, if the result of multiplying the corrected target deviation e ^* (t) by a predetermined control gain is below the lower limit value du _min of the operation change amount du, the operation change amount calculation unit 112 calculates du _min as the operation change amount. Let du(t) be

<Control loop using control device 10>
Here, FIG. 13 shows an example of a case where the control loop of the plant 20 to be controlled is configured using the control device 10 according to the present embodiment (more precisely, a closed loop is approximated). FIG. 13 is a diagram showing an example of a control loop of a plant 20 to be controlled using the control device 10 according to the first embodiment.

In the example shown in FIG. 13, the control loop is configured as follows.

・Real plant transfer function of controlled plant 20: P ^* (s)
・Controller: K (s) = k _I × 1/s
A filter B that generates a first response estimate y _n,B (t): F _B (s)
A filter A that produces a second response estimate y _n,A (t): F _A (s; T _p )
- A filter C that generates a look-ahead target value r(t+T _p ): _FC (s; T _p )
・Filter D for generating past reading correction value e _b ^* (t): F _D (s; T _q )
In the example shown in FIG. 13 , the filter C corresponds to the target value prefetching section 101 . The controller K(s) corresponds to the operation change amount calculator 112 and the adder 113 . Note that Filter A, Filter B, and Filter D represent calculations performed by the response correction unit 111. For Filter A, the look-ahead length _Tp affects the output, and for Filter D, the past read length _Tq affects the output.

At this time, the filter C receives the target value time series {r(t)} for each control period T _c and outputs the look-ahead target value r(t+T _p ).

The controller adds the calculated operation change amount du to the operation amount current value _u0 at each control cycle _Tc , and outputs the operation amount u(t). At this time, the controller multiplies the correction target deviation e ^* by a predetermined control gain to calculate the operation change amount du, as described above. Thus, the controller can be approximated by an integrator, K(s)=k _I ×1/s. Note that e ^* (t)=r(t+T _p )−(y _n,A (t)+w(t)+ _eb ^* (t)) is input to the controller. where w(t)=y(t)-y _n,B (t).

Also, the actual plant transfer function P ^* (s) inputs u(t)+v(t), where v(t) is the disturbance, and outputs the controlled variable y(t).

In the example shown in FIG. 13, as an overall operation, the controlled variable y(t) is set to the estimated value of the controlled variable after the time _Tp (that is, the second response prediction value y _{n , A} (t) and the first response estimate y _n,B (t)). In addition, the filter D corrects the past prediction error (that is, the error between the controlled variable measured value y ₀ (t) and the controlled variable predicted value y _b (t) corresponding to this controlled variable measured value y ₀ ). corrected by the value e _b ^* (t).

Then, the corrected control amount is controlled by the controller so as to match the look-ahead target value r(t+T _p ). Therefore, the control device 10 according to the present embodiment corrects the past prediction error, and corrects the operation amount such that the control amount after the lapse of _Tp from the current time t matches the target value, and the conventional model prediction It can be said to be determined without optimization operations such as control.

Here, for example, if the plant response model is P(s) and the look-ahead length T _p is a sufficiently long value that the step response converges, F _A (s; T _p ) is the steady-state gain P(0) equal to For this reason, in the control device 10 according to the present embodiment, the look-ahead length _Tp can be set to any value including infinity. is preferred. In particular, it is preferable to set the value up to about 80% where the step response reaches the target value. Furthermore, for example, when the target value of the step response changes periodically, it is preferable to set the value sufficiently small in accordance with this change period.

When the look-ahead length _Tp is set to 0, the target deviation, which is the difference between the current target value and the current control amount, is used instead of the look-ahead target deviation, which is the difference between the future target value and the future predicted value. A correction target deviation corrected by the past reading correction value is calculated. On the other hand, when the look-ahead length _Tp is set to a sufficiently large value, the corrected target deviation is calculated by correcting the terminal response target deviation, which is the difference between the target value and the convergence value of the future predicted value, with the past-read correction value. will be

As described above, the control device 10 according to the present embodiment uses the arbitrarily set look-ahead length T _p and the past read length T _q to look-ahead the target value r (t + T _p ) at the look-ahead length T _p , The manipulated variable that causes the controlled variable to follow the prefetched target value r(t+ _Tp ) can be determined after correcting the past prediction error using the past read length _Tq . Therefore, according to the control device 10 according to the present embodiment, it is possible to effectively suppress the influence of the disturbance on the control amount.

In particular, a higher disturbance suppression effect can be obtained by appropriately setting the past reading length _Tq according to the time constant, duration, etc. of disturbances that can actually occur during plant operation. For example, a long past reading length _Tq is set when a disturbance with a long time constant is assumed, and a short past reading length _Tq is set when a disturbance with a short time constant is assumed.

Note that the past reading length _Tq may be set in advance in accordance with an assumed disturbance, or may be appropriately adjusted in accordance with the degree of process fluctuation during actual plant operation.

<Hardware Configuration of Control Device 10>
Next, the hardware configuration of the control device 10 according to this embodiment will be described with reference to FIG. 14 . FIG. 14 is a diagram showing an example of the hardware configuration of the control device 10 according to the first embodiment.

As shown in FIG. 14, the control device 10 according to the present embodiment includes an input device 201, a display device 202, an external I/F 203, a communication I/F 204, a ROM (Read Only Memory) 205, and a RAM 206. , a CPU 207 and an auxiliary storage device 208 . These pieces of hardware are connected to each other via a bus 209 so as to be able to communicate with each other.

The input device 201 is, for example, various buttons, a touch panel, a keyboard, a mouse, etc., and is used to input various operations to the control device 10 . The display device 202 is, for example, a display, and displays various processing results by the control device 10 . Note that the control device 10 may not have at least one of the input device 201 and the display device 202 .

An external I/F 203 is an interface with an external device. The external device includes a recording medium 203a and the like. Through the external I/F 203, the control device 10 can read from and write to the recording medium 203a. The recording medium 203a includes, for example, an SD memory card, a USB memory, a CD (Compact Disk), a DVD (Digital Versatile Disk), and the like. Note that one or more programs for realizing each functional unit of the control device 10 may be stored in the recording medium 203a.

A communication I/F 204 is an interface for the control device 10 to perform data communication with other devices. It should be noted that one or more programs that implement each functional unit of the control device 10 may be acquired (downloaded) from a predetermined server or the like via the communication I/F 204 .

A ROM 205 is a non-volatile semiconductor memory that can retain data even when the power is turned off. A RAM 206 is a volatile semiconductor memory that temporarily holds programs and data. The CPU 207 is an arithmetic device that reads programs and data from, for example, the auxiliary storage device 208 and the ROM 205 to the RAM 206 and executes various processes.

The auxiliary storage device 208 is, for example, an HDD (Hard Disk Drive), an SSD (Solid State Drive), or the like, and is a non-volatile memory that stores programs and data. The programs and data stored in the auxiliary storage device 208 include, for example, one or more programs that realize each functional unit of the control device 10, an operating system (OS) that is basic software, and various applications that run on the OS. programs, etc.

The control device 10 according to the present embodiment has the hardware configuration shown in FIG. 14, and thus can implement the various processes described above. Although FIG. 14 shows a hardware configuration example in which the control device 10 is realized by one computer, the control device 10 may be realized by a plurality of computers.

[Second embodiment]
Next, a second embodiment will be described. As shown in FIG. 15, the control device 10 according to the second embodiment does not use the look-ahead length _Tp . This is the same as when T _p =0 in the control device 10 according to the first embodiment.

By not using the look-ahead length _Tp , as described above, the control device 10 according to the second embodiment corrects the target deviation, which is the difference between the current target value and the current control amount, using the past-read correction value. A corrected target deviation can be calculated.

[Example]
Next, examples will be described. Subsequent examples explain the case where control object plant 20 is controlled by control device 10 concerning a first embodiment. In this embodiment, in regulation control with a constant target value, addition of correction (past reading correction value) using the past reading length _Tq improves responsiveness to disturbance.

First, FIG. 16 shows a step response model of the plant 20 to be controlled. As shown in FIG. 16, this embodiment uses a step response model in which the response reaches a steady value at about t=10-12.

Moreover, in this embodiment, the control gain is designed by the control system design support device disclosed in International Publication No. 2015/060149. Each piece of information (information indicating control specifications) input to this control system design support device is shown in FIGS. 17(a) to 17(c). FIG. 17A shows the upper and lower limits of the amount of change in operation at each operation time. FIG. 17(b) shows restrictions on the allowable target deviation at each control time. FIG. 17(c) is a control response waveform. A region R shown in FIG. 18 is output by inputting each of these pieces of information into the control system design support device. A region R is a graph representing a region where the correction target deviation e ^* and the operation change amount du satisfy the information shown in FIGS.

Here, in this embodiment, with reference to the region R shown in FIG. 18, a value of 1.746 passing through the vicinity of the center of this region R is selected as the control gain _KI . That is, du(t)=1.746*e ^* (t).

At this time, FIG. 19 shows the result of controlling the controlled plant 20 by the conventional PID control with the control cycle T _c =0.6. As shown in FIGS. 19A to 19C, the target value r(t) is fixed at 0, and the disturbance v(t) becomes a constant value after changing like a ramp function. At this time, the controlled variable y(t) temporarily increases to around 0.6 due to the influence of the disturbance v(t), and the manipulated variable u(t) drops to around -12 to prevent further increase. there is After the disturbance v(t) becomes a constant value, the controlled variable y(t) converges to the target value r(t)=0.

Moreover, the result of controlling the controlled plant 20 by the control device 10 according to the first embodiment with the control period T _c =0.6, the look-ahead length T _p =1.8, and the past read length T _q =1.8 is shown in FIG. As shown in FIGS. 20(a) to 20(c), the target value r(t) is fixed at 0 as in FIG. 19, and the disturbance v(t) is constant after changing like a ramp function. value. At this time, the controlled variable y(t) temporarily increases to around 0.5 due to the influence of the disturbance v(t), and the manipulated variable u(t) drops to around -12 to prevent further increase. there is After the disturbance v(t) becomes a constant value, the controlled variable y(t) converges to the target value r(t)=0.

Comparing FIG. 19 and FIG. 20, the peak value of the controlled variable y(t) is about 0.8 smaller in FIG. 20, and the influence of disturbance can be reduced by about 62%.

Next, the target value r(t) and the disturbance v(t) are the same as in _FIGS . 21. As shown in FIGS. 21(a) to 21(c), the controlled variable y(t) once increases to around 0.6 due to the influence of the disturbance v(t), and the manipulated variable u(t) becomes - It has fallen to around 12 and is preventing further increases. After the disturbance v(t) becomes a constant value, the controlled variable y(t) converges to the target value r(t)=0. When compared with FIG. 19, the change cycle of the manipulated variable is shorter, but the control response itself does not change that much.

Next, the target value r ₍ t) and the disturbance v(t) are _the same as in _FIG . 22 shows the result of controlling the plant 20 to be controlled by the control device 10 according to the first embodiment. As shown in FIGS. 22(a) to 22(c), the controlled variable y(t) once increases to around 0.3 due to the influence of the disturbance v(t), and the manipulated variable u(t) becomes - It has fallen to around 12 and is preventing further increases. After the disturbance v(t) becomes a constant value, the controlled variable y(t) converges to the target value r(t)=0.

Comparing FIG. 21 and FIG. 22, the peak value of the controlled variable y(t) is about 0.4 smaller in FIG. 22, and the influence of disturbance can be reduced by about 57%.

As described above, according to the control device 10 according to the present embodiment, correction using the past reading length _Tq can effectively suppress the influence of disturbance on the control amount.

The invention is not limited to the specifically disclosed embodiments above, but various modifications and changes are possible without departing from the scope of the claims.

REFERENCE SIGNS LIST 10 control device 20 plant to be controlled 101 target value look-ahead unit 102 measurement unit 103 differentiator 104 operation amount update unit 111 response correction unit 112 operation change amount calculation unit 113 adder

Claims (7)

A control device that outputs a manipulated variable for a controlled object and causes the controlled variable of the controlled object to follow a target value,
calculating a predicted value of the controlled variable based on the plant response model of the controlled object and the amount of change in the manipulated variable in the past up to the present, and storing the calculated predicted value as a time series; predicted value calculation means to be stored in
Measured value acquisition means for acquiring measured values of the control amount and storing the acquired measured values in time series in a second storage means;
When the time series of the predicted values, the time series of the measured values, and the past reading length T _q indicating the time width for estimating the influence of the disturbance that can actually occur are input, the current time is t, the latest prediction value y _b (t) among the prediction values stored in the first storage means and the prediction value y _b (t−T _q ) at the time indicated by the past reading length T _q ; Using the latest measured value y ₀ (t) among the measured values stored in the storage means of 2 and the measured value y ₀ (t−T _q ) at the time indicated by the past reading length T _q , the The past reading correction value e _b ^* (t) representing the prediction error due to the influence of the disturbance from the past reading length T _q to the current time t is expressed as e _b ^* (t)=(y ₀ (t)−y _b ( t))-(y ₀ (t-T _q )-y _b (t-T _q )));
calculating a corrected target deviation obtained by correcting either the difference between the target value and the controlled variable or the difference between the future target value and the future controlled variable using the past reading correction value _eb ^* (t); 2 correction value calculation means;
manipulated variable calculation means for calculating the new manipulated variable based on the corrected target deviation;
has
The control device, wherein the past reading length can be set at any time to any time width allowed in the storage areas of the first storage means and the second storage means. The manipulated variable calculation means is
The amount of change in the manipulated variable is calculated based on the corrected target deviation, and the new manipulated variable is calculated by adding the amount of change in the manipulated variable to the current manipulated variable. Item 1. The control device according to item 1. The manipulated variable calculation means is
3. The control device according to claim 2, wherein the amount of change in the manipulated variable is calculated by multiplying the corrected target deviation by a predetermined gain. The predicted value calculation means is
updating the time series of predicted values stored in the first storage means each time the amount of change in the manipulated variable is calculated;
The measured value acquisition means is
4. The control device according to claim 2, wherein the time series of the measured values stored in said second storage means is updated each time the amount of change in said manipulated variable is calculated. The predicted value calculation means is
Based on the plant response model of the controlled object and the amount of change in the past manipulated variable up to the present, a generalized predicted value at each time s at the current time t is calculated, and the calculated generalized predicted value 5. The control device according to any one of claims 1 to 4, wherein a generalized predicted value with s=t is calculated as the predicted value at time t. A computer that outputs a manipulated variable for a controlled object and causes the controlled variable of the controlled object to follow a target value,
calculating a predicted value of the controlled variable based on the plant response model of the controlled object and the amount of change in the manipulated variable in the past up to the present, and storing the calculated predicted value as a time series; a predicted value calculation procedure to be stored in the
A measured value acquisition procedure for acquiring measured values of the control amount and storing the acquired measured values in time series in a second storage means;
When the time series of the predicted values, the time series of the measured values, and the past reading length T _q indicating the time width for estimating the influence of the disturbance that can actually occur are input, the current time is t, the latest prediction value y _b (t) among the prediction values stored in the first storage means and the prediction value y _b (t−T _q ) at the time indicated by the past reading length T _q ; Using the latest measured value y ₀ (t) among the measured values stored in the storage means of 2 and the measured value y ₀ (t−T _q ) at the time indicated by the past reading length T _q , the The past reading correction value e _b ^* (t) representing the prediction error due to the influence of the disturbance from the past reading length T _q to the current time t is expressed as e _b ^* (t)=(y ₀ (t)−y _b ( t))-(y ₀ (t-T _q )-y _b (t-T _q )));
calculating a corrected target deviation obtained by correcting either the difference between the target value and the controlled variable or the difference between the future target value and the future controlled variable using the past reading correction value _eb ^* (t); 2 correction value calculation procedure;
a manipulated variable calculation procedure for calculating the new manipulated variable based on the corrected target deviation;
and run
A control method according to claim 1, wherein the past reading length can be set at any time to an arbitrary time width allowed in the storage areas of the first storage means and the second storage means. A computer that outputs a manipulated variable for a controlled object and causes the controlled variable of the controlled object to follow a target value,
calculating a predicted value of the controlled variable based on the plant response model of the controlled object and the amount of change in the manipulated variable in the past up to the present, and storing the calculated predicted value as a time series; a predicted value calculation procedure to be stored in the
A measured value acquisition procedure for acquiring measured values of the control amount and storing the acquired measured values in time series in a second storage means;
When the time series of the predicted values, the time series of the measured values, and the past reading length T _q indicating the time width for estimating the influence of the disturbance that can actually occur are input, the current time is t, the latest prediction value y _b (t) among the prediction values stored in the first storage means and the prediction value y _b (t−T _q ) at the time indicated by the past reading length T _q ; Using the latest measured value y ₀ (t) among the measured values stored in the storage means of 2 and the measured value y ₀ (t−T _q ) at the time indicated by the past reading length T _q , the The past reading correction value e _b ^* (t) representing the prediction error due to the influence of the disturbance from the past reading length T _q to the current time t is expressed as e _b ^* (t)=(y ₀ (t)−y _b ( t))-(y ₀ (t-T _q )-y _b (t-T _q )));
calculating a corrected target deviation obtained by correcting either the difference between the target value and the controlled variable or the difference between the future target value and the future controlled variable using the past reading correction value _eb ^* (t); 2 correction value calculation procedure;
a manipulated variable calculation procedure for calculating the new manipulated variable based on the corrected target deviation;
and
A program according to claim 1, wherein the past reading length can be set at any time to an arbitrary time width allowed in the storage areas of the first storage means and the second storage means.

Images