JP7517874B2 - Control device and control method - Google Patents

Description

The present invention relates to a control device and a control method that combines feedback control and feedforward control.

A method has been proposed in which a feedforward component is added to PID control, which is a typical feedback control (hereinafter referred to as feedforward + feedback control) (see Patent Document 1).

In order to improve the practicality of such feedforward + feedback control, particularly when applied to a heating device such as that shown in FIG. 16, the inventors have proposed a feedforward method (Patent Document 2) in which the lower limit value OL and upper limit value OH of the manipulated variable MV asymptotically converge to normal values, and a feedforward method (Patent Document 3) in which the feedforward variable MV_P asymptotically converges to zero.

The heating device in FIG. 16 is composed of a heat treatment furnace 100 for heating the workpiece to be treated, an electric heater 101, a temperature sensor 102 for measuring the temperature inside the heat treatment furnace 100, a temperature regulator 103 for controlling the temperature inside the heat treatment furnace 100, a power regulator 104, a power supply circuit 105, and a PLC (Programmable Logic Controller) 106 for controlling the entire heating device. The temperature regulator 103 calculates the manipulated variable MV so that the temperature PV (controlled variable) measured by the temperature sensor 102 matches the temperature set value SP. The power regulator 104 determines the power according to the manipulated variable MV, and supplies the determined power to the electric heater 101 via the power supply circuit 105.

The feedforward + feedback control proposed by the inventor in Patent Document 3 is close to typical feedforward control. The technology proposed by the inventor in Patent Document 3 will be explained using the block diagram of the control system in Figure 17. P in Figure 17 indicates the controlled object.

The manipulated variable calculation unit 201 receives a set value SP and a controlled variable PV as input, and calculates a manipulated variable MV (which in this invention is referred to as a basic manipulated variable MV) by performing a PID control calculation, for example, using the following transfer function equation, so that the controlled variable PV coincides with the set value SP.
MV=(100/Pb) {1+(1/Tis)+Tds}(SP-PV)
... (1)
Pb is the proportional band, Ti is the integral time, Td is the differential time, and s is the Laplace operator.

When an operating volume addition value FF_P (FF_P≠0) which is a target value of the addition amount of the feedforward portion for the basic operating volume MV is input, the addition volume calculation unit 204 calculates an operating volume addition value MV_P which gradually converges to zero after approaching the operating volume addition value FF_P. Specifically, the addition volume calculation unit 204 calculates the operating volume addition value MV_P by the following transfer function formula.
MV_P={Kxs/(1+Tfs) ² }FF_P...(2)

In equation (2), Tf is a parameter that specifies the time for the manipulated variable addition amount MV_P to gradually converge. Kx is a parameter that specifies the magnitude of the feedforward. An example of the change in the manipulated variable addition amount MV_P is shown in FIG. 18. In the example of FIG. 18, the manipulated variable addition value FF_P = 50%, the parameter Tf = 100 sec., and the parameter Kx = 275.

When an operating volume subtraction value FF_M (FF_M ≠ 0) which is a target value of the subtraction amount of the feedforward portion for the basic operating volume MV is input, the subtraction amount calculation unit 205 calculates an operating volume subtraction amount MV_M which gradually converges to a zero value after approaching the operating volume subtraction value FF_M. Specifically, the subtraction amount calculation unit 205 calculates the operating volume subtraction amount MV_M by the following transfer function formula.
MV_M={Kxs/(1+Tfs) ² }FF_M...(3)

Tf in formula (3) is a parameter that specifies the time for gradually converging the manipulated variable subtraction amount MV_M. The manipulated variable change unit 206 adds the manipulated variable additional amount MV_P calculated by the additional amount calculation unit 204 to the basic manipulated variable MV calculated by the manipulated variable calculation unit 201, and further subtracts the manipulated variable subtraction amount MV_M calculated by the subtraction amount calculation unit 205 from the added amount MV_P, thereby calculating the manipulated variable MV_F (referred to as the actual manipulated variable MV_F in the present invention).
MV_F=MV+MV_P-MV_M...(4)

The limit processing unit 207 performs lower limit processing to limit the actual operation amount MV_F calculated by the operation amount change unit 206 to a value equal to or greater than a predetermined operation amount lower limit value OL, and upper limit processing to limit the actual operation amount MV_F to a value equal to or less than a predetermined operation amount upper limit value OH. The actual operation amount MV_F' limited by this limit processing unit 207 is output to the control object P.

The technology proposed in Patent Document 3 makes it possible to reduce problems that occur with discontinuous control, such as changing the basic operating variable MV by a feedforward amount and returning the feedforward amount to 0% after a certain period of time has elapsed.

However, it is rare for an ordinary operator (control technology user) who is not an expert in control technology to be able to properly evaluate the results of execution of feedforward control. The added amount of the manipulated variable for the feedforward portion (in this invention, to distinguish it from Patent Document 3, it is called the feedforward amount MV_X) must be adjusted while being properly evaluated. If the feedforward amount MV_X is inappropriate, the effect of the feedforward control deteriorates. However, it is difficult for an ordinary operator who is not an expert in control technology to properly adjust the feedforward amount MV_X to improve the effect of the feedforward control.

JP 2007-102816 A JP 2019-101846 A JP 2019-101847 A

The present invention has been made to solve the above problems, and aims to provide a control device and control method that can rationally estimate an appropriate feedforward amount.

The control device of the present invention includes a manipulated variable calculation unit configured to calculate a first manipulated variable for each control cycle using a set value and a controlled variable as inputs, a feedforward calculation unit configured to calculate a first feedforward amount required to suppress a disturbance for each control cycle in response to an input of a trigger variable that is a significant value before the disturbance is applied and becomes a non-significant value after the disturbance is applied, a feedforward execution unit configured to add the first feedforward amount to the first manipulated variable calculated by the manipulated variable calculation unit for each control cycle , a manipulated variable output unit configured to output a second manipulated variable obtained by adding the first feedforward amount to the first manipulated variable to a control target, and a disturbance recovery when only the first manipulated variable is output to the control target as the second manipulated variable in a state in which the first feedforward amount is zero during a specified first trial period in which the trigger variable is input. a first estimation unit configured to estimate a difference between the second manipulated variable at the time of settling before a response and the second manipulated variable after settling of disturbance recovery as a first parameter for calculating a second feedforward amount required for suppressing a step disturbance; a total amount calculation unit configured to calculate a total amount that is an integrated value of the second feedforward amount for each control period until the second manipulated variable is settling when it is assumed that the second feedforward amount is output during the first trial period, based on the first parameter; and a total amount calculation unit configured to subtract the total amount of the second feedforward amount from a total amount that is an integrated value of an amount of change in the second manipulated variable for each control period until the second manipulated variable is settling during the first trial period , and to estimate a value obtained by multiplying the control period by a result of the subtraction as a second parameter for calculating a third feedforward amount required for suppressing an impulse disturbance. a reactive amount calculation unit configured to calculate an reactive amount which is an integrated value of the third feedforward amount calculated by the feedforward calculation unit for each control period until the settling of the second manipulated variable in a specified second trial period in which the trigger variable is input, the reactive amount being an integrated value of the amount which is invalidated due to either output saturation in which the second manipulated variable exceeds an manipulated variable upper limit value or output saturation in which the second manipulated variable falls below an manipulated variable lower limit value; and a calculation unit configured to calculate a ratio between a total amount which is an integrated value of the third feedforward amount calculated by the feedforward calculation unit for each control period until the settling of the second manipulated variable in the second trial period and a total amount which is an integrated value of the third feedforward amount for each control period that has not been invalidated due to the output saturation, based on the reactive amount, and to calculate a ratio such that a corresponding portion of the invalid amount is allocated within a range of the second manipulated variable that is not invalidated due to the output saturation. and a correction unit configured to estimate a correction value obtained by multiplying the second parameter by a rate as a parameter for correcting the third feedforward amount, wherein the feedforward calculation unit, when the trigger variable is input during the second trial period , sets the first feedforward amount to a sum of the second feedforward amount calculated for each control cycle based on the first parameter and the third feedforward amount calculated for each control cycle based on the second parameter, and, when the trigger variable is input after the second trial period, sets the first feedforward amount to a sum of the second feedforward amount calculated for each control cycle based on the first parameter and the third feedforward amount calculated for each control cycle using the parameter for correction estimated by the correction unit.

Moreover, in one configuration example of the control device of the present invention, the third feedforward amount is calculated by [Kx2s/{(1+ATfs)(1+(2.0-A)Tfs)}]FF_X, where the second parameter is Kx2, a time constant defining a time for gradually converging the third feedforward amount is Tf, a coefficient for adjusting the time constant is A, the trigger variable is FF_X, and a Laplace operator is s.
Furthermore, in one configuration example of the control device of the present invention, the correction unit calculates the ratio RH by Kx2/(Kx2-FF_N×dt), where FF_N is the reactive amount, dt is the control period which is a calculation period of the first manipulated variable and the first, second and third feedforward amounts, and RH is the ratio, and calculates, as parameters for correcting the third feedforward amount, a correction value of the second parameter Kx2 by Kx2×RH and a correction value of the time constant Tf by Tf×RH.
Moreover, in one configuration example of the control device of the present invention, the second feedforward amount is calculated by [Kx1/{(1+ATfs)(1+(2.0-A)Tfs)}]FF_X, where Kx1 is the first parameter, Tf is a time constant that specifies the time for gradually converging the second feedforward amount, A is a coefficient that adjusts the time constant, FF_X is the trigger variable, and s is a Laplace operator.

Moreover, one configuration example of the control device of the present invention further includes a limit processing unit that performs limit processing to limit a second manipulated variable calculated by the feedforward execution unit to a value equal to or greater than the manipulated variable lower limit value and equal to or less than the manipulated variable upper limit value, the manipulated variable output unit outputs the second manipulated variable that has been limited to a control target, the first estimator, the total amount calculator and the second estimator perform processing using the second manipulated variable that has been limited to as an input, and the reactive amount calculator calculates an output saturation or pre-limitation state in which the second manipulated variable before the limit processing exceeds the manipulated variable upper limit value among the third feedforward amounts calculated by the feedforward calculation unit for each control cycle until the second manipulated variable that has been limited to is settling during a specified second trial period in which the trigger variable is input. the correction unit calculates an invalid amount which is an integrated value of the amount invalidated by any of the output saturation below the operating amount lower limit value, and the correction unit calculates a ratio between a total amount which is an integrated value of each control cycle of the third feedforward amount calculated by the feedforward calculation unit until the second operating amount subjected to the limit processing is settled in the second trial period, and a total amount which is an integrated value of each control cycle of the third feedforward amount which was not invalidated by the output saturation, based on the invalid amount, and estimates a correction value obtained by multiplying the second parameter by the ratio as a parameter for correcting the third feedforward amount so that a corresponding portion of the invalid amount is distributed within a range of the second operating amount before the limit processing which is not invalidated by the output saturation .

The control method of the present invention includes a first step of calculating a first manipulated variable for each control cycle using a set value and a controlled variable as inputs, a second step of calculating a first feedforward amount required for suppressing the disturbance for each control cycle in response to an input of a trigger variable that is a significant value before the application of the disturbance and becomes a non-significant value after the application of the disturbance, a third step of adding the first feedforward amount to the first manipulated variable for each control cycle , a fourth step of outputting a second manipulated variable obtained by adding the first feedforward amount to the first manipulated variable to a control target, and a fourth step of calculating the second manipulated variable at the time of settling before a disturbance recovery response and the second manipulated variable after settling of disturbance recovery when only the first manipulated variable is output to the control target in a state in which the first feedforward amount is zero during a first prescribed trial period in which the trigger variable is input. a fifth step of estimating the difference between the second manipulated variable and the second manipulated variable as a first parameter for calculating a second feedforward amount required to suppress a step disturbance; a sixth step of calculating a total amount which is an integrated value of the second feedforward amount for each control period until the second manipulated variable is settling when it is assumed that the second feedforward amount has been output during the first trial period, based on the first parameter; a seventh step of subtracting the total amount of the second feedforward amount from a total amount which is an integrated value of the amount of change of the second manipulated variable for each control period until the second manipulated variable is settling during the first trial period, and multiplying the result of the subtraction by the control period, as a second parameter for calculating a third feedforward amount required to suppress an impulse disturbance; an eighth step of calculating an invalid amount which is an integrated value of the third feedforward amount calculated in the second step for each control cycle until the second manipulated variable is settling during the second trial period and which is invalidated due to either output saturation in which the second manipulated variable exceeds an upper limit value or output saturation in which the second manipulated variable is below a lower limit value; and calculating a ratio between a total amount which is an integrated value of the third feedforward amount calculated in the second step for each control cycle until the second manipulated variable is settling during the second trial period and a total amount which is an integrated value of the third feedforward amount in each control cycle which is not invalidated due to the output saturation, based on the invalid amount , and calculating a correction value obtained by multiplying the second parameter by the ratio so that a corresponding amount of the invalid amount is distributed within a range of the second manipulated variable which is not invalidated due to the output saturation. and a ninth step of estimating the third feedforward amount as a parameter for correcting the third feedforward amount, wherein the second step includes a step of setting, when the trigger variable is input during the second trial period , the sum of the second feedforward amount calculated for each control cycle based on the first parameter and the third feedforward amount calculated for each control cycle based on the second parameter as the first feedforward amount, and when the trigger variable is input after the second trial period, setting, when the trigger variable is input after the second trial period, the sum of the second feedforward amount calculated for each control cycle based on the first parameter and the third feedforward amount calculated for each control cycle using the parameter for the correction estimated in the ninth step as the first feedforward amount.

Moreover, in one configuration example of the control method of the present invention, the third feedforward amount is calculated by [Kx2s/{(1+ATfs)(1+(2.0-A)Tfs)}]FF_X, where the second parameter is Kx2, a time constant defining a time for gradually converging the third feedforward amount is Tf, a coefficient for adjusting the time constant is A, the trigger variable is FF_X, and a Laplace operator is s.
Furthermore, in one configuration example of the control method of the present invention, the ninth step is characterized in including a step of calculating the ratio RH by Kx2/(Kx2-FF_N×dt), where FF_N is the reactive amount, dt is the control period which is a calculation period of the first manipulated variable and the first, second and third feedforward amounts, and RH is the ratio, and calculating a correction value of the second parameter Kx2 by Kx2×RH and calculating a correction value of the time constant Tf by Tf×RH as parameters for correcting the third feedforward amount.
Moreover, in one configuration example of the control method of the present invention, the second feedforward amount is calculated by [Kx1/{(1+ATfs)(1+(2.0-A)Tfs)}]FF_X, where Kx1 is the first parameter, Tf is a time constant that specifies a time for gradually converging the second feedforward amount, A is a coefficient that adjusts the time constant, FF_X is the trigger variable, and s is a Laplace operator.

Moreover, one configuration example of the control method of the present invention includes a tenth step of performing limit processing to limit the second manipulated variable calculated in the third step to a value equal to or greater than the manipulated variable lower limit value and equal to or less than the manipulated variable upper limit value, the fourth step includes a step of outputting the limit-processed second manipulated variable to a control target, the fifth, sixth, and seventh steps perform processing using the limit-processed second manipulated variable as an input, and the eighth step includes a step of outputting an output saturation where the second manipulated variable before the limit processing exceeds the manipulated variable upper limit value or falls below the manipulated variable lower limit value among the third feedforward amounts calculated in the second step for each control cycle until the limit-processed second manipulated variable is settling during a specified second trial period in which the trigger variable is input. the ninth step includes a step of calculating an invalid amount which is an integrated value of the amount invalidated by any one of the above sums, and the ninth step includes a step of calculating a ratio between a total amount which is an integrated value of each control cycle of the third feedforward amount calculated by the second step until the second operation amount which has been limited processed is settling in the second trial period, and a total amount which is an integrated value of each control cycle of the third feedforward amount which has not been invalidated by the output saturation, based on the invalid amount, and estimating a correction value obtained by multiplying the second parameter by the ratio as a parameter for correcting the third feedforward amount so that a corresponding portion of the invalid amount is distributed within a range of the second operation amount before the limit processing which is not invalidated by the output saturation .

According to the present invention, the first estimation unit estimates a first parameter for calculating the second feedforward amount required to suppress step disturbances, the second estimation unit estimates a second parameter for calculating the third feedforward amount required to suppress impulse disturbances, the invalid amount calculation unit calculates an invalid amount which is the total amount of the third feedforward amount that has been invalidated due to output saturation, and the correction unit calculates the ratio between the total amount of the third feedforward amount calculated by the feedforward calculation unit and the total amount of the third feedforward amount that has not been invalidated due to output saturation based on the invalid amount, and estimates a parameter for correcting the third feedforward amount based on the ratio so that a corresponding amount of the invalid amount is distributed within the range of the second operation amount that is not invalidated due to output saturation. This makes it possible to rationally estimate an appropriate first feedforward amount based on the execution result (trial result) of disturbance recovery control. In this invention, it is possible to prevent the feedforward amount from being invalidated due to the influence of output saturation. As a result, this invention makes it possible to enhance the effect of feedforward control.

FIG. 1 is a block diagram showing the configuration of a control device according to an embodiment of the present invention. FIG. 2 is a block diagram of a control system according to an embodiment of the present invention. FIG. 3 is a flowchart illustrating the operation of the control device according to the embodiment of the present invention. FIG. 4 is a flowchart illustrating the operation of the control device according to the embodiment of the present invention. FIG. 5 is a diagram showing an example of a change in the feedforward amount according to the embodiment of the present invention. FIG. 6 is a diagram showing an example of a change in the feedforward amount according to the embodiment of the present invention. FIG. 7 is a diagram showing an example of a change in the feedforward amount according to the embodiment of the present invention. FIG. 8 is a diagram showing a simulation result in the case where feedforward control is not executed and only feedback control is executed in the embodiment of the present invention. FIG. 9 is a diagram showing a simulation result in the case where feedforward control is not executed and only feedback control is executed in the embodiment of the present invention. FIG. 10 is a diagram showing a simulation result in the case where feedforward control is not executed and only feedback control is executed in the embodiment of the present invention. FIG. 11 is a diagram showing a simulation result when feedforward control and feedback control are executed in the embodiment of the present invention. FIG. 12 is a diagram showing a simulation result of the feedforward amount when feedforward control and feedback control are executed in the embodiment of the present invention. FIG. 13 is a diagram showing a simulation result when feedforward control and feedback control are executed in the embodiment of the present invention. FIG. 14 is a diagram showing a simulation result of the feedforward amount when feedforward control and feedback control are executed in the embodiment of the present invention. FIG. 15 is a block diagram showing an example of the configuration of a computer that realizes a control device according to an embodiment of the present invention. FIG. 16 is a block diagram showing the configuration of the heating device. FIG. 17 is a block diagram of a feedback+feedforward control system. FIG. 18 is a diagram showing an example of a change in the added amount of the manipulated variable in the feedback+feedforward control.

[Principle of the Invention]
The present invention is applied to a disturbance recovery response, which is particularly applicable to the technique proposed in Patent Document 3. Disturbances to which the present invention is applied are roughly divided into two types: impulse disturbances and step disturbances.

An impulse disturbance is a disturbance in which the actual operating amount MV_F at settling time changes little before and after the disturbance recovery response, and is a temporary, shock-like disturbance. A step disturbance is a disturbance in which the actual operating amount MV_F at settling time changes significantly before and after the disturbance recovery response, and is a disturbance in which the equilibrium point itself, known as the settling state, changes. The technology proposed in Patent Document 3 is particularly suitable for impulse disturbances.

If there is only an impulse disturbance, the feedforward amount MV_X can be optimized using the total amount of change ΔMV_imp in the actual operation amount MV_F in disturbance recovery control by feedback control alone (control that returns the control amount PV to the set value SP) as a guide. Also, if there is only a step disturbance, the feedforward amount MV_X can be optimized using the actual operation amount difference ΔMV_stp before and after the disturbance recovery response by feedback control alone as a guide. Disturbances in which both impulse disturbances and step disturbances occur simultaneously (referred to as a mixed disturbance in this invention) are particularly difficult to adjust.

As a result of extensive research, the inventors have discovered that the feedforward amount MV_X can be estimated by the following steps (I) and (II), and have proposed a control method that follows steps (I) and (II) (Patent Application No. 2020-016891).

(I) At least execute disturbance recovery control using feedback control (or disturbance recovery control using only feedback control), and estimate the feedforward amount MV_X1 required for the step disturbance based on the actual control variable difference ΔMV_stp before and after the disturbance recovery response. Theoretically, the appropriate feedforward amount MV_X1 is proportional to the actual control variable difference ΔMV_stp.

(II) In the disturbance recovery control executed in (I) above, the total amount of change ΔMV_all in the actual control variable MV_F is calculated, and the feedforward amount MV_X2 required for the impulse disturbance is estimated based on the total amount of change ΔMV_imp in the actual control variable MV_F obtained by subtracting the total amount of feedforward amount MV_X1 required for the step disturbance estimated in (I) MV_X1_all. Theoretically, the appropriate feedforward amount MV_X2 is roughly proportional to the total amount of change ΔMV_imp in the actual control variable MV_F.

The above steps (I) and (II) make it possible to estimate the feedforward amount MV_X appropriately. However, because the feedforward amount as shown in FIG. 18 causes the actual operation amount MV_F to fluctuate significantly temporarily, there is a high possibility that output saturation will occur that exceeds the upper limit OH or lower limit OL of the actual operation amount MV_F. When such output saturation occurs, the expected feedforward amount cannot be secured. In other words, adjustments are particularly difficult when the feedforward amount required for impulse disturbances is affected by output saturation.

As a result of intensive research, the inventors have found that by adding the following procedure (III) to the above procedures (I) and (II), it is possible to estimate a more accurate feedforward amount MV_X than the control method proposed in Japanese Patent Application No. 2020-016891. Note that this invention is premised on the fact that the timing of applying the disturbance is known, and therefore the timing of applying the feedforward operation can be automatically determined. In addition, this invention is premised on the fact that, by employing, for example, PID control as feedback control, the parameter Tf for the time at which the feedforward amount MV_X converges can be automatically and uniquely determined based on the PID parameters, which are controller parameters.

(III) Disturbance recovery control is executed again using the feedforward control for step disturbances (feedforward amount MV_X1) estimated in the above steps (I) and (II), the feedforward control for impulse disturbances (feedforward amount MV_X2), and feedback control, and a correction amount MV_X2H for the feedforward amount MV_X2 is estimated based on the ratio RH calculated from the total amount of feedforward amounts invalidated due to output saturation (invalid amount FF_N) and the feedforward amount MV_X2 for impulse disturbances, so that the equivalent amount of the invalid amount FF_N is distributed within the range of the actual operating amount MV_F that is not invalidated due to output saturation.

In other words, in the present invention, in order to make the feedforward amount MV_X particularly accurate, it is sufficient to check the disturbance recovery response two times as described above. The fact that only two times is sufficient makes it a rational and efficient procedure.

[Example]
Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Fig. 1 is a block diagram showing the configuration of a control device according to an embodiment of the present invention. The control device includes a manipulated variable calculation unit 1 that calculates a basic manipulated variable MV (first manipulated variable) by PID calculation using a set value SP and a controlled variable PV as inputs, a feedforward calculation unit 2 that calculates a feedforward amount MV_X (first feedforward amount) required to suppress a disturbance in response to an input of a trigger variable FF_X that is a significant value before the application of the disturbance and becomes a non-significant value after the application of the disturbance, and a feedforward execution unit 3 that adds the feedforward amount MV_X to the basic manipulated variable MV calculated by the manipulated variable calculation unit 1 to obtain an actual manipulated variable MV_F (second manipulated variable).

The control device also includes a step response estimation unit 4 (first estimation unit) that estimates a parameter Kx1 (first parameter) for calculating a feedforward amount MV_X1 (second feedforward amount) required to suppress a step disturbance based on the actual operation amount difference ΔMV_stp before and after the disturbance recovery response when the actual operation amount MV_F including at least the basic operation amount MV is output to the control object during a specified first trial period in which the trigger variable FF_X is input, and a step response estimation unit 5 (first estimation unit) that estimates a parameter Kx1 (first parameter) for calculating a feedforward amount MV_X1 (second feedforward amount) required to suppress a step disturbance based on the actual operation amount difference ΔMV_stp before and after the disturbance recovery response when the actual operation amount MV_F including at least the basic operation amount MV is output to the control object during a specified first trial period in which the trigger variable FF_X is input, and The system includes a total amount calculation unit 5 that calculates the total amount MV_X1_all of the feedforward amount MV_X1 when the actual operation amount MV_F is equal to or greater than the first trial period based on a parameter Kx1, and an impulse response estimation unit 6 (second estimation unit) that subtracts the total amount of the feedforward amount MV_X1 from the total amount of change ΔMV_all of the actual operation amount MV_F during the first trial period, and estimates a parameter Kx2 (second parameter) for calculating the feedforward amount MV_X2 (third feedforward amount) required to suppress the impulse disturbance based on the subtraction result ΔMV_imp.

Furthermore, the control device includes an invalid amount calculation unit 7 that calculates an invalid amount FF_N, which is a total amount of the feedforward amount MV_X2 calculated by the feedforward calculation unit 2 during a specified second trial period in which the trigger variable FF_X is input, that is invalidated due to either output saturation in which the actual operation amount MV_F exceeds the operation amount upper limit value OH or output saturation in which the actual operation amount MV_F falls below the operation amount lower limit value OL, and calculates a ratio between the total amount of the feedforward amount MV_X2 calculated by the feedforward calculation unit 2 during the second trial period and the total amount of the feedforward amount MV_X2 that was not invalidated due to output saturation based on the invalid amount FF_N, and calculates the invalid amount FF_N based on the invalid amount FF_N. the step-response estimator 4, the impulse-response estimator 6, and the impulse-response corrector 8 for estimating parameters for correcting the feedforward amount MV_X2 based on a ratio so that a corresponding portion of the reactive amount FF_N is distributed within a range of the actual manipulated variable MV_F that is not subject to the influence of the reactive amount FF_N; an estimation result output unit 9 for setting the estimation results of the step-response estimator 4, the impulse-response estimator 6, and the impulse-response corrector 8 in the feedforward calculation unit 2; a limit processing unit 10 for performing limit processing to limit the actual manipulated variable MV_F calculated by the feedforward execution unit 3 to a value not less than a manipulated variable lower limit value OL and not more than a manipulated variable upper limit value OH; and a manipulated variable output unit 11 for outputting the limited actual manipulated variable MV_F' (second manipulated variable) to a controlled object.
2 is a block diagram of the control system of this embodiment, in which P indicates a controlled object.

Next, the operation of the control device of this embodiment will be described with reference to Figures 3 and 4. The set value SP (e.g., temperature set value) is set by an operator of the control device, etc., and input to the operation amount calculation unit 1 (Figure 3, step S100).

The controlled variable PV (e.g., a temperature measurement value) is measured by a measuring device (not shown) (e.g., a temperature sensor that measures the temperature of the heated object) and input to the manipulated variable calculation unit 1 (step S101 in FIG. 3).

The manipulated variable calculation unit 1 receives the set value SP and the controlled variable PV as input, and calculates the basic manipulated variable MV by performing a PID calculation, for example, using the following transfer function formula, so that the controlled variable PV coincides with the set value SP (step S102 in FIG. 3).
MV=(100/Pb) {1+(1/Tis)+Tds}(SP-PV)
...(5)
Pb is the proportional band, Ti is the integral time, Td is the differential time, and s is the Laplace operator.

The feedforward calculation unit 2 calculates the feedforward amount MV_X when feedforward control is executed, but if the trigger variable FF_X, which is 1 (significant value) before the application of the disturbance and 0 (insignificant value) after the application of the disturbance, is 0 (NO in step S103 in FIG. 3), it is assumed that only feedback control is performed and feedforward control is not executed, and the feedforward amount MV_X is set to 0 (step S104 in FIG. 3).

The feedforward execution unit 3 calculates an actual manipulated variable MV_F by adding the feedforward variable MV_X calculated by the feedforward calculation unit 2 to the basic manipulated variable MV calculated by the manipulated variable calculation unit 1 (step S105 in FIG. 3).
MV_F=MV+MV_X...(6)

Here, since MV_X = 0, MV_F = MV. The limit processing unit 10 performs lower limit processing to limit the actual manipulated variable MV_F calculated by the feedforward execution unit 3 to a value equal to or greater than a predetermined manipulated variable lower limit value OL, and upper limit processing to limit the actual manipulated variable MV_F to a value equal to or less than a predetermined manipulated variable upper limit value OH (step S106 in FIG. 3).
IF MV_F<OL THEN MV_F'=OL...(7)
IF MV_F>OH THEN MV_F'=OH...(8)
That is, when the actual manipulated variable MV_F is smaller than the manipulated variable lower limit OL, the limit processing unit 10 sets the actual manipulated variable MV_F'=OL, and when the actual manipulated variable MV_F is larger than the manipulated variable upper limit OH, the limit processing unit 10 sets the actual manipulated variable MV_F'=OH.

The manipulated variable output unit 11 outputs the actual manipulated variable MV_F' that has been subjected to limit processing by the limit processing unit 10 to the controlled object (step S107 in FIG. 3). The actual manipulated variable MV_F' is output to an operating unit (not shown) such as a heater or a valve. In the case of a heater, the actual output destination of the actual manipulated variable MV_F' is a power regulator (not shown) that supplies power to the heater.

The control device executes steps S100 to S107 in FIG. 3 every control cycle until control is terminated, for example, by an operator's instruction (YES in step S108 in FIG. 3).

Next, we will explain the operation when the trigger variable FF_X becomes 1 (a significant value). In this embodiment, it is assumed that the timing of applying the disturbance is known, and therefore the timing of executing feedforward control can be automatically determined. In a system to which the control device of this embodiment is applied, it is assumed that the trigger variable FF_X = 1 is automatically input from an external device to the control device at a specified timing in order to suppress disturbances that are anticipated during control.

For example, in a pharmaceutical manufacturing device, there is a situation where the temperature inside the pharmaceutical manufacturing furnace fluctuates when the door of the furnace is opened. In this case, the control device (external device) that controls the temperature of the furnace sends a trigger variable FF_X = 1 to the control device of this embodiment at a point in time before the door of the furnace opens (the timing of applying the disturbance).

Similarly, in a reflow furnace with a constant set value SP (temperature set value), there is a situation where the temperature fluctuates due to the periodic introduction of printed circuit boards to be soldered. In this case, the control device (external device) that controls the transportation of the printed circuit boards sends a trigger variable FF_X = 1 to the control device of this embodiment at a point in time prior to the timing at which the printed circuit boards are introduced into the reflow furnace (timing at which the disturbance is applied).

Furthermore, the external device sets the trigger variable FF_X to 0 (insignificant value) a predetermined time after the application of the disturbance ends. The timing for setting the trigger variable FF_X to 0 needs to be set to a timing later than the timing at which the actual manipulated variable MV_F′ is settled after the application of the disturbance.
The extent to which the trigger variable FF_X should be set to 1 relative to the timing of application of the disturbance will be described later.

When the trigger variable FF_X changes from 0 to 1 (YES in step S103 in FIG. 3), during the first disturbance recovery control period (first trial period) in which the trigger variable FF_X becomes 1 (YES in step S109 in FIG. 3), the step correspondence estimation unit 4 determines that only feedback control is performed and no feedforward control is performed, as when FF_X=0, and causes the feedforward calculation unit 2 to output a feedforward amount MV_X=0 (step S110 in FIG. 3).

The processing of steps S111 to S113 in FIG. 3 is the same as the processing of steps S105 to S107. Then, the step response estimation unit 4 and the impulse response estimation unit 6 store the actual operation amount MV_F' calculated by the feedforward execution unit 3 and processed by the limit processing unit 10 together with the time (step S114 in FIG. 3).

When the step correspondence estimation unit 4 determines that the actual operation amount MV_F' has settled in the first disturbance recovery control when the trigger variable FF_X becomes 1 (YES in step S115 in FIG. 3), it calculates the actual operation amount difference ΔMV_stp before and after the disturbance recovery response (step S116 in FIG. 3). The actual operation amount difference ΔMV_stp is the difference between the settling value of the actual operation amount MV_F' when it is determined in step S115 that it has settled and the actual operation amount MV_F' before the application of the disturbance (the settling value of the actual operation amount MV_F' immediately before the trigger variable FF_X becomes 1).

Then, the step correspondence estimator 4 estimates an appropriate value of the parameter Kx1 for calculating the feedforward amount MV_X1 necessary for suppressing the step disturbance according to the following equation (step S117 in FIG. 3).
Kx1=ΔMV_stp...(9)
That is, in this embodiment, the step correspondence estimator 4 sets the actual manipulated variable difference ΔMV_stp as the parameter Kx1 as it is.

The estimation result output unit 9 sets the value of the parameter Kx1 estimated by the step corresponding estimation unit 4 to the feedforward calculation unit 2 (step S118 in FIG. 3). In this way, the estimated value of the parameter Kx1 may be automatically set in the feedforward calculation unit 2, or the value of the parameter Kx1 may be displayed on a screen for the operator to confirm and then manually set in the feedforward calculation unit 2 by the operator. Setting the parameter Kx1 enables the feedforward calculation unit 2 to sequentially calculate the feedforward amount MV_X1, as described below.

The feedforward amount MV_X1 required for suppressing the step disturbance is calculated as follows:
MV_X1
= [Kx1/{(1+ATfs)(1+(2.0-A)Tfs)}]FF_X
...(10)

The feedforward amount MV_X2 required for suppressing the impulse disturbance is calculated as follows:
MV_X2
= [Kx2s/{(1+ATfs)(1+(2.0-A)Tfs)}]FF_X
...(11)

In equations (10) and (11), Tf is a parameter (time constant) that specifies the time for gradually converging the feedforward amounts MV_X1 and MV_X2. The time constant Tf can be set in advance based on information about the feedback control system, etc.

In the formulas (10) and (11), A is a coefficient (A is a real number greater than 0) that adjusts the balance of the time constant Tf. In this embodiment, A=1.0.
As described above, the parameter Kx1 is a value estimated by the step correspondence estimator 4, and is substantially a target value of the feedforward addition/subtraction amount for the basic manipulated variable MV. According to equation (10), the feedforward amount MV_X1 gradually converges from zero to the value of the parameter Kx1.

Furthermore, Kx2 in formula (11) is a parameter for calculating the feedforward amount MV_X2 required to suppress impulse disturbances, and is substantially a target value for the total amount of addition and subtraction of the feedforward portion for the basic manipulated variable MV (the integrated value of the addition and subtraction amounts for each control cycle). According to formula (11), after the total amount (the integrated value for each control cycle) of the feedforward amount MV_X2 approaches the value of the parameter Kx2, the feedforward amount MV_X2 itself gradually converges to zero.
It should be noted that the formulas (10) and (11) are stored in the feedforward calculation unit 2.

Next, the total amount calculation unit 5 calculates the total amount MV_X1_all (the integrated value of the addition and subtraction amounts for each control cycle) of the feedforward amount MV_X1 until the actual manipulated variable MV_F' is settled, assuming that the feedforward amount MV_X1 required to suppress the step disturbance is output in the first disturbance recovery control when the trigger variable FF_X becomes 1 (step S119 in FIG. 3). Since the value of the parameter Kx1 has been estimated by the step correspondence estimation unit 4, the feedforward amount MV_X1 output from the point when the trigger variable FF_X becomes 1 can be calculated by the above formula (10).

The impulse response estimation unit 6 calculates the total amount of change ΔMV_all (the integrated value of the amount of change in each control cycle) of the actual operation amount MV_F' in the first disturbance recovery control when the trigger variable FF_X becomes 1, relative to the actual operation amount MV_F' before the application of the disturbance (the set value of the actual operation amount MV_F' immediately before the trigger variable FF_X becomes 1), based on the stored time series data of the actual operation amount MV_F' (step S120 in FIG. 3). Note that in this embodiment, in the case of the first disturbance recovery control when the trigger variable FF_X becomes 1, the actual operation amount MV_F includes only the basic operation amount MV, so MV_F = MV.

Next, the impulse-responsive estimator 6 calculates the total amount of change ΔMV_imp by subtracting the total amount MV_X1_all of the feedforward amount MV_X1 calculated by the total amount calculator 5 from the total amount of change ΔMV_all of the actual manipulated variable MV_F' calculated in step S120 (step S121 in FIG. 3).
ΔMV_imp=ΔMV_all−MV_X1_all (12)

Furthermore, the impulse response estimation unit 6 estimates an appropriate value of a parameter Kx2 for calculating a feedforward amount MV_X2 necessary for suppressing the impulse disturbance, as shown in the following equation (step S122 in FIG. 3).
Kx2=ΔMV_imp×dt...(13)

In equation (13), dt is the control period (the calculation period for the basic manipulated variable MV and the feedforward variables MV_X1, MV_X2, and MV_X) in seconds.

The estimation result output unit 9 sets the value of the parameter Kx2 estimated by the impulse response estimation unit 6 in the feedforward calculation unit 2 (step S123 in FIG. 3). In this way, the estimated value of the parameter Kx2 may be automatically set in the feedforward calculation unit 2, or the value of the parameter Kx2 may be displayed on a screen for the operator to confirm and then manually set in the feedforward calculation unit 2 by the operator. Setting the parameter Kx2 enables the feedforward calculation unit 2 to sequentially calculate the feedforward amount MV_X2.

The process from step S109 to step S123 corresponds to the above-mentioned steps (I) and (II). This makes it possible to estimate appropriate values for the parameters Kx1 and Kx2 that are practically acceptable. Note that the process from step S116 to step S123 only needs to be performed once when the actual manipulated variable MV_F' has settled.
As described above, the trigger variable FF_X returns from 1 to 0 at a prescribed timing after the application of the disturbance ends.

Next, when the trigger variable FF_X changes from 0 to 1 (YES in step S103), the feedforward calculation unit 2 executes feedforward + feedback control during the second disturbance recovery control period (second trial period) in which the trigger variable FF_X becomes 1 (YES in step S124 in FIG. 4), and calculates the feedforward amount MV_X1 required to suppress the step disturbance using equation (10) (step S125 in FIG. 4). At the same time as calculating this feedforward amount MV_X1, the feedforward calculation unit 2 calculates the feedforward amount MV_X2 required to suppress the impulse disturbance using equation (11) (step S126 in FIG. 4).

Then, the feedforward calculation unit 2 calculates the feedforward amount MV_X by adding the feedforward amounts MV_X1 and MV_X2 according to the following formula (step S127 in FIG. 4).
MV_X=MV_X1+MV_X2...(14)

The processing in steps S128 to S130 in FIG. 4 is the same as the processing in steps S105 to S107. The reactive amount calculation unit 7 and the impulse-compatible correction unit 8 store the actual operation amount MV_F' calculated by the feedforward execution unit 3 and processed by the limit processing unit 10, and the feedforward amounts MV_X2 and MV_X calculated by the feedforward calculation unit 2, together with the time (step S131 in FIG. 4).

When the actual control input MV_F' settles during the second disturbance recovery control in which the trigger variable FF_X becomes 1 (YES in step S132 in FIG. 4), the invalid amount calculation unit 7 calculates the invalid amount FF_N, which is the total amount of the feedforward amount MV_X2 calculated by the feedforward calculation unit 2 until the actual control input MV_F' settles, which is invalidated due to either output saturation in which the actual control input MV_F exceeds the control input upper limit value OH or output saturation in which the actual control input MV_F falls below the control input lower limit value OL (step S133 in FIG. 4).

Specifically, in the case of output saturation where the actual manipulated variable MV_F exceeds the manipulated variable upper limit value OH, the reactive quantity calculation unit 7 may take the integrated value of the positive difference of the difference MV_X-OH between the feedforward variable MV_X obtained for each control cycle and the manipulated variable upper limit value OH as the reactive quantity FF_N. In addition, in the case of output saturation where the actual manipulated variable MV_F is below the manipulated variable lower limit value OL, the reactive quantity calculation unit 7 may take the integrated value of the negative difference of the difference MV_X-OL between the feedforward variable MV_X obtained for each control cycle and the manipulated variable upper limit value OH as the reactive quantity FF_N.

In addition, in the case of output saturation, not only the feedforward amount MV_X2 but also the basic operation amount MV or the feedforward amount MV_X1 may be partially invalidated. However, in the feedforward + feedback control of the present invention, most of the steep rise or fall of the actual operation amount MV_F that causes output saturation is due to the feedforward amount MV_X, and most of that is due to the feedforward amount MV_X2. The basic operation amount MV and the feedforward amount MV_X1 change later than the feedforward amount MV_X2. Therefore, the invalid amount FF_N may be calculated simply by a simple comparison of the feedforward amount MV_X and the operation amount upper limit values OH and OL as described above.

Next, the impulse-responsive correction unit 8 calculates the ratio RH between the total amount Kx2 of the feedforward amount MV_X2 calculated by the feedforward calculation unit 2 until the actual operating amount MV_F' is settled and the total amount of the feedforward amount MV_X2 that has not been invalidated due to output saturation among the total amount Kx2, based on the invalid amount FF_N, as shown in the following equation (step S134 in FIG. 4).
RH=Kx2/(Kx2-FF_N×dt)...(15)

The total amount of the feedforward amount MV_X2 coincides with the parameter Kx2 set in the feedforward calculation unit 2. Therefore, in the calculation of equation (15), the total amount is not calculated by integrating the feedforward amount MV_X2, but the parameter Kx2 is set as the total amount of the feedforward amount MV_X2.

Then, the impulse correction unit 8 estimates the correction amount of the feedforward amount MV_X2 so that the equivalent of the invalid amount FF_N is allocated within the range of the actual manipulated variable MV_F that is not invalidated due to output saturation. Specifically, the impulse correction unit 8 estimates the correction value Kx2H of the parameter Kx2 and the correction value TfH of the time constant Tf based on the ratio RH (step S135 in FIG. 4).
Kx2H=Kx2×RH...(16)
TfH=Tf×RH...(17)

The estimation result output unit 9 updates the value of the parameter Kx2 set in the feedforward calculation unit 2 to the correction value Kx2H newly estimated by the impulse response correction unit 8, and updates the value of the time constant Tf set in the feedforward calculation unit 2 to the correction value TfH newly estimated by the impulse response correction unit 8 (step S136 in FIG. 4). In this way, the values of the parameter Kx2 and the time constant Tf may be automatically updated, or the values of the correction values Kx2H and TfH may be displayed on the screen for the operator to confirm and then the operator may manually update them.

The processing from steps S124 to S136 corresponds to the above procedure (III). Note that the processing from steps S133 to S136 only needs to be performed once when the actual manipulated variable MV_F' has stabilized.

The processing of steps S137 to S142 in the third or subsequent disturbance recovery control operations in which the trigger variable FF_X becomes 1 is the same as steps S125 to S130, and therefore will not be described here.

Figure 5 shows an example of the change in the feedforward amount MV_X1, Figure 6 shows an example of the change in the feedforward amount MV_X2, and Figure 7 shows an example of the change in the feedforward amount MV_X. In the examples of Figures 5 to 7, Kx1 = 20.0, Kx2 = 3304.0, and Tf = 15.0 sec. The total amount (integrated value for each control cycle) mentioned above is the amount equivalent to the area enclosed by the curves in Figures 5 to 7.

The effect of this embodiment will be verified by a simulation below. In the following example, the controlled object is a control system that can be approximated by a first-order lag transfer function with a process gain of 10.0, a process time constant of 400.0 sec, and a process dead time of 20.0 sec. In other words, the model equation Gp of the controlled object can be written as follows:
Gp=10.0exp(-20.0s)/(1+400.0s)...(18)

In addition, the upper limit value OH of the actual manipulated variable MV_F is set to 75% so that output saturation occurs due to the feedforward control, which means that the amount exceeding 75% is invalidated.
The PID parameters set in the operation amount calculation unit 1 are proportional band Pb=60%, integral time Ti=120.0 sec, and differential time Td=10.0 sec. The control period dt is 1.0 sec.

The start timing of feedforward control, i.e., the timing at which the trigger variable FF_X changes from 0 to 1, can be set based on the differential time Td. Specifically, it is appropriate that the start timing is αTd before the time point at which the disturbance is applied (the coefficient α is a real number greater than 0, for example 0.7). When the differential time Td = 10.0 sec., the start timing of feedforward control is 7.0 sec. before the time point at which the disturbance is applied. However, the coefficient α is a value that can be fine-tuned as appropriate.

Similarly, the time constant Tf can be set based on the differential time Td. Specifically, it is appropriate for Tf to be βTd (coefficient β is a real number greater than 0, e.g., 0.4). When differential time Td is 10.0 sec., the time constant Tf is 4.0 sec. However, coefficient β is a value that can be fine-tuned as appropriate.

Figures 8 to 10 are diagrams showing examples of disturbance recovery responses to which this embodiment is applied, and are diagrams showing examples of changes in the controlled variable PV and the actual manipulated variable MV_F' (= manipulated variable MV) when temperature control is performed using only feedback control without performing feedforward control.

Figure 8 shows the simulation results of the disturbance recovery response by feedback control when an impulse disturbance is applied at 100 seconds. The actual manipulated variable MV_F' (= manipulated variable MV) at the time of settling remains almost unchanged before and after the disturbance recovery response.

Figure 9 shows the simulation results of the disturbance recovery response by feedback control when a step disturbance is applied at 100 seconds. There is a noticeable change in the actual manipulated variable MV_F' (= manipulated variable MV) during settling before and after the disturbance recovery response.

Figure 10 shows the simulation results of the disturbance recovery response by feedback control when a mixed disturbance is applied at 100 seconds. This simulation result corresponds to the result of the first control when the trigger variable FF_X becomes 1. Because the mixed disturbance includes a step disturbance, a significant change occurs in the actual manipulated variable MV_F' (= manipulated variable MV) at the time of settling before and after the disturbance recovery response.

In the example of FIG. 10, the actual control amount MV_F' before the disturbance application is 30.0%, while the set value of the actual control amount MV_F' after the disturbance recovery response is 45.0%, so the actual control amount difference ΔMV_stp = 15.0%. The step correspondence estimation unit 4 estimates the parameter Kx1 = 15.0 using equation (9) (step S117 in FIG. 3).

In the example of FIG. 10, the total change amount ΔMV_imp obtained by subtracting the total amount of feedforward amount MV_X1 from the total change amount ΔMV_all of the actual manipulated variable MV_F' is 768.2%. The impulse response estimation unit 6 estimates the parameter Kx2 to be 768.2 using equation (13) (step S122 in FIG. 3).

Figure 11 shows the simulation results of the disturbance recovery response by feedforward control and feedback control when a mixed disturbance is applied at 100 sec., and Figure 12 shows the simulation results of the feedforward amount MV_X in the case of Figure 11. These simulation results correspond to the results of the second control when the trigger variable FF_X becomes 1.

The parameters Kx1 = 15.0 and Kx2 = 768.2 are set by the first control when the trigger variable FF_X becomes 1. The timing when the feedforward control starts, i.e., the timing when the trigger variable FF_X changes from 0 to 1, is 93 seconds before the disturbance is applied, 7.0 seconds before. As mentioned above, the time constant Tf = 4.0 seconds.

In the example of FIG. 11, the total change amount ΔMV_imp obtained by subtracting the total amount of feedforward amount MV_X1 from the total change amount ΔMV_all of the actual manipulated variable MV_F' is 768.2%. In addition, the invalid amount calculation unit 7 detects the invalid amount FF_N = 142.3 (step S133 in FIG. 4).

The impulse correction unit 8 calculates the ratio RH = 768.2/(768.2-142.3 x 1.0) = 1.227 using equation (15) (step S134 in FIG. 4). Furthermore, the impulse correction unit 8 estimates the correction value Kx2H of the parameter Kx2 = 768.2 x 1.227 = 942.6 using equation (16), and estimates the correction value TfH of the time constant Tf = 4.0 x 1.227 = 4.9 using equation (17) (step S135 in FIG. 4).

Figure 13 shows the simulation results of the disturbance recovery response by feedforward control and feedback control when a mixed disturbance is applied at 100 sec., and Figure 14 shows the simulation results of the feedforward amount MV_X in the case of Figure 13. These simulation results correspond to the results of the third and subsequent controls in which the trigger variable FF_X becomes 1 after Kx1 = 15.0, Kx2 = Kx2H = 942.6, and Tf = TfH = 4.9 sec. are set by the first and second controls in which the trigger variable FF_X becomes 1.

To compare the control results, FIG. 11 shows the results of feedforward control affected by output saturation, while FIG. 13 shows the results of feedforward control corrected in this embodiment. In the example of FIG. 11, the control amount PV drops significantly due to the mixed disturbance, whereas in the example of FIG. 13, it can be seen that the effect of the mixed disturbance is suppressed.

In this embodiment, an example of a disturbance in which the controlled variable PV decreases relative to the set value SP has been described, but the present invention can also handle a disturbance in which the controlled variable PV increases relative to the set value SP. When a disturbance occurs in which the controlled variable PV increases, the actual controlled variable difference ΔMV_stp, the total amount of change ΔMV_all of the actual controlled variables MV_F', the total amount MV_X1_all of the feedforward amount MV_X1, the parameters Kx1, Kx2, the feedforward amounts MV_X1, MV_X2, and the reactive amount FF_N become negative values.

In addition, in this embodiment, a case is described in which the feedforward amount MV_X1 is not 0 in the second or subsequent disturbance recovery control in which the trigger variable FF_X becomes 1, but MV_X1 may be set to 0. Whether or not MV_X1 is set to 0 may be appropriately determined depending on the disturbance.

The control device described in this embodiment can be realized by a computer equipped with a CPU (Central Processing Unit), a storage device, and an interface, and a program that controls these hardware resources. An example of the configuration of this computer is shown in FIG. 15.

The computer includes a CPU 300, a storage device 301, and an interface device (I/F) 302. A temperature sensor or a power regulator, for example, is connected to the I/F 302. In such a computer, a program for implementing the control method of this embodiment is stored in the storage device 301. The CPU 300 executes the processing described in this embodiment according to the program stored in the storage device 301.

The present invention can be applied to control devices.

1...operational quantity calculation section, 2...feedforward calculation section, 3...feedforward execution section, 4...step response estimation section, 5...total quantity calculation section, 6...impulse response estimation section, 7...reactive quantity calculation section, 8...impulse response correction section, 9...estimation result output section, 10...limit processing section, 11...operational quantity output section.

Claims (10)

a manipulation amount calculation unit configured to calculate a first manipulation amount for each control period using a set value and a control amount as input;
a feedforward calculation unit configured to calculate a first feedforward amount necessary for suppressing a disturbance for each control period in response to an input of a trigger variable that is a significant value before application of the disturbance and is a non-significant value after application of the disturbance;
a feedforward execution unit configured to add the first feedforward amount to the first manipulated variable calculated by the manipulated variable calculation unit for each control period ;
a manipulated variable output unit configured to output a second manipulated variable obtained by adding the first feedforward variable to the first manipulated variable to a control target;
a first estimation unit configured to estimate, as a first parameter for calculating a second feedforward amount necessary for suppressing a step disturbance, a difference between the second manipulated variable at a time of settling before a disturbance recovery response and the second manipulated variable after settling of a disturbance recovery when only the first manipulated variable is output to a control target as the second manipulated variable in a state in which the first feedforward amount is zero during a first prescribed trial period in which the trigger variable is input;
a total amount calculation unit configured to calculate, based on the first parameter, a total amount that is an integrated value of the second feedforward amount for each control period until the second manipulated variable is settled, assuming that the second feedforward amount is output during the first trial period;
a second estimation unit configured to subtract a total amount of the second feedforward amount from a total amount that is an integrated value of an amount of change in each control cycle of the second manipulated variable until the second manipulated variable is settled during the first trial period, and to estimate a value obtained by multiplying the control cycle by a result of the subtraction as a second parameter for calculating a third feedforward amount required to suppress impulse disturbances;
an invalid amount calculation unit configured to calculate an invalid amount which is an integrated value of the third feedforward amount calculated by the feedforward calculation unit for each control cycle until the second manipulated variable is settling during a specified second trial period in which the trigger variable is input, the invalid amount being invalidated due to either output saturation in which the second manipulated variable exceeds an upper limit value or output saturation in which the second manipulated variable falls below a lower limit value;
a correction unit configured to calculate a ratio between a total amount that is an integrated value of each control cycle of the third feedforward amount calculated by the feedforward calculation unit until the second manipulated variable is settling during the second trial period and a total amount that is an integrated value of each control cycle of the third feedforward amount that has not been invalidated due to the output saturation, based on the invalid amount, and to estimate a correction value obtained by multiplying the ratio by the second parameter as a parameter for correcting the third feedforward amount so that a corresponding portion of the invalid amount is allocated within a range of the second manipulated variable that is not invalidated due to the output saturation,
the feedforward calculation unit, when the trigger variable is input during the second trial period , sets the first feedforward amount to a sum of the second feedforward amount calculated for each control period based on the first parameter and the third feedforward amount calculated for each control period based on the second parameter, and, when the trigger variable is input after the second trial period, sets the first feedforward amount to a sum of the second feedforward amount calculated for each control period based on the first parameter and the third feedforward amount calculated for each control period using a parameter for correction estimated by the correction unit. 2. The control device according to claim 1,
The third feedforward amount is calculated by [Kx2s/{(1+ATfs)(1+(2.0-A)Tfs)}]FF_X, where Kx2 is the second parameter, Tf is a time constant that specifies a time for gradually converging the third feedforward amount, A is a coefficient that adjusts the time constant, FF_X is the trigger variable, and s is a Laplace operator. 3. The control device according to claim 2,
the correction unit calculates the ratio RH by Kx2/(Kx2-FF_N×dt), where FF_N is the reactive amount, dt is the control period which is a calculation period of the first manipulated variable and the first, second and third feedforward amounts, and RH is the ratio, and calculates, as parameters for correcting the third feedforward amount, a correction value of the second parameter Kx2 by Kx2×RH and a correction value of the time constant Tf by Tf×RH. The control device according to any one of claims 1 to 3,
The second feedforward amount is calculated by [Kx1/{(1+ATfs)(1+(2.0-A)Tfs)}]FF_X, where Kx1 is the first parameter, Tf is a time constant that specifies a time for gradually converging the second feedforward amount, A is a coefficient that adjusts the time constant, FF_X is the trigger variable, and s is a Laplace operator. The control device according to any one of claims 1 to 4,
a limit processing unit that performs limit processing to limit the second manipulated variable calculated by the feedforward execution unit to a value that is equal to or greater than the manipulated variable lower limit value and equal to or less than the manipulated variable upper limit value,
the operation amount output unit outputs the second operation amount subjected to the limit processing to a control target;
the first estimator, the total amount calculator, and the second estimator perform processing using the second manipulated variable subjected to the limit process as an input;
the invalid amount calculation unit calculates an invalid amount which is an integrated value of the third feedforward amount calculated by the feedforward calculation unit for each control period until the second manipulated variable subjected to the limit processing is settling during a specified second trial period in which the trigger variable is input, the invalid amount being an integrated value of the amount invalidated due to either output saturation in which the second manipulated variable before the limit processing exceeds the manipulated variable upper limit value or output saturation in which the second manipulated variable before the limit processing falls below the manipulated variable lower limit value;
The control device is characterized in that the correction unit calculates a ratio between a total amount that is an integrated value of each control cycle of the third feedforward amount calculated by the feedforward calculation unit until the second operation amount that has been limited processed is settling during the second trial period, and a total amount that is an integrated value of each control cycle of the third feedforward amount that has not been invalidated by the output saturation, based on the invalid amount, and estimates a correction value obtained by multiplying the second parameter by the ratio as a parameter for correcting the third feedforward amount so that a corresponding portion of the invalid amount is distributed within a range of the second operation amount before the limit processing that is not invalidated by the output saturation. A first step of calculating a first manipulated variable for each control cycle using a set value and a controlled variable as input;
a second step of calculating a first feedforward amount necessary for suppressing the disturbance for each control period in response to an input of a trigger variable that is a significant value before the application of the disturbance and is a non-significant value after the application of the disturbance;
a third step of adding the first feedforward amount to the first manipulated variable for each control period ;
a fourth step of adding the first feedforward amount to the first manipulated variable to output a second manipulated variable to a control target;
a fifth step of estimating , as a first parameter for calculating a second feedforward amount necessary for suppressing a step disturbance, a difference between the second manipulated variable at a time of settling before a disturbance recovery response and the second manipulated variable after settling of disturbance recovery when only the first manipulated variable is output to a control target as the second manipulated variable in a state in which the first feedforward amount is zero during a first prescribed trial period in which the trigger variable is input;
a sixth step of calculating, based on the first parameter, a total amount that is an integrated value of the second feedforward amount for each control period until the second manipulated variable is settled, assuming that the second feedforward amount is output during the first trial period;
a seventh step of subtracting a total amount of the second feedforward amount from a total amount that is an integrated value of an amount of change of the second manipulated variable in each control cycle until the second manipulated variable is settling during the first trial period, and estimating a value obtained by multiplying the control cycle by a result of the subtraction as a second parameter for calculating a third feedforward amount required for suppressing impulse disturbances;
an eighth step of calculating an invalid amount, which is an integrated value of the third feedforward amount calculated for each control cycle by the second step until the second manipulated variable is settling during a specified second trial period in which the trigger variable is input, which is invalidated due to either output saturation in which the second manipulated variable exceeds a manipulated variable upper limit value or output saturation in which the second manipulated variable falls below a manipulated variable lower limit value;
a ninth step of calculating, based on the invalid amount, a ratio between a total amount that is an integrated value of each control cycle of the third feedforward amount calculated by the second step until the second manipulated variable is settling during the second trial period and a total amount that is an integrated value of each control cycle of the third feedforward amount that has not been invalidated due to the output saturation, and estimating, as a parameter for correcting the third feedforward amount , a correction value obtained by multiplying the second parameter by the ratio so that a corresponding portion of the invalid amount is allocated within a range of the second manipulated variable that is not invalidated due to the output saturation;
the second step includes a step of setting, when the trigger variable is input during the second trial period , the sum of the second feedforward amount calculated for each control cycle based on the first parameter and the third feedforward amount calculated for each control cycle based on the second parameter as the first feedforward amount, and, when the trigger variable is input after the second trial period, setting, when the trigger variable is input after the second trial period, the sum of the second feedforward amount calculated for each control cycle based on the first parameter and the third feedforward amount calculated for each control cycle using the parameter for correction estimated in the ninth step as the first feedforward amount. 7. The control method according to claim 6 ,
The third feedforward amount is calculated by [Kx2s/{(1+ATfs)(1+(2.0-A)Tfs)}]FF_X, where Kx2 is the second parameter, Tf is a time constant that specifies the time for gradually converging the third feedforward amount, A is a coefficient that adjusts the time constant, FF_X is the trigger variable, and s is a Laplace operator. 8. The control method according to claim 7 ,
the ninth step includes a step of calculating the ratio RH by Kx2/(Kx2-FF_N×dt), where FF_N is the reactive amount, dt is the control period which is a calculation period of the first manipulated variable and the first, second and third feedforward amounts, and RH is the ratio, and calculating a correction value of the second parameter Kx2 by Kx2×RH and a correction value of the time constant Tf by Tf×RH as parameters for correcting the third feedforward amount. 9. The control method according to claim 6 ,
The second feedforward amount is calculated by [Kx1/{(1+ATfs)(1+(2.0-A)Tfs)}]FF_X, where Kx1 is the first parameter, Tf is a time constant that specifies a time for gradually converging the second feedforward amount, A is a coefficient that adjusts the time constant, FF_X is the trigger variable, and s is a Laplace operator. 10. The control method according to claim 6,
a tenth step of limiting the second manipulated variable calculated in the third step to a value equal to or greater than the manipulated variable lower limit value and equal to or less than the manipulated variable upper limit value;
the fourth step includes a step of outputting the second manipulated variable subjected to the limit process to a control target;
the fifth, sixth and seventh steps are performed by using the second manipulated variable that has been subjected to the limit process as an input;
the eighth step includes a step of calculating an invalid amount which is an integrated value of the third feedforward amount calculated for each control cycle by the second step until the second manipulated variable subjected to the limit processing is settling during a specified second trial period in which the trigger variable is input, the invalid amount being an integrated value of the amount invalidated due to either output saturation in which the second manipulated variable before the limit processing exceeds the manipulated variable upper limit value or output saturation in which the second manipulated variable before the limit processing falls below the manipulated variable lower limit value;
the ninth step includes a step of calculating, based on the invalid amount, a ratio between a total amount that is an integrated value of each control cycle of the third feedforward amount calculated by the second step until the second operation amount that has been limited processed is settling during the second trial period, and a total amount that is an integrated value of each control cycle of the third feedforward amount that has not been invalidated due to the output saturation, and estimating a correction value obtained by multiplying the second parameter by the ratio as a parameter for correcting the third feedforward amount so that a corresponding portion of the invalid amount is distributed within a range of the second operation amount before the limit processing that is not invalidated due to the output saturation.

Images