JPS63101902A - Controller - Google Patents

Abstract

PURPOSE:To obtain target value follow-up characteristics free from vibration while keeping disturbance suppressing characteristics at an optimum state by setting up the objective value follow-up characteristics independently of the disturbance suppressing characteristics. CONSTITUTION:A target value change detecting means 12 is constituted of a model 16 with gain '1' approximating to the response characteristics of a controlled system 2, a subtraction operating element 17, a level deciding part 18 for outputting a switching signal when a deviation outputted from the element 17 exceeds a prescribed value, etc. A change in a target value signal inputted to an target value feedforward control means 11 is detected by a target value change detecting means 12, and when the change in the target value signal is large, i.e. when a transient state that target value feedforward control is acting is formed, a feedback controlled variable is restricted and target value feedforward control is mainly executed. At the time of small change in the target value signal, the feedback control is normally driven. Consequently, the target value follow-up characteristics free from vibration can be rapidly obtained in accordance with a change in the target value signal.

Description

DETAILED DESCRIPTION OF THE INVENTION [Purpose of the Invention (Industrial Application Field) The present invention relates to a control device used for controlling various plants, and particularly to a control device with excellent vibration-free target value tracking characteristics.

(Prior Art) A conventional general PID control system is constructed by a control block diagram as shown in FIG.

That is, this control system consists of subtraction calculation element 1. It consists of a control device 3 equipped with adjustment means 2, etc., and a controlled object 4, and has a target value S.
After calculating the deviation e (=SV - PV) by subtracting V and the controlled variable Pv obtained from the controlled object 4 using the subtracting calculation element 1, the adjusting means 2 receives this deviation e and adjusts the controlled variable Pv. The configuration is such that the manipulated variable MV is obtained by performing a PI calculation operation or a PID calculation operation so that the deviation e matches the target value SV, that is, the deviation e becomes zero, and this is given to the controlled object 4 for operation. .

Therefore, when the adjustment means 2 performs a PI calculation operation, that is, a (proportional+integral) calculation operation, the transfer function Gc (s) of the adjustment means 2 can be expressed by the following equation.

Gc (s)=Kp (1+1/Tz-s)-(1
) Here, KP is the proportional gain, TX is the integration time, and S is the Laplace operator.

By the way, the adjusting means 2 is adjusted so that the deviation e becomes zero (1)
Although the adjustment calculation is performed based on the formula, the target value Sv
When MV changes, the manipulated variable MV is not uniquely determined even if the parameters KP and T are determined, and continues to change until the deviation e becomes zero. The basis of feedback control is to judge the validity of the manipulated variable MV by the resulting deviation e, and then determine the next manipulated variable, but this is an inevitable result of the basic idea of feedback control. That's a problem. In other words, the adjusting means 2 generally sets the optimum PI parameter for the change in the target value and performs the PI calculation operation based on it, but it changes the target value Sv as shown in FIG. Then, the controlled variable PV inevitably responds with a delay due to the dead time 9 time constant of the controlled object 4 itself. Due to this response delay, that is, the deviation in the shaded area in Figure 9, if the PI parameters are matched to the optimum disturbance suppression, the control amount P will greatly overshoot with respect to the change in the target value, causing oscillatory changes. will be presented. FIG. 10 shows a target value follow-up characteristic when the PID parameter is adjusted to optimally suppress disturbance and the target value is changed in a general PID control system. As is clear from this figure, the target value tracking characteristic greatly overshoots and exhibits oscillatory changes, and if a two- or three-stage cascade control system is configured as is, sustained vibration will occur, and this vibration will be suppressed. If the gain is lowered in an attempt to stop this, the disturbance suppression characteristics deteriorate, and this is a major obstacle to improving plant control characteristics.

(Problem to be Solved by the Invention) Therefore, when the control system as described above performs adjustment control by setting the PI or PID parameters so as to match the optimum disturbance suppression, There was a contradictory problem in that overshooting caused repeated oscillatory changes, and if the gain was lowered to stop this oscillation, the disturbance suppression characteristics deteriorated, making it difficult to improve the target value tracking characteristics. .

The present invention has been made in view of the above-mentioned circumstances, and allows the target value tracking characteristic to be set independently from the disturbance suppression characteristic. It is an object of the present invention to provide a control device that can obtain follow-up characteristics and can also cope with energy saving, high quality, and flexible production of plants.

[Structure of the Invention] (Means for Solving the Problems) A control device according to the present invention performs calculation using a feedback signal of a controlled variable obtained from a controlled object and a preset target value signal of the controlled object. In the control device, the control device obtains an operation signal and applies it to the controlled object, including target value feedforward control means for outputting the target value signal directly or by multiplying it by a predetermined coefficient as the operation signal, and detecting a change in the target value signal. after passing the target value signal through a model that approximates the response of the controlled object,
In addition to determining the deviation from the feedback signal, a calculation to make this deviation zero is performed in a non-limiting and limited manner depending on the magnitude of the target value change by the target value change detection means, and based on the calculation output. and feedback amount limit adjustment means for modifying the operation signal output from the target value feedforward means.

(Effect) Therefore, by using the above-mentioned means, the target value “
When a change in the target value signal input to the feedforward control means is detected by the target value change detection means, and as a result, the change in the target value signal is large.

In other words, in a transient state where the target value feedforward control is operating, the feedback control amount is limited and the target value feedforward control is used as the main control, and when the change in the target value signal is small, the feedback control operates normally to control the target value. It is possible to obtain vibration-free target tracking characteristics in response to signal changes.

(Example) Hereinafter, an example of the apparatus of the present invention will be described with reference to FIG. In the figure, reference numeral 10 denotes a control device according to the present invention, and this control device 10 includes target value feedforward control means 11. It consists of target value change detection means 12, feedback amount limit adjustment means 13, and the like. This target value feedforward control means 11 is configured to obtain a manipulated variable MV by passing the target value Sv through a lead/lag compensating means 14 with a gain of "1" and an addition calculation element 15. The target value change detection means 12 is a model 16 degree subtraction calculation element 17 with a gain of "1" that approximates the response characteristic of the controlled object 2, and is switched when the deviation output from the subtraction calculation element 17 exceeds a predetermined value. It is composed of a level determining section 18 that outputs a signal, and the like. The feedback amount limit adjusting means 13 includes the model 16, and further includes the model 16.
a subtraction calculation element 19 for calculating the deviation between the output of and the control amount Pv;
Gap means 20, a switch circuit 21 that selectively takes in and outputs the output of the subtraction calculation element 19 and the output of the gap means 20 based on the output of the level determination section 18, and a switch circuit 21 for each of proportional, integral, and differential calculations. It is constituted by an adjustment means 22 etc. that performs at least one calculation among them, and the output of this adjustment means 22 is given to the addition calculation element 15 to correct the manipulated variable MV obtained by the target value feedforward control means 11. It has the function of The model 16 is generally approximated by dead time + second-order lag, dead time + first-order lag, or only first-order lag. Furthermore, when the deviation signal from the subtraction calculation element 19 is e, and the output of the gap means 20 is eO, when 1e1>δ, the gap means 20 calculates 1) When let≦δ, the calculation eQme is performed.

Therefore, the target value SV is introduced as a main signal into the addition calculation element 15 after being compensated for speeding up or slowing down the response of the controlled object 2 through the lead/lag compensation means 14. Further, the output Svo is obtained through a model 16 that approximates the target value Sv to the response characteristic of the controlled object 2, and the deviation e between this output Svo and the control amount Pv is extracted by a subtraction calculation element 19.
This is introduced into the adjusting means 22 to perform calculations such that the deviation e becomes zero. Furthermore, after the difference between the signals on the input and output side of the model 16 is extracted by the subtraction operation element 17, this difference signal is led to the level determination section 18, where the level is determined. Now, if the predetermined judgment level set in advance is δ, when lhl≦δ, the switch circuit 21 is on the b side 1hl>δ
When the switch circuit 21 is set to the a side, the level determination section 1
The switch circuit 21 is switched and controlled by the output of 8. In other words,
When the change in the target value is small, the deviation e obtained by the subtraction calculation element 19 is directly introduced into the adjustment means 22 to perform continuous control, and so-called feedback control in control ff1PV is normally operated. On the other hand, when the change in the target value is large, the gap means 2o subjects the deviation e to discontinuous processing and introduces it into the adjustment means 22, where adjustment calculations are performed to obtain the output Mvo, which is guided to the addition calculation element 15. In this configuration, the feedforward signal based on the target value signal is corrected and outputted as the manipulated variable MV, and then supplied to the controlled object 2.

In the above embodiment, the level is determined from the signal difference between the input and output of the model 16, but as shown by the dotted line in the figure, the signal difference between the input and output sides of the lead/lag compensation means 14 is subtracted by the calculation element 17'. The switching control signal for the switch circuit 21 may be obtained by extracting the signal and determining the level in the level determining section 18'.
The following explanation deals with the case where it is obtained from the input/output difference of .

Therefore, according to the configuration of the embodiment as described above, the lead/lag compensating element 14 is connected in series to the output line of the target value signal Sv, and the control response characteristic is changed to a desired response model, for example, in response to a change in the target value. Functionally, it plays a very important role because it speeds up the control response characteristics, and in the case of parallel operation, it matches the individual response characteristics to the characteristics of the reference boiler. Here, the transfer function GH(S) of the lead/lag compensation means 14 will be explained. Let the transfer function of controlled object 2 be kOIIGp (s) (ko is gain),
If the desired response model of the controlled object 2 is G'(s), the response characteristic of the control system including the transfer function GH(S) of the lead/lag compensation means 14 is approximately GH(S) ・Gp(s) From this, G'(s) of the lead/delay compensating means 14 may be determined so that it is equal to the desired response model G'(s).

In other words, GH (S) ・Gp (s) - G' (s)
−(2), and from equation (2), the lead/lag compensation means 14
GH(s) is Go(s)=G(s)/GP(S)
−(3). Both Gp (s) and G (S) are generally approximated by dead time + second-order lag, dead time + first-order lag, or first-order lag. When approximated by the most frequently used dead time + first-order lag, it becomes, and if the desired response characteristic for this is G (s), the dead time cannot be changed, so it becomes. By these equations (3) to (5), we get that advance/
This becomes a delay element. In other words, it is a proportional + differential operation.

Therefore, when ■, "r-"rp...GH(s
)-1, and ■, when Tゝ<”rp...G
H(S) becomes an advance element and can speed up the overall response. T − (Tp/3) to (TP/4) is suitable. ■When "r>"rp...Go
(s) becomes a delay element, and the overall response is delayed. This is applied when adjusting the fast controlled object 2 to the reference during parallel operation. By specifying desired response characteristics in this manner, the response of the entire control system can be freely manipulated, resulting in a practically effective function.

Then, as described above, the target value signal sv is added with necessary compensation through the lead/lag compensation means 14 and introduced into the addition calculation element 15, where it is added and combined with the output signal of the feedback amount limit adjustment means 13. It is supplied to the controlled object 2 as the operation signal MV.

Here, in the device of the present invention, the feedforward control signal based on the target value signal Sv is the main signal, and the output of the feedback amount limit adjustment means 13 is mainly for removing steady-state deviation, and its composition ratio in the operation signal is small. .

On the other hand, the output signal of the lead/lag compensation means 14 is
The subtraction calculation element 1 is passed through the model 16 that approximates the response characteristics of
9 together with the feedback signal Pv, and the difference between them is taken as the deviation signal e.

When the change in the target value signal S is smaller than a predetermined value, the deviation signal e is introduced into the adjusting means 22 through the b side of the switch circuit 21, and the deviation signal e is adjusted to zero. The addition calculation element 15 adds the calculation output to the target value feedforward signal and supplies the result to the controlled object 2.

On the other hand, when the signal level difference between the input and output sides of the lead/lag compensation means 14 is taken out, and the level determination section 18' determines that the change in the target value signal is larger than a predetermined value, the switch circuit 21 is switched to the a side. Can be switched. As a result, the deviation signal e is introduced into the adjustment means 22 through the gap means 20 of the nonlinear element.

The reason for this configuration is that in target value feedforward control, the manipulated variable is theoretically predicted and output, but in an excessive state where the controlled object 2 does not follow the target value, feedback control is activated. This disturbs the predicted value of the target value feedforward control, disrupts the control, causes it to overshoot, and becomes oscillatory. Therefore, when the change in the target value signal Sv is larger than a predetermined value, the controlled object 2 is in a transient state in which it is about to respond. Then, feedback control is suppressed or put into a hold state to prevent prediction of the manipulated variable from going out of order. Therefore, the intention is to include a model 16 that approximates the response characteristics of the controlled object 2 so that the deviation e becomes zero in a transient state, but in actual cases, it is necessary to always make the control characteristics and the model 16 match. This is an extremely difficult technique, and this slight deviation has a negative effect on control. Therefore, in a transient state, the effect of deviation due to the deviation between the actual characteristics of the controlled object 2 and the model 16 is ignored through the nonlinear element filling means 20. ing.

Next, FIGS. 2 and 3 are block diagrams showing other embodiments of the feedback amount limit adjustment means 13, particularly regarding a combination of continuous calculation and non-continuous calculation. In FIG. 2, the output side of the subtraction calculation element 19 is branched into two, a continuous PID adjustment means 31 is connected to one branch side, and a sample PI adjustment means 32 is connected to the other branch side. Similarly, when the change in the target value is small, the continuous PID adjustment means 31 is selected to output the continuous calculation result,
If the change in the target value is large, the sample PI adjustment means 3
2 and outputs the discontinuous calculation results, and also converts these calculation results into the velocity shadow signal-position signal conversion means 33.
The configuration is such that the signal is converted into a positional signal and applied to the addition calculation element 15. That is, in contrast to FIG. 1, in which the magnitude of the deviation (or equivalently, the gain) is processed in response to changes in the target value using linear calculations and nonlinear calculations by the gap means 20, in FIG. Depending on the magnitude of change in the target value, the speed type adjustment calculation output signal is the output of the continuous PID adjustment means 31, which can be considered to be equivalent to continuous in time, and the output of the sample PI adjustment calculation means 32, which can be considered to be temporally discontinuous. The speed type adjustment calculation output signal and the switch circuit 21
At the same time, this speed shadow signal is converted into a position signal by the signal conversion means 33, and then the addition calculation element 1
5, corrects the target value feedforward signal, extracts it as a manipulated variable MV, and supplies it to the controlled object 2.

FIG. 3 shows a configuration in which a sample period generating section 42 is operated based on the output of the normal adjusting means 41 and the determination result of the level determining section 18'. That is, when lhl≦δ,
...The switch circuit 21' is closed and continuous control is performed.

When 1h1>δ...The switch circuit 21' is turned on and off by the output of the sampling period generating section 42 for discontinuous control. Then, the calculation output outputted from the switch circuit 21' from the adjustment means 41, that is, the speed type calculation output, is converted into a position type signal by the signal conversion means 33, and then introduced into the addition calculation element 15 to feed the target value. This configuration is such that the forward signal is corrected, extracted as a manipulated variable, and supplied to the controlled object 2.

Next, FIG. 4 is a block diagram showing another embodiment of a portion that combines linear calculation and nonlinear calculation in accordance with the magnitude of change in target value signal SV. In contrast to the method shown in FIG. 1 in which the gain is switched in two stages according to the magnitude of change in the target value signal Sv, in FIG. After extracting the absolute value 1hl, this is guided to the function calculation means 52 to obtain a gain corresponding to the magnitude of change in the target value signal Sv, and this is led to the multiplication calculation element 53 to be multiplied by the deviation signal e and output. is introduced into the adjustment means 22, and adjustment calculation is performed here so that the deviation signal e becomes zero to obtain the output MV0, which is added to the addition calculation element 1.
5, corrects the target value feedforward signal, extracts it as a manipulated variable MV, and supplies it to the controlled object 2.

Next, in FIG. 5, the device shown in FIGS. 1 to 4 multiplies the target value feedforward signal by a predetermined coefficient to obtain the manipulated variable MV, whereas the device in FIGS. 1 to 4 considers the coefficient to be "1". That is, in FIG. 5, as the characteristic of the controlled object 2, that is, the gain changes, the output M V o of the adjusting means 22 and thus the correction amount increases. In this state, the target value signal Sv changes, resulting in over-correction or under-correction, which disrupts the control due to the change in the target value signal Sv, and requires a long time to stabilize. As a means to improve this drawback, the feedforward gain of the target value signal Sv may be modified so that the magnitude of the output MV0 of the adjusting means 22 is always near zero. This embodiment has been realized based on this viewpoint, and 61 is a coefficient calculation means.

62 indicates coefficient multiplication means. That is, in this embodiment, the output FFn of the coefficient multiplication means 62 and the output M V n of the addition calculation element 15 are taken in, and the output MVo of the adjustment means 22 is made close to zero by the coefficient calculation means 61 using the following equation. The coefficients of the coefficient multiplication means 62 are corrected by performing the calculation shown in FIG.

Kn = (Kn-1) X (MVn /FFn)
-(7) However, Kn is the coefficient of the current coefficient multiplication means 62.

Kn-1 indicates the coefficient of the previous coefficient multiplication means 62, MVn indicates the current manipulated variable, and FFn indicates the current target value feedforward amount. The calculation of this equation (7) is an example, and the MV
o as an example (or may be an intermittent periodic calculation). This makes it possible to provide a control system that can also adapt to changes in the gain of the controlled object 2.

Next, the overall effects of the embodiments described above will be explained. In a conventional general PID control system, when the target value signal Sv changes, the manipulated variable is not determined even if the parameter is specified, and the manipulated variable continues to change until the control deviation e becomes zero. Therefore, the response characteristics of the controlled object 2 become oscillatory as a whole.

On the other hand, in the embodiment device of the present invention, when the target value signal changes, the manipulated variable changes directly and is applied to the controlled object 2, and the feedback signal is connected to the target value that has passed through the response model 16 of the controlled object 2. The system is configured to calculate the deviation, ignore the deviation in transient conditions where the change in the target value is large, and correct the target value feedforward signal for steady deviations by the output of the normally operating adjustment means. be. For this reason, as shown in FIG. 6, in the device of the present invention, once the parameters are determined, the amount of operation relative to the change in the target value is approximately determined. In addition, limiting the amount of feedback has the role of eliminating steady control deviations, and has a characteristic that its component ratio in the manipulated variable is small, and vibration components are significantly reduced.

Incidentally, in order to see the effect of the device of the present invention, the PID parameters for both the transfer functions of the controlled object 2 are set to the disturbance suppression characteristic optimum values KP-2, 59, Tl-3, 41 seconds.

If To is set to 0.56 seconds and the response characteristics to changes in the target value are taken, the result will be as shown in FIG.

In other words, the conventional general PID control system is thick (overshoots and takes time to settle).In contrast, the device of the present invention has a vibration-free response characteristic.In the actual process, the target value signal has a step shape. There are cases where the target value signal changes occasionally and cases where the target value signal changes randomly, but stable control can be obtained in both cases.Especially in the latter case, conventional PID control includes a resonant frequency. Control oscillates or is disturbed, but with the device of the present invention, such vibrations are completely eliminated and stable control can be achieved.

Although the above embodiment has been described using position shadow signals as signal calculation processing, it can be similarly applied to velocity shadow signal processing calculations that are frequently used in digital control systems.

Further, it can be similarly applied to a mixing process such as a heat exchanger. Now, the amount of heat Qi to be given by the heat exchanger is Qi - ((Ts - Ti) + f e (Ts - To) l XF i. However, Ts is the target value of temperature, Ti is the heat exchanger inlet temperature, To
is the heat exchanger outlet temperature, Fi is the flow rate of the heated fluid, fe
(Ts-T□) represents the feedback control output. Therefore, the target value feedforward output is not Ts, but is for the controlled object of the mixing process such that (Ts - Ti), that is, (target value signal Ts - control amount Ti before mixing). In addition, the present invention can be implemented with various modifications without departing from the gist thereof.

[Effects of the Invention] As described in detail above, according to the present invention, the target value tracking characteristic can be set independently from the disturbance suppression characteristic. Therefore, it is possible to provide a control device that can achieve target value follow-up characteristics without any problems, and can also cope with energy saving, high quality, and flexible production of plants.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing one embodiment of a control device according to the present invention, FIGS. 2 to 5 are block diagrams showing other embodiments of the device of the present invention, and FIGS. 6 and 7 are block diagrams, respectively. FIG. 3 is a diagram illustrating the suggested price follow-up characteristic of the embodiment effect of the device of the present invention. DESCRIPTION OF SYMBOLS 10... Control device, 11... Target value feed forward control means, 12... Target value change detection means, 13.
... Feedback amount limit adjustment means, 14... advance/
Delay compensation means, 15... Addition calculation element, 16... Model that approximates the response characteristics of the controlled object, 17° 17';
19... Subtraction calculation element, 18.18'... Level determination section, 20... Gap means, 21.21'... Switch circuit, 22.41... Adjustment means, 31... Continuous Target PID adjustment means, 32... Sample PI adjustment means, 33... Signal conversion means, 42... Sample period generation section, 51... Absolute value calculation means, 52... Function calculation means, 53. . . . Multiplication calculation element, 61 . . . Coefficient calculation means, 62 . . . Coefficient calculation means. Applicant's agent Patent attorney Takehiko Suzue Figure 6 Figure 7

Claims (6)

[Claims] (1) In a control device that calculates an operation signal by calculation using a feedback signal of a controlled variable obtained from a controlled object and a preset target value signal of the controlled object and supplies it to the controlled object, the control device target value feedforward control means for outputting a signal directly or by multiplying it by a predetermined coefficient as the operation signal; target value change detection means for detecting a change in the target value signal; After passing through the approximated model, the deviation from the feedback signal is determined, and the calculation to make this deviation zero is carried out in a non-restrictive or restrictive manner depending on the magnitude of the target value change by the target value change detection means. and feedback amount limit adjustment means for correcting the operation signal output from the target value feedforward means based on the calculated output. (2) The target value feedforward control means has lead/lag compensating means, outputs the target value signal to the feedforward control system through the lead/lag compensating means, and
2. The control device according to claim 1, wherein the control device supplies the feedback amount to the feedback amount limit adjustment means. (3) The coefficient used in the target value feedforward control means is obtained by calculating the difference between the output signal obtained by multiplying the target value signal by a predetermined coefficient and the operation signal corrected by the feedback amount limiting adjustment means, and when this difference is near zero. 2. The control device according to claim 1, which uses coefficients modified so that (4) The target value change detection means detects a change in the target value signal using any one of the input/output difference of the model and the input/output difference of the lead/lag compensating means. The control device according to item 1. (5) The feedback amount limit adjusting means is performed using any one of linear operation and non-linear operation, continuous operation and discontinuous operation, function non-multiplication and function multiplication as described in claim 1. control device. (6) The lead/lag compensation means adjusts the time constant of the controlled object to Tp
, the desired response time constant of the controlled variable to the change in the target value signal is T^*, and the Laplace operator is s, then (1+Tp・s
)/(1+T^*·s).