JPH06230820A - Device and method for controlling feedback - Google Patents

Abstract

PURPOSE:To improve the control performance of a feedback controller having a system for determining the manipulated variable of a controlled object such as a PWM inverter by DSP, and in detail to improve the deterioration of the control performance caused by disturbance. CONSTITUTION:The feedback controller for determining the manipulated variable V(n) of a controlled object 10 based upon an error signal between the detection value of a controlled variable for the object 10 and a command value is provided with a means for finding out the predictive value of a change in the controlled variable based upon the manipulated variable V(n) and adding the forecasting value of the change to the detection value of the controlled variable to obtain the forecasting value of the controlled variable and a means for finding out a forecasting error consisting of a difference between the predictive value of the controlled variable and the practical controlled variable detection value and correcting the predictive value of the change based upon the predictive error. A predictive error V1'(n) to be a disturbance component is added to the manipulated variable V(n) to find out a new manipulated variable.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control device and a control method for controlling a controlled object such as an inverter and a converter.

[0002]

2. Description of the Related Art FIG. 1 shows in principle a feedback control device in a system for supplying electric power from a variable voltage source 2 composed of a PWM converter or the like to a load 1 composed of a reactor having an inductance value L. This system is configured to supply a load current I corresponding to a digital command current value Ir generated from a digital command value (reference value or target value) generator 3 from a variable voltage source 2 to a load 1. The load current I is a control amount in the feedback control method, and is detected by the current detector (control amount detector) 4 and the ADC (analog / digital converter) 5 connected thereto. AD
C5 samples the analog detection current detected by the current detector 4 at a constant sampling period, converts this sample into a digital signal, and outputs it. ADC here
Both the input and the output of 5 will be denoted by I. The digital subtractor 6 as a comparison means forms an output (error output) of the difference between the command value Ir and the detection value I. The digital controller 7 connected to the subtractor 6 is a control element including a proportional integral means or a proportional compensating means in the feedback control system, is composed of a microcomputer (microcomputer or microprocessor), and has an error given from the digital subtractor 6. Form a digital manipulated variable V for operating the variable voltage source 2 to bring the signal to zero. The command value generator 3, the ADC 5, and the subtractor 6 can be included in the microcomputer together with the digital controller 7. A hold circuit 8 is connected between the digital controller 7 and the variable voltage source 2 to form an analog manipulated variable corresponding to the digital manipulated variable V, hold it for one sample period, and output it. Since this hold circuit 8 includes a DAC (digital-analog converter) and forms an analog operation amount corresponding to the digital operation amount V, it can be called an analog operation amount generator. Here, both the digital operation amount and the analog operation amount are indicated by V. The variable voltage source 2 includes, for example, a circuit that generates a PWM control pulse in response to the analog operation amount V and a switching element that responds to the PWM control pulse, performs an operation corresponding to the analog operation amount V, and obtains an output current I. Output the voltage for. In addition, here
In FIG. 1, the hold circuit 8, the variable voltage source 2 and the load 1 are collectively referred to as a controlled object. But load 1
It is also possible to call only the controlled object, and the digital controller 7, the hold circuit 8 and the variable voltage source 2 are collectively called the control element. In this case, the output voltage kV of the variable voltage source 2 (where k is a constant) is the manipulated variable.

1a in FIG. 1 is a current Il due to disturbance.
Is a disturbance current source that schematically shows the occurrence of
The disturbance current Il is also included in the current I detected at.

Even when the load shown in FIG. 1 is replaced with an inductance L and a resistance R, the same control system as in FIG. 1 can be formed. In addition, the load 1 in FIG. 1 is a capacitor C, and a current is supplied to the load from the variable current source to load voltage Va.
Is basically controlled in the same manner as in FIG. 1 when the voltage follows the voltage command value.

With respect to FIG. 1, the transfer function for the controlled object is obtained by inputting the manipulated variable and outputting the controlled variable, and the following equation (1) is obtained. G _A (s) = 1 / Ls (1) Further, in the case where a capacitor C is provided instead of L in FIG. 1 and a variable current source is connected to this, the transfer function is similarly calculated and the following equation (2) is obtained. Becomes G _B (s) = 1 / Cs (2) According to equations (1) and (2), the controlled object is simulated by an integral element (element in which the output of the indial response increases with time) for both the L load and the C load. Both the manipulated variable and the controlled variable are similar circuits except that the voltage and current are exchanged. Therefore, only the case of the L load in FIG. 1 will be described below.

FIG. 2 shows the transfer function z in the control system of FIG.
It is expressed in the discrete time domain by conversion. Here, the digital controller 7 including a proportional integrator is ki / (z-
The element 7a of 1), the element 7b of Kp, and the adder 7c are shown. Further, the hold circuit 8, the variable voltage source 2 and the load 1 in FIG. 1 are (kTs / L) {(m (z-1) +1) /
It is shown as a controlled object 10 including an element 10a composed of (z-1)} and an element 10b of Z- ¹ . Further, the control amount I of the controlled object 10 is shown as including the disturbance component Il.

FIG. 3 shows a control amount (load current) I detected at each Ts and its detected value and a digital value when the sampling period (interval) and the control interval of the controller 7 in the control device of FIG. 1 are both Ts. Alternatively, the relationship of the analog operation amount V is shown. Note that n-2, n-1, n, and n + 1 indicate sampling points, and I (n-2), I (n-1), I (n), and I (n
+1) indicates the control amount (sample) detection value at each sampling time, and V (n-2), V (n-1), V (n), and V (n + 1) are manipulated variables at each sampling interval. Indicates. If the sampling time point n is the current time point, the manipulated variable V (n) output at the current time point n
Is the control amount detection value I at the previous sampling time point n-1
It is decided based on (n-1). Therefore, in the feedback control system including the digital controller 7, a control delay of one sampling time interval Ts and a control error occur.

The problem concerning the control delay and the control error is a problem peculiar to the control system constituted by the digital controller, and it is known to use the control amount predictor as shown in FIG. 4 as an improvement measure. FIG. 4 is a block diagram in which a control amount predictor 9 is added to the control system of FIG. In FIG. 1, the transfer function of the hold circuit 38 can be expressed by the following equation (3). G _H (s) = (1−e ^{−s · Ts} ) / s (3) Further, the variable voltage source 2 can be represented by the proportional element k. Further, since the transfer function of the controlled load 1 is the expression (1), the total transfer function G (s) from the hold circuit 8 to the controlled load 1 is the following expression (4). _{G (s) = G H (} s) · k · G A (s) = ((1-e -s · Ts) / s) · k · (1 / Ls) = k · (1-e -s · ^Ts ) / (s ² L) (4) Here, when G (s) is represented in the discrete time domain by Z transformation, the following equation (5) is obtained by the Z transformation formula. G (z) = (k · Ts / L) · (1 / (Z−1)) · Z ⁻¹ (5) In FIG. 4, 7 is a digital controller represented in the discrete time domain, and 10 is an equation ( 5) transfer function from the hold circuit 8 to the controlled load 1, 11 is a one-sample delay element, 9 is a controlled variable predictor, 12 is an adder, 6 is a subtractor,
3 is a command value for the controlled variable. The line 13 corresponds to the output of the ADC 5 of FIG. 1, and the control amount detection value is obtained here. One input terminal of the adder 12 is connected to the line 13, and the other input terminal is connected to the predictor 9. The predictor 9 is connected to the digital controller 7 via the delay element 11, and the predicted value Δi (n-1) is calculated based on the digital manipulated variable V.
To occur.

More specifically, the predictor 9 calculates and predicts only the increment (change) of the control amount when a certain manipulated variable V is given to the controlled object. Therefore, since the controlled object of this example is the reactor 4, the function for obtaining the increment can be expressed by (k / Ln) .multidot.Ts (where Ln represents the nominal value of L). FIG. 5 shows the state of control. (N-1)
At the time point, the controlled variable I (n-1) is detected on line 13.
Further, the manipulated variable v (n-1) output at the time (n-1) is obtained by the delay element 11, and this is input to the predictor 9 to obtain T
The predicted value Δi (n-1) for the increase in the control amount in s time is obtained. By adding the control amount increase predicted value and the detected control amount in the adder 12, the control amount predicted value i (n) at the time point n is added.
Is obtained. The subtracter 6 obtains a control error from the obtained control amount predicted value and command value at the time n, and the digital controller 7 outputs the control error in one sample period from the time n so that the control error converges to zero. The manipulated variable V (n) is determined and output to the controlled object during the period from n to n + 1. By this method, the control error caused by one sample time is compensated. However, when a constant (for example, L) to be controlled changes, an error occurs. The longer the prediction time, the larger the error. Here, the variation of the constant to be controlled means that the inductance value L varies in the example of FIG. 1, and that the load 1 has a resistance component together with the inductance L.

In order to solve the above-mentioned drawbacks, the applicant of the present invention shortens the prediction time and makes the error less likely to occur, and when a prediction error occurs, the value obtained by integrating the error between the predicted value and the actual value is used. Japanese Patent Application No. 4-26436 discloses a method of converging the prediction error to 0 by adding it to the input or output of the predictor.
Proposed in No. 3. Next, the feedback control device according to this method will be described with reference to FIGS. 6 and 7. However, in FIGS. 6 and 7, the same parts as those in FIGS. 1 to 5 are designated by the same reference numerals and the description thereof will be omitted. As is clear from the comparison between FIG. 6 and FIG. 4, the system of FIG. 6 is the system of FIG. 4 with the addition of the prediction error detector 14, the prediction error compensator 15, and the adder 16. That is, this compensation means is provided in addition to the control amount change prediction means 9a. The newly provided adder 16 adds the output V (n-1) of the delay element 11 and the prediction error compensation amount obtained from the prediction error compensator 15 and sends it to the change (increase) predictor 9b. The prediction error detector 14 is
It comprises a delay element 17 for one sampling period, an exp (-mTs) circuit 18, and a subtractor 19. The delay element 17 is connected to the output line of the adder 12 and outputs the predicted control value i (n-1) at (n-1), which is one sample period Ts before the predicted control value i (n). The circuit 18 outputs the control amount detection value I (n-1) at the time of (n-1). Therefore, the subtractor 1
From 9, the output Δie (n-1) of the difference between the predicted value i (n-1) and the detected value I (n-1) is obtained as the prediction error. The prediction error compensator 15 is composed of a kz / (z-1) circuit and sends the prediction error compensation amount to the adder 16. In FIG. 6, the control amount predictor 9a, the prediction error detector 14, the prediction error compensator 15, the adder 12,
16 is functionally shown as a block, but actually D
It is composed of an SP (digital signal processor) or a microcomputer, and is processed by software according to a predetermined program.

In this embodiment, as shown in FIG. 7, m points are provided during each sampling time interval Ts, and the timings of control amount detection and operation amount output are set as follows. An example will be described between the points (n-1) and (n).
The control amount I (n-1, m) at the m point between n-1 and n, that is, the (n-1, m) Ts point is detected and output at the n point. Here, m takes a value between 0 and 1. When m = 1, there is no control delay. When m = 0, a control delay of 1 sample occurs. Therefore, the prediction time is (1-m) Ts
This is less than 1 sample time.

Next, the transfer function in the discrete time domain of the controlled object at the above timing can be expressed by the equation (6) by the extended Z transform. G (z, m) = (k.Ts / L). (M (Z-1) +1) / (Z-1)). Z- ¹ (6) Further, the predictor 9b is used to increase the control amount. The desired function is (1
-M) To predict the time after Ts, the following is performed. (K / Ln) ・ (1-m) Ts

By introducing the timing m, FIG.
Controlled object 10 is (kTs / L) {[m (z-1) +1] /
A block 10a indicated by (z-1)} and a delay element 10b of one sample are obtained, and a control amount detection value indicated by I (n-1, m) is obtained as the output. In addition, the predictor 9b uses (k /
Ln) (1-m) Ts, and (1-m) Ts time control amount increase predicted value Δi (n-1, m) from (n-1, m) Ts time to (n) Ts time Is output.

6 and 7, the manipulated variables v (n-1) and (n-1) Ts output to the controlled object at (n-1) Ts.
The value obtained by adding the output value of the prediction error compensator 15 at the time of addition by the adder 16 is input to the predictor 9b to calculate the control amount increase predicted value Δi (n-1, m) after (1-m) Ts. . In addition, (n-1,
m) The controlled variable I (n-1, m) at Ts is detected and its value is Δi (n
-1, m) is added by the adder 12 to calculate the control amount predicted value i (n) at (n) Ts. Also, (n-1) Ts
Predicted value i (n-1) and (n-1, m) Ts
MTs time delay element of the control amount detection value at time, that is, (n
-1) Difference Δie (n-1) from the detected control amount I (n-1) at Ts
Is the prediction error of the controlled variable at (n-1) Ts.
In order to converge this prediction error to 0, the prediction error is input to the prediction error compensator 15 composed of integral elements, and its output, which is the prediction error compensation amount, is manipulated variable V, which is the input value of the predictor 9b. By adding to (n-1), the predicted value of the control amount increase is compensated. Predicted control amount i at (n) Ts
(n) is compared with the command value Ir (n) to obtain a control error, and based on this error, the digital controller 7 determines the manipulated variable v (n) such that the error converges to 0, and (n ) Output at Ts.

By the above operation, when the constant L of the load 1 to be controlled in FIG. 1 changes from the nominal value Ln, the command value Ir is accurately followed even if the load 1 includes the resistance R together with L. Control is realized.

The disturbance current Il shown in FIG. 6 is converted into a voltage Vl, which is applied to the adder 20 on the input side of the controlled object 10 as shown in FIG. 8 to add the manipulated variable V (n). Can also be shown. The control amount I is the disturbance voltage Vl.
Alternatively, since it changes depending on the current Il, the prediction error compensator 15
The post-compensation prediction error obtained from Eq.

[0017]

By the way, in the method shown in FIG. 6 or FIG. 8, the predicted value of the control amount including the disturbance component reaches the subtracter 6 from the adder 12 and further includes the proportional-integral element or the proportional element. Pass through vessel 7. The time constant of the integral element in the controller 7 is set to a long time of about several times the time (estimated time constant) at which the predicted value (estimated value) is determined in the controlled variable predicting means 9a in order to enhance the stability of the control system. It Therefore, it is difficult to satisfactorily compensate for a fast-changing disturbance (a disturbance changing at a high frequency).

Therefore, it is an object of the present invention to provide a feedback control device and method capable of favorably compensating for a fast-changing disturbance.

[0019]

In order to achieve the above object, the present invention detects a control amount of a controlled object or a corresponding amount at a predetermined sampling time interval to detect a control amount as a digital amount. Digital control amount detecting means for obtaining a value, digital command value generating means for generating a command value that the control amount detection value should follow as a digital amount, and the control so that the control amount detection value follows the command value In a digital feedback control device comprising a digital control means for controlling the controlled object by determining an operation amount for the object by a digital operation at a predetermined control time interval, the digital control means comprises a comparing means and a proportional-plus-integral compensation. Means or proportional compensation means, addition means, control amount prediction means, prediction error detection and compensation means, and the comparison means The output of the proportional-integral compensating means or the proportional-compensating means corresponds to the difference between the control amount predicted value corrected by the prediction error compensation amount and the command value. Proportional-integral compensation or proportional-compensation, wherein the addition means adds the prediction error compensation amount obtained from the prediction error detection and compensation means to the output of the proportional-integral compensation means or the proportional compensation means to control the controlled object. The control amount predicting means outputs the operation amount for controlling, and the control amount predicting means calculates the amount of change between the current control amount and the control amount at the next sampling based on the operation amount output from the adding means. The predicted value is calculated and the predicted value for the next sample is obtained by adding the predicted value for the change to the control amount detected at the current sample. The means obtains a prediction error based on the difference between the control amount predicted value and the actual control amount detected value, obtains a prediction error compensation amount compensated so that the prediction error converges to 0, and the prediction error compensation amount is used to calculate the prediction error compensation amount. The present invention relates to a digital feedback control device characterized in that it corrects a predicted control amount value. The method described in claims 2 to 4 can be adopted.

[0020]

According to the invention of each claim. Since the prediction error indicating the disturbance component is added to the output of the proportional-plus-integral compensation means or the proportional-compensation means, it is possible to compensate for the disturbance component that changes rapidly without delay, and obtain good control characteristics.

[0021]

[First Embodiment] Next, a feedback control system according to a first embodiment of the present invention will be described with reference to FIG. However, in FIG. 9, the same parts as those in FIGS. 1 to 8 are designated by the same reference numerals and the description thereof will be omitted. FIG. 9 shows the circuit of FIG. 8 with an adder 21 added. That is, in FIG. 9, the adder 21 is added to the output stage of the controller 7 as the proportional-plus-integral compensating means or the proportional-compensating means.
Is provided, the output of the controller 7 and the compensated prediction error obtained from the prediction error compensator 15 are added, and the manipulated variable V
(N) is formed.

In FIG. 9, the relationship between the disturbance Vl and the controlled variable Il is expressed by the following equation. VL = (L / kTs) (z-1) Il / {m (z-1)
+1} i (n-1) and I (n-1) are matched by the action of the prediction error compensator 15, the output Vl '(n) of the prediction error compensator 15 is controlled by the predictor 9a. Since the model is 10, it corresponds to the disturbance component Vl (n) in the controlled object. Therefore, the output Vl 'of the prediction error compensator 15
The output of the controller 7 is added or subtracted so that (n) cancels the disturbance Vl (n). This makes it possible to compensate for a fast-changing disturbance regardless of the delay in the controller 7 as the proportional-plus-integral compensation means.

[0023]

[Second Embodiment] A second embodiment of the present invention will now be described with reference to FIG.
The feedback control device of the embodiment will be described. However, in FIG. 10, the same parts as those in FIG. 9 are designated by the same reference numerals and the description thereof will be omitted. In this embodiment, the output of the prediction error compensator 15 is added to the output of the predictor 9b using the adder 12. Others are the same as those in FIG. 9, and thus the same effects can be obtained.

[0024]

[Third Embodiment] FIG. 11 shows a three-phase UPS to which the present invention is applied.
2 shows an inverter device for a vehicle This device is configured by connecting a bridge type three-phase inverter circuit 31 to a DC power supply 30 and connecting the inverter circuit 31 to a load 34 via a reactor 32 and a transformer 33.
A switching element (not shown) of the inverter circuit 31 is controlled by the PWM control circuit 35. This PWM control is performed so that the voltage applied to the load 34 becomes a constant value.

In this embodiment, the inverter circuit 31, the reactor 32, the transformer 33 and the load 34 can be regarded as the controlled object 10 in FIG. PW of the inverter 31
The means for controlling the M control circuit 35 includes an output voltage detection transformer 36, an AD converter or ADC 37, a filter capacitor 38, a current detector 39, an ADC 40, a clock signal generator 41, and a 1/4 frequency division. And a digital signal processor or DSP 43.

As shown in FIG. 11, the DSP 43 includes a reference voltage generator 44, first and second predicting means 45 and 46,
The first and second comparators 47 and 48, the first and second proportional compensators 49 and 50, and the third and fourth comparators 51 and 5
2, first and second proportional-integral compensators or proportional compensators 53 and 54, first and second adders 55 and 56, proportional gain circuits 57 and 58, and a subtractor 59, The U, V, and W phase control data are output to the output lines 60, 61, and 62, and these are supplied to the PWM control circuit 35. Figure 9
11 and the correspondence relationship between FIG. 11 and FIG.
4, the first and second comparators 47 and 48, and the first and second proportional compensators 49 and 50 correspond to the command value generator 3 in FIG. The third and fourth comparators 51 and 52 correspond to the subtractor 6 in FIG. First and second predicting means 45,
Reference numeral 46 corresponds to the prediction means 9a, the prediction error detector 14, and the prediction error compensator 15 in FIG. The first and second adders 55 and 56 correspond to the adder 21 of FIG.

The digital reference voltage generator 44, which corresponds to the command value generator 3 of FIG. 9, is composed of a ROM in which sine wave data is stored, and generates two-phase reference signals composed of digital sine wave data. U-phase and W-phase inverter output voltages are detected by the transformer 36 and the ADC 37 to form a constant-voltage control major loop, and are sent to the first and second comparators (error amplifiers) 47 and 48. An error output between the reference voltage and the detected voltage is obtained from the comparators 47 and 48, and this becomes the command value Ir (n) through the proportional compensators 49 and 50. The comparators 51 and 52 use the command value Ir
(N) and the control prediction value i obtained from the prediction means 45 and 46
The value of the difference from (n) is calculated. The proportional-plus-integral compensator or the proportional compensators 53 and 54 are the same as the controller 7 in FIGS. 2 and 9. The adders 55 and 56 add manipulated variables V (n) by adding the prediction error Vl '(n) as the disturbance component obtained from the predicting means 45 and 46 to the outputs of the proportional-plus-integral compensators or proportional-compensators 53 and 54. Is generated, and the manipulated variables Vu and Vw of the U phase and the W phase are sent to the lines 60 and 62. Subtractor 59 uses U phase and W
The V-phase manipulated variable Vv is created based on the phase manipulated variable and sent to the line 61.

The predicting means 45 and 46 are provided in the minor loop. U-phase and W-phase of the Y-connected capacitor 38
The phase current is detected by the current detector 39 and then detected by the ADC 40.
Is AD-converted and sent to the predicting means 45 and 46 as the control amount detection value I (n-1, m). The prediction means 45 and 46 also accept Vu and Vw as the manipulated variables V (n), and the prediction means 9a, the prediction error detector 14, and the prediction error compensator 1 of FIG.
The same processing as in step 5 is executed to output the control amount prediction value i (n) and the prediction error (disturbance component) Vl '(n).

The PWM control circuit 35 controls the manipulated variables Vu, Vv,
The data indicating Vw is converted into a PWM pulse and sent to the switching element of the inverter circuit 31. The clock pulse generator 41 includes a PWM control circuit 35, a DSP 43, and an A
The frequency divider 42 which is connected to the DC 40 and divides the output of the clock pulse generator 41 into quarters includes a DSP 43 and an A
It is connected to the DC 37. Clock pulse is DSP4
3 becomes an interrupt signal, and a program for detecting a control amount and performing arithmetic processing is activated and executed by this interrupt signal.

The operation of this device is shown by a DSP surrounded by a chain line.
It is realized by software 43. FIG. 12 shows a flow chart for one phase for only the portion of the software relating to the present invention. A, C, E, and G in the figure are work registers. The register values are A, C,
It will be indicated by E and G. k, Ln, m and Ts are constants. V, I, i, Δi and Ir are variables. D (z)
Is a function. The flowchart will be described below.

Initialization is performed in the first step 70 after the start. This initialization is a reset of working registers and variables. Next, in step 71, an interrupt is awaited. Next, at step 72, the manipulated variable V calculated in the previous processing is output. Next, at step 73, the control amount I is read from the ADC 40. Then in step 74 the ADC
The control amount predicted value calculated in the previous process is subtracted from the control amount read from 40 and stored in the register A. That is, the control amount prediction error detection value is stored in the register A. Next, in step 75, the value of register C is added to the value obtained by multiplying the value of register A by a constant k, and the answer is stored again in register C. As a result, the prediction error compensation amount is stored in the register C.
Next, at step 76, the predicted error compensation amount C of the register C is added to the manipulated variable V output at step 72 and stored in the register E. Next, as shown in step 77, the value of the register E is multiplied by a constant K and further multiplied by (1-m) Ts and divided by Ln, and the answer is stored as the control amount increase predicted value Δi. Next, as shown in step 78, the control amount I is set to the ADC
Read from 40. Then, as shown in step 79,
The control amount increase predicted value Δi is added to the control amount I read from the ADC 40 to obtain the control amount predicted value i, and this is stored in the register. Next, as shown in step 80, the command value ir
The control amount predicted value i is subtracted from this and stored in the register G.
Next, as shown in step 81, the manipulated variable V is obtained from the value of the register G and the function D (z) and stored in the register. Next, as shown in step 82, the value of register C is added to V of step 81 and stored in the register. Then, the process returns to step 71. Here, the time from step 71 to the start of step 78 is mTs time. This program is activated and executed by an interrupt at Ts time intervals.

From the above, the prediction error converges to 0 at the time constant of the integral element unless there is a sudden change in the constant of the controlled object between the sampling points (change in L in this example). As a result, even if there is an approximation error or variation in the model (modeled by Ln in this example) for the controlled object in the controlled variable predictor, the controlled variable is accurately predicted, and the sampling interval and control interval peculiar to digital control are determined. It is possible to compensate for the deterioration of the control performance caused by this. This method is a method in the case where there is a one-sample delay when m is 0. Therefore, it can be applied to the conventional prediction method. Further, the disturbance is compensated without delay of the proportional-plus-integral compensator or the proportional compensators 53 and 54.

[Brief description of drawings]

FIG. 1 is a block diagram showing a conventional feedback control device.

FIG. 2 is a block diagram equivalently showing the feedback control device of FIG.

FIG. 3 is a diagram showing a relationship between a control amount and an operation amount in FIG.

FIG. 4 is a block diagram showing another conventional feedback control device.

5 is a diagram showing a control amount, a control amount predicted value, and an operation amount of FIG.

FIG. 6 is a block diagram showing still another conventional feedback control device.

FIG. 7 is a diagram showing the control amount, the control amount predicted value, and the operation amount of FIG.

FIG. 8 is a block diagram functionally obtained by modifying the feedback control device of FIG.

FIG. 9 is a block diagram showing a feedback control device according to the first embodiment of the present invention.

FIG. 10 is a block diagram showing a feedback control device of a second embodiment.

FIG. 11 is a block diagram showing a three-phase inverter device according to a third embodiment.

12 is a flow chart for determining a manipulated variable with the device of FIG.

[Explanation of symbols]

 3 command value generator 7 digital controller 10 controlled object 9a controlled variable predicting means 14 prediction error detector 15 prediction error compensator 21 disturbance adder

Claims (4)

[Claims] 1. A digital control amount detecting means for detecting a control amount of a controlled object or a corresponding amount at a predetermined sampling time interval to obtain a control amount detection value as a digital amount, and the control amount detection value. A digital command value generating means for generating a command value to be followed as a digital amount, and a digital operation of the operation amount for the controlled object at a predetermined control time interval so that the control amount detection value follows the command value. In the digital feedback control device comprising the digital control means for controlling the controlled object determined by the above, the digital control means includes a comparison means, a proportional integral compensation means or a proportional compensation means,
It has an addition means, a control amount prediction means, and a prediction error detection and compensation means, and the comparison means outputs an output corresponding to a difference between the control amount predicted value corrected by the prediction error compensation amount and the command value. The proportional-integral-compensating means or the proportional-compensating means performs proportional-integral-compensation or proportional-compensation on the output obtained from the comparing means, and the adding means includes the proportional-integral-compensating means or the proportional-compensating means. A predictive error compensation amount obtained from the predictive error detection and compensation means is added to an output to output an operation amount for controlling the controlled object, and the controlled variable predicting means outputs from the adding means. Based on the manipulated variable, the predicted value for the change between the current controlled variable and the controlled variable at the next sample is calculated by calculation, and the predicted value for the changed amount is used as the controlled variable detected at the current sample. In order to obtain the predicted value at the next sample by adding, the prediction error detecting and compensating means calculates the prediction error by the difference between the control amount predicted value and the actual control amount detected value, and the prediction error A digital feedback control device, characterized in that a prediction error compensation amount compensated so as to converge to 0 is obtained, and the control amount prediction value is corrected by this prediction error compensation amount. 2. A digital control amount detecting means for detecting a control amount of a controlled object or a corresponding amount at a predetermined sampling time interval to obtain a control amount detection value as a digital amount, and the control amount detection value. A digital command value generating means for generating a command value to be followed as a digital value, and a control time interval for presetting an operation amount for obtaining a control amount detection value corresponding to the command value from the digital control amount detecting means. In the method for controlling the controlled object by a digital feedback control device comprising a digital control means for controlling the controlled object, which is determined by a digital operation, the current controlled variable based on the manipulated variable output to the controlled object and The predicted value for the change from the control amount at the next sample is calculated and the predicted value for the change is calculated at the current sample. By adding to the detected control amount, a predicted value of the control amount at the next sampling is obtained, and a prediction error composed of the difference between the control amount predicted value and the actual control amount detected value is calculated, and this prediction error is set to 0. A prediction error compensation amount that is compensated so as to converge to the control amount prediction value is corrected by this prediction error compensation amount, and the difference between the control amount prediction value corrected by the prediction error compensation amount and the command value is corrected. A digital value characterized in that the operation amount is obtained by adding the prediction error compensation amount to the proportional-integral compensation or proportional-compensated value. Feedback control method. 3. The predicted value of the control amount is obtained by adding 1 to the manipulated variable based on a value obtained by adding the predicted error compensation amount to the manipulated variable.
The step of obtaining the predicted value of the change of the sample period, and the step of adding the predicted value of the change to the control amount detection value to obtain the predicted value of the control amount, Detecting a prediction error based on a detected value and the control amount prediction value, and compensating the prediction error obtains a prediction error compensation amount through a prediction error compensator including an integral element for the prediction error. The feedback control method according to claim 2. 4. The step of obtaining the predicted value of the control amount includes the step of obtaining a predicted value of a change in one sample period based on the manipulated variable, and the predicted value of the change and the prediction error compensation amount. The step of obtaining a control amount predicted value by adding a control amount detected value, and obtaining the prediction error is to detect a prediction error based on the control amount detected value and the control amount predicted value. The feedback control method according to claim 2, wherein compensating the prediction error is to obtain the prediction error compensation amount through a prediction error compensator including an integral element.