JP2001075607A - Electropneumatic positioner - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve valve control by calculating and feeding back the difference between at least one state variable among predicted state variables and measured data. SOLUTION: A controller 26 of a control arithmetic part 22 inputs signals relating to an input signal SP and a signal from a displacement sensor 15 which detects the displacement (y) of a system 14 to generate a controlled variable U(t), a feedback gain part 27 inputs the signal from the displacement sensor 15 and a state variable predicted value to send a feedback gain signal out to the controller 26, and a state observer 28 inputs the signal from the displacement sensor 15 and the controlled variable U(t) from the controller 26 to generate a state variable predicted value. The input signal to a valve and the displacement signal of a stem are inputted to output a controlled variable, and the flow rate of air supplied to an actuator is controlled. Then the state variables of the air flow rate and the displacement quantity of the stem are predicted and the difference between at least one state variable out of the predicted state variables and measured data is calculated and fed back.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to an electropneumatic positioner, and more particularly to an electropneumatic positioner which calculates a predicted value of valve control based on a stem displacement signal for controlling a valve.

[0002]

2. Description of the Related Art Conventionally, valve control algorithms have been studied in various ways by positioner vendors, and are greatly reflected in the competitiveness of products. The most difficult factor in controlling a valve is how to absorb the nonlinearity of the valve to be controlled, and set the valve opening more precisely, faster, and more stable. Value. Among the non-linearity of the valve, the characteristic that is most difficult to control is
This is a hysteresis caused by a frictional force of a gland packing provided to prevent the fluid in the valve from being transmitted to the stem and leaking.

[0003] In recent years, attention has been paid to efficient operation of a plant, and the pressure of a fluid flowing through piping of the plant has increased, and the temperature of the fluid has also increased. In order to prevent such high-pressure, high-temperature fluid leakage, asbestos-based gland packing, which is considered to have a large frictional force, is used, and the asbestos-based packing tends to be tightly tightened against the valve stem. Also, due to environmental problems, similar measures are taken to reduce leakage of fluid in the valve. That is, the frictional force generated between the valve stem and the gland packing increases, which increases the hysteresis of the valve. Even if the frictional force increases, the valve hysteresis can be suppressed by increasing the force of the air actuator of the valve.However, it is difficult to change the air actuator under the current situation where capital investment is suppressed. It is often operated with the structure as it is. As a result, the hysteresis of the valve increases, and it is desired that this wrinkle be absorbed by the electropneumatic positioner.

Various methods have been studied to control a valve having a large valve hysteresis. However, the history of digital control in this field is short, and as a control algorithm, PID control (proportional control) for controlling output based on an input signal is used. , Integration, and differential operation) are the mainstream.

[0005] An electropneumatic positioner 11 that controls based on the PID control, that is, a field-back only the so-called stem (valve stem) displacement, as shown in FIG.
Detecting the displacement y of the stem 14 of No. 3 and feeding it back,
It controls the air flow rate Q to the actuator 12. The electropneumatic positioner 11 includes a displacement sensor 15 for detecting a displacement y of the stem 14, an A / D converter 16 for converting a signal from the displacement sensor 15 from analog to digital,
A digital control circuit 17 for inputting the input signal SP and the stem displacement signal y from the A / D converter 16 and executing a control operation, and a D / A converter for converting the operation signal of the digital control circuit 17 into an analog signal 18, an input module 19, which is a current / pressure converter for controlling the air pressure PN supplied based on the converted analog signal, and the signal of the input module 19 and the air flow rate Q to the actuator 12 from the supply pressure PN. Control relay 20 to control
It is composed of

The electropneumatic positioner 11 having such a configuration is
The control operation is executed based on the state of the stem displacement y due to the pressure of the compressed air and the applied input signal SP so that a normal valve operation can be performed.

However, in the valve control by the PID control, since only the displacement (valve opening degree) y of the stem 14 is fed back to calculate the operation signal, it is difficult to absorb the nonlinearity of the controlled object. There is a disadvantage that there is.

[0008] Japanese Patent Publication No. 9-502292 discloses a solution to this problem. In this technique, an output pressure of an air supply pressure of a valve position control device (electro-pneumatic positioner) is detected by a pressure sensor, and a differential value thereof is used for control. Further, this technique is characterized in that not only the stem displacement y but also the differential value of the air flow signal p related to the operation signal to the actuator 12 is used for feedback control.

As shown in FIG. 9, the valve position control device (electro-pneumatic positioner) 11A having the above-described characteristics provides feedback by both the displacement y of the stem 14 and the value obtained by differentiating the pressure p to the actuator 12. This controls the air flow rate Q to the actuator 12.
That is, the input signal SP, the displacement sensor 15 and the A / D
The feedback of the signal corresponding to the stem displacement y given from the converter 16 and the data obtained by giving the pressure signal p measured by the pressure sensor 21 to the differentiating circuit 22 and converting it into the air flow rate signal q to the digital control circuit 17A. It is a feature. The output signal of the digital control circuit 17A is supplied to the D / A converter 18 in the same manner as in FIG. 8, and further supplied to the actuator 12 via the input module 19 and the control relay 20 as an air flow rate Q as an operation signal. You.

[0010] The electropneumatic positioner 11A having such a configuration.
Takes into account the air flow signal Q in the input signal SP and the stem displacement y, performs a control operation using these three signals, and executes a normal valve operation.

[0011]

However, since the valve control described with reference to FIG. 9 is configured to feed back not only the stem displacement y but also the air flow rate p, the provision of the pressure sensor 21 is essential. For this reason, the pressure sensor 21 has to be installed in the electropneumatic positioner, and there is a fundamental problem that the cost increases accordingly.

Further, by adding the pressure sensor 21, if the pressure sensor 21 does not operate normally, or if a failure that it does not operate at all occurs, the valve control becomes abnormal. There is also a danger that the electro-pneumatic positioner itself may be degraded, and the reliability of the electropneumatic positioner itself may be reduced if an accurate control is attempted by adding the pressure sensor 21.

Further, a field device including an electropneumatic positioner generally operates at 4 to 20 mA.
There is a problem that an internal circuit (a hardware that consumes current, such as a CPU, an actuator, and a sensor) must be driven by A. Therefore, naturally, a basic design of the current consumption due to the increase in the pressure sensors 21 is required, and the design becomes extremely strict. In other words, in recent years, circuit devices for reducing current consumption have been
There is a background that it becomes stricter with the addition of arithmetic functions such as PUs, which is causing trouble for engineers.

Further, by newly installing the pressure sensor 21, new problems such as a basic design, a pneumatic circuit design, and a reliability design are not limited to the design of the electropneumatic positioner alone. There is also a spillover problem.

Further, even if the pressure sensor 21 is mounted, a physical quantity that cannot be calculated by simple detection or a physical quantity that cannot be measured due to the structure of the device cannot be measured, so that the use range is limited.

Therefore, in the control of the valve, not only the displacement of the stem but also the pressure of the electropneumatic positioner is fed back, so that the nonlinearity of the valve can be easily absorbed and the controllability of the electropneumatic positioner can be improved. In addition to being able to do so, there is a problem that it is necessary to solve a control method that does not hinder the design without separately mounting a pressure sensor.

[0017]

In order to solve the above-mentioned problems, an electro-pneumatic positioner according to the present invention is an electro-pneumatic positioner for controlling a valve by supplying gas pressure to an actuator to displace a stem. Control operation means for inputting an input signal to the valve and a displacement signal of the stem and outputting a predetermined control amount; and electropneumatic conversion means for controlling a gas flow rate supplied to the actuator based on the control amount. Inputting the control amount, predicting the state variables of the gas flow rate and the displacement amount of the stem, and calculating and feeding back a difference between at least one of the predicted state variables and the measured data. And a state observing means.

Further, the state observation means includes a predetermined mathematical model in advance, and converts the control data by the mathematical model to calculate a predicted value of a state variable; A first-order lag system model that integrates a gas flow value and has a time constant determined by the gas flow rate; the mathematical model is approximated to a first-order lag system model having a time constant determined by the gas volume capacity of the actuator. The state observation means receives the control amount and the stem displacement signal and calculates a predicted value of a state variable of the gas flow rate and the displacement amount of the stem; Outputting the controlled variable based on the predicted value obtained by the state observation means as self-diagnosis data of the control calculation means and self-diagnosis of the valve; Serial predicted state variable is to set changing the eigenvalues by changing the gain of the feedback as appropriate.

As described above, the electropneumatic positioner is provided with the state observation means for predicting the output pressure of the air pressure, and the predicted pressure value is used for controlling the valve, so that the non-linear valve can be provided without providing the pressure sensor on the stem. The stem can be controlled accurately.

Further, since the predicted value is calculated by the state observation means based on the value of the stem displacement originally measured by the displacement sensor or the like, the accuracy of the predicted value can be improved.

Further, the state observing means can also predict physical quantities that could not be detected due to various constraints, so that parameters that could not be used for control can be used for control.

[0022]

Next, an embodiment of an electropneumatic positioner according to the present invention will be described with reference to the drawings. still,
The same components as those described in the related art will be described with the same reference numerals.

As shown in FIG. 1, the electropneumatic positioner 11B of the present invention is a positioner that controls the valve 13 by displacing the stem 14 by supplying an air flow rate Q to the actuator 12. The configuration includes a control operation unit 22 that calculates a control amount U (t) based on a mathematical model,
And an electropneumatic converter 23 for controlling the air flow rate Q (t) supplied to the actuator 12 based on (t).

The control operation unit 22 receives the input signal SP and a signal related to the signal from the displacement sensor 15 for detecting the displacement y of the stem 14 to generate a control amount U (t), A feedback gain unit 27 that receives the signal from the sensor 15 and the predicted value of the state variable and sends a feedback gain signal to the controller 26
And a state observer 28 that receives a signal from the displacement sensor 15 and a control amount U (t) calculated by the controller 26 to generate a predicted state variable value. The controller 26 calculates the control amount U (t) based on the input signal SP and the gain signal from the feedback gain unit 27, and outputs the control amount U (t) to the state observer 28 and the electromagnetic actuator 29 in the electropneumatic conversion unit 23 at the subsequent stage. It has a configuration.

The electropneumatic converter 23 controls an air flow rate based on the control amount U (t) from the control operation section 22 and an air flow rate Q to the air actuator 12 under the control of the electromagnetic actuator 29. (T).

The controlled object inputs the air flow rate Q (t) controlled by the pilot relay 30 and
, A stem 14 that is displaced by the operation of the air actuator 12, and a valve 13 whose opening is controlled by the displacement of the stem 14. The air actuator 12 includes an air chamber 12a that accommodates air with a predetermined volume V, a diaphragm 12b that operates with pressure in the air chamber 12a, and a diaphragm 1b.
And a spring 12c for resisting the movement of the diaphragm 12b to displace the stem 14 in conjunction with the movement of the diaphragm 12b.

The electropneumatic positioner of the present invention thus configured functionally inputs an input signal SP (target value V (t)) as a set point and controls the stem 14 of the valve 13. Therefore, a 1-input, 1-output control system is basically configured. Further, the parameter that can be detected can be only the stem displacement. The control system in such a case can be virtually represented by functional blocks as shown in FIG. In this case, the control targets include the electropneumatic converter 23 in the electropneumatic positioner 11B and the valve side (the air actuator 12, the valve 13, and the stem 14). As shown in FIG.
Is supplied to the control target side, and the stem displacement Y (t) from the control target is supplied to the state observer 28. The state observer 28 calculates a predicted value X ＾ (t) based on the state variable X (t) based on the stem displacement Y (t), and supplies the calculated value X ＾ (t) to the calculation unit 27 together with the target value V (t). The output of the calculation unit 27 is output to the controller 26.

On the other hand, the state equation and the output equation of the mathematical model to be controlled in this control system can be assumed as in the following equations (1) and (2).

DX (t) / dt = A * X (t) + B * U (t) (1) Y (t) = C * X (t) (2)

If the state feedback matrix is K and the controller 26 is a proportional controller having a gain of one, the state equations (1) and (2) of the control system described above are used.
Is as follows: dX (t) / dt = A * X (t) + B * (-K * X (t) + V (t)) (3) Y (t) = C * X (t) (4) Expression (3) can be modified as in the following expression (5). dX (t) / dt = (A−B * K) X (t) + B * V (t) (5)

X (t) is determined by solving equation (5), and the behavior of the state variable X (t) is determined by the eigenvalue of A-BK. Further, by setting the state feedback matrix K, the value of A-BK can be changed arbitrarily.

As described above, if the mathematical model mounted on the state observer 28 is composed of blocks as shown in FIG. 2, the above-described assumed state equation (Equation (1))
And the output equation (Equation (2)) can be expressed by approximating Equations (6) and (7) below. dX ＾ (t) / dt = A * X ＾ (t) + B * U (t) + G * {Y (t) -W (t)} (6) W (t) = C * X ＾ (t) ( 7)

From the equations (6) and (7), dX ＾ (t) / dt = (A−G * C) * X ＾ (t) + G * Y (t) + B * U (t) (8) Holds. Here, the matrix G is a state observer feedback gain matrix.

Therefore, by solving this equation (8),
The predicted value X ＾ (t) of the state variable X (t) can be obtained. In addition, the characteristic of the state observer 28 can be determined based on the eigenvalue of [A-G * C], and [AG-C]
Can be arbitrarily varied by setting a matrix G. Furthermore, by multiplying the predicted value X ＾ (t) by the feedback matrix K, the arithmetic unit 27
To the feedback signal. That is,
Instead of the state variable X (t) in equation (5), a predicted value X ＾
By using (t), a similar control system can be established.

The problem here is how to determine the A, B, and C matrices. That is, the electropneumatic positioner 11
How to control the electropneumatic conversion unit 23 for converting the electric signal in B into the air flow rate is controlled by various methods. The most common one is that an electromagnetic actuator 29 that drives a nozzle flapper converts a current signal into a nozzle back pressure, which is an air signal, and a pilot relay 30 that converts an air pressure corresponding to the nozzle back pressure into an output pressure to output air. It determines the flow rate.

As shown in FIG. 3, the electro-pneumatic converter 23 extracts a main component of the current i (t) corresponding to the control amount U (t) from the controller 26, as shown in FIG. An electromagnetic actuator 29 to be input, a pilot relay 30 that inputs a nozzle back pressure pn (t) from the electromagnetic actuator 29 to output an air flow q (t), and an air flow q (t) from the pilot relay 30 Pneumatic actuator 12 which outputs as control pressure p (t)
It is composed of

In such a configuration, assuming a model below, the conversion from the current i (t) to the control pressure pn (t) is a first-order lag system determined by a time constant determined by the air capacity in the configuration. is there. Further, this air capacity is determined by the opening area of the exhaust valve or the intake valve inside the pilot relay 30,
The value is determined by the air capacity of the valve air actuator 12 and the like.

Here, the electropneumatic conversion gain of the electromagnetic actuator 29 is Ga, and the time constant of the nozzle back pressure Pn (t) is Tn.
When expressed by a Laplace-transformed equation, Pn (S) / I (S) = Ga / (I + Tn * S) (9)

On the other hand, the output pressure p of the electropneumatic positioner 11B
(T) is given by Laplacian conversion assuming that the air pressure amplification gain of the pilot relay 30 is Gp and the time constant of the first-order lag determined by the capacity of the air actuator 12 on the valve 14 side is Tp, and P (S) / Pn (S) = Gp / (1 + Tv * S) (10)

As shown in FIG. 4, the air actuator 12 for controlling the valve includes an air chamber 12a for inputting an air pressure p (t), a diaphragm 12b (area a) provided on the bottom surface of the air chamber 12a. And a spring 12c (spring constant k) which resists the pressure of the diaphragm 12b.

In the air actuator 12 having such a configuration, the stem displacement y is calculated from the air pressure p (t).
When the equation of motion up to (t) is expressed by a Laplace transform equation, Y (S) / P (S) = a / (mS * S + cS + k) (11)

Accordingly, as shown in FIG. 5, the entire block including the electro-pneumatic converter 23 has the above-mentioned equations (9) and (1).
0) and (11) can be represented by a mathematical model.

Here, when formulas of internal variables are prepared in more detail, y (t) = x1 (t) (12) x2 (t) = dx1 (t) / dtp (t) = x3 (t Assuming that pn (t) = x4 (t), dx1 (t) / dt = x2 (t) dx2 (t) / dt = -c / m * x2 (t) -k / m * x1 (t) + a / M * x3 (t) dx3 (t) / dt = -1 / Tv * x3 (t) + Gp / Tv * x4 (t) dx4 (t) / dt = -1 / Tn * x4 (t) + Ga / Tn * I (t).

To summarize these equations, the state equation of the system can generate the following equation (13).

[0045]

(Equation 13)

As an output equation, the following equation (14) can be generated from the equation (12).

[0047]

(Equation 14)

As described above, the above equations (13) and (1)
According to 4), the matrices A, B, and C can be obtained, and the block diagram of the mathematical model at this time is as shown in FIG.

In the above equations (13) and (14), four state variables of the system are set, but a more elaborate model is constructed by further increasing the state variables to create a detailed model. can do. Conversely, there is a disadvantage in that the amount of computation increases if there are too many state variables. However, if Tn << Tv in Tn and Tv in FIG. 6 described above, the operation of x3 (t) is dominated by Tv, and x4 (t) becomes x4 (t).
(T) can be omitted.

As shown in FIG. 7, when the gain from the current i (t) to the pressure x3 (t) is G1, the state equation and the output equation of the system are as follows.
5) and equation (16) can be obtained.

[0051]

(Equation 15)

[0052]

(Equation 16)

In this way, even when there are three state variables, the matrices A, B, and C can be determined. If such a state variable is set to a predetermined value in advance, it can be used as self-diagnosis data of the control operation unit 22 and self-diagnosis data of the valve 13.

[0054]

As described above, by providing the electropneumatic positioner with the state observation means, it is possible to predict a state variable for controlling the valve, and use the predicted value for controlling the valve. There is an effect that control of the valve can be improved.

Conventionally, an air pressure sensor had to be separately provided in order to measure some of the state variables used for valve control. However, if a state observation means is provided, a mathematical model is required. This makes it possible to predict and process the state variables by using the parameters, and has the effect that accurate valve control can be performed without providing a separate air pressure sensor.

Further, in the related art, among the state variables of the valve control, some parameters cannot be measured and cannot be used for the valve control. However, the state variables can be predicted by installing the state observation means. Since this predicted value can be used for control, the following advantages are obtained. (1) In order to measure a state variable of a control target, a new air pressure sensor is not necessary as described above. (2) Since there is no need to attach an air pressure sensor, a low-cost electropneumatic positioner can be realized. (3) The reliability of the electropneumatic positioner can be improved by control based on predicted values without providing an air pressure sensor. (4) If the air pressure sensor is not provided, the design of the device becomes easier. (5) Parameters that could not be detected until now can be controlled, and more precise and accurate control can be performed. (6) Since the internal model can be simplified by approximating the model inside the electropneumatic positioner with a first-order lag element having a time constant determined by the air actuator capacity of the valve, the state observation means is introduced. Also, the load on the calculation unit can be reduced.

[Brief description of the drawings]

FIG. 1 is a block diagram schematically showing an electropneumatic positioner according to the present invention.

FIG. 2 is a block diagram showing a control between the electropneumatic positioner and a control target in a mathematical model.

FIG. 3 is a block diagram showing a schematic configuration of a control target connected to the electropneumatic positioner.

FIG. 4 is an explanatory view schematically showing a mechanism of the valve.

FIG. 5 is a block diagram showing control in the valve mechanism by a mathematical model.

FIG. 6 is a block diagram showing a mathematical model based on four parameters in the state observer.

FIG. 7 is a block diagram showing a mathematical model based on three parameters in the same state observer.

FIG. 8 is a schematic block diagram of an electropneumatic positioner that performs valve control only by displacement of a stem in the related art.

FIG. 9 is a schematic block diagram showing a valve position control device for controlling a valve based on a displacement of a stem and a signal from a pressure sensor of air pressure according to the related art.

[Explanation of symbols]

11; electropneumatic positioner, 11A; electropneumatic positioner, 11
B; electropneumatic positioner, 12; actuator, 13; valve, 14; stem, 15; displacement sensor, 16;
D converter, 17; digital control circuit, 17A; digital control circuit, 18; D / A converter, 19; input module, 20; control relay, 21; differentiating circuit, 22; control operation unit (control operation means), 23; electro-pneumatic converter (electro-pneumatic converter), 24; pipe, 25; fluid flow rate detector, 26; controller, 27; feedback gain unit, 28; state observer (state observing means), 29; 30; pilot relay

Claims (8)

[Claims] 1. An electro-pneumatic positioner for controlling a valve by displacing a stem by supplying gas pressure to an actuator, wherein a predetermined control amount is inputted by inputting an input signal to the valve and a displacement signal of the stem. Control arithmetic means for outputting the control amount; electropneumatic conversion means for controlling a gas flow rate supplied to the actuator based on the control amount; and inputting the control amount to obtain a state variable of the gas flow rate and the displacement amount of the stem. A valve electro-pneumatic positioner comprising: a state observing means for predicting and calculating a difference between at least one of the predicted state variables and the measured data and feeding back the difference. 2. The state observation unit according to claim 1, wherein the state observation means includes a predetermined mathematical model in advance, and calculates the predicted value of the state variable by converting the data of the control amount using the mathematical model. Valve position control device. 3. The electropneumatic device according to claim 2, wherein the mathematical model is a first-order lag system model that integrates a gas flow value to the actuator and has a time constant determined by the gas flow. Positioner. 4. The electropneumatic positioner according to claim 2, wherein the mathematical model is approximated to a first-order lag system model having a time constant determined by a gas volume capacity of the actuator. 5. The apparatus according to claim 1, wherein the state observation means inputs the control amount and the stem displacement signal and calculates a predicted value of a state variable of a gas flow rate and a displacement amount of the stem. Electro-pneumatic positioner as described. 6. The electropneumatic positioner according to claim 1, wherein the control calculation means outputs the control amount based on the predicted value. 7. The predicted value obtained by the state observation means is:
2. The electropneumatic positioner according to claim 1, wherein the self-diagnosis data of the control calculation means and the self-diagnosis of the valve are used. 8. The electro-pneumatic positioner according to claim 1, wherein the predicted state variable changes a setting of an eigenvalue by appropriately changing a gain to be fed back.