JPS6323561B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a process control device, and more particularly to an improvement in a process control device with a reset limiter.

Process control equipment with reset limiters have
There are two types of devices: devices in which the output limiter is released when the deviation decreases (hereinafter referred to as integral feedback type) and devices in which the output limiter is not released until the polarity of the deviation changes (hereinafter referred to as deviation feedback type). The integral feedback type is used for batch control that requires the prevention of reset wind up, and the deviation feedback type is used for controlling safety valves that are always closed and open in an emergency.

The present invention realizes a novel process control device with a reset limiter that can operate as either an integral feedback type or a deviation feedback type by selecting the magnitude of the reset bias value.

FIG. 1 is a block diagram showing one embodiment of the apparatus of the present invention. In the figure, reference numeral 1 denotes a control calculation unit that performs a proportional integral calculation on the deviation e between the measured value PV and the set value SV, which has a proportional calculation element P and an integral calculation element I, and the output K _P of the proportional calculation element P. e (K _P is a proportional gain) and the output B of the integral calculation element I (hereinafter referred to as integral term B) are added to generate the calculation output YV. Either the output K _{Pe of the proportional calculation element P} or the output FV of the adder/subtractor 4, which will be described later, is selectively applied to the integral calculation element I by the changeover switch 5. When the selector switch 5 is connected to the a side, the integral calculation element I outputs the output K _P e of the proportional calculation element P.
is integrated to determine the integral term B, and when connected to the b side, the integral term B is set to a certain time constant (for example, integration time T _I )
The configuration is such that the value is set to a value determined by the output FV of the addition/subtraction section 4. The integral term B can be easily set to a value determined by FV, for example, by monitoring the integral by the output FV of the adder/subtractor 4 and inhibiting the integral operation when B=FV. 2 is an output limiter, when the PI calculation output YV exceeds the limit value LM (upper limit value H or lower limit value L),
This limits the manipulated output MV to LM. 3 is a circuit for setting the reset bias value RB, and a potentiometer is shown. The addition/subtraction unit 4 adds and subtracts the operation output MV, the reset bias value RB, and the output K _Pe of the proportional calculation element, and calculates MV+RB-
Generates an output FV of value K _P e. The changeover switch 5 is normally connected to the a side, and the control calculation section 1
When the output YV exceeds the limit value LM of the limiter 2, the connection is switched from the a side to the b side, and when YV becomes within the limit value of the limiter 2, the connection is switched to the a side again.

The operation of the present invention configured in this way will be explained below using FIGS. 2 to 4. Now, PV=SV is controlled by the proportional-integral operation of the control calculation section 1, and at this time, the proportional bands PB are arranged above and below the set value SV as shown in FIG. In this state, the proportional calculation output K _p e of the control calculation section 1 is zero, so
The integral term B has the same value as the manipulated output MV (=YV). When the measured value PV decreases due to a process shutdown or the like (at time _t0 ) and a deviation e occurs, the manipulated output MV and the integral term B increase based on the proportional calculation output _Kpe . If this continues, reset wind up will occur, but when the PI calculation output YV reaches the limit value LM of the output limiter (at t ₁ ), the output limiter 2 limits the manipulated output MV to LM, and
Switch the changeover switch 5 from side a to side b. As a result, the output of the addition/subtraction unit 4 is added to the input of the integral calculation element I.
FV (= MV + RB - K _p e) is added, I is FV
Integrate. Here, when the reset limiter value RB is set to zero, since MV=LM, the integral term B decreases according to the integration time T _I , and finally becomes LM
−K _p e _s . Note that e _s is the steady-state deviation. Therefore, the proportional body PB moves as shown in Fig. 2B, and unless the deviation e decreases, the manipulated output will decrease.
MV is also kept at the limit value LM. Therefore, when the measured value PV increases and the deviation e decreases due to process start-up, etc., YV < LM, the output limiter 2 is released, the manipulated output MV begins to decrease based on the proportional calculation output K _p e, and the changeover switch 5 is switched to the a side, and the integral calculation element I becomes K _p
Start integrating e. That is, when RB=0 is set, the proportional-integral operation is started at the same time as the start-up of the process, and the same operation as the conventional integral feedback type is performed.

Next, set the reset bias value RB to 0<RB<K _Pe
When the manipulated output MV is limited by the output limiter 2, the integral term B finally becomes
LM + RB - K _P e is set, the proportional band PB moves as shown in Figure 3 B, and when the deviation e decreases to the value determined by the reset bias value RB, the proportional/integral operation is started. Become. In other words, even if the measured value PV begins to change, the output limit is released with a delay equal to the reset bias value RB.
Startup response can be accelerated. For example, in batch control, the output limit starts with being held for a long time, so if the reset bias value RB is zero, the manipulated output MV quickly escapes from the output limiter as the measured value PV changes, and the measurement Although the rise of the value PV is slow and it takes time to reach the target value, by selecting the reset bias value RB, the exit time of the output limit is adjusted and the optimum rise characteristic can be obtained. Therefore, the rise characteristics at startup in batch control can be easily set using the reset bias value RB.

Furthermore, if the reset bias value RB is selected to be RB=K _p e _s , while the manipulated output MV is limited by the output limiter 2, the integral term B is finally set to LM, and the proportional band PB is set as shown in Fig. 4. It moves as shown in (b) and does not start the proportional/integral operation until the polarity of the deviation c is reversed, and performs the same operation as the conventional deviation feedback type, which can be used for safety valve control.

In the above description, a case has been exemplified in which the control calculation section 1 performs proportional integral (PI) calculation, but if a differential calculation element D is added as shown in FIG. 5, proportional integral differential (PID) calculation is performed. Furthermore, in FIG. 5, the first-order delay circuit 6 is provided on the output side of the adder/subtracter 4 in order to prevent a pullback phenomenon caused by differentiation and proportionality near the saturation point of the PID calculation output.

FIG. 6 is a conceptual block diagram of an embodiment of the present invention using a processor. In FIG. 6, 10 is a processor (for example, a microprocessor), 21 to
27 is an analog comparator provided on the input side;
30 is a digital-to-analog converter (DA converter) provided on the output side, 40 is a semiconductor switch,
50 is an amplifier, 60 is a hold capacitor provided at its input end, 70 is a manually operated switch,
80 is a manual/automatic changeover switch, and 90 is a setting potentiometer.

Analog comparators 21, 22, 23, 24, 2
The output of the amplifier 50, the measured value PV of the process variable, the set value SV, the proportional gain K _p , the integral time T _I , and the differential time are input to one input terminal of 5, 26, and 27, respectively.
T _D and reset bias value RB are applied as analog voltages, and the output voltage of the DA converter 30 is commonly applied to the other input terminal. These constitute a mechanism for converting an analog input signal into a digital signal and inputting it into the processor 10.
That is, the processor 10 sequentially sets the digital output signal to "1" starting from the MSB, and sends the digital output signal to DA one by one.
The converter 30 converts it into an analog voltage, and the analog comparators 21 to 27 compare this voltage with the respective input voltages, and the logical value of each bit of the digital signal is sequentially determined according to the comparison output of a desired one analog comparator. will be finalized. The value of the established digital signal is equal to the desired analog comparator input voltage. The analog comparators 21 to 27 are switched in order each time one input signal is completed, and each input signal is sequentially converted into a digital signal and then fetched into the processor 10. Such a configuration has the advantage of eliminating the need for an analog-to-digital converter.

The processor 10 calculates a position-type control output based on the received input signal, and converts this control output into
It is converted into an analog voltage by a DA converter 30 and then connected to a hold capacitor 60 through a semiconductor switch 40.
is charged. The voltage of the capacitor 60 is the voltage of the amplifier 50
is amplified and given to the controlled object. amplifier 50
The input impedance of is set sufficiently high so that the attenuation of the charge on the hold capacitor 60 is not a problem.

The acquisition of input signals and the calculation of control outputs are repeated at a constant sampling period. The sampling period is set to about 0.1 sec, for example.

The voltage of the hold capacitor 60 can be increased or decreased as desired by the operator during manual control. That is, when the manual operation switch 7 is turned on the + side or the - side, current from a DC constant current source (not shown) flows into or out of the hold capacitor 60, changing the voltage of the hold capacitor 60. Therefore, this allows manual control of the controlled object. When switching to manual control, the latest value of the control output of the processor 10 is held in the hold capacitor 60, and manual control can be started using that value as a starting point, so switching from automatic control to manual control is bumpless. I can do it.

Now, in the device configured as described above, the calculation of the control output in the processor 10 is performed by the following equation (in the case of PI operation).

y _V = K _p {e+t/T _I Σe} (1) Here, y _V ... Control output e ... Deviation Δt ... Sampling period K _p ... Proportional gain T _I ... Integral time In other words, the processor 10 is Produces position type PI control output. In equation (1), if the integral term is B, the following equation is obtained.

y _V =K _p e+B (2) In formula (2), when y _V >LM, the value of B is set appropriately when switching between manual and automatic switching, and when switching proportional gain. The settings are made by the processor 10 program.

When the control calculation output y _V exceeds the limit value LM of the output limiter, the processor 10 uses the reset bias value RB, the limit value LM, the proportional gain K _p and the steady-state error e _s , and the integral term B finally becomes B= LM+
Set with a certain time constant so that RB−K _p e _s .

Adjustment of B during manual-automatic switching is performed as follows. When switching from manual control to automatic control, the processor 10 uses the manual control output (output of the amplifier 50) y _M and the proportional gain read immediately before.
Using K _p and deviation e, determine the value of B using the following equation.

B=y _M −K _p e _o (3) And if we set the value of equation (3) as B of the first control output y _A after automatic switching, y _A = K _p e _o + B=K _p e _o +y _M −K _p e _o =y _M (4) Therefore, no fluctuation in the control output occurs before and after manual-automatic switching. In other words, switching from manual control to automatic control can be performed bumplessly. Note that switching from automatic control to manual control can be performed bumplessly using the capacitor 60 as described above. Therefore, manual-automatic switching can be performed bumplessly in both directions.

Adjustment of B when switching the proportional gain K _p is performed as follows. When the proportional gain is switched to K _p ', the processor 10 uses the control output y _A read immediately before, the new proportional gain K _p ', and the deviation e _o to determine the value of B according to the following equation.

B=y _A −K _p ′e _o (5) And the first control output after proportional gain switching
If we use the value of equation (5) as B of y _A ′, then y _A ′=K _p ′e _o +B=K _p ′e _o +y _A −K _p ′e _o =y _A (6), and the proportional gain No fluctuation occurs in the control output before and after switching _Kp . i.e. proportional gain
K _p can be switched bumplessly.

As explained above, in the present invention, by selecting the magnitude of the reset bias value,
A novel process control device with a reset limiter that can operate as either an integral feedback type or a deviation feedback type is obtained.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing one embodiment of the device of the present invention, FIGS. 2 to 4 are explanatory diagrams of its operation, and FIG. 5 is a block diagram showing another embodiment of the device of the present invention.
FIG. 6 is a conceptual block diagram of an embodiment of the present invention using a processor. 1... Control calculation section, 2... Output limiter, 3...
...Reset bias value setting circuit, 4... Addition/subtraction section, 5... Changeover switch, 10... Processor,
30...Digital-to-analog converter, 50...
amplifier.

Claims (1)

[Claims] 1 After performing at least proportional and integral calculations on the deviation, it is output as an operation output via an output limiter, and
In a process control device equipped with a reset limiter that resets the value of the integral term when the manipulated output reaches the limit value of the output limiter, there is a means for setting the reset bias value, the limit value LM of the manipulated output, the proportional gain Kp, and the deviation e. and a reset bias value RB, when the manipulated output exceeds the limit value, a means is provided to set the integral term B to a value such that B = LM + RB - Kpe with a certain time constant, and when the deviation decreases below a certain value. A process control device characterized in that switching between an integral feedback type in which an output limiter is released and a deviation feedback type in which an output limiter is released when the polarity of the deviation changes is performed by selecting the reset bias value.