JPS6118202B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a digital calculation type process control device. More specifically, the present invention includes:
The present invention relates to a process control device which performs proportional and integral calculations in a positional manner using a processor, and which can effectively prevent reset wind-up.

(Prior Art) A digital arithmetic type process control device is constructed by utilizing the arithmetic functions of a computer or an arithmetic unit microprocessor (hereinafter simply referred to as a processor).
The processor determines the control output by performing PID (proportional, integral, differential) calculations, etc., but conventionally, the calculations in the processor are based on ease of manual/automatic bumpless switching, measures to prevent reset/windup, and measures to prevent processor failure. Velocity type calculations were often adopted due to their ease of calculation.
However, since velocity type calculation is based on I operation, it is not suitable for performing only P operation or PD operation. Further, even in the case of PID operation, an output pullback phenomenon based on P/D operation occurs near the saturation point of the control output, which is disadvantageous.

A position type calculation process control device is considered to be suitable for P operation and PD operation and does not cause the output pullback phenomenon.

(Problems to be Solved by the Invention) However, when performing positional calculations using a digital processor, there are problems such as manual/automatic bumpless switching and measures to prevent reset/windup. be.

SUMMARY OF THE INVENTION An object of the present invention is to provide a digital arithmetic process control device that uses positional arithmetic operations to prevent reset and wind-up, and provides an optimal response at startup.

(Means for Solving the Problems) The present invention allows a processor to perform at least PI calculations using positional calculations, and
The feature is that when an excessive deviation continues for a long period of time, the value of the integral operation term is reset to the preload value.

(Example) The present invention will be explained below with reference to the drawings. FIG. 1 is a conceptual block diagram of an embodiment of the present invention. In FIG. 1, 1 is a processor (for example, a microprocessor), 21 to 26 are analog comparators provided on its input side, 3 is a digital-to-analog converter (DA converter) provided on its output side, and 4 5 is a semiconductor switch, 5 is an amplifier, 6 is a hold capacitor provided at the input end of the amplifier 5, 7 is a manually operated switch, and 8 is a manual/automatic changeover switch.

Analog comparators 21, 22, 23, 24, 2
Amplifier 5 is connected to one input terminal of 5 and 26, respectively.
output, measured value Mn of process variable, set value SP,
Proportional gain K _p , integral time T _I , and derivative time T
_D is given as an analog voltage, and the output voltage of the DA converter 3 is commonly given 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 1. That is, the processor 1 sequentially sets the digital output signal to "1" starting from the MSB, and converts it one by one into an analog voltage using the DA converter 3.
This voltage is compared with the input voltage in analog comparators 21 to 26, respectively, and the logic value of each bit of the digital signal is determined one after another according to the comparison output of a desired one analog comparator. The value of the established digital signal is equal to the desired analog comparator input voltage. Analog comparators 21 to 26 are 1
Each time one input signal is taken in, the switching is performed in order, and each input signal is sequentially converted into a digital signal and taken in by the processor 1. Such a configuration has the advantage of eliminating the need for an analog-to-digital converter.

Processor 1 receives measured values Mn and set values
A position type control output is calculated based on each input signal such as SP, proportional gain K _p , integral time T _I , differential time, etc., and this control output is converted to an analog voltage by the DA converter 3, and then passed through the semiconductor switch 4. hold·
Charge capacitor 6. The voltage of the capacitor 6 is amplified by the amplifier 5 and applied to the controlled object. The input impedance of the amplifier 5 is set sufficiently high so that attenuation of the charge on the hold capacitor 6 does not become 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 6 can be amplified 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 6, changing the voltage of the hold capacitor 6. Therefore, this allows manual control of the controlled object. When switching to manual control, the latest value of the control output of processor 1 is held in the hold capacitor 6, 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 1 is performed by the following equation (in the case of PI operation).

y _o = K _p {e _o + Δt/T _I Σe _o } (1) Here, y _o ...control output e _o ...deviation Δt...sampling period K _p ...proportional gain T _I ...integration time In other words, processor 1 is the position produces a PI control output of the form. In equation (1), if the integral term is B, then
We get the following equation.

y _o =Kpe _o +B (2) In equation (2), the value of B is adjusted as appropriate to prevent reset/windup. In adjusting the value of B, the value of the integral term is actually adjusted. Such calculations are performed by the program of the processor 1, and for convenience of explanation, they can be shown conceptually as shown in FIG. 2.

Adjustment of B during reset/windup prevention is performed as follows. If a large deviation continues for a long time and the value of the integral term continues to change in one direction, the proportional/integral output will reach the limit value of the PI limiter. Then processor 1 has a proportional gain K _p and a deviation e _o
, the limit value L of the PI limiter reached at that time (H _L
or L _L ), the value of B is determined by the following formula.

B=L−Kpe _o (3) Then, the control output becomes _yo = Kpe _o +B = Kpe _o +L−Kpe _o =L (4) and is limited to match the limit value L of the PI limiter. As long as the proportional/integral output reaches the limit value L, such an operation is repeated for each sampling. As the deviation e _o becomes larger, the integral term B becomes smaller and smaller according to equation (3). this is,
Reset windup can be prevented by moving the proportional band in the opposite direction to the normal integral operation that is not affected by the PI limiter. e
When _o becomes excessive, if we control with B=0%,
When Kpe _o >100%, the output y=L, and the proportional operation restarts from Kpe _o =100%, and when Kpe _o <100%, the proportional term is utilized.

In the present invention, for example, in the case of batch control, a preload value F is introduced, and when the above e _o is excessive, control is performed using B=F to appropriately move the starting point of the proportional operation and obtain an optimal response. That's what I do.

In addition, in the above explanation, when limiting, the integral term is
= L−Kpe _o I have described the method of instantaneously setting it, but
The integral term B may be integrated in the opposite direction using an integral time constant to move the proportional band. The specific method is shown in FIG. That is, when applied to the PI limiter, the polarity of the deviation signal e _o is inverted and integrated as shown in (c).
Then, when it stabilizes, it will stop as shown in the image below. This method of backward integration using an integration time constant has the advantage of being resistant to noise and can prevent reset windup.

FIGS. 3 and 4 are operation waveform diagrams showing such operations assuming a case of shutdown and restart after completion of batch control.

First, as shown in Fig. 3A, when the measured value Mn is controlled to be equal to the set value SP by the PI(D) operation (the state where the deviation e _o is zero), the control output Y is maintained at 70%, for example. Further, in this state, the proportional bands PB are arranged above and below the set value SP as shown in FIG. 3B. Also, in this state, the proportional term (K _p・e _o )
is zero since e _o =0, and the integral term B
As shown in Figure 3 (c), the ratio is maintained at approximately 70%. Now, when a deviation e _o occurs due to a process shutdown or the like, the control output Y increases as shown in FIG. 3A, and the integral term B also increases as shown in FIG. 3C. If left as is, a reset windup will occur, but when the control output Y is saturated, an integral term reset operation is performed. That is, when the control output Y is saturated, it is applied to the PI limiter and -K _p ·e _o is integrated. As a result, the content of the integral term B is corrected in a decreasing direction as shown in FIG. 3C, and the proportional band PB moves as shown in FIG. 3B. The time constant at this time uses the integration time T _I . When the content of integral term B becomes zero, this inverse integration is stopped. After this, even if the deviation increases further, the integral term is maintained at zero.

At start-up, the deviation e _o decreases, and at point P where the proportional term (K _p · e _o ) is 100%, integration starts as shown in Figure 3 C and A, and the output based on the proportional term Y begins to decrease.

FIG. 4 is an operational waveform diagram when a bridle value, which is a feature of the present invention, is introduced. The position of the proportional band at start-up is set above the briload setting value to ensure optimal response at start-up.

In the above description, an example of a processor that saves analog/digital converters has been given, but the processor may have a general configuration using an analog/digital converter and digital/analog conversion as shown in FIG. Furthermore, although the example of a single-loop controller has been given, the present invention can also be applied to a device that controls multiple loops.

(Effects of the Invention) As explained above, the present invention uses a processor to perform at least the P _I operation by position type calculation, and resets the value of the integral operation term to the briload value when an excessive deviation continues for a long time. I did it like that.
Therefore, according to the present invention, it is possible to realize a digital calculation type process control device that takes advantage of position type calculations, can effectively prevent reset and windup, and can obtain an optimal response at startup.

[Brief explanation of the drawing]

Figure 1 is a conceptual diagram of an embodiment of the present invention, Figure 2 is a conceptual diagram of calculation, Figures 3 and 4 are operational waveform diagrams, and Figure 5 is a conceptual diagram of another embodiment of the invention. It is a diagram. 1...Processor, 21-26...Analog comparator, 3...Digital-to-analog converter, 4...
...Semiconductor switch, 5...Amplifier, 6...Capacitor, 7...Manual operation switch, 8...A/M conversion switch.

Claims (1)

[Claims] 1. A processor, comprising a signal reading means for reading into the processor at least a measured value, a set value, a proportional gain _Kp , and an integral time T _I ; The automatic control output Y _A = K p { using the proportional gain K _p read at the sampling period Δt through the signal reading means by proportional/integral calculations, the deviation e _o between the measured value and the set value, and the integration time _{T I} e _o +Δt/T _I Σ e _o }, and when the automatic control output Y _A exceeds a predetermined limit L, the integral term B (=Δt/T _I K _p Σ
A process device characterized in that the value of e _o ) is reset to a preload value F. 2. The process control device according to claim 1, wherein the integral term B is reset with a delay due to an integral time constant.