JPH0618006B2 - Process control method - Google Patents

Description

Description: TECHNICAL FIELD OF THE INVENTION The present invention relates to a process control method in a control system having a dead time.

[Technical background of the invention]

Generally, in process control, the dead time may be constant or the dead time may change. FIG. 1 shows an example of a conventional process control system when the dead time changes, that is, when the dead time is variable. In this process, a fluid A having a certain component is mixed with a fluid B having a certain component to form a fluid C having a desired component X _S. Therefore, the fluid B is injected in proportion to the flow rate of the fluid A, the component of the fluid C is measured, and this is adjusted and calculated to be a predetermined value X _S, and the injection flow rate of the fluid B with respect to the flow rate of the fluid A is calculated. Correct the ratio.
In this process, the dead time is also changed by changing the flow rate of the fluid A (controlled by changing the transfer time).

Next, the structure and operation of the conventional process control system shown in FIG. 1 will be described. First, the fluid A is poured into the mixing tank 3 through the pipe 1. The flow rate of the fluid A is detected by the flow rate detector 2 provided in the middle of the pipe 1. On the other hand, the fluid B is the pipe 9
Is injected into the mixing tank 3 via. The flow rate of fluid B is
It is detected by the flow rate detector 10 provided in the middle of the pipe 9,
It is adjusted by the control valve 8. The fluid C generated by the mixing is sent out to the next step through the pipe 4.

The flow rate signal f _A obtained by the flow rate detector 2 is the ratio coefficient part 5
Is multiplied by the ratio coefficient α (theoretical input ratio of the input amount (flow rate) of the fluid B to the flow rate of the fluid A necessary for the mixed fluid C to have a predetermined component), and the ratio is corrected by the multiplication unit 6. After that, it is led to the flow controller 7. Then, in the flow rate controller 7, the flow rate signal f _B from the flow rate detector 10 and the signal 6a from the multiplication unit 6 are compared and adjusted, and the operation signal 7a formed as a result adjusts the opening degree of the control valve 8. , Thereby adjusting the injection flow rate of the fluid B.

On the other hand, the component of the fluid C is detected by the component detector 11, the detected component measurement signal X is guided to the subtraction unit 12, and the deviation e with respect to the component set value X _S is obtained. The deviation e is multiplied by the coefficient K _I in the intermittent integration coefficient unit 13 to obtain a signal K _I · e. This signal K _I · e is the contact 1 that closes intermittently.
Signal _K I · _{e n} becomes through 4 'is guided to the integrating unit 15. Integrator 15, the _{MV n = MV n-1 +} K I · e n becomes operational. Here, MV _n-1 is the result of the same calculation up to the previous time, and forms the output of the integrator 15 until the current calculation result MV _n is obtained. The intermittent integration is performed here in order to obtain stable control with respect to the control target as shown in FIG. 1 in which the control deviation is oscillatory and the dead time is long. That is, the operation amount (operation amount of the fluid B) is updated, and the effect is the control amount (component of the fluid C).
This is for faithfully executing the basic of feedback control, that is, the control deviation is obtained at the time when it is completely reflected in, and the operation amount is changed by performing the adjustment calculation so that the control deviation becomes zero.

The output of the integrator 15 is guided to the calculator 16.

On the other hand, the contact 14 ′ is closed when the pulse P ₂ is generated from the variable sampling period generator 14 which receives the flow rate signal f _A. The generation period T _p2 of this pulse P ₂ is the dead time for fluid transfer from the injection point of the fluid B to the detection point of the component detector 11 t _d1 , and the dead time until the component detection signal X is obtained from the detection point. t _d2 , the time constant τ of the component detector 11,
_Is a proportional constant, T _p2 = t _d1 + t _d2 + K · τ (4). However, K is about 3 to 5. This is because it can be determined that the response is almost complete after the elapse of 3 to 5 times the time constant τ. V is the total volume from the injection point of fluid B to the component detection point, F
_{When A (MAX)} is the maximum value of the measurement range of the fluid A and f _A is the current flow rate (%) of the fluid A, td ₁ is It is obtained by finding t _d1 that satisfies On the other hand, t
_d2 and τ are constant values. If t _d1 cannot be used as a measurement value or if it is difficult,
It is also possible to have a fixed value for t _d1 , and thus a fixed value for T _p2 . When the cycle is fixed, the fixed cycle is determined by taking the longer dead time in the normal operation state. For example, when the dead time is determined by the transfer time of the fluid A as in the conventional process control system, the cycle is fixed according to the case where the flow rate is small.

Next, in the memory 17, the variable sampling period generator 14
Component measurement signal X at the time of generation of pulse P ₂ (at the time of sampling)
Is sampled and its value is stored as X _on until the next sample. And the component measurement signal X at each instant
Deviation and the sample at the time of the value _{X on} the previous _(X on -X)
Is calculated in the subtracting unit 18. In the proportional gain section 19, the output of the subtraction section 18 is multiplied by the proportional gain Kp to obtain Kp · (X _on −X). This signal is guided to the adder 16 via the contact '.

Here, the reason for performing proportional control will be described. The control deviation when the control is executed should be almost zero at the next control. However, since the control deviation that occurs during the control pause period after the control is executed is not corrected at all, the value that is proportional to the deviation amount from this point is added to the manipulated variable to correct the control deviation based on the control deviation when the control is executed. This is because the control deviation during the control suspension period will be compensated.

Contact 20 'is then set with pulse P ₂ and pulse P ₁
The memory 20 that is reset by is closed while it is set. Pulse _{P 1} is generated after the lapse of the dead time from the generation of the pulse _{P 2 T p1 = t d1 +} t is given by _{d2 T p1,} i.e. from the injection of fluid B to obtain the component detection signal.

The _sum of the output MV _n of the integrator 15 and the output Kp (X _on -X) of the proportional gain unit 19 transmitted via the contact 20 ′ is obtained in the adder 16, and this sum is sent to the multiplier 6 as an operation signal. Given. The multiplication unit 6 multiplies the operation signal by the output of the ratio coefficient unit 5 and supplies it to the controller 7. As a result, the injection rate of the flow rate f _B of the fluid B with respect to the flow rate f _A of the fluid _A is adjusted, and the component X of the fluid C is brought close to the set value X _S.

[Problems of background technology]

In the above configuration, there is no problem when the component set value X _S is constant, but the set value X _S is between one sample time and the next sample time (the interval between the samples is the actual plant, No control corrective action is taken. Therefore, (i) the settling time of the control response with respect to the change in the set value becomes long, resulting in process transient loss and transient quality deterioration, and the product homogeneity cannot be maintained.

(ii) Control response is delayed, so process and equipment safety,
It becomes a big obstacle for protection.

Thus, there was a big problem in process control.

[Object of the Invention]

An object of the present invention is to provide a process control method capable of promptly responding to a change in set value at mutual intervals during sampling in a control system performing intermittent integration.

[Outline of Invention]

The process control method of the present invention is designed to perform a special proportional operation between sample times. That is, the deviation e _O at each sample is stored, and the difference Δe = e between the deviation e at each instant and the deviation e _O at each sample until the dead time elapses from each sample. The sum of the signal Kp · Δe proportional to −e _O and the intermittent integral output signal MV _n is used as the operation signal, and after the dead time has elapsed, the intermittent integral output signal MV _n is used as the operation signal until the next sampling. It is a thing.

Example of Invention

FIG. 2 shows an embodiment of the present invention. In the figure, the same reference numerals as those in FIG. 1 denote the same members. The memory 27 provided in place of the memory 17 stores the output e = X _S −X of the subtraction unit 12. The stored contents are updated when the pulse P ₂ is generated. The subtraction unit 28 obtains a difference Δe = e−e _O between the deviation e at each instant and the deviation e _O at the previous sampling. Here, the meaning of this difference Δe will be considered. First, e = X _S −X (1), where X _S and X are set values and measured values at any time points (since the last sample time), that is, at each instant, while e _O = X _OS -X _O (2), where X _OS and X _O are the set values and measured values at the previous sample. Therefore, Δe = e−e _O is Δe = e−e _O = (X _S −X) − (X _OS −X _O ) = (X _O −X) + (X _S −X _OS ) ... (3) Here, (X _O −X) represents the amount of change in the measured value, that is, the control amount after the last sample, and (X _S −X _OS ) represents the amount of change in the set value after the last sample. Therefore, △ e
Means the sum of the amount of change in the control amount after the last sample and the amount of change in the set value after the last sample.

The proportional gain section 19 outputs the proportional gain K to the output of the subtraction section 28.
Multiply by p to obtain Kp · Δe. This Kp · Δe is guided to the addition unit 16 via the contact 20 ′. The contact 20 'is closed only during the dead time as in the case of FIG. Therefore, only during this period, Kp · Δe is added to the output MV _n of the integrator 15.

In this way, the total dead time T from each sample
Until _p1 elapses, the contact 20 'is closed and Kp ·
The operation signal is obtained by adding Δe to MV _n . That is, when either the set value or the control amount changes significantly after the previous sampling, compensation by Kp · Δe is performed. On the other hand, after the elapse of T _p1 , the contact 20 ′ is opened until the next sampling. This is the compensation by Kp · Δe that cancels the action of the integrator after the influence of the manipulated variable of the integrator determined based on the deviation e = X _S −X at the previous sample begins to appear in the control amount. This is done so that the influence of Kp · Δe will not be exerted at the time of the next sampling without performing.

〔The invention's effect〕

As described above, in the control system having a large variable dead time, the following effects can be obtained by adding the special proportional operation to the variable intermittent integration operation only during the dead time.

(i) Regardless of whether the control amount or the set value changes between the samples, the control operation corresponding to the change can be performed, and the control response is greatly improved.

(ii) Process safety, equipment protection, and product homogenization can be ensured because of the quick response to process abnormalities and large changes in set values.

[Brief description of drawings]

FIG. 1 is a schematic diagram showing an apparatus used for carrying out a conventional process control method together with a controlled process, and FIG. 2 is a schematic view showing an apparatus used for carrying out a process control method according to an embodiment of the present invention. FIG. 3 is a schematic view showing the process of the above. 11 ... Detector, 12 ... Subtraction unit, 13 ... Coefficient unit, 1
4 ... Sampling cycle generator, 15 ... Intermittent accumulator, 16
...... Adding unit, 19 …… Proportional gain unit, 27 …… Memory,
28 ... Subtraction unit.

Claims (1)

[Claims] 1. In a process control method for a control system having a dead time, a deviation (e) between a set value (X _S ) and a control amount of the control system is obtained, and a signal (K _I ·
e) is sampled at a variable cycle or a fixed cycle equal to the time proportional to the sum of the dead time and the time constant of the control amount detection end and intermittently integrated to obtain an intermittent integrated output signal (MV _n ) at each sampling time. From the time until the dead time elapses, the signal (Ke) proportional to the value (Δe) obtained by subtracting the deviation (e _O ) at each sampling from the deviation (e) at each instant.
p · Δe) and the intermittent integration output signal (MV _n ) as the operation signal, and only the intermittent integration output signal (MV _n ) is the operation signal until the next sampling after the dead time has elapsed. And process control method.