JPS6250901A - Process control device - Google Patents

Abstract

PURPOSE:To automate the process control operation, to save the labor of opera tion, to rationalize, to prevent the misoperation and to shorten the start/stop time of the plant by making concident the changing quantity of the control quantity to the target value and correcting the control rule automatically. CONSTITUTION:When a control rod is operated and generator outputs y1 and DELTAyr1 only are changed, an operation equipment determining part 11 selects the operation equipment and determines a changing quantity 'yr1 of a control quantity y1. An operation quantity determining part 12 determines an operation quantity mui by using the corresponding contro rule and outputs DELTAfr1(=muiDELTAyr1) to a sub-loop control device 2, based upon the information such as the coordinates of a control rol (i) operated, the pulling-out position and the average pulling-out position unclear furnace output of the ambient control rod. As the result, fri is changed by DELTAfri only and a generator output y1 is changed by DELTAy1 only. A learning parat 13 corrects the control rule used for determining the operation quantity mui so that the deviation of DELTAy1 and DELTAyr1 can be smaller by the next operation based upon the operation quantity mui, a target quantity DELTAy1 and the changing quantity DELTAy1.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to a device that collectively controls a system having a plurality of operating devices, and particularly relates to a device for controlling a system having a plurality of operating devices, and in particular, a control device for controlling a controlled variable to match a target value. The present invention relates to a process control device suitable for appropriately determining.

[Background of the invention]

As a conventional technology, Proceedings of the Society of Instrument and Control Engineers, Volume 20,
"Automatic learning fuzzy controller" described in issue 8
There is. As shown in FIG. 2, this controller performs direct feedback sterilization so that the control amount f matches its set value fr. The learning department is
K so as to improve the dynamic braking characteristics (stability, responsiveness) of the control amount f.

The control rule used to calculate the amount of change ΔV in the manipulated variable V is modified. However, this controller does not describe how to change the set value fr. The present invention relates to a process control device suitable for controlling these control devices and appropriately changing the set value fr in a system having a plurality of such direct feedback control devices. Such functionality is
Follow.

Previously, operators made appropriate changes in response to changes in process conditions.

[Purpose of the invention]

An object of the present invention is to provide an indirect foot check control method suitable for appropriately changing the setting value fr of each control device for a control device including a plurality of direct feedback control devices and open rule control devices. The purpose of the present invention is to provide a control device.

[Summary of the invention]

The present invention utilizes operating know-how to more easily automate operations that have traditionally been performed by operators.
It was created by focusing on the need for a means to determine the set values of each side device and the amount of operation at that time, as well as a means to learn how to operate it in the same way as an operator. .

The first feature of the present invention is that the setting value fr of a plurality of restraint devices is
In order to change i in accordance with changes in the state of the process, an operation f' is performed to determine the operation timing of the operating device (control device t) 1 and the target value Δyrk of the amount of change in the controlled variable that changes with the operation.
¥ equipment determining means, determining a manipulated variable uI for making the change amount Δyk of the controlled variable match the target value Δyrk;
The set value fri of each control device is set to Δfri(:ui・Δyr
h) together with the operation amount determining means that changes the amount of change Δyk.
A learning means is provided for correcting the manipulated variable ui when the deviation between the target value Δyrk and the target value Δyrk is greater than or equal to a predetermined value. Here, the correction amount Δu+ for the manipulated variable is a value proportional to the product of the manipulated variable ui and the control characteristic evaluation value Pk (=Δyrk/Δyk−1).

In order to converge the manipulated variable to a certain target value, the proportional gain ξ is set to a value between 0 and 10.

A second feature of the present invention is that in order to quickly converge the manipulated variable ui to a certain target value and to reduce the influence of observation noise, the The purpose is to change the proportional gain ξ.

[Embodiments of the invention]

An embodiment of the present invention will be described below with reference to FIG. 1g
In FIG. 1, numeral 1 is a general control device which is the main part of the present invention, 2 is a sub-loop control μs device, and 3 is a plant consisting of a plurality of processes. In a boiling water nuclear power plant, for example, process A31 is the control rod drive system, process B32 is the recirculation system, process C33 is the water supply system, and process D34 is the reactor and power generation system. f, , f2
is the control rod withdrawal position, f is the recirculation pump speed, f4
is the water supply flow rate, Yt is the generator output, and Yt is the reactor water level. Although an actual plant has many processes and is more complex than those shown in FIG. 1, only representative processes are shown here because it is a WJ system. In the subloop control device 2 having a plurality of control devices, 21 is a recirculation flow rate control device, and 22 is a water supply flow rate control device. f
rl and fr2 are set values for the control rod withdrawal position, frB is a recirculation pump speed set value, and fr4 is a reactor water level set value.

The integrated control device 1 of the indirect feedback control method that does not directly and continuously feedback control amount yk is frl
, frt, fr3, etc. to Δfri
Only outputs a signal for changing.

Step 11 determines which of fr, to fr4 to change, i.e., i, and roughly determines the amount of restraint yk at that time (here,
1 or 2) is an operating device determining unit that determines the target value Δyrk of the amount of change. Reference numeral 12 denotes a manipulated variable determination unit that determines the manipulated variable u1 at that time using a control rule, and outputs the change amount Δfrl of the set value to the subloop control device M2 as the product of ui and Δyrk. Subloop control device fa
12 When the signal TEH, Δfrl is input, the corresponding setting value fri is changed by Δfri. As a result, the state of the process changes and the corresponding constraint +1JiLyk becomes Δyk
only changes.

13 is the manipulated variable ur so that the difference between the amount of change Δyk of the controlled variable at this time and its target value Δyrk becomes smaller.
This is a study section for modifying the six-constraint rule used for determining . For example, by operating the control rod, the generator output y, Δyr
Assuming that only , is changed, the operating device determining unit 11
, the operating device is selected (i is 1 or 2), and the control f
The amount of change Δyr1 (h=x) in y is determined. In the operation amount determining section 12.

Controls to be operated - coordinates of rod i, its pulling position, surrounding controls -
Based on information such as the average rod withdrawal position and reactor output, the manipulated variable u1 is determined using the corresponding control rule, and Δ
fr+(=uiΔyr+) is output to the subloop control device 2. As a result, fri changes by Δfri and becomes 1
The generator output y changes by Δy. The learning unit 13 calculates the manipulated variable u+ and the target value ΔYr+.

In the first operation based on the amount of change Δy1, Δy1 and Δ
The control rule used to determine the manipulated variable ui is corrected so that the deviation from the manipulated variable ui becomes smaller. The above is an overview of the present invention. This will be explained in detail below.

The operating device determination unit 11 stores the following driving rules.

Ruler 42 (Target value Δy, is 0.5ch) Ruler 45 The operating device determining unit 11 determines Δyrk and i by applying such an operation rule according to a change in the state of the process.

Next, the operation of the manipulated variable determining section 12 will be explained. The control rule used to determine the amount of operation is 1. For example, when it concerns the operation of a control rod, it is as follows.

control rule 0) control rule φ(j+1) where PB is positive Big, PS is p
It means O3itjVeSmall.

A method of determining the manipulated variable uj using such a control rule will be briefly explained using FIG. 3. The rule shown in FIG. 3 takes out only two conditions of the above-mentioned rule. This is a simplified rule for explanation. Based on the control rod withdrawal sequence and control rod pattern, the coordinates of the control rod i to be operated and its withdrawal position fi are known, so the current xl, x
The value of 2 x10+"20 is determined.

Degree to which condition 1 of rule (j) is satisfied μH, condition 2
Calculate the degree to which μj2 is satisfied, and find its minimum value μ
Let j be the membership of rule ÷ (j).

The manipulated variable is uj. Since there are a plurality of such rules, the control rod operation iu, ie, % u i , is determined by taking the weighted average value of the degree μj of each rule being satisfied and the operation amount uj. and.

The actual withdrawal amount Δfri is output as the product of U and ΔyrI, and the state of the process is changed.

Next, the operation of the learning section 13 will be explained using FIG. 4. Assume that the control 1k yk changes by Δyk due to a change in the state of the process. The learning unit 13 calculates the amount of change Δyk, and calculates the control custom evaluation value Pk (=Δyr
k/Δyk−1). If pk is 0, the amount of change Δyk coincides with the target value Δyrk, which indicates that the manipulated variable ui was appropriate. If the absolute value of Pk is large, it means that the manipulated variable uj of the control rule is inappropriate, so uj is corrected. Here, the correction amount ΔU of the manipulated variable U (or ui) is set to be linear.

ΔU=ξU(Δyrk/Δyk-1)=ξuPc...(1) Here, Xl. , Also, the manipulated variable U is.

Since the decision was made based on several rules, the correction amount Δuj of each rule is set to a value proportional to the degree of contribution (membership μj) to the determination of the manipulated variable U. As a result, Δuj
becomes as follows.

Δuj =ξuPkμjΣμj/Σμj”
-(2) Here, ξ is changed depending on the size of Pk.

The reason is as follows. If the absolute value of Pk is large, it is obvious that uj was inappropriate, so
By increasing ξ, the correction amount Δuj of uj is set to dog. on the other hand,
If the absolute value of Pk is small, this may be due to the influence of observation noise when measuring Δyk, so ξ is set to 0 or a small value, and the correction amount Δuj of uj is made small.

In the manner described above, the manipulated variable U is determined while learning.

The characteristics of this example described above were confirmed through a simulation test. The results are described next. The number of hook control rules used in the test was 625 (5 x 5 x 5 x 5).

Figure 7 shows the results of the test with xlozx4° constant. Before learning (initial value), the ratio of Δy+/ΔYr+ was about 0.2, but after 3 learnings, ΔY+ /ΔYr+
is almost l and the rule operation i! By confirming that kuj has converged to its target value, and by changing ξ according to Pk (using the adaptive rule shown in Figure 4), it is possible for uj to converge even if noise is applied. Recognize.

In Figure 6, X30 and X40 are fixed, and X10 and X2
These are the results when learning the manipulated variable uj while changing 0. As in the case of FIG. 5, it was found that uj almost converged to the target value. Note that here, Δyl/
The reason why Δy′+ does not necessarily become 1 or 0 but fluctuates is that an interpolation error occurs as a result of calculating the manipulated variable u using a small number of rules.

In order to bring Δy1/ΔyrI closer to “1.0,” it is sufficient to increase the number of control rules.

As described above, according to this embodiment, in order to make the amount of change Δyk of the control amount coincide with its target value Δyrk, it is possible to automatically correct the manipulated variable uj of the control rule while learning it. .

1. In the explanation of this embodiment, the control rod withdrawal operation was mainly explained, but it is also possible to change the generator output y by changing the recirculation pump speed using the corresponding operation rules and control rules. It is easy to realize this.

According to this embodiment, it is possible to provide a process control device using an indirect feedback control method that can appropriately change the set value fr of a plurality of control devices. Therefore, operations that were conventionally performed by operators can be automated.

As a result, driving is labor-saving, streamlined, and prevents operational errors.

This has the effect of reducing plant startup/stopping time.

In the above embodiment, the explanation was centered on the control rod withdrawal operation of a boiling water nuclear power plant.

The present invention is not limited thereto, and is applicable to automation of the operation of various control devices, such as operation of a recirculation flow control system, operation of a feedwater flow rate system, operation of a turbine control system, and operation of a feedwater heater bacteriostatic system. can. In this way, various operations can be automated by the present invention.

This has the effect of saving labor, streamlining driving, and preventing erroneous operation.

Furthermore, the present invention is applicable not only to boiling water nuclear power plants but also to the automation of general industrial plants. As a result, the same effect as described above is obtained.

〔Effect of the invention〕

According to the present invention, a plurality of control devices can be integrated.

The operation amount Δfr (the amount of change in the set value) for changing the amount of change Δy in the control amount by the target value Δyr can be determined while learning. Therefore, it becomes possible to easily automate operations that have conventionally been performed by an operator, which has the effect of saving labor, streamlining operation, and preventing erroneous operations.

[Brief explanation of drawings]

FIG. 1 is a system diagram of one embodiment of the present invention, FIG. 2 is a block diagram for explaining the prior art, and FIGS. 3 to 6 are supplementary explanatory diagrams of one embodiment of the present invention.

Claims (1)

[Claims] 1. In a device that centrally controls a system having a plurality of operating devices, operating device determining means determines the operation timing of the operating device and the target value Δyr of the amount of change in the controlled variable, and the amount of change Δy in the controlled variable. a manipulated variable determining means for determining a manipulated variable u for making the amount of change Δy coincide with the target value Δyr; 1. A process control device characterized by being provided with learning means for making corrections.