JPH01114902A - Process controller - Google Patents

Abstract

PURPOSE:To attain parallel processing of determination for an optimum parameter and a control variable by separating a control object unit, the determination of a model of controlled system, a parameter determining part, and a control variable determining part from each other and setting up a bus for inputting/ outputting the quantity of state of the controlled system to the shortest length. CONSTITUTION:The model of controlled system which may be changed in accordance with an operating condition is corrected and determined by a controlling computer 6 and obtained parameters and gain are transmitted to process control parts 1-1, 1-2 through a network 8. The control parts 1-1, 1-2 calculate control output vectors based on the quantity 27 of state inputted from the controlled system 7 through a process unit device 2 and an optimum gain matrix applied from the computer 6 and output the calculated vectors to the controlled system 7 through a process output device 3. Said processing is executed by a processing unit 21 in the control part 1-1 and a functional process input device 4 calculates an offset value, etc., and outputs the calculated result to the controlled system 7 as a control signal 24.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a process control device that takes in a process quantity, performs calculations to bring the quantity to be controlled close to a target value, and outputs a controlled quantity to the process. , relates to a process control device suitable for control that handles multiple variables.

[Conventional technology]

Conventionally, various efforts have been made to achieve higher efficiency and higher quality control, but recently control systems using modern control theory have been attracting particular attention. Examples include ``Application to industrial plants'' described in Fuji Jiho J (1987, No. 6, pp. 418-421).
This prior art example is a document regarding the application of a programmable controller (PC) to an industrial plant. In such a conventional example, an application example of a hierarchical control system and an optimal regulator method is shown.

This optimal regulator is one of modern control theories.

Modern control theory is characterized by treating multiple variables and parameters as a set, such as optimal control theory based on state equations and multivariable control theory in the frequency domain. Therefore, a control system using these theories needs to be able to efficiently process multivariate data. In order to meet this demand, it is necessary to handle larger amounts of data at higher speeds, and the conventional approach has been to deploy multiple microcontrollers and minicomputers using microprocessors and connect them via a network. There is a multi-controller system tied together.

[Problem that the invention seeks to solve]

The above conventional technology does not take into account the configuration of the controller, bus, or process input/output device that realizes the reception and passing of data according to the control content, and the processing speed may decrease due to excessive data traffic on the bus. There were problems such as concentration of processing load on the controller.

The present invention has been made to pay attention to these points and to construct a control system that is more economical and easier to implement. That is, an object of the present invention is to provide a control system that economically realizes multivariable control.

[Means to solve the problem]

The present invention analyzes the control content when applying modern control theory such as the optimal regulator method, and analyzes the control target input section,
It is divided into a controlled object model determination section, a parameter determination section, a controlled variable determination section, a process input section, and a process output section, and a controller that is suitable for each of these sections is constructed, and data transfer between these components is possible as long as possible. The data bus is arranged to simplify it. In other words, it is possible to independently and parallelly process the determination of the optimal parameters used for each control calculation and the determination of the control amount, and it is also possible to shorten the bus from taking in the state quantity of the controlled object to outputting the control amount, making it ideal for high-speed processing. A multivariable control system can be realized.

[Effect]

According to the present invention, the determination of optimal parameters and the determination of control amounts can be processed independently and in parallel.

〔Example〕

The optimal regulator method, which is one of the modern control theories that is the subject of the present invention, will be described.

In the optimal regierator method, first, the controlled object is modeled as shown in equation (1).

'X-f (Y, L9), X (te) -
No... (1) Here, X is a vector of n×1.1 rXl.

An evaluation function F is introduced for the controlled object in (1).

Then, the controlled object is controlled so that the evaluation function F is minimized.

For example, when considering a case where the controlled object can be linearly modeled, it is known that the control quantity is L9=R-'B'PX-(3). Here, -R-'B'P is the optimal gain matrix, U is the control output vector, ! is the state quantity vector.

The following example relates to a control system for a linear controlled object under the optimal leggierator approach.

FIG. 1 shows a block diagram of a process control system which is an embodiment of the present invention.Functional input/output device 4 and vector processing unit 211 are connected as necessary. The role and function of this device 4 and unit 211 will be made clear in the following description. Also,
Connecting as necessary includes the case where both are connected as shown in the figure and whether they are activated or not is switched depending on whether they are used or not.

Control computer 6, network 8, process control section 1-
1.1-2. A control system consisting of a linkage device 22 and a process input/output device 2.3.4 controls a control object 7.
The output signal and status signal 27 are taken in, and the control signal 24 is
Output. The three control computers 6 represent an example of parallel processing, and basically only one computer is required.

Further, although there are three vector processing units 21, this is an example of parallel processing, and one may be sufficient.

Here, the linkage device 22 is a device that performs data transfer at high speed, and each process input device 2. Output device 3. All the data of the input/output device 4 can be cyclically supplied and the data can be updated. Further, the present linkage device 22 includes the process control section 1-1.
The control data held by 1-2 is also transferred to the process input/output device 2,
Similar to the data in 3.4, other process control units 1-1.1
-2 can be cyclically supplied and the data can be updated. (A specific example of this linkage device is JP-A-58-123148.) Now, the control computer 6 determines and modifies a controlled object model that changes depending on the operating conditions of the controlled object 7, and parameters, gains, etc. are determined and transmitted to the process control unit 1-1.1-2 via the network 8.

The process control unit 1-1.1-2 calculates a control output vector from the state 127 input from the controlled object 7 via the process input device 2 and the optimal gain matrix given from the control computer 6, and outputs a process output. It is output to the controlled object 7 via the device 3.

Furthermore, when the state quantity of the controlled object 7 is unknown, the process control device 1-1.1-2 estimates the state quantity.

As described above, the process control unit 1-1.1-2 calculates the vector quantity, and the calculation is performed by the vector processing unit (hereinafter abbreviated as VPU-A) 21. (The configuration and operation inside the process control unit 1-1.1-2 will be described later.) On the other hand, a vector processing unit (hereinafter referred to as VPU) directly connected to the linkage device 22
-B) 211 is used for closed-loop system calculations that do not involve state estimation or state quantity conversion. That is, VPU-
8211 is the state quantity 27 of the controlled object 7 taken in from the controlled object 7 via the process input device 2 and the process control unit 1
-1.1-2 or including the optimum gain matrix and target value offset amount given from the control computer 6, and outputs it to the controlled object 7 as a control signal 24 via the process output device 3.

(The internal configuration and operation of the VPU-8211 will be described later.) Furthermore, the functional process input/output device 4 itself has a vector calculation function, and calculates the state quantity 27 of the controlled object 7 that it has taken in. It performs calculations with the gain matrix and the offset amount including the target value given from the process control unit 1-1.1-2 or the control computer 6, and outputs it to the controlled object 7 as a control signal 24. (The internal configuration and operation of the functional process input/output device 4 will be described later.)
Configuration and operation of process control unit 1-1.1-2) Process control unit 1-1.1-2 includes a CPU 20, a plurality of vector processing units (VPU-A) 21, and a data link memory 23 connected via a bus 28. has been done. One of these operations is as follows.

(Operation 1) The CPU 20 receives a control parameter matrix such as an optimal gain matrix from the control computer 6 via the network 8 and distributes it to the VPU-A 21. On the other hand, the state quantity vector of the controlled object 7 input from the process input device 2 is stored on the data link memory 23.

On each VPU-A 21, a control parameter matrix such as an optimal gain matrix is multiplied by a state quantity matrix, and the result is stored in the data link memory 23.

The stored contents are stored in the linkage device 22. Control signal (control output vector) of the controlled object 7 via the process output device 3
It is output as 24.

(Operation 2) As another operation, within the controller 1-1 or 1-2,
There is an operation to calculate parameters such as an optimal gain matrix.

For example, in order to solve the Riccati equation using a recurrence formula (for example, Kano, Nishimura: Matrix Riccati equation for optimal estimation control problems "Measurement and Control J VO120
, No. 6 (Refer to the June 1986 issue)) The matrix given from the control computer 6 is transferred from the CPU 20 to the VPU-A.
21, performs calculations cyclically, and stores the calculation results in the data link memory 23 periodically. The stored contents are provided to the functional process input/output device 4 via the linkage device 22.

(The operation of the functional process input/output device will be described later.) (Operation 3) Furthermore, as another operation, the state of the controlled object 7 may be estimated.

For example, to solve the observer using recurrence formulas, see "Observer for Multivariable Systems" (written by Loinburger) D, G', Luenburger;
0bserver for Mult; v
ariable Systems, IEEEB Tran
s, ON Automatic Control,
AC-11-2190/197 (1966)) The matrix given from the control calculation m6 and the known state vector stored in the data link memory 23 from the process input device 2 via the linkage device 22 are
The PU-A 21 performs calculations and periodically stores the results in the data link memory 23. The stored contents are handled in the same way as the state quantity of the controlled object described in (operation 1).

(Internal configuration and operation of VPU-8211) Next, VPU-8
The configuration and operation of 211 will be explained. Figure 2 shows VPU-82
11 shows the configuration of No. 11.

The data link memory 2112 receives data from the host computer 6 or the process control unit 1-1.1-2 from the data link device 22.
．． Control parameters such as an optimal gain matrix and offset vector quantities including target values inputted via the data link interface 2111 and estimated state quantity vectors of the controlled object given from the process input device 2 are stored.

The vector multiplier 2113 multiplies the state vector of the controlled object by the optimal gain matrix, and the vector adder 2114 adds the result to the offset amount including the target value.
The control output vector is stored in the data link memory.

The stored results are transferred to data link 22. It is output to the controlled object 7 via the process output 3. As a result, the controlled object 7 is controlled to reach the target value.

(Internal configuration and operation of functional process input/output device 4) Part 3
1 is an internal configuration diagram of a functional process input/output device 4 connected to the control system of FIG. In the figure, 22 is a linkage device, and 230 is a linkage interface device. 101 is a data link memory that stores an optimal gain matrix; 102 is a buffer memory that stores process state quantities; and 103 is a vector processing unit (hereinafter referred to as VPU-C
), 104 is an A/D converter, 106 is a D/A converter, 110 is an input interface, 111 is an output interface, and 120 to 123 are system buses consisting of a data bus, an address bus, and a control bus. Yes, each device is connected as shown in the figure.

One of these operations is as follows.

Status signals 27 from the process are input to the input interface 11
0 into the process input/output device, A/
It is converted into a digital signal by a D converter 104. This digital signal is sent from the A/D converter 104 to the system bus 1.
20 to the buffer memory 102 and linkage interface device 230.

The linkage interface device 230 sends this process status signal to the linkage device 22, and the process control device 1 (1-1 or 1-2 in FIG. 1) determines the optimum gain matrix. The process control unit 1 transmits the optimum gain matrix through the linkage device 22, the linkage interface device 230, and the system bus 121.
Write to data link memory 101. After finishing writing the optimum gain matrix to the data link memory, the process control unit 1 sends an activation signal to the vPU-0103 via the control bus within the system bus 121. V.P.
Upon receiving this activation signal, the U-C 103 reads the optimal gain matrix and the process state quantity matrix from the data link memory IJIOI and the buffer memory 102, respectively, performs matrix operations to determine the process control quantity vector, and through the system bus 123, the D/A
At the same time as transmitting it to the converter 106, a process state quantity readout signal is transmitted to the multiplexer 105 in the A/D converter 104 via the system bus 122.

The D/A converter 106 converts the process control amount into an analog value and outputs it through the output interface 111.

According to this embodiment, VPU-A, VPU-B, VP
By using U-C, it is possible to separate the part that determines the optimal parameters and the part that determines the control amount, and furthermore, it is possible to minimize the bus from taking in the state amount of the controlled object to outputting the control amount. , multivariable control can be realized economically.

Also, depending on the configuration of the control system, for example, the VP
Among the three types, U-A, VPU-B, and vpu-c, V
For example, it is equipped with only PU-A and VPU-B.
Not necessarily the relevant VPU-A, VPU-B, VPU-C
(7) It goes without saying that the same effect can be obtained even without the three types.

〔Effect of the invention〕

According to the present invention, a part for inputting a control target, determining a controlled object model, determining parameters, and a part for determining a controlled amount are separated, and furthermore, there is a path for taking in the state quantity of the controlled object and outputting the controlled amount. can be made in the shortest possible time, allowing for highly accurate control.

Furthermore, the vector processing unit placed in the process control section can speed up multivariable calculations, so it can be applied to the field of high-speed control.

Furthermore, by connecting multiple process input/output devices and multiple process control units with a linkage device as necessary, the outputs and state quantities of the control target can be shared between each process control unit, reducing the overhead of data transfer. It can be reduced.

Furthermore, if necessary, vector
By directly connecting the processing units, simple calculations can be performed via the process control section, thereby reducing the overhead of the process control section.

Furthermore, by connecting a functional process input/output device having an output calculation section to the linkage device, simple calculations can be performed without going through the process control section, so that the overhead of the process control section can be reduced.

Furthermore, a functional process input/output device having the above-mentioned functions can shorten the data transfer path, thereby realizing high-speed control.

Furthermore, by employing the above configuration, control can be easily distributed, so system design can be simplified.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a process control system that is an embodiment of the present invention, FIG. FIG. 3 shows a block diagram of a functional process input/output device 4 which is also connected to the linkage device 22. As shown in FIG. 1.1-1.1-2... Process control device, 2...
Process input device, 3... Process output device, 4...
・Functional process input/output device, 21,103.211・
...Vector processing unit, 22...Linkage device, 6...Control computer. Agent Patent Attorney Tadashi Akimoto Minoru 1 Figures 1-1 to 1-27 People with their mouths! 'JGpL, j 6 battle control battleships, 11 ships, 22 ringoon commanders! 21.211 Vector 7゛ Rosebununbu Unit 7 Cape Cape It
20CPU 2-41 Seven people sword UI 8 spears - notebook work 28 people Figure 2 Figure 3

Claims (1)

[Claims] 1. In a process control device that takes in a process quantity, performs calculations to bring the quantity to be controlled close to the target value, and outputs the controlled quantity to the process, output calculation calculates the controlled quantity to be output by performing calculations on the process quantity taken in. A process characterized by having independent means and a parameter calculation means for determining the parameters used for the calculation based on the control target and the captured process quantity, and determining the control amount and determining the parameters in parallel with both of the above means. Control device.