JPH0448301A - Knowledge generator for system to be controlled - Google Patents

Abstract

PURPOSE:To generate the subject knowledge including the relative constitution information on the constitutional members by estimating the relation between a corresponding constitutional member and another by the attributes and the array of basic words. CONSTITUTION:A knowledge base control part 9 is connected to a subject knowledge 11 which is referred to at production of a program of a CPU 5 and a basic word dictionary 13 which shows the meanings of the basic words related to the machine names of the controlled subjects. Furthermore the part 9 is connected to a production rule 15 which estimates the constitutional element of each machine based on the array of basic words of each machine name and a method 17 which estimates the types of the constitutional relations among machines based on the meanings of basic words. Then the part 9 controls and retrieves the knowledge information on those knowledge 11, dictionary 13, rule 15, and method 17 respectively. Thus it is possible to automatically generate the subject knowledge including the constitution relative information among the constitutional members of facilities.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial Application Field) The present invention relates to a control object knowledge generation device in a control execution program automatic generation system.

(Prior Art) For example, in programmable controllers that control various plants, etc., in order to operate them effectively,
A control execution program must be generated. Therefore, we aim to improve the efficiency of work [first, by collecting the equipment configuration and operating conditions of the controlled object, its constraints, timing of each operation, etc. as target knowledge in a knowledge base, and specifying only the outline of the operation, Automatic generation of control execution programs has come to be practiced. Furthermore, in order to improve the versatility of the automatic program generation system, a method for automatically generating target knowledge has also been proposed. In this method, the machine name written in the equipment information is input, the machine name is divided into basic words, and the upper-level constituent machines are inferred from the arrangement of each basic word according to the attributes of the basic word. The parts that represent the machine configuration are automatically generated.

(Problem to be Solved by the Invention) However, in the conventionally proposed methods for automatically generating subject knowledge, it is not possible to estimate that a given component is a component of another higher-level component. However, there is a problem in that it is not possible to estimate the mutual relationship between arbitrary constituent members.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a control object knowledge generation device that generates object knowledge from which configuration relationships can be estimated.

[Structure of the invention]

(Means for Solving the Problems) In order to solve the above problems, the present invention uses object knowledge that defines relationships related to controlled objects, and generates a control execution program based on input operation information. This control object knowledge generation device includes a decomposition means for decomposing manually input names of component parts into basic words, and a method for classifying the relevant component and other component parts from the attributes and arrangement of the basic words decomposed by the decomposition means. The present invention is characterized by comprising an estimating means for estimating the relationship between the components, and a generating means for generating target knowledge expressing the configuration of the constituent members from the relationship between the constituent members estimated by the estimating means.

(Operation) According to the controlled object knowledge generation method of the present invention, the relationship between the relevant constituent member and other constituent members is estimated from the attributes and arrangement of basic words, and the object including compositional relationship information between the constituent members is estimated. It becomes possible to generate knowledge.

(Example) Hereinafter, the present invention will be explained in detail using the drawings.

FIG. 1 is a block diagram showing the system configuration of a control execution program automatic generation system according to an embodiment of the present invention.

In FIG. 1, a bus 3 includes a central processing unit (hereinafter simply referred to as CPU) 5, which serves as a means for automatically generating programs such as a control execution program and generating target knowledge.
A program storage device 7 stores programs manually generated via the bus 3, a knowledge base management section 9, a database interface 21, and a program storage device 7 for storing programs manually generated via the bus 3; a knowledge base management section 9; a database interface 21; A terminal device 27 and a programmable controller interface 29 are connected for presenting information to the designer by inputting or displaying a generated program.

The knowledge base management unit 9 also stores object knowledge 11 generated by the CPU 5 and referred to when generating a program, a basic word dictionary 13 representing the meanings of basic words related to machine names to be controlled, and a basic word sequence of machine names. It is connected to a production rule 15 for estimating which machine a machine is a component of, and a method 17 for estimating the type of configuration relationship between each machine from the meaning of the basic word. Knowledge of dictionaries 13, production rules 15, methods 17, etc. ↑Manage and search information.

Further, the database interface 21 is used for managing and managing database information stored in the database 25.
A database management unit 23 that performs searches is connected.

Furthermore, a programmable controller interface 29
A programmable controller 31 is connected to the programmable controller 31, a controlled object 33 is connected to the programmable controller 31, and the programmable controller 31 controls the controlled object 33 according to a generated program.

In FIG. 1, the portion surrounded by dotted lines constitutes the control execution program automatic generation system 1.

FIG. 2 is an example of a description of equipment information input by a designer using the terminal 27, and FIG. 2(a) is a sensor list,
Figure (b) shows the actuator list. Equipment information input using the terminal 27 is stored in the database 25 through the database interface 21 and the database management section 23, and can be referenced by the CPU 5.

Next, we will compare the case of extracting machine names from the equipment information accumulated in this database 25 and automatically generating the part expressing the equipment configuration from the target knowledge as shown in Figure 3 with the conventional method. I will explain.

First, conventionally proposed methods will be explained.

First, the machine name in the equipment information is divided into basic words using the basic word dictionary 13 shown in FIG. 6, and the category of each divided basic word is known.

Next, by applying the production rule 15 shown in FIG. 14, it is estimated which machine each machine is a component of based on the sequence of basic words, and four pieces of target knowledge are created.

However, depending on this conventionally proposed method, although it is presumed that it is a certain machine or a component of a certain machine, the r has-unit-copp in FIG.
Onents J-related machine called “Tane 1 Ancoira Clamp DS” and “Arashi 1 Ancoira Clamp WS”
It is not possible to estimate the type of structural relationship between machines, such as the fact that these machines are driven simultaneously.

Next, a description will be given of an automatic target knowledge generation method that includes a procedure for estimating the type of configuration relationship between machines in order to destroy the target knowledge shown in FIG. 3 from the equipment information shown in FIG. 2.

FIG. 4 is a flowchart showing the procedure.

First, a machine name is extracted from the equipment information, and a frame shown in FIG. 5 is generated (step 5AI). Next, each machine name is divided into basic words based on the basic word dictionary 13 as shown in FIG. 6 (step 5A3).

For example, “No, 1 uncoiler clamp ws” is “No, 1 uncoiler clamp ws”
"+"Uncoiler 2+"Clamp→-"ws", and the division result is added to the slot value as a list as shown in Figure 7.The same goes for other machine names.After performing the above processing , all mechanical frames are added to the working memory of the development system (step 5A5).

In this example, the frame [Nα1 uncoiler clamp WSJ
and r No, 1 Amphira Clamp DSJ and “Yang, 1
Ankoira Mota”. The Broda Crayon system is activated targeting this frame in working memory.

First, it is checked whether there is a machine name frame that matches the condition part of rule 1 (step SA, 7).

For example, frame r No, 1 uncoiler clamp W
S” slot r 5eparate-deserip
Looking at tlonJ, the last word is "WS", which can be seen from the basic word dictionary as a category or a "location modifier". Also, the word before “WSo” is °clamp.

This is also found to be a "machine name" from the dictionary.

Therefore, frame "No. 1 uncoiler clamp WSJ complies with the condition part of rule 1. Similarly, frame r
Since the N11.1 uncoiled club/bu DSJ also matches and there is a machine name that matches the condition part of rule 1, the action part is executed and the frame shown in FIG. 8 is generated (step 5AII). The machine name is “No. 1 Uncoiler Clamp” with “DS” or “WS” removed.
The result of dividing into basic words is (No, 1 uncoiled clamp).

Next, the upper machine r No, 1' generated in this way
7 Coil Clamp J and its subordinate machine rNo. 1N07
A procedure for estimating the structural relationship between "1 uncoiler clamp r No. 1 uncoiler clamp WS" will be explained.

Step SA] In the previous process, rNctlN07
1 Ancoiler clamp DSJ, 1 Ancoiler clamp W
It is presumed that SJ is a component of "r No. 1 Uncoiler Clamp". However, as things stand, we do not know what kind of compositional relationship they have with each other.

Therefore, the slot r5eparaLe-
Refer to the slot value of descr 11onJ, that is, each of the divided basic words. Furthermore, among the attributes of the 91-word frame, there is a slot r fu that expresses the machine function.
Check whether the slot value of nctfonJ has a method for estimating the type of configuration relationship (Step 5A
13). In this example, the slot value of slot r f'unctionJ in the word frame "uncoiler" is "rewind".
and the thrower of the word frame "clamp") r fune
The presence or absence of the method is checked for the slot values "Material fixation" and "Sandwiching" of tionJ.

As shown in FIG. 9, since this method exists in the frame "Hamamu", this method is activated in order to estimate the type of configuration relationship. In this case, the lower machine is the "magnetic 1 uncoiler clamp DS" in contrast to the upper machine "rNa, l uncoiler clamp" which has a "pinching" function.
Since there are two or more such as J, rk, and 1 uncoiler clamp WSJ, it matches the if part and therefore executes the then part. That is, in order to express that the lower machine rE+lL1 uncoiler clamp DSJ and [No. 1 uncoiler clamp WSJ are driven simultaneously, the slot r has- of the frame rNo. 1 uncoiler clamp" is shown in FIG. unit-coBonents J value frame r No, 1 uncoiled clamp WSJ
and frame ``seed 1 ancoir clamp DSJ'' (step 5A15).

By describing the method in the function frame in this way, it is possible to estimate the same type of machine configuration relationship for machines having the same function even if the words are different.

Next, delete the frames “Seed 1 Ancoiler Clamp WSJ and Frame r No. 1 Ancoiler Clamp DSJ” that conformed to Rule 1 this time from the working memory, and create “NO, 1 Ancoiler Clamp” newly generated as a higher-level machine. Add to working memory (step 5A19)
.

Therefore, the contents of the current working memory are frame "Haruka 1 uncoiler clamp" and frame rlllo, 1 uncoiler motor.

Again, in step SA7, it is checked whether there is a machine name that matches the condition part of rule 1, but since it does not exist, the process proceeds to the next step SA9, and it is checked whether there is a machine name frame that matches the condition part of rule 2.

At this time, the frame "No. 1 uncoiler clamp" and the frame "Magic 1 uncoiler motor" are both compatible, and the 11th
A frame as shown in the figure is generated (step 5A11).

Next, in order to estimate the structural relationship between the upper machine "Y, 1 uncoiler J" and its subordinate machines "R No, 1 uncoiler motor", the upper machine frame "r NO, l uncoiler" is I check whether there is a method for the function frame "rewind" in the divided word frame "uncoiler", but since it does not exist (step 5
A13), default configuration relationship r has-coIl
Relationships are established using ponentsJ.

That is, as shown in FIG.
Uncoiler's slot r has-componen
Frame the value of LsJ "No, 1 amp Iraq Clamp"
and the frame is set as "1st uncoiler motor" (step 5A17).

Next, frame r No. 1 that complies with rule 2 this time
"Uncoiler Clamp" and frame "No. 1 Uncoiler Motor J" are deleted from the working memory, and "r No. 1 Uncoiler" newly generated as a host machine is added to the working memory (step 70).

The current working memory is only r No. 1 Uncoiler. Check the contents of this working memory for machine names that match the condition part of rule 1 (step 5).
A7) does not exist, so proceed to the next step SA9, check the case of rule 2, and similarly do not exist. Therefore, the process ends.

By the above procedure, the structural relationship between the machines shown in FIG. 3 is generated. Frame r No, 1 uncoiler clamp's slot rhas-unit-component
s J value or frame rNo,] Ancoiler clamp D
The structure of SJ and frame "Nα1 ancoiler lamp WSJ" expresses that the two machines mentioned above are driven at the same time.

For example, after closing both conventional r No. 4 Ancoiler Clamp DS and N091 Ancoiler Clamp WS,
.

1. The operation of "rotating the uncoiler motor" is the first
It had to be written as a control specification as shown in Figure 5.

Program start here SC], outline of target operation of controlled object SC3, SC5, 5CII, transition condition of target operation S
C7, 3C9, program end 5C13. Also L
indicates parallel branching.

However, in the object knowledge as shown in FIG. 3 generated by the control object knowledge automatic generation method in the control execution programm automatic generation system of the present invention, rNc + l uncoiled clamp DSJ and rNn, 1 uncoiled clamp WSJ are simultaneously generated. It includes the knowledge of driving. Therefore,
By using subject knowledge to supplement what previously had to be described, such as the control specifications shown in Figure 15,
As shown in the figure, it can be written more simply, and the designer's effort can be significantly reduced.

The automatically generated object knowledge is not only used to generate the control execution program, but also used for the control of the controlled object itself, as well as for specification verification, stain correction, etc. It is also effective for planning, etc.

〔Effect of the invention〕

As described above in detail in I7, according to the present invention, it is possible to automatically generate target knowledge that also includes configuration relationship information between equipment constituent members.

Therefore, in a control execution program automatic generation system using this, it becomes possible to more schematically describe the input and the operation information at 1 to 1, and the designer's labor is significantly reduced.

[Brief explanation of the drawing]

Figure 1 is a block diagram schematically showing the configuration of the present invention, Figure 2 is a block diagram schematically showing the configuration of the present invention.
The figure shows an example of equipment information, Figure 3 is a diagram showing the expression format of the part of the target knowledge that expresses the machine configuration, and Figure 4 is a flowchart showing the sequence of automatic generation of target knowledge.
Figures 7, 8, and 10 to 12 are diagrams showing the knowledge representation for the machine in the subject knowledge, Figure 6 is the memory format of the basic word dictionary, and Figure 9 is A diagram showing an example of a method expression for estimating the type of configuration relationship between machines, a diagram showing an example of a more schematic description of human power information than Figure 13, and a diagram showing an example of the description of the basic human power information from the sequence of 191 words in Figure 14. FIG. 15 is a diagram showing an example of the second rule expression for estimating whether it is a component of a machine, and FIG. 15 is a diagram showing input information input manually by the designer. 1-Control execution program automatic generation system 3・Bus 5・CPU 7-・Program storage device 9...Knowledge management unit] 1-Object knowledge 13・Basic word dictionary 15・Broadcast rules 17・Method 2-To database interface 23 Data storage management unit 25... Data storage 27 - Terminal 29 Programmable controller interface 31 Programmable controller 33 Controlled object 11 Figure

Claims (2)

[Claims] (1) Using object knowledge that defines the relationships related to the controlled object,
A control object knowledge generation device for generating a control execution program based on input operation information, comprising a decomposition means for decomposing an input component name into basic words, and a basic word decomposed by the decomposition means. estimating means for estimating the relationship between the constituent member and other constituent members from the attributes and arrangement of the constituent members; and generation for generating target knowledge expressing the configuration of the constituent member from the relationships between the constituent members estimated by the estimating means. 1. A controlled object knowledge generation device comprising: means. (2) Controlled object knowledge generation according to claim 1, wherein the generating means generates knowledge describing the relationship between the constituent members estimated by the estimating means for a constituent member above the constituent member concerned. Device.