JP2015035174A - Control program dividing apparatus, control program dividing method and recording medium therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a control program division device determining dependence of an input signal or a temporary variable on an output signal in advance, collectively dividing a source code to source codes each having similar dependence, and achieving reduction in control program execution time.SOLUTION: A control program division device includes: an input processing unit 10 reading a symbol string of a source code; a lexical analysis processing unit 11 to which the symbol string is input and that generates a token string; a syntax analysis processing unit 12 to which the token string is input, and that generates an input/output signal table, a dependence table, and a syntax tree from the token string; a division processing unit 13 collectively dividing the source code to a plurality of source codes each having dependence on an output signal from meaning of syntax rules described in the input/output signal table, the dependence table, and the syntax tree; and an output processing unit 14 outputting the divided source codes as a symbol string, the source code of a control program being collectively divided to the source codes each having similar dependence on the output signal.

Description

  Embodiments described herein relate generally to a control program dividing device, a control program dividing method, and a recording medium thereof used in a cloud control system.

  In order to facilitate disaster recovery, attention has been focused on remote control technology for remotely controlling field devices via a network. A system that performs control from a plurality of cloud servers that are remotely located is referred to herein as a cloud control system.

  In the cloud control system, since the control program is held in a plurality of physically separated servers, there is no possibility that all the control programs are lost in the event of a disaster. In addition, when recovering from a disaster, it is possible to shorten the time required for rebuilding the control program and replacing the control server.

  From such a background, in a cloud control system, a precompiler that divides a control program so that it can be executed in parallel by a plurality of servers is required.

  By the way, in a general multiprocessor compiler, there is a method of dividing loop processing so that it can be executed in parallel (see, for example, Patent Document 1).

  For the development of compilers for many programming languages, Lex is used as a program for generating a lexical analyzer (lexical analyzer), and Yacc is used as a program for generating a parser (parser generator). Yes (see Non-Patent Document 1, for example).

Japanese Patent No. 3729644

Authors John R Levine, Tony Mason, Doug Brown, Book Title Lex & Yacc, Paperback-306 pages 2nd / updated edition, Publication O’Reilly & Associates, Publication Date October 1992, International Standard Book Number: ISBN 1565920007

  In the above-described compiler of Patent Document 1, in order to determine the dependency of the program data and execute the loop processing in parallel, after dividing the loop, the calculation is performed by a plurality of different servers for each loop. Find the output.

  Therefore, since a single output is calculated while exchanging a plurality of data, communication between servers increases, so the network between servers is an ultra-high speed network.

  In other words, the compiler of Patent Document 1 requires a calculation time for obtaining one output for one day, such as a scientific and technical calculation including multiple loops in which there is a data dependency in which the calculation time is extremely longer than the communication time. This is applied to a program and has an effect of shortening the processing time of one output.

  However, the target program structure and network are different from a cloud control system that has a large number of input / output points, the ratio of the calculation time and communication time is the same, and the calculation time is in milliseconds. It cannot be applied as a compiler for high-speed processing of control programs.

  In the cloud control system configured by the network using the Internet, the present application determines the control program required to process a large number of input / output signals in advance by determining the dependency of the input signals and temporary variables on the output signals. Control program dividing apparatus, control program dividing method, and recording medium for reducing the execution time of the control program by dividing the source code having dependencies together and executing the divided control programs in parallel The purpose is to provide.

  In order to achieve the above object, the control program dividing apparatus according to the present embodiment includes a control program for a field device, which is a control target, in a cloud control system that controls a plurality of servers provided in remote locations. A control program dividing device as a precompiler that is divided into a plurality of servers in advance and executed in parallel. The control program source code generated in a predetermined programming language is input, and a symbol string of the source code is input. An input processing unit for reading, and a lexical analysis process for inputting a rule string of the programming language that has been read and made machine-readable, via a program Lex, and generating a token string (vocabulary string) of the source code And the token string analyzed by the lexical analysis processing unit, The token sequence syntax rule definition of the source code is given through the program Yacc, and the source code signal name and the input signal from the field device from the token sequence that matches the source code syntax rule definition, primary variables, and Input / output signal table described by determining the type of output signal, dependency table described by determining which output signal depends on which input signal and primary variable, and structure of the source code From the meaning of the syntax rule definition described in the syntax analysis processing unit for generating the syntax analysis processing unit, the input / output signal table, the dependency table, and the syntax tree, A division processing unit that divides each of the plurality of divided source codes depending on the output signal, and an output that outputs the divided source code as the symbol string Comprising a processing section, a, the source code of the control program, characterized by dividing together with the source code together with the same dependence on the output signal.

  In order to achieve the above object, the method for dividing a control program according to the present embodiment is a cloud control system in which a control program for a field device that is a control target is controlled by a plurality of servers provided in remote locations. A method of dividing a control program as a precompiler that is divided into a plurality of servers in advance and executed in parallel, wherein the processor as the precompiler inputs a source code of the control program generated in a predetermined programming language An input processing step for reading a symbol string of the source code; a rule definition of the programming language that is input to the read symbol string and made machine-readable is given via a program Lex, and a token string (vocabulary of the source code) Lexical analysis processing step for generating a sequence) and the lexical analysis processing The token sequence parsed in the above is input, the syntax rule definition of the token sequence of the programming language is given via the program Yacc, and the signal name of the source code from the token sequence that matches the syntax rule definition of the source code Input / output signal table describing and describing the types of input signals, primary variables, and output signals with the field equipment, and determining which output signals depend on which input signals and primary variables A syntax analysis processing step for generating a described dependency table, a syntax tree described by determining the structure of the source code, the input / output signal table, the dependency table, and the syntax tree; Is divided into a plurality of divided source codes that depend on the output signal. And an output processing step for outputting the divided source code as the token string, and the source code of the control program is divided into source codes having the same dependency on the output signal. It is characterized by that.

The block block diagram of the control program division | segmentation apparatus of embodiment. Example of source code before dividing the control program. An example of an input / output signal table. An example of an input / output signal dependency table. An example of a syntax tree. An example of the source code of a divided control program. The detailed block diagram of an output process part. The other detailed block diagram of an output process part. The block diagram of the control program division | segmentation apparatus which compiles the source code after a division | segmentation.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the control program dividing apparatus 1 shown in the present embodiment, a control program used in a cloud control system that controls a control program for a field device to be controlled by a plurality of servers in a remote place is parallelized by a plurality of servers. It is a precompiler for dividing the control program in advance so that it can be executed.

  The control program dividing device 1 may be configured as a dedicated device, or may be configured so that any server connected to the network becomes a precompiler before executing the control program.

  First, the configuration of the control program dividing device 1 will be described with reference to FIG. The configuration of the control program dividing device 1 includes an input processing unit 10 that inputs a source code of the control program generated in a predetermined programming language and reads a symbol string of the source code, and a symbol string read by the input processing unit 10 A lexical analysis processing unit 11 for generating a machine-readable rule definition of the programming language via a program Lex (see Non-Patent Document 1) and generating a token string (vocabulary string) of a source code. Prepare.

  Further, the token string analyzed by the lexical analysis processing unit 11 is input, and the syntax rule definition of the programming language token string is given via the program Yacc (see Non-Patent Document 1), and matches the syntax rule definition of the source code. An input / output signal table 20 in which the signal name of the source code and the input signal of the field device, the primary variable, and the type of the output signal are determined and described from the token sequence, which output signal is which input signal and primary It includes a dependency analysis table 12 that generates a dependency table 21 that is described by determining whether it depends on a variable, and a syntax tree 22 that is described by determining the structure of the source code.

  Further, from the meaning of the syntax rule definition described in the input / output signal table 20, the dependency table 21, and the syntax tree 22, it is divided into a plurality of divided source codes that depend on the output signal. A division processing unit 13 and an output processing unit 14 that outputs the divided source code as a token string are provided.

  Here, the programming language of the control program will be described assuming the programming language defined in the international standard IEC61131-3 and this existing IEC61131-3, but may include the expanded international standard IEC61499, and others It may be a high-level programming language such as FORTRAN.

  The cloud control system network here is composed of a general-purpose Internet and a plurality of servers and field devices connected to the network. The field devices and this network are connected by a dedicated network. You may comprise as follows.

  Next, before describing the configuration of each unit, the source code 30 of the control program for describing the present embodiment will be described with reference to FIG.

  The source code 30 of the control program shown in FIG. 2 shows the source code 30 composed of a declaration part and an instruction part before division based on IEC61131-3. The declaration part describes input / output signal variables, primary variables and functions POU1 to POU4, and the instruction part describes their functions POU1 to POU4. The meaning of each token string is omitted because it is a common knowledge.

  Next, the detailed configuration of each unit will be described. The input processing unit 10 reads the symbol string described from the source code 30 before the division, and sends the read symbol string to the lexical analysis processing unit 11.

  Further, the lexical analysis processing unit 11 inputs a symbol string read by the input processing unit 10 and gives a rule definition of a programming language that is machine-readable via a program Lex (see Non-Patent Document 1). A token string (vocabulary string) is generated.

  Here, the token string of the source code 30 of the control program is a word or signal name indicating a control syntax defined in a programming language, a signal name arbitrarily described by the user, a signal type, or the like.

  For example, in the case of the source code 30 of the control program described in the control program language IEC61131-3 shown in FIG. 2, “PROGRAM” indicating the start of the program, “POU0” indicating the program name, and “ “VAR”, “X1” indicating the signal name, “INT” indicating the signal type, and the like are token strings of the program source code.

  Such processing in the lexical analysis processing unit 11 gives a syntax rule definition of a preset programming language that is machine-readable in advance using a program such as Lex or Flex that is a lexical analyzer generation tool. Then, lexical analysis is performed on the input symbol string, and a token string is generated.

  Next, the syntax analysis processing unit 12 gives a syntax rule definition of the token sequence analyzed by the lexical analysis processing unit 11 in advance through the program Yacc, and the input token sequence is checked against the syntax rule definition given in advance. Then, it is determined which syntax rule is matched, and the meaning of the source code 30 of the control program is described as a table or a structure diagram.

  For example, it is described in the input / output signal table 20 in which the signal name of the source code 30 and the type of the input signal, temporary variable, and output signal of the field device are described, and which output signal depends on which input signal and temporary variable Is described in the dependency table 21 that describes whether or not the source code 30 is structured. The structure of the source code 30 is described as a syntax tree 22 described later.

  In the example shown in FIG. 2, “PROGRAM to END_PROGRAM” is one source code 30, and “VAR to END_VAR” is a part that describes the definition of the signal name. Is determined.

  Similarly, “X1 AT% IW61.1: INT;” is also determined that the input signal X1 of the integer variable is stored in the memory address 61.1.

  Based on such token string determination information of the source code 30, an input / output signal table 20, a dependency table 21, and a syntax tree 22 are generated.

  In the generated input / output signal table 20, as shown in FIG. 3, a signal name 23 and an input / output signal type 24 are described.

  The number of input / output signals in the source code 30 is unknown unless the source code 30 is read, so the input / output signal table 20 is held in a table structure in which the number of data can be dynamically added. You may do it.

  In the dependency table 21, as shown in FIG. 4, a signal name 25, an input / output signal type 26, a direct dependency 27, and an indirect dependency 28 are described.

  In the case of the source code 30 in FIG. 2, this dependency table 21 is collated with the syntax rule definition related to the dependency set in the syntax analysis processing unit 12 and “POU1 (X: = X1, Y => T1);” In the description of the line, since it can be determined that the function POU1 takes X1 as input and outputs T1, T1 is determined to depend on X1, and this determination is based on the directivity dependency 27 of the signal name T1. X1 is described in the column.

  Similarly, it is determined that Y1 depends on T1, Y2 depends on T2 and X2, and T2 depends on X3, and this determination is described in the dependency table 21 as a direct dependency 27.

  Furthermore, an indirect dependency 28 can be generated from the direct dependency 27 for each signal name.

  For example, since Y1 depends on T1 and T1 depends on X1, it is described in the dependency table 21 as indirect dependency 28 that Y1 also depends on X1. Therefore, X1 is described in the column of the indirect dependency 28 of the signal name Y1. Similarly, X3 is described in the column of indirect dependency 28 of Y2.

  In other words, the dependency is a variable that is directly used to obtain the corresponding output is a direct dependency. If a variable that has a direct dependency is obtained by another variable, this other variable is used. Is determined to have indirect dependency.

By describing such a dependency table 21 in the table, the dependency of the input signal and the temporary variable on the output signal can be referred to in a short time.
The syntax tree 22 is a graphic description of the structure of the source code 30 in comparison with the syntax rule definition related to the program structure. The parsing processing unit 12 describes a tree structure in which one line of the source code 30 is a node, and a line subordinate to a certain line is a lower node of that node.

  For example, for the source code 30 shown in FIG. 2, graphic data having a tree structure as shown in FIG. 5 is described.

  Thus, by graphing the structure of the source code 30 with a tree structure, it is possible to set the indent of the output source code 30 and output the source code 30 with good visibility.

  Next, the division processing unit 13 reads the input / output signal table 20, the dependency table 21, and the syntax tree 22, and checks which output signal each node of the syntax tree 22 depends on against the syntax rule definition related to the program structure. Judgment.

  For example, in the example of FIG. 5, nodes such as “PROGRAM POU0” and “VAR” are nodes necessary for configuring a program regardless of the output signal. Further, according to the dependency table 21, the node “POU1 (X: = X1, Y => T1);” depends on the output signal Y1.

  Therefore, “POU1 (X: = X1, Y => T1);” is a node necessary only for the program that outputs Y1.

  Similarly, the node “POU2 (X: = T1, Y => Y1);” also depends on the output signal Y1 according to the dependency table 21.

  Therefore, “POU2 (X: = T1, Y => Y1);” is a node necessary only for the program that outputs Y1, and it is determined that POU1 and POU2 have the same dependency.

  On the other hand, according to the dependency table 21, the node “POU3 (X: = T2, X2 = X2 Y => Y2); is Y => Y2;” depends on the output signal Y2.

  Therefore, the node “POU3 (X: = T2, Y => Y2);” is a node necessary only for the program that outputs Y2.

  Similarly, the node “POU4 (X: = X3, Y => T2);” also depends on the output signal Y2 according to the dependency table 21.

  Therefore, “POU4 (X: = X3, Y => T2);” is a node necessary only for the program that outputs Y2, and it is determined that POU3 and POU4 have the same dependency.

  As a result, as shown in FIG. 6A, the source code 30 shown in FIG. 2 includes the divided source code 31 having the same dependency on the instruction part and the instruction code shown in FIG. The divided source code 32 having the same dependency is output.

  That is, in the division, the source code is divided into the number of output signals.

  The output processing unit 14 scans the generated syntax tree 22 and outputs the nodes related to the output signals as divided source codes 41 and 42. As shown in FIG. 21, a signal name acquisition unit 15 that acquires a signal name, a file name setting unit 16 that sets a file name for outputting the divided source codes 41 and 42, and the divided source codes 41 and 42 as text POU_Y1 and POU_Y2. A text output unit 17 for outputting as follows may be provided.

  By providing the signal name acquisition unit 15 and the file name setting unit 16, it is possible to output the file of the divided source codes 41 and 42 with the names of the output signals. It is possible to determine which output signal is a program for calculating from the file name.

  As shown in FIG. 8, the output processing unit 14 includes a serial number generation unit 18 that generates serial numbers 1 to N, a file name setting unit 16a that sets a file name to be output, and a divided source. It can also be configured to include a text output unit 17a that outputs the code as text.

  By providing the serial number generation unit 18, the file name setting unit 16a, and the text output unit 17a, it is possible to output the files 51 and 52 with serial numbers, and how many output signals are present. Judgment can be made only from the file name.

  As shown in FIG. 9, the control program dividing apparatus 1 further includes a compiler 19 for converting the source code 30 of the control program into the object program, and the divided control programs 34 to 36 may be output.

  Also in this case, similarly, an output signal name may be added to the file names of the divided control programs 34 to 36, or serial numbers may be added.

  According to the present embodiment, the source code of a control program having a plurality of input / output signals is divided into source codes having the same dependency on output signals and divided into a plurality of divided control programs that can be executed in parallel. Can be divided into

  Therefore, according to the present embodiment, in a cloud control system configured by a network using the Internet, a control program required to process a plurality of input / output signals is the same as an input signal or temporary variable for an output signal in advance. It is possible to provide a control program dividing device, a control program dividing method, and a recording medium thereof, which are divided into source codes having dependencies and reduced by a plurality of servers to reduce the execution time of the control program.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalents thereof.

DESCRIPTION OF SYMBOLS 1 Control program division | segmentation apparatus 10 Input processing part 11 Lexical analysis processing part 12 Syntax analysis processing part 13 Division processing part 14 Output processing part 15 Signal name acquisition part 16 File name setting part 17 Text output part 18 Serial number generation part 19 Compiler 20 In Output signal table 21 Dependency table 22 Syntax tree 23, 25 Signal name 24, 26 Input / output 27 Direct dependency 28 Indirect dependency 30 Source code 31, 32, 33 before division Source code 34, 35, 36 after division Later source program

Claims (13)

In a cloud control system in which a control program for field devices to be controlled is controlled by a plurality of servers provided in remote locations, the control program is divided as a precompiler that divides the control program into the plurality of servers and executes them in parallel A device,
An input processing unit for inputting a source code of the control program generated in a predetermined programming language and reading a symbol string of the source code;
The lexical analysis processing unit that inputs the read symbol string and gives the rule definition of the programming language made machine-readable via the program Lex, and generates a token string (vocabulary string) of the source code;
The token string analyzed by the lexical analysis processing unit is input, the syntax rule definition of the token string of the programming language is given via the program Yacc, and the source from the token string that matches the syntax rule definition of the source code The input / output signal table describing the signal name of the code and the input signal of the field device, the primary variable, and the output signal, and which output signal depends on which input signal and the primary variable A syntax analysis processing unit for generating a dependency table described by determining whether or not, a syntax tree described by determining the structure of the source code,
Divided for each of a plurality of divided source codes having a dependency on the output signal from the meaning of the syntax rule definition described in the input / output signal table, the dependency table, and the syntax tree. A split processing unit to
An output processing unit for outputting the divided source code as the symbol string;
With
An apparatus for dividing a control program, comprising: dividing the source code of the control program into source codes having the same dependency on the output signal. The dependency table is determined based on the syntax rule definition of the token sequence given in advance, the input signal with respect to the output signal, and the direct dependency of the primary variable;
An indirect dependency which is determined based on the syntax rule definition of the token sequence given in advance and indicates an indirect dependency described from the direct dependency of all the source codes of the control program;
Is described as a table,
2. The control program dividing apparatus according to claim 1, wherein the dependence of the input signal and the primary variable on the output signal can be referred to. The syntax tree is determined based on the syntax rule definition of the token sequence given in advance, and describes one line of the source code of the control program as a node, and a line subordinate to a certain line is subordinate to that node. Describe the source code in a tree structure described as a node,
2. The control program dividing apparatus according to claim 1, wherein the source code can be output in a tree structure. The output processing unit sets a signal name acquisition unit that acquires the signal name from the dependency table for the plurality of source codes divided by the division processing unit, and sets the file name of the divided source code A file name setting section and a text output section for outputting the file
With
2. The control program dividing device according to claim 1, wherein it is possible to know which output signal is a program for calculating the output signal from the file name of the divided source code. The output processing unit is a file that sets a serial number generation unit that generates a serial number of the plurality of divided source codes divided by the division processing unit, and a file name that sets the file name of the source code given the serial number A name setting part and a text output part for outputting the file,
With
2. The control program dividing apparatus according to claim 1, wherein the number of the output signals can be known from the file name of the divided source code.   The control program dividing apparatus according to claim 1, further comprising a compiler that converts the source code of an executable control program into an object code for the plurality of divided source codes. In a cloud control system in which a control program for field devices to be controlled is controlled by a plurality of servers provided in remote locations, the control program is divided as a precompiler that divides the control program into the plurality of servers and executes them in parallel A method,
The processor as the precompiler is:
An input processing step of inputting a source code of the control program generated in a predetermined programming language and reading a symbol string of the source code;
The lexical analysis processing step of inputting the read symbol string and providing the machine language readable rule definition of the programming language via the program Lex to generate a token string (vocabulary string) of the source code;
The token string analyzed by the lexical analysis processing unit is input, the syntax rule definition of the token string of the programming language is given via the program Yacc, and the source from the token string that matches the syntax rule definition of the source code The input / output signal table describing the signal name of the code and the input signal of the field device, the primary variable, and the output signal, and which output signal depends on which input signal and the primary variable A syntax analysis processing step for generating a dependency table described by determining whether or not, and a syntax tree described by determining the structure of the source code,
Divided for each of a plurality of divided source codes having a dependency on the output signal from the meaning of the syntax rule definition described in the input / output signal table, the dependency table, and the syntax tree. Split processing steps to
An output processing step of outputting the divided source code as the token string;
With
A method for dividing a control program, comprising: dividing the source code of the control program into source codes having the same dependency on the output signal. The generation of the dependency table in the parsing processing step is determined based on the syntax rule definition of the token sequence given in advance, the input signal with respect to the output signal, and the direct dependency of the primary variable; An indirect dependency which is determined based on the syntax rule definition of the token sequence given in advance and indicates an indirect dependency described from the direct dependency of all the source codes of the control program; To describe it,
8. The control program dividing method according to claim 7, wherein the dependence of the input signal and the primary variable on the output signal can be referred to. The generation of the syntax tree in the previous parsing step is determined based on the syntax rule definition of the token sequence given in advance, and describes one line of the source code of the control program as a node. The source code is described in a tree structure in which subordinate lines are described as subordinate nodes of the node,
8. The control program dividing method according to claim 7, wherein the source code can be output in a tree structure. The output processing step obtains the signal name from the dependency table for the plurality of source codes divided by the division processing unit, names the file names of the divided source codes, and Output text to file,
8. The control program dividing device according to claim 7, wherein it is possible to know which output signal is a program for calculating the output signal from the file name of the divided source code. The output processing step generates a serial number of the plurality of divided source codes divided by the division processing unit, sets a file name of the source code given the serial number, and outputs the file And
8. The control program dividing method according to claim 7, wherein the number of output signals can be known from the file name of the divided source code.   8. The control program dividing method according to claim 7, further comprising the step of converting the source code of the executable control program into object code for the plurality of divided source codes.   A recording medium storing a precompiler that realizes the method for dividing the source code of the control program according to claim 7.

Images