Screen information generation device, machine learning device, and screen information generation method

Abstract

The present disclosure generates screen information constituting a screen on which screen components related to status information of a control target can be confirmed with less labor by eliminating the need for screen transitions.　This screen information generation device is provided with: a signal description information acquisition unit that acquires signal description information including at least one of an address of a signal, a comment added to the signal, the structure of a sequence program, and a screen component type, which are included in the sequence program; a status information acquisition unit that acquires, from a control target controlled by the sequence program, status information of the control target; a screen component group classification unit that, on the basis of the signal description information and the status information, classifies screen components into a plurality of screen component groups related to the status information and including the signal description information; and a screen information output unit that outputs screen information of the screen component groups to a display device.

Description

Screen information generating device, machine learning device, and screen information generating method

[0001] The present disclosure relates to a screen information generating device, a machine learning device, and a screen information generating method, and in particular to a screen information generating device, a machine learning device, and a screen information generating method that generate screen information for configuring the screen of a display device.

[0002] Screen information generating devices for generating screen information for configuring a display device screen are described in, for example, Patent Documents 1 to 7. Patent Document 1 describes a screen information generating device that reduces the workload when creating screen information for individual operation screens for operating controlled equipment in operation units. Specifically, Patent Document 1 describes that a computer serving as a screen information generating device generates screen information showing the display content of a monitor device connected to a programmable controller that controls equipment in accordance with a chart-style program that indicates the execution order of multiple steps provided for each operation unit. When generating screen information for each individual operation screen for operating the equipment individually in operation units, the computer generates screen information for displaying multiple operation buttons corresponding to each of the multiple steps on the monitor device based on the chart-style program.

[0003] Patent Document 2 describes a PLC control panel that displays multiple lamps and switches to enable PLC operation, can display the name of an abnormality occurring from the PLC, and allows the operator to easily grasp the details of the abnormality. Specifically, Patent Document 2 describes an LCD display panel for the PLC control panel that displays multiple soft lamps SL for checking the PLC's operating status and multiple soft switches SS for operating the PLC. Part of this display screen is provided with abnormality name display sections ABW1-ABW4 that can display several names of abnormalities transmitted from the PLC. When the operator touches one of these sections ABW1-ABW4 with his or her finger, the display can instantly switch to a detailed screen for the abnormality displayed in that section. Furthermore, when multiple abnormalities occur simultaneously, the abnormality names are displayed in the abnormality name display sections ABW1-ABW4 in order of priority, based on importance and urgency.

[0004] Patent Document 3 describes a programmable display device that can analyze display processing in real time and easily identify the cause and location of a malfunction. Specifically, Patent Document 3 describes that the programmable display device includes a debug device I / F that communicates with a debug device, a PLC I / F that receives status data transmitted by a PLC, a status data reception control unit that writes the status data received by the PLC I / F to a received data memory, a display control unit that references the status data in the received data memory and displays a screen based on the referenced status data, and a stop condition storage unit that stores a stop condition for the screen display. When the display control unit stops display processing based on the stop condition, the status data reception control unit stops writing the status data to the received data memory, and the debug device I / F transmits the status data in the received data memory to the debug device after stopping writing the status data.

[0005] Patent Literature 4 describes a plant diagnostic device capable of detecting a predictive state on the way from a normal state to an abnormal state. Specifically, Patent Literature 4 describes a plant diagnostic device that diagnoses the operating state of a plant based on measurement signals measuring state quantities from the plant and displays the diagnosis results on an image display device. The plant diagnostic device includes a learning unit that constructs a model to be used for diagnosis using the measurement signals measuring the plant state quantities in the plant diagnostic device, and a diagnostic unit that diagnoses the operating state of the plant using the model constructed by the learning unit. The learning unit includes a classification unit that classifies data having similar values ​​into the same category, a period determination unit that evaluates differences in trends of the measurement signals based on the category classification results of the classification unit and determines normal periods, predictive periods, and abnormal periods, and a model construction unit that constructs a normal model using the measurement signals of the normal period. The diagnostic unit determines whether the current measurement signal is classified into the normal model constructed by the model construction unit, and if the current measurement signal is included in the normal model, diagnoses the plant as being in a normal state and displays this on the image display device. However, if the current measurement signal is not included in the normal model, diagnoses the plant as being in an unknown state that the plant has never experienced before and displays this on the image display device.

[0006] Patent Literature 5 describes a programmable display that can reduce the effort required to select a destination screen by sequentially displaying multiple screens that are candidate destinations in a predetermined order. Specifically, Patent Literature 5 describes that the programmable display includes a display control unit, a priority determination unit, and a destination determination unit. The display control unit displays one of multiple screens. When a predetermined operation is performed on the currently displayed screen, the priority determination unit determines the priority of the destination screen candidates based on past performance data. The destination determination unit displays the destination screen candidates in a predetermined format in accordance with the priority determined by the priority determination unit, and determines a screen arbitrarily specified for the display as the destination screen.

[0007] Patent Document 6 describes a program input device that can indent and fold programs with simple operations. Specifically, Patent Document 6 describes that the program input device includes an input unit for inputting programs and comments, a storage unit for storing the input programs and comments, hierarchical information indicating the hierarchical relationship between the comments, and folding information indicating whether the program of the comment and programs and comments in lower hierarchical levels are to be folded, a display unit for indenting the comments based on the hierarchical information and folding information, and selecting and displaying the programs and comments, and a control unit for storing the programs and comments in the storage unit and editing the hierarchical information and folding information.

[0008] Patent Document 7 describes a screen pattern classification device that improves productivity and ensures quality in program development. Specifically, Patent Document 7 describes that the screen pattern classification device, which processes a plurality of designed screen patterns and classifies these screen patterns into a plurality of groups, comprises: an acquisition means that acquires, for each screen pattern to be processed, information on program parts that can be identified from the screen pattern; and a classification means that measures the distance between the screen patterns to be processed in accordance with the program part information acquired by the acquisition means, and classifies the screen patterns to be processed into a plurality of groups based on the distance.

[0009] JP 2016-212802 A JP 2001-154709 A JP 2007-109120 A International Publication No. 2012 / 073289 International Publication No. 2014 / 125587 JP 2011-86118 A JP 2007-193482 A

[0010] When arranging screen components corresponding to each step of a sequence program on a screen, the screen components related to the state of the controlled object may be distributed across multiple screens, making it impossible to view them all at once, and it may take time to transition from one screen to another. For example, when an abnormality occurs, it is preferable to be able to check the various abnormal lamps that make up the screen components on one screen with minimal effort, and it is desirable to eliminate the time required for screen transitions.

[0011] Therefore, there is a need for a screen information generation device, a machine learning device, and a screen information generation method that eliminate the hassle of screen transitions and generate screen information that configures a screen on which screen components related to the status information of the object to be controlled can be checked with minimal effort.

[0012] A first representative aspect of the present disclosure is a screen information generating device that generates screen information that constitutes a screen of a display device, comprising: a signal explanatory information acquisition unit that acquires multiple pieces of signal explanatory information including at least one of the addresses of signals included in a sequence program, comments added to the signals, the structure of the sequence program, and screen component types; a status information acquisition unit that acquires status information of a control object controlled by the sequence program from the control object; a screen component group classification unit that classifies the multiple pieces of signal explanatory information into multiple screen component groups associated with the status information based on the multiple pieces of signal explanatory information and the status information; and a screen information output unit that generates at least one screen component based on at least one piece of signal explanatory information in each screen component group, and outputs the generated screen information of each screen component group including at least one screen component to the display device.

[0013] A representative second aspect of the present disclosure is a machine learning device including: an input data acquisition unit that acquires, as input data, first signal explanatory information including at least one of an address of a signal included in a first sequence program, a comment added to the signal, a structure of the first sequence program, and a screen component type, and first state information of a control object controlled by the first sequence program; a similarity calculation unit that calculates a similarity between the first signal explanatory information and the first state information; and a learning unit that performs unsupervised learning using the input data and the similarity, and when inputting a plurality of second signal explanatory information included in a second sequence program, each including the same type of information as the first signal explanatory information, and second state information of a control object controlled by the second sequence program, generates a trained model that classifies the plurality of second signal explanatory information into a plurality of screen component groups associated with the second state information.

[0014] A third representative aspect of the present disclosure is a screen information generating method in which a computer as a screen information generating device that generates screen information that constitutes the screen of a display device executes the following processes: a process of acquiring multiple pieces of signal explanatory information, each piece including at least one of the addresses of signals included in a sequence program, comments added to the signals, the structure of the sequence program, and screen component types; a process of acquiring status information of a control object controlled by the sequence program from the control object; a process of classifying the multiple pieces of signal explanatory information into multiple screen component groups associated with the status information based on the multiple pieces of signal explanatory information and the status information; and a process of generating at least one screen component based on at least one piece of signal explanatory information in each screen component group, and outputting the generated screen information of each screen component group including at least one screen component to the display device.

[0015] 1 is a block diagram showing an example of a configuration of a screen information generation system including a screen information generation device according to a first embodiment of the present disclosure. FIG. 1 is a diagram showing an example of a sequence program using a ladder language. FIG. 2 is a diagram showing signal explanation information and machine status information. FIG. 3 is a diagram showing a classification result by a screen part group classification unit. FIG. 4 is a diagram showing screen information of two screens for "abnormality occurring" generated by a screen information output unit. FIG. 5 is a diagram showing screen information of one screen for "abnormality occurring" as another example generated by the screen information output unit. FIG. 6 is a diagram showing screen information of two screens for "processing in progress" generated by the screen information output unit. FIG. 7 is a diagram showing a case where a pop-up display is performed on one screen and a case where the pop-up display is closed. FIG. 8 is a flowchart showing an example of a screen information generation method. FIG. 9 is a block diagram showing an example of a configuration of a machine learning device according to a second embodiment of the present disclosure. FIG. 10 is a diagram showing a configuration of a screen information generation system during machine learning, including the machine learning device according to the second embodiment. FIG. 11 is a diagram showing a configuration of a screen information generation system after machine learning, including the machine learning device according to the second embodiment. FIG. 12 is a diagram showing the operation of a similarity calculation unit of the machine learning device according to the second embodiment. FIG. 13 is a diagram showing similarities calculated by the similarity calculation unit. FIG. 14 is a flowchart showing a machine learning method by the machine learning device according to the second embodiment. Fig. 10 is a block diagram showing a configuration example of a screen information generating system including a screen information generating device of a third embodiment of the present disclosure. Fig. 11 is a diagram showing an operation in which an output order determination unit of the screen information generating device of the third embodiment determines the output order of a group of screen parts based on an average of similarities. Fig. 12 is a block diagram showing a configuration example of a screen information generating system including a screen information generating device of a fourth embodiment of the present disclosure. Fig. 13 is a diagram showing an operation in which an output order determination unit of the screen information generating device of the fourth embodiment determines the output order of a group of screen parts based on an operation history.

[0016] An embodiment of the present disclosure will be described in detail below with reference to the accompanying drawings. In the following description, "signal description information" refers to at least one of a signal address, a comment added to a signal, a sequence program contact, a structure such as a coil type, and a screen component type such as a button or a lamp. "Status information" refers to information indicating the status of a control target, such as "machining in progress," "abnormality occurring," or "maintenance in progress." A "screen component" refers to one of multiple screen components of a screen. For example, screen component E1 shown in FIG. 5 indicates a comment "coolant shortage abnormality" and represents a lamp, which is a screen component type, as a rounded rectangle. Furthermore, screen component E2 shown in FIG. 5 indicates a comment "abnormality reset" and represents a button, which is a screen component type, as a flat panel.

[0017] (First embodiment) Fig. 1 is a block diagram showing an example configuration of a screen information generating system including a screen information generating device according to a first embodiment of the present disclosure. As shown in Fig. 1, the screen information generating system 10 includes a PLC (Programmable Logic Controller) 100, a control target 200, a screen information generating device 300, and a display device 400. The screen information generating device 300 may be provided in the PLC 100 or the control target 200. The PLC 100 may be connected to a numerical control device and control a peripheral device that is a control target in response to a request from the numerical control device. The PLC 100 may be provided within the numerical control device. The display device 400 may be provided as a display unit within the screen information generating device 300, the PLC 100, or the numerical control device.

[0018] The PLC 100 controls the control target 200 by executing a sequence program. The PLC 100 includes a storage unit 101 that stores the sequence program and a program execution unit 102 that executes the sequence program. The sequence program used by the PLC 100 to perform sequence control is created using a ladder language, and a sequence program created using a ladder language is called a ladder program. FIG. 2 is a diagram showing an example of a sequence program using a ladder language. The sequence program used by the PLC 100 to perform sequence control is not limited to the ladder language and may be created using other languages. For example, the sequence program may be created using a structured text (ST) language, a sequential function chart (SFC) language, a function block diagram (FBD) language, or an instruction list (IL) language. The sequence program includes, as signal explanation information, at least one of the following: the address of the signal, a comment added to the signal, and the structure of the sequence program, such as contact points and coil types.

[0019] In addition to the sequence program, the storage unit 101 stores a correspondence table including signal addresses, comments added to signals, and screen component types such as lamps and buttons associated with at least one of the structures of the sequence program, such as contacts and coil types. The screen component type is one piece of signal description information. Although FIG. 3 does not show the structures of the sequence program, such as contacts and coil types, the correspondence table may include this structure. In the signals of FIG. 3, "R" indicates an internal relay signal, "K" indicates a hold-type relay signal, and "G" indicates a signal from the PLC to the CNC.

[0020] The control target 200 is a machine such as a machine tool or a robot, or a peripheral device of a machine, and the control target 200 is controlled by the program execution unit 102 executing a sequence program.

[0021] The screen information generating device 300 generates screen information that constitutes the screen of the display device 400. The screen information generating device 300 acquires signal explanatory information and machine status information, classifies the signal explanatory information into a plurality of screen component groups associated with the machine status information based on the signal explanatory information and the machine status information, and outputs the screen component groups. As shown in FIG. 1 , the screen information generating device 300 includes a signal explanatory information acquiring unit 301, a machine status information acquiring unit 302, a screen component group classifying unit 303, and a screen information output unit 304.

[0022] The signal explanatory information acquisition unit 301 acquires signal addresses, comments added to signals, and structures such as contacts of the sequence program and coil types from the sequence program stored in the storage unit 101. The signal explanatory information acquisition unit 301 acquires screen component types from the storage unit 101. The acquired signal addresses, comments added to signals, contacts of the sequence program, structures such as coil types, and screen component types become signal explanatory information. The signal explanatory information may be at least one of the signal addresses, comments added to signals, contacts of the sequence program, structures such as coil types, and screen component types. The signal explanatory information acquisition unit 301 acquires multiple pieces of signal explanatory information.

[0023] The machine state information acquisition unit 302 acquires machine state information from the control target 200. The machine state information is, for example, "machining in progress", "abnormality occurring", "maintenance in progress", etc., as shown in FIG. 3 , and corresponds to the state information. The machine state information acquisition unit 302 corresponds to the state information acquisition unit. The machine state information acquisition unit 302 may acquire the machine state information based on signal description information such as a signal address.

[0024] The screen component group classification unit 303 classifies the signal description information into screen component groups related to the machine status information based on the signal description information and the machine status information. FIG. 4 shows the classification results. For example, as shown in FIG. 4 , when the machine status information is "abnormality occurring," the screen component group classification unit 303 classifies the comments related to the abnormality occurrence, such as "coolant shortage abnormality," "workpiece floating abnormality," and "abnormal reset," into screen component group G1 so that they are arranged on the same screen. The screen component group classification unit 303 classifies the comments "automatic operation in progress" and "paused" into screen component group G2. The screen component group classification unit 303 classifies the comments "power saving ON," "automatic operation started," and "emergency stop" into screen component group G3. FIG. 4 also shows the classification of screen component groups G1, G2, and G3 when the machine status information is "machining in progress" and "maintenance in progress." The screen component group classification unit 303 may insert a message indicating the machine state, which is a screen component, into the screen component group G1, the screen component group G2, and the screen component group G3.

[0025] One example of a method for classifying screen component groups by the screen component group classification unit 303 is to calculate the similarity between signal explanation information and machine status information, and then divide the screen component groups into screen component groups G1, G2, and G3 based on the similarity. The similarity is calculated by vectorizing the natural language comments to obtain sets of multiple numerical values, vectorizing the natural language machine status information to obtain sets of multiple numerical values, and then using the two sets of values ​​to calculate the cosine distance as the similarity. The cosine distance, also known as cosine similarity, is a measure of how similar two vectors are, and is the cosine value of the angle between the two vectors. Details of the method for calculating the similarity will be described in the second embodiment below.

[0026] Another example of a method for classifying screen component groups by the screen component group classification unit 303 is a method for dividing the screen component groups into screen component groups G1, G2, and G3 based on an operation history that stores the operation contents that a user performed while viewing screen components on the screen of the display device 400. The method for dividing the screen component groups G1, G2, and G3 based on the operation history can be the same as the method for determining the output order of screen component groups based on the operation history in the fourth embodiment described later.

[0027] The screen information output unit 304 generates at least one screen component using at least one signal explanation information, such as a comment, for each screen component group classified by the screen component group classification unit 303, and outputs one or more screen component groups including the generated at least one screen component as screen information. As shown in FIG. 5 , the screen information output unit 304 generates screen information for a screen S1 on which a screen component group G1 is arranged when the machine status information indicates "abnormality occurring." Also, as shown in FIG. 5 , the screen information output unit 304 generates screen information for a screen S2 on which screen component groups G2 and G3 are arranged when the machine status information indicates "abnormality occurring." A screen component is one of multiple screen components of a screen. For example, as shown in FIG. 5 , screen component E1 represents a lamp indicating a "low coolant abnormality" and screen component E2 represents a button indicating a "reset abnormality." Screen component E1 and screen component E2 represent a comment and a screen component type, respectively. The rounded rectangle of screen component E1 represents a lamp, and the flat shape of screen component E2 represents a button.

[0028] As described above, the screen component group classification unit 303 may insert messages indicating the machine status as screen components into the screen component group G1, screen component group G2, and screen component group G3. In this case, the screen information output unit 304 outputs the screen information of the screens S1 and S2 in which the messages have been inserted into the screen component group G1, screen component group G2, and screen component group G3 to the display device 400. Fig. 6 is a diagram showing the screen information of the screen S1 in which a message with the message content "An abnormality has occurred" has been inserted as a screen component into the screen component group G1. The rectangle of the screen component E3 represents the message.

[0029] The screen information output unit 304 outputs the screen information of the screen S1 and the screen S2 when the machine status information is "abnormal" to the display device 400. Similarly, the screen information output unit 304 generates screen information of the screen S1 on which the screen component group G1 is arranged and screen information of the screen S2 on which the screen component groups G2 and G3 are arranged in the cases of "processing" and "maintenance", and outputs these to the display device 400. Fig. 7 is a diagram showing the screen information of the screen S1 and the screen information of the screen S2 in the case of "processing".

[0030] The display device 400 is, for example, a liquid crystal display device. The display device 400 may be a display / input device for displaying images and inputting information, such as a liquid crystal display device with a touch panel. The display device 400 receives screen information for the screen S1 on which the screen component group G1 is arranged and screen information for the screen S2 on which the screen component groups G2 and G3 are arranged, output from the screen information output unit 304, and first displays the screen S1 on the display screen. When the user instructs switching from the screen S1 to the screen S2, the display device 400 displays the screen S2.

[0031] The display device 400 can switch screen information on a screen S1 relating to a certain machine state to screen information on a screen S1 relating to another machine state by pop-up display, depending on the machine state. As shown in Fig. 8, image information for the "machining in progress" state is displayed on the screen S1, and when an abnormality occurs, image information for the "abnormality occurring" state is displayed as a pop-up display on a pop-up screen P1 that covers the screen components of the screen S1. When the user presses the close button P11 in the upper right corner of the pop-up screen P1, the screen returns to the screen S1 on which the image information for the "machining in progress" state is displayed.

[0032] 9 is a flowchart showing an example of a screen information generating method. In the following description, an example will be described in which the screen information generating method of the present disclosure is executed by the screen information generating device 300, but the screen information generating method of the present disclosure can also be executed by a configuration other than the screen information generating device 300.

[0033] In step ST11, the signal description information acquisition unit 301 acquires signal description information from the memory unit 101, the signal description information including at least one of the signal address, comments added to the signal, contacts of the sequence program, structure such as coil type, and screen component type.

[0034] In step ST12 , the machine state information acquisition unit 302 acquires machine state information from the control target 200 .

[0035] In step ST13, the screen component group classification unit 303 classifies the signal explanatory information into a plurality of screen component groups related to the machine state information based on the signal explanatory information and the machine state information.

[0036] In step ST14, the screen information output unit 304 generates at least one screen component using at least one signal explanation information, such as a comment, for each screen component group classified by the screen component group classification unit 303, and outputs one or more generated screen component groups each including at least one screen component as screen information.

[0037] According to the screen information generating device and screen information generating method of this embodiment described above, it is possible to eliminate the need for screen transitions and generate screen information that constitutes a screen on which screen components related to machine status information can be checked with minimal effort.

[0038] Second Embodiment In the first embodiment, an example was described in which the screen component group classification unit 303 classifies screen components into a screen component group G1, a screen component group G2, and a screen component group G3 based on similarity. In the present embodiment, a machine learning device that creates a trained model that can be used by the screen component group classification unit 303 and that classifies screen components into a screen component group G1, a screen component group G2, and a screen component group G3 based on similarity will be described. FIG. 10 is a block diagram illustrating an example configuration of a machine learning device according to a second embodiment of the present disclosure. FIG. 11 is a diagram illustrating a configuration of a screen information generation system 10A during machine learning. FIG. 12 is a diagram illustrating a configuration of the screen information generation system 10A after machine learning. As shown in FIG. 11 , the machine learning device 500 is provided outside the screen information generation device 300, but may also be provided within the screen information generation device 300.

[0039] 10 and 11 , during machine learning, the machine learning device 500 is connected to the signal explanatory information acquisition unit 301, the machine state information acquisition unit 302, and the screen component group classification unit 303. The dashed lines in Fig. 11 indicate non-connection. The machine learning device 500 generates a learning model to be used in the screen component group classification unit 303 based on the signal explanatory information output from the signal explanatory information acquisition unit 301 and the machine state information output from the machine state information acquisition unit 302, and outputs the learning model to the screen component group classification unit 303.

[0040] As shown in Fig. 12, after machine learning, the machine learning device 500 is separated from the signal explanation information acquisition unit 301, the machine state information acquisition unit 302, and the screen component group classification unit 303, and the screen information generation system 10A has the same configuration as the screen information generation system 10 shown in Fig. 1. The dashed lines in Fig. 12 indicate that they are not connected.

[0041] As shown in FIG. 10, the machine learning device 500 includes an input data acquisition unit 501, a similarity calculation unit 502, a learning unit 503, and an output unit 504.

[0042] The input data acquisition unit 501 acquires signal explanatory information from the signal explanatory information acquisition unit 301 and acquires machine state information from the machine state information acquisition unit 302, and outputs the acquired information to the similarity calculation unit 502. The signal explanatory information and machine state information serve as input data. The similarity calculation unit 502 calculates the similarity between the signal explanatory information and the machine state information.

[0043] The learning unit 503 performs unsupervised learning using the signal explanation information, machine status information, and the similarity calculated by the similarity calculation unit 502, and generates a trained model that classifies the signal explanation information included in the sequence program into screen component groups related to the machine status information based on the magnitude of the similarity. When generating n screen component groups, for example, the similarity between the signal explanation information and the machine status information can be input and the n screen component groups can be generated using the k-means method. The output unit 504 outputs the trained model to the screen component group classification unit 303.

[0044] The operation of the similarity calculation unit 502 will be described with reference to FIG. 13 . The similarity calculation unit 502 vectorizes the natural language comments using a vectorization model such as word2vec to obtain a set of multiple numerical values ​​(hereinafter referred to as a first set). Similarly, the similarity calculation unit 502 vectorizes the natural language machine status information using a vectorization model such as word2vec to obtain a set of multiple numerical values ​​(hereinafter referred to as a second set). The similarity calculation unit 502 then calculates the cosine distance as the similarity using the first set and the second set. The cosine distance, also known as cosine similarity, is a measure of how similar two vectors are, and is the cosine value of the angle between the two vectors.

[0045] 13, the similarity calculation unit 502 represents the comment "coolant shortage abnormality" as a set of multiple numerical values ​​(0.1, 0.4, ..., -0.2) (first set) using a vectorization model such as word2vec. Also, the machine status information "abnormality occurring" as a set of multiple numerical values ​​(0.2, 0.5, ..., -0.1) (second set) using a vectorization model such as word2vec.

[0046] The corpus used for natural language processing of comments and machine status information in similarity calculation can be a general-purpose corpus or a special-purpose corpus. The general-purpose corpus can be a corpus that is widely used in the natural language field, such as the Balanced Corpus of Contemporary Written Japanese (BCCWJ). The special-purpose corpus can be created from manufacturing-related papers and trade journals, or specifications provided by CNC manufacturers and machine manufacturers.

[0047] The similarity calculation unit 502 obtains a first set for comments other than "insufficient coolant abnormality" by the method shown in Fig. 13, obtains a second set for the machine status information "abnormality occurring", and calculates the cosine distance as the similarity using the obtained first set and second set. Fig. 14 is a diagram showing the similarity calculated by the similarity calculation unit 502.

[0048] 15 is a flowchart showing the operation of the machine learning device. Steps ST21 and ST22 indicate the start and end of a loop process for calculating the similarity between the signal explanation information and the machine state information from the beginning to the end of a program. In the loop process between steps ST21 and ST22, a program is input, and steps ST211 and ST212 are repeated from the beginning to the end of the program.

[0049] In step ST211, the input data acquisition unit 501 acquires signal explanatory information from the signal explanatory information acquisition unit 301, and acquires machine state information from the machine state information acquisition unit 302. In step ST212, the similarity calculation unit 502 calculates the similarity between the signal explanatory information and the machine state information.

[0050] In step ST23, the learning unit 503 performs unsupervised learning using the signal explanation information, the machine status information, and the similarity calculated by the similarity calculation unit 502, and generates a trained model that classifies the signal explanation information included in the sequence program into screen component groups related to the machine status information. In step ST24, the output unit 504 outputs the trained model to the screen component group classification unit 303.

[0051] According to this embodiment, a trained model for classifying each signal into a plurality of screen component groups associated with the machine status information can be obtained by machine learning based on the signal description information and the machine status information. This reduces the effort required to configure the screen component group classification unit to classify the signals into a plurality of screen component groups.

[0052] 16 is a block diagram showing a configuration example of a screen information generating system including a screen information generating device according to a third embodiment of the present disclosure. The screen information generating system 10B shown in Fig. 16 includes a screen information generating device 300B in which the screen information generating device 300 is further equipped with an output order determination unit 305, as compared with the screen information generating system 10 shown in Fig. 1.

[0053] The output order determination unit 305 calculates the similarity between the comments included in the signal explanation information and the machine status information. The calculation of the similarity is the same as the method shown in FIG. 13 . The output order determination unit 305 determines the output order of the screen component groups classified by the screen component group classification unit 303 based on the calculated statistical quantities of similarity. The statistical quantities are, for example, the mean value, variance, standard deviation, median, and mode. The following explanation will be given for the case where the mean value is used as the statistical quantity.

[0054] 17 is a diagram showing the operation of the output order determination unit 305 in determining the output order of screen component groups based on the average similarity. As shown in Fig. 17, the output order determination unit 305 determines the average similarity of the screen component group G1 to be 0.89, the average similarity of the screen component group G2 to be 0.21, and the average similarity of the screen component group G3 to be 0.46. Since (average similarity of the screen component group G1) > (average similarity of the screen component group G3) > (average similarity of the screen component group G2), the output order of the screen information of the screen component groups is the screen component group G1, the screen component group G3, and the screen component group G2.

[0055] According to this embodiment, in addition to the effects of the first embodiment, the output order of the screen information is determined based on the similarity between the signal explanation information and the machine status information, so that a group of screen components with high similarity is preferentially displayed on the screen, thereby reducing the time and effort required for screen transitions to display a group of screen components with high similarity.

[0056] (Fourth embodiment) Fig. 18 is a block diagram showing an example configuration of a screen information generating system including a screen information generating device according to a fourth embodiment of the present disclosure. A screen information generating system 10C shown in Fig. 18 includes an operation device 600 compared to the screen information generating system 10B shown in Fig. 16. In addition, the screen information generating device 300B including the output order determination unit 305 is replaced with a screen information generating device 300C including an output order determination unit 305C.

[0057] The operation device 600 outputs to the output order determination unit 305C the operation content operated by the user while viewing screen components on the screen of the display device 400. The operation device 600 and the display device 400 may be integrated, and may be configured as, for example, a liquid crystal display device with a touch panel. The output order determination unit 305C stores the operation content output from the operation device 600 as an operation history, and determines the output order of the screen components based on the operation history corresponding to the machine status information.

[0058] 19 is a diagram showing the operation of the output order determination unit 305C to determine the output order of the screen component groups based on the operation history. As shown in Fig. 19, in the case of "an abnormality occurring," the output order determination unit 305C calculates the percentage of the time that the user checked various abnormality lamps (which become the respective screen components) on the screen of the display device 400 and operated the operation device 600, the percentage of the time that the user checked various operation status lamps (which become the respective screen components) on the screen of the display device 400 and operated the operation device 600, and the percentage of the time that the user checked the power saving ON button (which becomes the screen component) on the screen of the display device 400 and operated the operation device 600. The magnitude of these proportions is (proportion of operations to check various abnormality lamps) > (proportion of operations to check various operating status lamps) > (proportion of operations on the power saving ON button), so the output order of the screen information of the screen component groups is screen information of screen component group G1, screen information of screen component group G3, and screen information of screen component group G2, and as shown in Figure 5, screen component group G1 is displayed on screen S1, which is displayed first, and screen component group G2 and screen component group G3 are displayed on screen S2, which is displayed next.

[0059] According to this embodiment, in addition to the effects of the first embodiment, the output order of screen information is determined based on the percentage of operations, so that screen parts with a large percentage of operations are preferentially displayed on the screen, thereby reducing the time and effort required for screen transitions to display screen parts with a large percentage of operations.

[0060] In order to realize the functional blocks included in the screen information generating device in each embodiment described above, the screen information generating device can be realized by hardware, software, or a combination of these. The screen information generating method can also be realized by hardware, software, or a combination of these. Here, "realized by software" means that the method is realized by a computer reading and executing a program.

[0061] To realize the components included in the screen information generating device by software or a combination thereof, the screen information generating device includes a processing unit such as a CPU (Central Processing Unit). The processing unit functions as an execution unit. The screen information generating device also includes an auxiliary storage device such as an HDD (Hard Disk Drive) that stores various control programs such as application software or an OS (Operating System), and a main storage device such as a RAM (Random Access Memory) that stores data temporarily required for the processing unit to execute the programs.

[0062] The screen information generating device then performs arithmetic processing based on the application software or OS while the arithmetic processing unit reads the application software or OS from the auxiliary storage device and deploys the loaded application software or OS in the main storage device. Furthermore, based on the results of this calculation, the screen information generating device controls various pieces of hardware. This realizes the functional blocks of this embodiment. The screen information generating method can also be realized with a configuration similar to that of the screen information generating device.

[0063] The components included in the screen information generating device can be realized by hardware including electronic circuits, etc. When the screen information generating device is configured by hardware, some or all of the functions of the components included in the screen information generating device can be configured by an integrated circuit (IC), such as an ASIC (Application Specific Integrated Circuit), a gate array, an FPGA (Field Programmable Gate Array), or a CPLD (Complex Programmable Logic Device).

[0064] The program can be stored and supplied to a computer using various types of non-transitory computer-readable media. Non-transitory computer-readable media include various types of tangible storage media. Examples of non-transitory computer-readable media include magnetic recording media (e.g., hard disk drives), magneto-optical recording media (e.g., magneto-optical disks), CD-ROMs (Read Only Memory), CD-Rs, CD-R / Ws, and semiconductor memories (e.g., mask ROMs, PROMs (Programmable ROMs), EPROMs (Erasable PROMs), flash ROMs, and RAMs (Random Access Memory)). The program may also be supplied to the computer by various types of transient computer readable media.

[0065] According to the screen information generation device, machine learning device, and screen information generation method of the present disclosure, including the embodiments described above, it is possible to eliminate the need for screen transitions and generate screen information that configures a screen on which screen components related to the status information of the object to be controlled can be checked with minimal effort.

[0066] Although the above-described embodiments are preferred embodiments of the present invention, the scope of the present invention is not limited to the above-described embodiments, and the present invention can be implemented in various modified forms within the scope that does not deviate from the gist of the present invention.

[0067] The following supplementary note is further disclosed regarding the above embodiment: (Supplementary Note 1) A screen information generating device (300) that generates screen information constituting a screen of a display device (400), comprising: a signal explanatory information acquiring unit (301) that acquires a plurality of pieces of signal explanatory information, each piece including at least one of an address of a signal included in a sequence program, a comment added to the signal, a structure of the sequence program, and a screen component type; a state information acquiring unit (302) that acquires, from a control object controlled by the sequence program, state information of the control object; a screen component group classification unit (303) that classifies the plurality of pieces of signal explanatory information into a plurality of screen component groups associated with the state information, based on the plurality of pieces of signal explanatory information and the state information; and a screen information output unit (304) that generates at least one screen component based on at least one piece of signal explanatory information in each screen component group, and outputs the generated screen information of each screen component group including at least one screen component to the display device.

[0068] (Supplementary Note 2) The screen information generating device according to Supplementary Note 1, further comprising: an output order determination unit (305) that determines an output order of the screen information of each of the screen part groups classified by the screen part group classification unit (303) based on statistics of similarity between at least one of the signal address, the comment, and the structure and the state information.

[0069] (Supplementary Note 3) The screen information generating device according to Supplementary Note 1, further comprising: an output order determination unit (305C) that determines an output order of the screen information of each of the screen component groups classified by the screen component group classification unit (303) based on an operation history corresponding to the state information.

[0070] (Supplementary Note 4) A machine learning device (500) comprising: an input data acquisition unit (501) that acquires, as input data, first signal explanatory information including at least one of an address of a signal included in a first sequence program, a comment added to the signal, a structure of the first sequence program, and a screen component type, and first state information of a control object controlled by the first sequence program; a similarity calculation unit (502) that calculates a similarity between the first signal explanatory information and the first state information; and a learning unit (503) that performs unsupervised learning using the input data and the similarity, and when inputting a plurality of second signal explanatory information included in a second sequence program, each of which includes information similar to the first signal explanatory information, and second state information of a control object controlled by the second sequence program, generates a trained model that classifies the plurality of second signal explanatory information into a plurality of screen component groups associated with the second state information.

[0071] (Appendix 5) A screen information generating device described in any one of Appendices 1 to 3, wherein the screen part group classification unit (303) classifies the plurality of pieces of signal explanation information into a plurality of screen part groups associated with the status information using a trained model generated by the machine learning device (500) described in Appendices 4.

[0072] (Supplementary Note 6) A screen information generating method in which a computer as a screen information generating device (300) that generates screen information that constitutes the screen of a display device executes the following processes: a process of acquiring multiple pieces of signal explanatory information including at least one of the addresses of signals included in a sequence program, comments added to the signals, the structure of the sequence program, and screen component types; a process of acquiring status information of a control object controlled by the sequence program from the control object; a process of classifying the multiple pieces of signal explanatory information into multiple screen component groups associated with the status information based on the multiple pieces of signal explanatory information and the status information; and a process of generating at least one screen component based on at least one piece of signal explanatory information in each screen component group, and outputting the generated screen information of each screen component group including at least one screen component to the display device.

[0073] 10, 10A, 10B, 10C Screen information generating system 100 PLC 101 Memory unit 102 Program execution unit 200 Control object 300, 300B, 300C Screen information generating device 301 Signal explanation information acquisition unit 302 Machine state information acquisition unit 303 Screen component group classification unit 304 Screen information output unit 305, 305C Output order determination unit 400 Display device 500 Machine learning device 600 Operation device

Claims

1. A screen information generating device that generates screen information that constitutes the screen of a display device, comprising: a signal explanation information acquisition unit that acquires multiple pieces of signal explanation information including at least one of the addresses of signals included in a sequence program, comments added to the signals, the structure of the sequence program, and screen component types; a status information acquisition unit that acquires status information of a control object controlled by the sequence program from the control object; a screen component group classification unit that classifies the multiple pieces of signal explanation information into multiple screen component groups related to the status information based on the multiple pieces of signal explanation information and the status information; and a screen information output unit that generates at least one screen component based on at least one piece of signal explanation information in each screen component group, and outputs the generated screen information of each screen component group including at least one screen component to the display device.

2. A screen information generating device as described in claim 1, further comprising an output order determination unit that determines the output order of the screen information of each screen part group classified by the screen part group classification unit based on statistics of the similarity between at least one of the signal address, the comment, and the structure and the state information.

3. A screen information generating device as described in claim 1, further comprising an output order determination unit that determines the output order of the screen information of each screen component group classified by the screen component group classification unit based on the operation history corresponding to the state information.

4. A machine learning device comprising: an input data acquisition unit that acquires as input data first signal explanatory information including at least one of the addresses of signals included in a first sequence program, comments added to the signals, the structure of the first sequence program, and screen component types, and first state information of a control object controlled by the first sequence program; a similarity calculation unit that calculates the similarity between the first signal explanatory information and the first state information; and a learning unit that performs unsupervised learning using the input data and the similarity, and when inputting a plurality of second signal explanatory information included in a second sequence program, each of which includes the same type of information as the first signal explanatory information, and second state information of a control object controlled by the second sequence program, generates a trained model that classifies the plurality of second signal explanatory information into a plurality of screen component groups associated with the second state information.

5. A screen information generating device as described in any one of claims 1 to 3, wherein the screen part group classification unit classifies the plurality of pieces of signal explanation information into a plurality of screen part groups associated with the status information using a trained model generated by the machine learning device as described in claim 4.

6. A screen information generation method in which a computer as a screen information generation device that generates screen information that constitutes the screen of a display device executes the following processes: a process of acquiring multiple pieces of signal explanatory information including at least one of the addresses of signals included in a sequence program, comments added to the signals, the structure of the sequence program, and screen component types; a process of acquiring status information of a control object controlled by the sequence program from the control object; a process of classifying the multiple pieces of signal explanatory information into multiple screen component groups related to the status information based on the multiple pieces of signal explanatory information and the status information; and a process of generating at least one screen component based on at least one piece of signal explanatory information in each screen component group, and outputting the generated screen information of each screen component group including at least one screen component to the display device.

Images