KR20220096342A - Application recommendation methods for machine tools and the systems used there - Google Patents

Abstract

The present invention relates to a method for recommending an application for a machine tool, and a system used therefor and, more specifically, to a method for recommending an application for a machine tool, capable of analyzing characteristics between machine tool NC machining programs through a cooperative-filtering recommendation method, and recommending an application suitable for a new NC machining program made by a user based on signal data thereof, and a system used therefor. The method for recommending an application for a machine tool, includes the flowing steps of: (a) generating signal data of a plurality of machine tool NC machining programs into a matrix; (b) analyzing and processing the acquired signal data; (c) calculating similarity between the NC machining programs based on the analyzed and processed signal data; (d) calculating an application ranking matrix by applying the similarity between the NC machining programs as a weighted value; and (e) recommending an application for a specific NC machining program based on the calculated application ranking matrix.

Description

Application recommendation methods for machine tools and the systems used there

The present invention relates to a method for recommending an application for a machine tool and a system used therefor, and in particular, analyzes characteristics between machine tool NC machining programs using a collaborative filtering recommendation method, and creates a user-created method based on the signal data. It relates to an application recommendation method for a machine tool that can recommend an application suitable for a new NC machining program, and a system used therefor.

In general, in order to process a workpiece using a machine tool, an NC machining program that automatically controls the tool path, cutting conditions, and peripheral devices using a control device such as a computer is required. It is created using the part program generation system.

And, when creating a machining program using the machine tool machining program generation system, information on the shape of the workpiece must be input to create a machining path.

The shape of the workpiece is provided in the form of a drawing file created by a CAD program, and the operator loads the drawing file from the machining program generation system and inputs the shape of the workpiece by selecting the target line segment.

Here, the method for selecting the target line segment includes a "direct selection method" in which an operator directly selects all line segments, and automatically selects line segments connected to a newly added line segment starting from the currently selected line segment by the operator, but a branch point There is a "branch point automatic tracking selection method" that asks the operator to select a branch line segment when it encounters .

However, since there is no method of recommending an application suitable for an NC machining program in the prior art, there is a problem in that workability is deteriorated because it is necessary to find and connect (connect) an application suitable for a new NC machining program every time.

In addition, in the prior art, in order to find and connect an application suitable for the NC machining program, many manipulations are required by the operator, so that when the machining is complicated, the working time increases and the possibility of erroneous manipulation increases.

Publication No. 10-2019-0105957

Accordingly, the present invention has been devised to solve the above problems, and an object of the present invention is to automatically recommend an application suitable for a machine tool NC machining program created by a user, thereby improving workability. It is to provide a method and a system used therefor.

In order to achieve the above object, the present invention provides an application recommendation method for a machine tool, comprising the steps of: (a) generating a plurality of machine tool NC machining program signal data in a matrix; (b) analyzing and processing the acquired signal data; (c) obtaining a degree of similarity between each NC machining program based on the analyzed and processed signal data; (d) calculating the degree of similarity between the respective NC machining programs as a weight; (e) recommending an application for a specific NC machining program based on the calculated application ranking matrix; is made including

The present invention is an application recommendation method for a machine tool, wherein the signal data is an M signal generated when an M code command is a program code representing an auxiliary function of a machine tool, and a G code is a program code representing a preparation function of the NC machine tool. It is characterized in that it can be used as any one of a G signal generated when a command of , and a T signal generated when a command to use a tool is commanded.

The present invention is characterized in that in the method for recommending an application for a machine tool, the signal data in the step (a) is the number of uses.

In the present invention, in the method for recommending an application for a machine tool, among the signal values displayed in a matrix (matrix) in step (a), the values for which the correlation value between the NC machining program and the signal is not determined are subjected to matrix factorization. It is characterized in that it is predicted through

In the present invention, in the application recommendation method for a machine tool, each of the NC machining programs 1,2,3,4 through the Pearson correlation coefficient of each of the NC machining programs 1,2,3,4,5 in step (c) It is characterized by finding the similarity between ,5.

The present invention is characterized in that in the method for recommending an application for a machine tool, the calculation in step (d) multiplies the value of the number of executions of the application by the value of the similarity by a weight.

In the present invention, in the method for recommending an application for a machine tool, the recommendation in step (e) is based on the similarity value obtained as a weight, the number of times each application is executed is multiplied by the weight, and the sum of the values is the highest. characterized in that

In addition, the present invention provides a system used in an application recommendation method for a machine tool, comprising: a CNC machine tool; a data collecting unit serving to collect and transmit data that is an NC machining program signal generated from a computer numerical control (CNC) and programmable logic control device of the machine tool through an application programming interface (API); an analysis and storage unit for analyzing, storing, and storing data collected and transmitted from the data collection unit; an application recommendation unit for recommending an application based on the data acquired through the data collection unit and the analysis and storage unit; It is characterized in that it comprises a.

The present invention is a system used in a method for recommending an application for a machine tool, wherein the machine tool further comprises an application for recommending a machine tool application and an HMI device (Humanmachine Interface).

The present invention is a system used for an application recommendation method for a machine tool, wherein the application programming interface (API) of the data collection unit is NC of various machine tool manufacturers such as FANUC, Siemens, HEIDENHAIN. It is characterized in that it is configured to enable communication support.

The present invention relates to a system used in a method for recommending an application for a machine tool, wherein the analysis and storage unit analyzes the collected data in real time and stores time series data, and an analysis module for storing the time series data classifies the data to obtain necessary information from the collected data and a data processing module for processing, a data storage module for storing classified and processed data, and a system management module for managing a server system.

The present invention is a system used for an application recommendation method for a machine tool, wherein the HMI device includes a display unit for indicating an NC machining program stored inside the machine tool, an input unit for selecting the NC machining program, and an internal storage Including a display unit indicating the application being processed, an input unit for application execution, a display unit indicating the storage path of the program currently designated as the main NC part program, and a display part showing the contents of the program currently designated as the main NC part program characterized by being made.

As described above, the present invention automatically recommends an application suitable for a new NC machining program every time, so that the connection path for using the application is simple, thereby improving the user's work efficiency.

In addition, the present invention analyzes characteristics between NC machining programs through deep learning, and automatically connects applications suitable for NC machining programs created by users based on the data, thereby reducing work time and achieving quality improvement. It works.

In addition, since the application is recommended by applying an algorithm inside the system based on the program information written by the user, there is an effect that the application can be used more easily without a separate connection process for using the application.

1 is a flowchart for an application recommendation method for a machine tool according to a preferred embodiment of the present invention.
2 is a block diagram showing an embodiment of a system for application recommendation for a machine tool according to the present invention.
3 is a block diagram schematically illustrating a state in which signal data is collected through a data collection unit according to the present invention.
4 is a diagram schematically showing the configuration of an HMI device according to the present invention.

Hereinafter, a preferred embodiment according to the present invention will be described in more detail based on the accompanying drawings.

Here, in all the drawings below, components having the same function are omitted by using the same reference numerals, and repeated descriptions are omitted, and the terms to be described below are defined in consideration of the functions in the present invention, which have their own and commonly used meanings. should be interpreted as

As shown in Figure 1, the method for recommending an application for a machine tool according to the present invention includes (a) generating a plurality of machine tool NC machining program signal data in a matrix (S10); (b) analyzing and processing the acquired signal data (S20); (c) obtaining a degree of similarity between each NC machining program based on the analyzed and processed signal data (S30); (d) calculating by applying the degree of similarity between the respective NC machining programs as a weight (S40); (e) recommending an application for a specific NC machining program based on the calculated application ranking matrix (S50); is divided into

First, the system capable of implementing the application recommendation method for a machine tool according to the present invention, as shown in FIGS. A CNC machine tool (computerized numerically controlled machine tool) 100 that converts NC data, holds it in a pulse signalized state, and processes a workpiece as needed, and a computer numerical control (CNC) and program of the machine tool 100 A data collection unit 200 that collects and transmits data that is an NC processing program signal generated from a programmable logic controller through an application programming interface (API), and the data collection An analysis and storage unit 300 that analyzes, stores, and stores data collected and transmitted from the unit 200, and an application based on the data acquired through the data collection unit 200 and the analysis and storage unit 300 and a recommended application recommendation unit 400 .

The machine tool 100 is a programmable logic controller, also called a programmable controller, as a general-purpose control device used for automatic control and monitoring of the machine as a target for receiving data that is an NC machining program signal. is equipped with

Accordingly, the machine tool 100 generates and provides a machining program signal (data) from a computer numerical control and programmable logic control device when machining a workpiece.

And, the machine tool 100 according to the present invention further includes an application and HMI device (Humanmachine Interface) 110 for machine tool application recommendation as shown in FIG. 4 .

The HMI device 110 includes a display unit 111 indicating an NC machining program stored in the machine tool, an input unit 112 for selecting the NC machining program, and a display unit 113 indicating applications stored therein. ), an input unit 114 for application execution, a display unit 115 indicating the storage path of the program currently designated as the main NC machining program, and a display unit 116 indicating the contents of the program currently designated as the main NC machining program ) is included.

In addition, the application programming interface (API) of the data collection unit 200 is configured to support NC communication of various machine tool manufacturers such as FANUC, SIEMENS, and HEIDENHAIN. do.

In addition, the analysis and storage unit 300 includes an analysis module 310 that analyzes the collected data in real time and stores time series data, and a data processing unit that classifies and processes data to obtain necessary information from the collected data. processing) module 320 , a data storage module 330 for storing classified and processed data, and a system management module 340 for managing a server system.

Since the analysis and storage unit 300 is commonly applied in various forms as needed, a detailed description thereof will be omitted.

A method for recommending an application for a machine tool according to the present invention using the system 1000 according to the present invention having such a configuration is described as follows.

In the step (a), a plurality of machine tool NC machining program signal data is generated as a matrix. (S10)

That is, signal data generated by a computer numerical control unit and a programmable logic control device when machining a workpiece in the machine tool 100 is collected and acquired through an application programming interface (API) of the data collection unit 200 . to create a matrix.

Here, the signal data is generated when an M signal, which is a program code indicating an auxiliary function of the machine tool 100, is commanded, and a G code, which is a program code indicating a preparation function of the NC machine tool 110, is commanded. A G signal and a T signal generated when a tool is commanded to be used may be used.

In other words, signals such as T signal, M signal (M code), and G signal (PLC signal) are also related to specific processing, so they can be used as data for application recommendation.

As an example, in the present invention, an NC machining program is created for machining a workpiece in the machine tool 100 .

That is, the user (operator) of the machine tool 100 creates an NC machining program according to the workpiece to be machined.

A user (operator) of the machine tool 100 sees a production drawing of the workpiece to be machined, and through the input unit 112 for selecting the NC machining program constituting the HMI device 110 of the machine tool 100, the object to be machined By setting the coordinate value of , an NC machining program is created to process the workpiece easily.

In addition, all the created NC machining programs are displayed through the display unit 111 indicating the NC machining programs of the HMI device 110 .

The storage path of the program currently designated as the main NC machining program is displayed by the display unit 115 , and the content of the program currently designated as the main NC machining program is displayed by the display unit 116 .

Then, the T signal generated between each created NC machining program and the tool used in each NC machining program is collected and acquired as signal data.

Specifically, when various workpieces are processed in the machine tool 100, appropriate processing is performed by utilizing each NC processing program and the characteristics of each tool according to the workpiece to be processed, so the computer numerical control unit and the programmable logic control device are used When a tool is selected, a T signal is generated.

In addition, the application programming interface (API) of the data collection unit 200 collects and acquires the number of times of use of the generated T signal as data.

For example, when machining a workpiece using the NC machining programs 1,2,3,4,5 of the machine tool 100, each tool TOO1, T002, T003, T004, T005 is used as shown in Table 1 below. The number of times is generated by the T signal.

TOO1 T002 T003 T004 T005 09001 2 N/A One N/A One 09002 One One N/A N/A 2 09003 3 One N/A One N/A 09004 One N/A One One One 09005 One N/A One N/A One

Here, 09001, 09002, 09003, 09004, and 09005 are NC machining programs 1,2,3,4,5, and TOO1, T002, T003, T004, and T005 are tool numbers used in the machine tool 110, and numbers 1,2,3 are the number of called tools, and N/A is the signal value of the uncalled tool.

And, when the values shown in Table 1 are expressed as a matrix (matrix) in Table 2, it is as follows.

2 0 One 0 One One One 0 0 2 3 One 0 One 0 One 0 One One One One 0 One 0 One

This process is a step for analyzing the relationship between each NC part program, and the value of the uncalled signal (N/A) is estimated using matrix factorization for the corresponding combination matrix.

That is, even if a specific signal is not called in a specific program, it cannot be considered as irrelevant, so it is a process for obtaining an estimated value for an implicit data set.

As such, among the signal values displayed in the matrix (matrix), the values for which the correlation value between the NC processing program and the signal is not determined are predicted through matrix factorization.

In addition, the signal data acquired through the data collection unit 200 is transmitted to the analysis and storage unit 300 .

In step (b), the acquired signal data is analyzed and processed. (S20)

That is, the signal data collected and acquired from the data collection unit 200 is analyzed and processed through the analysis and storage unit 300 of the system 1000 according to the present invention.

Since the analysis and processing of such signal data is performed by pre-programmed deep learning, a detailed description thereof will be omitted.

In step (c), the degree of similarity between each NC machining program is obtained based on the processed signal data. (S30)

That is, it is possible to know the degree of similarity between the programs obtained based on the T signal data (the number of times the T signal is used) of the tool obtained through the above process.

As an example, among the NC machining programs, a Pearson correlation coefficient between NC machining program 1 and NC machining program 5 is applied, and the degree of similarity between 2 NC programs is obtained using the corresponding calculation formula below.

That is, let X be the NC machining program 09001, the average value of X is 0.800, let Y be the NC machining program 09005, and the average value of Y is 0.600.

The standard deviation value of X, which is the NC machining program 09001, is 0.837, and the standard deviation value of Y, which is the NC machining program 09005, is 0.548.

If this is calculated using the above formula, the result is 0.873.

As described above, the results of obtaining the degree of similarity between the remaining NC machining programs 2 to 4 and also between the NC machining programs by applying the Pearson correlation coefficient are shown in Table 3 below.

09001 09002 09003 09004 09005 90001 One 0.286 0.488 0.535 0.873 90002 0.286 One 0 -0.134 0.327 09003 0.488 0 One 0 0 09004 0.535 -0.134 0 One 0.612 09005 0.873 0.327 0 0.612 One

That is, the similarity created based on the tool signal data between the NC part program 1 and the NC part program 2 is 0.286, and the similarity created between the NC part program 1 and the NC part program 3 based on the tool signal data is 0.488, and the NC The similarity between part program 1 and NC part program 4 based on tool signal data is 0.535, and between NC part program 1 and NC part program 5 based on tool signal data is 0.873.

And, the similarity between each of the NC machining programs 1,2,3,4,5 obtained through the Pearson correlation coefficient of each of the NC machining programs 1,2,3,4,5 shown in Table 3 is shown in Table 4 below. shown in

NC machining program NC machining program Similarity

09005 09001 0.873 09002 0.327 09003 0 09004 0.612

As described above, in the present invention, NC machining program 09005 is selected as a specific NC machining program, and the relationship between the two programs is analyzed and processed using similarity values with other NC machining programs 09001, 09002, 09003, and 09004.

Then, the corresponding similarity value obtained through the above process is used as a weight in the next process.

In this case, when the obtained corresponding similarity value is a negative number, it is preferable to substitute an absolute value for weighting with a positive value.

In step (d), the calculation is performed by applying the degree of similarity between each NC machining program as a weight. (S40)

To this end, as shown in Table 5 below, a matrix was created based on the number of executions of applications used in each of the NC machining programs 09001, 09002, 09003, and 09004.

App.1 App.2 App.3 App.4 09001 9 0 One 0 09002 0 0 6 0 09003 8 0 0 One 09004 3 0 One 0

Here, each of the above applications App.1, App. 2, App. 3, App. When 4 .... etc. are input through the input unit 114 for application execution of the HMI device 110 , the application stored in the machine tool 100 is displayed by the display unit 113 .

Then, a rating matrix is generated as shown in Table 6 below by multiplying the execution count value of the application by the similarity value obtained in Table 4 by a weight.

For example, the number of executions of the application used in the NC part program 90001 is 9, and the similarity between the NC part program 09005 and the NC part program 09001 is 0.873 as shown in Table 4.

If we calculate this, we get 9×0.873 = 7.857.

In addition, the number of executions of the application used in the NC part program 90002 is 0, and the similarity between the NC part program 09005 and the NC part program 09002 is 0.327 as can be seen from Table 4.

If you calculate this, 0×0.327 = 0.

Also, the number of executions of the application used in the NC machining program 09003 is 8, and the similarity between the NC machining program 09005 and the NC machining program 09003 is 0, as shown in Table 4.

If we compute this, we get 8×0 = 0.

Also, the number of executions of the application used in the NC machining program 09004 is 3, and the similarity between the NC machining program 09005 and the NC machining program 09004 is 0.612, as shown in Table 4.

If you calculate this, 3×0.612 = 1.836.

In step (e), an application for a specific NC machining program is recommended based on the application ranking matrix. (S50)

That is, the application recommendation unit 400 recommends an application based on the data acquired through the data collection unit 200 and the analysis and storage unit 300 .

For example, assuming that a new NC machining program 09005 is created, the similarity value obtained in Table 4 above is used as a weight, and the weight is given to the number of times each application (App.1, App.2, App.3, App.4) is executed. The sum of multiplying each and adding the corresponding values is derived as a result value as shown in Table 6.

NC machining program application result value

09005 App.1 9.693 App.2 0 App.3 3.447 App.4 0

That is, 9×0.873 = 7.857 obtained by multiplying the number of executions of the application executed by the NC part program 09001 by 9 and the weight of the NC part program 09005 and 0.873, which is the similarity between the NC part program 09001, and the application executed by the NC part program 90002 0×0.327 = 0 obtained by multiplying 0, the number of executions of NC part program 09005 and 0.327, which is the degree of similarity between NC part program 09005 and NC part program 09002, and the number of executions of the application executed in NC part program 09003 8, NC part program 09005 and NC 8×0 = 0, which is obtained by multiplying 0, which is the similarity between part program 09003, 3, which is obtained by multiplying 3, the number of executions of the application executed in NC part program 09004, and 0.612, which is the similarity between NC part program 09005 and NC part program 09004 If ×0.612 = 1.836 is added, the application with the highest recommendation rank is Application 1 (App.1) with 7.857 + 0 + 0 + 1.836 = 9.693.

Next, if the recommended rankings are listed, application 3 (App.3) with 0.873 + 1.9628 + 0 + 0.612 = 3.447, application 2 with 0 + 0 + 0 + 0 = 0 (App.2), 0 + 0 + 0 + 0 = 0, the application 4 (App.4) is ranked and recommended through the application recommendation unit 400 .

Therefore, in the present invention, when a machine tool user creates a new NC machining program through such a method, an appropriate application is automatically recommended based on the data in the program, thereby improving workability and quality.

1000: system 100: machine tool
110: HMI device
111: Display unit for NC machining program
112: Input part for NC part program selection
113: display unit indicating the application
114: input for application execution
115: display unit indicating the program storage path
116: display unit indicating the contents of the program
200: data collection unit 300: analysis and storage unit
400: application recommendation unit

Claims (12)

A method for recommending an application for a machine tool, comprising:
(a) generating a plurality of machine tool NC machining program signal data in a matrix (matrix);
(b) analyzing and processing the acquired signal data;
(c) obtaining a degree of similarity between each NC machining program based on the analyzed and processed signal data;
(d) calculating the degree of similarity between the respective NC machining programs as a weight;
(e) recommending an application for a specific NC machining program based on the calculated application ranking matrix; Application recommendation method for machine tools, including. According to claim 1,
The signal data includes an M signal generated when an M code command, which is a program code representing an auxiliary function of a machine tool, and a G signal generated when a G code command, which is a program code representing the preparation function of an NC machine tool, is commanded and a tool Application recommendation method for machine tools, characterized in that it can be used as any one of the T signals generated when the use command of According to claim 1,
In the step (a), the signal data is an application recommendation method for a machine tool, characterized in that the number of uses. According to claim 1,
Among the signal values displayed in a matrix (matrix) in step (a), the values for which the correlation value between the NC machining program and the signal is not determined are predicted through matrix factorization. Way. According to claim 1,
In the step (c), the similarity between each of the NC machining programs 1,2,3,4,5 is obtained through the Pearson correlation coefficient of each of the NC machining programs 1,2,3,4,5. How to recommend applications for machines. According to claim 1,
The calculation in step (d) is a method for recommending an application for a machine tool, characterized in that the value of the number of executions of the application is multiplied by the value of the similarity by a weight. According to claim 1,
Application recommendation method for a machine tool, characterized in that the recommendation in step (e) is based on the obtained similarity value as a weight, the number of times each application is executed is multiplied by the weight, and the sum of the values is the highest. A system used in an application recommendation method for a machine tool, the system comprising:
CNC machine tools;
a data collecting unit serving to collect and transmit data that is an NC machining program signal generated from a computer numerical control (CNC) and programmable logic control device of the machine tool through an application programming interface (API);
an analysis and storage unit for analyzing, storing, and storing data collected and transmitted from the data collection unit;
an application recommendation unit for recommending an application based on the data acquired through the data collection unit and the analysis and storage unit; A system used in an application recommendation method for a machine tool, characterized in that it comprises a. 9. The method of claim 8,
The machine tool is a system used in an application recommendation method for a machine tool, characterized in that it further comprises an application for machine tool application recommendation and an HMI device (Humanmachine Interface). 9. The method of claim 8,
The application programming interface (API) of the data collection unit is configured to support NC communication of various machine tool manufacturers such as FANUC, SIEMENS, and HEIDENHAIN. Application recommendation for machine tools, characterized in that The system used in the method. 9. The method of claim 8,
The analysis and storage unit includes an analysis module that analyzes the collected data in real time and stores time series data, a data processing module that classifies and processes data to obtain necessary information from the collected data, and A system used in an application recommendation method for a machine tool, characterized in that it consists of a data storage module for storing data, and a system management module for managing the server system. 10. The method of claim 9,
The HMI device includes a display unit displaying an NC machining program stored in the machine tool, an input unit for selecting the NC machining program, a display unit displaying an application stored therein, an input unit for executing an application, and a current main A system used in an application recommendation method for a machine tool, characterized in that it comprises a display unit indicating the storage path of the program designated as the NC machining program and a display unit displaying the contents of the program currently designated as the main NC machining program.

Images