Data collection plan generating device and data collection plan generating method
The data collection plan generation device efficiently addresses the inefficiencies in data collection for industrial machines by generating a plan that evenly covers the machine's specification range, enhancing machine learning performance and reducing development time.
Patent Information
- Application Number
- JP2022570013
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2020-12-18
- Filing Date
- 2021-12-14
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2041-12-14
AI Technical Summary
Existing data collection methods for industrial machines are inefficient in covering the specification range without excess or deficiency, leading to increased development steps and time, particularly when using machine learning for thermal displacement correction and operational status determination.
A data collection plan generation device and method that includes a specification setting unit, constraint extraction unit, parameter selection unit, and operating condition generation unit to create a data collection plan that evenly covers the machine's specification range, reducing development time and improving machine learning performance.
The solution enables the generation of an appropriate data collection plan, stabilizing machine learning performance and reducing development man-hours by ensuring data coverage without excess or deficiency, applicable to thermal displacement correction and operational status determination.
Smart Images

Figure 0007586931000001 
Figure 0007586931000002 
Figure 0007586931000003
Abstract
Description
[Technical field]
[0001] The present invention relates to a data collection plan generating device and a data collection plan generating method for generating a data collection plan consisting of a combination of operating conditions when acquiring data from a machine in operation. [Background technology]
[0002] At manufacturing sites such as factories, industrial machines such as machine tools and robots are installed. At manufacturing sites, a plurality of workers operate these industrial machines to manufacture products. The components of the industrial machines expand or contract due to heat generated by the operation of each part and changes in the environmental temperature. The expansion or contraction of each part of the industrial machine affects the positioning of the moving parts of the industrial machine. Therefore, when the industrial machine is required to operate precisely, for example, when precision machining is being performed or parts are being assembled by a robot, it is necessary to perform so-called thermal displacement correction, which is a correction of the position of the moving parts according to the thermal state of the industrial machine. As one method of thermal displacement correction, there is a method of estimating the correction amount using machine learning based on operation information acquired from the industrial machine and the temperature of each part of the industrial machine detected by a sensor or the like (for example, Patent Document 1, etc.). This method of performing thermal displacement correction using machine learning is also known.
[0003] Furthermore, when industrial machinery continues to operate at a manufacturing site, an abnormality may occur in the operation of the industrial machinery due to external factors such as aging of the components that make up the industrial machinery, changes in outside temperature, vibrations, and setting errors made by operators, etc. When it is determined that an abnormality has occurred in the operation of the industrial machinery, the operator stops the operation of the industrial machinery where the abnormality occurred, removes the cause of the abnormality, and then resumes operation of the industrial machinery to continue work.
[0004] The operating status of industrial machines, including normal / abnormal, is monitored by, for example, a management device. The device for managing the operating status of industrial machines monitors, for example, via a network, time series data such as the position, speed, and torque of a motor detected by each industrial machine, data showing a change in a signal at a predetermined time, and time series data such as vibration, sound, and moving images detected by a sensor attached to the industrial machine, and detects the change state of each data over time, thereby managing the operating status of the industrial machines (for example, Patent Documents 2 and 3, etc.). Note that one method for determining the operating status of an industrial machine is a method based on physical quantities detected by a sensor or the like from the industrial machine. A method using machine learning as a method for such determination is publicly known. In this way, machine learning is beginning to be used in various aspects of manufacturing sites. [Prior art documents] [Patent documents]
[0005] [Patent Document 1] JP 2019-166603 A [Patent Document 2] Patent No. 6453504 [Patent Document 3] JP 2019-012473 A Summary of the Invention [Problem to be solved by the invention]
[0006] In order to utilize machine learning functions to perform thermal displacement correction and determine the operating state of industrial machinery, it is necessary to perform machine learning using data collected from the industrial machinery. At this time, in order to generate a learning model that can be used to estimate the appropriate thermal displacement correction amount and to appropriately determine the operating state of the industrial machinery, that is, a good learning model, data that can cover the specification range of the industrial machinery is required. Therefore, when collecting data, data must be collected under various operating conditions so as to cover the specification range of the industrial machinery. However, if the number of data is increased in an attempt to cover the specification range evenly, the development man-hours will increase significantly. In particular, if one data collection takes a long time, the development man-hours will be enormous. In other words, data collection must be planned so that the collected data can cover the specification range of the industrial machinery without excess or deficiency. Such issues arise not only in machine learning but also in operation experiments using other industrial machines and industrial systems. Therefore, there is a need for a mechanism that can generate an appropriate data collection plan for collecting appropriate data from general settings such as machine functions and specifications. [Means for solving the problem]
[0007] One aspect of the present invention is a data collection plan generation device that generates a data collection plan consisting of a combination of operating conditions for acquiring data from a machine in operation, the data collection plan generation device comprising: a specification setting unit that sets specifications of the machine; a constraint condition extraction unit that extracts constraint conditions related to the operation of the machine; an operating condition generation unit that generates a plurality of operating conditions including sets of parameter values related to the specifications set by the specification setting unit based on the specifications and the constraint conditions; a plan generation unit that generates a data collection plan from the operating conditions generated by the operating condition generation unit; and an output unit that outputs the data collection plan generated by the plan generation unit.
[0008] Another aspect of the present invention is a data collection plan generation method by a data collection plan generation device that generates a data collection plan consisting of a combination of operating conditions when acquiring data from a machine in operation, the data collection plan generation method including a specification setting step of setting specifications of the machine, a constraint condition extraction step of extracting constraint conditions related to the operation of the machine, and a specification setting step of extracting constraint conditions related to the operation of the machine based on the specifications and the constraint conditions. procedure an operating condition generating step of generating a plurality of operating conditions including a set of parameter values related to the specifications set in the operating condition generating step; procedure and an output step of outputting the data collection plan generated in the plan generation step. Effect of the Invention
[0009] One aspect of the present invention makes it possible to generate an appropriate data collection plan for collecting appropriate data from general settings such as machine functions and specifications, which is expected to achieve stable performance improvements in machine learning functions and reduced development labor. [Brief description of the drawings]
[0010] [Figure 1] FIG. 2 is a schematic hardware configuration diagram of a data collection plan generating device according to an embodiment. [Diagram 2] FIG. 1 is a schematic functional block diagram of a data collection plan generating device according to a first embodiment. [Diagram 3] FIG. 11 is a diagram illustrating an example of operating conditions. [Figure 4] 3 is a diagram illustrating a method for generating operating conditions by an operating condition generating unit in the data collection plan generating device of FIG. 2. [Diagram 5] 11A and 11B are diagrams illustrating another method for generating operating conditions by the operating condition generating unit. [Figure 6] 11A and 11B are diagrams illustrating another method for generating operating conditions by the operating condition generating unit. [Figure 7] 11A and 11B are diagrams illustrating another method for generating operating conditions by the operating condition generating unit. [Figure 8] FIG. 11 is a diagram illustrating an example of selection rules. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0011] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a schematic hardware configuration diagram showing the main part of a data collection plan generating device according to an embodiment of the present invention. The data collection plan generating device 1 of the present invention can be implemented, for example, as a control device that controls industrial machinery, and can also be implemented on a personal computer attached to the control device that controls industrial machinery, or on a personal computer, cell computer, fog computer, or cloud server connected to the control device via a wired / wireless network. In this embodiment, an example is shown in which the data collection plan generating device 1 is implemented on a personal computer connected to a control device that controls industrial machinery via a network.
[0012] The CPU 11 provided in the data collection plan generation device 1 according to this embodiment is a processor that controls the entire data collection plan generation device 1. The CPU 11 reads out a system program stored in the ROM 12 via the bus 22, and controls the entire data collection plan generation device 1 according to the system program. The RAM 13 temporarily stores temporary calculation data, display data, various data input from outside, and the like.
[0013] The non-volatile memory 14 is composed of, for example, a memory backed up by a battery (not shown) or an SSD (Solid State Drive), and the memory state is maintained even when the power supply of the data collection plan generating device 1 is turned off. The non-volatile memory 14 stores data read from the external device 72 via the interface 15, data input via the input device 71, data acquired from the control device 3 via the interface 20, and the like. The data stored in the non-volatile memory 14 may be expanded in the RAM 13 at the time of execution / use. In addition, various system programs such as known analysis programs are written in advance in the ROM 12.
[0014] The interface 15 is an interface for connecting the CPU 11 of the data collection plan creation device 1 to an external device 72 such as a USB device. For example, data acquired by other industrial machines can be read from the external device 72. In addition, data processed in the data collection plan creation device 1 can be stored in an external storage means (not shown) via the external device 72.
[0015] The interface 20 is an interface for connecting the CPU 11 of the data collection plan creation device 1 to a wired or wireless network 5. The network 5 is connected to the control device 3, a fog computer, a cloud server, etc., and exchanges data with the data collection plan creation device 1.
[0016] The display device 70 displays various data loaded into the memory, data obtained as a result of executing programs, etc., outputted via the interface 17. An input device 71 consisting of a keyboard, pointing device, etc., transfers commands, data, etc. based on operations by an operator to the CPU 11 via the interface 18.
[0017] FIG. 2 is a schematic block diagram showing functions of the data collection plan generating device 1 according to the first embodiment of the present invention. Each function of the data collection plan generating device 1 according to the present embodiment is realized by the CPU 11 of the data collection plan generating device 1 shown in FIG. 1 executing a system program and controlling the operation of each part of the data collection plan generating device 1. The data collection plan generating device 1 according to the present embodiment has a function of generating an appropriate data collection plan for collecting appropriate data from general setting values such as the function and specifications of the machine to be operated. Note that, in the present embodiment, the generation of a data collection plan in the case of performing thermal displacement correction of a machine tool using machine learning is shown as an example, but the data collection plan generating device 1 according to the present embodiment can also be applied to the case of, for example, determining the operating state of a machine using machine learning.
[0018] The data collection plan generating device 1 of this embodiment includes a specification setting unit 100, a constraint condition extraction unit 110, a parameter selection unit 120, an operating condition generating unit 130, a data collection plan generating unit 135, and an output unit 140. The RAM 13 to the non-volatile memory 14 of the data collection plan generating device 1 are provided with a specification storage unit 200 as an area for storing data related to the specifications acquired and set from the input device 71, the external device 72, the control device 3, etc., a data collection plan storage unit 210 as an area for storing data related to the generated data collection plan of the machine, and a selection rule storage unit 220 as an area in which rules for selecting parameters as operating conditions are stored in advance.
[0019] The specification setting unit 100 is realized by the CPU 11 included in the data collection plan generating device 1 shown in FIG. 1 executing a system program read from the ROM 12, and mainly performing arithmetic processing by the CPU 11 using the RAM 13 and the non-volatile memory 14, and input control processing by the interface 15, 18, or 20. The specification setting unit 100 acquires and sets information related to specifications related to the operation of the industrial machine 4 controlled by the control device 3. The information related to the specifications may be, for example, information that specifies a predetermined range of parameters related to the operation of the machine, such as the movement range of each axis, the maximum movement speed of each axis, the maximum movement acceleration of each axis, the maximum torque of each axis, the maximum load weight, the maximum spindle rotation speed, the maximum spindle torque, the maximum coolant flow rate, the maximum coolant temperature, the minimum coolant temperature, etc., information that specifies a predetermined range of data related to the operating environment of the machine, such as the maximum environmental temperature, the minimum environmental temperature, etc. These information related to the specifications may be acquired from information stored in the control device 3, a fog computer not shown, a cloud server, etc., or may be read from an external device 72. If necessary, the operator may set the specifications through the input device 71. The data acquired by the specification setting unit 100 is stored in the specification storage unit 200.
[0020] The constraint condition extraction unit 110 is realized by the CPU 11 included in the data collection plan generation device 1 shown in FIG. 1 executing a system program read from the ROM 12, and the CPU 11 mainly performing arithmetic processing using the RAM 13 and the non-volatile memory 14. The constraint condition extraction unit 110 extracts constraint conditions related to the operating conditions of the machine from the specifications set by the specification setting unit 100 and other information input by the operator from the input device 71. The constraint condition extraction unit 110 may extract maximum and minimum values of parameters related to the operation of each machine as constraint conditions, for example, based on information related to the specifications set by the specification setting unit 100. The constraint condition extraction unit 110 may extract maximum and minimum values of data related to the operating environment of each machine as constraint conditions, for example, based on information related to the specifications set by the specification setting unit 100. In addition, the constraint conditions may be a list of values that a predetermined parameter or operating environment can take, information indicating a distribution, or a conditional expression that constrains the relationship between predetermined parameters. The constraint condition extraction unit 110 may prompt an operator to input, for example, restrictions on operations associated with the size of the machine, restrictions on continuous operation time, restrictions on the number of operations, etc., and extract constraint conditions based on the information input from the input device 71. The constraint condition extraction unit 110 may extract, as constraint conditions, information such as typical spindle rotation speed and spindle rotation time from, for example, a machining program executed by the control device 3. The constraint conditions extracted by the constraint condition extraction unit 110 are stored in the specification storage unit 200 together with information related to the specifications.
[0021] The parameter selection unit 120 is realized by the CPU 11 included in the data collection plan generation device 1 shown in FIG. 1 executing a system program read from the ROM 12, and the CPU 11 mainly performing arithmetic processing using the RAM 13 and the non-volatile memory 14. The parameter selection unit 120 selects parameters to be used for generating the operating conditions of the machine. The parameter selection unit 120 may display parameters selectable as the operating conditions of the industrial machine 4 on the display device 70 via the output unit 140 based on the specifications set by the specification setting unit 100, for example, and allow the worker to select a parameter to be used as the operating condition of the industrial machine 4 from the displayed parameters. The parameter selection unit 120 may also select parameters to be used as the operating conditions of the industrial machine 4 according to the selection rules stored in the selection rule storage unit 220 according to the purpose related to the operating conditions of the industrial machine 4 by the worker, for example. In this case, as illustrated in FIG. 8, a selection rule in which a list of parameters to be selected is associated with the purpose of data collection is stored in the selection rule storage unit 220 in advance. The parameter selection unit 120 displays the purpose of data collection included in the selection rules on the display device 70 and allows the operator to select the purpose, thereby making it possible to easily select appropriate parameters. Note that, for parameters whose specifications are set by the specification setting unit 100, the parameter selection unit 120 is not necessarily a required component when all of them are to be selected as operating conditions. In such a case, even a data collection plan generation device 1 that does not include the parameter selection unit 120 can fully enjoy the effects of the present invention.
[0022] The operating condition generating unit 130 is realized by the CPU 11 included in the data collection plan generating device 1 shown in Fig. 1 executing a system program read from the ROM 12 and performing calculation processing mainly by the CPU 11 using the RAM 13 and the non-volatile memory 14. The operating condition generating unit 130 generates operating conditions for the industrial machine 4. As exemplified in Fig. 3, each operating condition generated by the operating condition generating unit 130 includes a set of parameter values selected by the parameter selecting unit 120.
[0023] The data collection plan generating unit 135 is realized by the CPU 11 included in the data collection plan generating device 1 shown in Fig. 1 executing a system program read from the ROM 12, and the CPU 11 mainly performing calculation processing using the RAM 13 and the non-volatile memory 14. The data collection plan generating unit 135 creates an appropriate data collection plan for collecting appropriate data from the industrial machinery 4. That is, in a vector space of vectors whose elements are the parameters included in the operating conditions, a combination of operating conditions to be used for data collection is generated so that the values of the parameters included in each operating condition are evenly scattered in the vector space.
[0024] In order to distribute the operating conditions evenly, the data collection plan generation unit 135 may, for example, define values for each parameter included in the operating conditions at a predetermined interval between the minimum and maximum values of the parameter, and generate a set of the defined values as a data collection plan. Fig. 4 shows an example of generating a data collection plan by selecting an environmental temperature and a spindle rotation speed as parameters in order to collect appropriate learning data for machine learning of a machine learning device that predicts a thermal displacement compensation value. In the example of Fig. 4, values are set at predetermined intervals between the minimum and maximum values for each of the environmental temperature and the spindle rotation speed, and a set of these set values is generated as a data collection plan.
[0025] In order to distribute the operating conditions evenly, the data collection plan generating unit 135 may, for example, determine values for each parameter included in the operating conditions at a predetermined interval between the minimum and maximum values of the physical quantity related to the parameter, and generate a set of the determined values as a data collection plan. Fig. 5, like Fig. 4, shows an example of generating a data collection plan by selecting an environmental temperature and a spindle rotation speed as parameters in order to collect appropriate learning data for machine learning of a machine learning device that predicts a thermal displacement compensation value. In the example of Fig. 5, a value is set for the environmental temperature at a predetermined interval between a minimum value and a maximum value, and a value for the spindle rotation speed is set so that the amount of heat generated from the spindle is at a predetermined interval, and a set of these set values is generated as a data collection plan.
[0026] The data collection plan generating unit 135 may thin out the operating conditions when the number of operating conditions included in the data collection plan to be generated is greater than the limit of the number of operations extracted by the constraint condition extracting unit 110. For example, as illustrated in Fig. 6, the operating conditions may be thinned out uniformly, or the operating conditions may be thinned out so as to be evenly distributed according to a predetermined rule.
[0027] Contrary to the above example, the data collection plan generating unit 135 may generate operating conditions iteratively within the range of the limit of the number of operations extracted by the constraint condition extracting unit 110. For example, as illustrated in FIG. 7, after generating a set of the minimum and maximum values of each parameter as an operating condition, a point on the vector space of the parameters where the density of the operating condition is the thinnest may be selected, and a set of values at the position of that point may be generated as an operating condition. At this time, the density of the operating condition may differ depending on the axis on the vector space. By thinning the density of the axis of a parameter whose value change does not have much effect on the machine learning result, it is possible to prevent the number of operating conditions from increasing unnecessarily. By selecting in this way, a set of parameter values of an arbitrary number of operations that are evenly scattered on the vector space of the parameters can be generated.
[0028] The data collection plan extracted by the data collection plan generating unit 135 is stored in the data collection plan storage unit 210. The data collection plan stored in the data collection plan storage unit 210 may be displayed on the display device 70 via the output unit 140. The output unit 140 may be configured to perform not only display output to the display device 70, but also output processing to the external device 72 or output processing via the network 5. In this case, the data collection plan stored in the data collection plan storage unit 210 can be stored in an external storage device (not shown) via the external device 72 for use, or transmitted to the control device 3, a fog computer (not shown), or a cloud server via the network for use.
[0029] By using the data collection plan generation device 1 according to this embodiment having the above-described configuration, an operator can generate an appropriate data collection plan for collecting appropriate data from general setting values such as machine functions and specifications.
[0030] Although one embodiment of the present invention has been described above, the present invention is not limited to the above-described embodiment, and can be embodied in various forms by making appropriate modifications. For example, in the above embodiment, the data collection plan generating device 1 generates the operating conditions based on the machine specifications and constraint conditions, but the specifications and constraint conditions handled by the data collection plan generating device 1 are not limited to machine specifications and constraint conditions. For example, the operating conditions of a system composed of multiple machines and devices, such as a production system or an experimental system, may be generated based on the specifications and constraint conditions of each component of the system. Even in such a case, the functions of the data collection plan generating device 1 of the present application can be effectively used. [Explanation of symbols]
[0031] 1 Data Collection Plan Generator 3. Control device 4. Industrial machinery 5. Network 6 Sensors 11 CPU 12 ROM 13 RAM 14 Non-volatile memory 15,17,18,20 Interface 22 Bus 70 Display device 71 Input Devices 72 External equipment 100 Specification Setting Section 110 Constraint condition extraction part 120 Parameter selection section 130 Operating condition generation section 135 Data Collection Plan Generation Division 140 Output section 200 Specification memory unit 210 Data Collection Plan Storage Section 220 Selection rule memory unit
Claims
1. A data collection plan generating device for generating a data collection plan consisting of a combination of operating conditions when acquiring data from a machine in operation, A specification setting unit that sets the specifications of the machine; A constraint condition extraction unit that extracts constraint conditions related to the operation of the machine; an operating condition generating unit that generates a plurality of operating conditions including a set of parameter values related to the specifications set by the specification setting unit, based on the specifications and the constraint conditions; a data collection plan generating unit that generates a data collection plan from the operating conditions generated by the operating condition generating unit; an output unit that outputs the data collection plan generated by the data collection plan generation unit; A data collection plan generating device comprising:
2. a parameter selection unit that selects parameters to be included in the operating conditions of the machine based on the specifications and the constraint conditions; the operating condition generating unit generates a plurality of operating conditions including a set of parameter values selected by the parameter selecting unit based on the specification and the constraint conditions; The data collection plan generator of claim 1 .
3. A data collection plan generation method by a data collection plan generation device that generates a data collection plan consisting of a combination of operating conditions when acquiring data from a machine in operation, comprising: A specification setting procedure for setting the specifications of the machine; A constraint condition extraction procedure for extracting constraint conditions related to the operation of the machine; an operating condition generating step of generating a plurality of operating conditions including a set of parameter values related to the specifications set in the specification setting step, based on the specifications and the constraints; a data collection plan generation step of generating a data collection plan from the operating conditions generated in the operating condition generation step; an output step of outputting the data collection plan generated in the data collection plan generation step; A method for generating a data collection plan.
4. After the constraint condition extraction step, a parameter selection step is further performed to select parameters to be included in the operating conditions of the machine based on the specifications and the constraint conditions; In the operating condition generation step, a plurality of operating conditions including a set of parameter values selected in the parameter selection step are generated based on the specification and the constraint conditions. The method of claim 3 .
Citation Information
Patent Citations
Overload fusible resistor
JP1989053504A
Data filtering device and method
JP2019012473A
Thermal displacement correction system and computer
JP2019166603A
Data collection system and data collection method
JP2020064390A
Data generation device, data generation method, learning device, and learning method
WO2020184240A1
