DATA EXTRACTING RULE CREATION DEVICE, DATA EXTRACTING RULE CREATION METHOD, AND STORAGE MEDIUM

Abstract

A data extraction rule preparation device comprises: a data acquisition unit that acquires first data which indicates the state of a factory; an extraction range acquisition unit that acquires the extraction range of the first data; an extraction rule candidate preparation unit that, on the basis of the extraction range of the first data, prepares a data extraction rule candidate which includes a condition for extracting the first data; an extraction rule evaluation unit that calculates an evaluation value of the extraction rule candidate; and an extraction rule optimization unit that, on the basis of the evaluation value calculated by the extraction rule evaluation unit, prepares a data extraction rule obtained by optimizing the data extraction rule candidate.

Description

[Technical Field]

[0001] The present invention relates to a data extraction rule creation device, a data extraction rule creation method, and a storage medium. [Background technology]

[0002] Currently, there is technology that collects data indicating the state of a factory, analyzes the collected data, and uses it for factory management. In this technology, data showing the operating status of industrial machines installed in a factory is stored, and data having a predetermined time span is extracted from the stored observation data and subjected to statistical processing to grasp the factory status (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-15573 Summary of the Invention [Problem to be solved by the invention]

[0004] Factory condition monitoring involves collecting data from multiple machines and sensors. Some of the data collected in factories is sampled at high frequencies, requiring a database with a large memory capacity to store the data. In addition, sending and receiving large amounts of data places a heavy burden on communications.

[0005] Furthermore, if data is accumulated indiscriminately and used to understand the operating status of industrial machinery, not only will the analysis be costly, but the accuracy of the analysis will also be reduced.To improve the accuracy of the analysis, it is necessary to extract only the data related to the operating status.

[0006] To address this issue, there is a conventional technique for extracting only necessary information and recording or transmitting / receiving only the extracted information. However, the rules for determining which range of data should be extracted are not simple, and require the judgment of an operator with knowledge and experience.

[0007] In the manufacturing industry, there is a demand for technology that can extract necessary data. [Means for solving the problem]

[0008] A data extraction rule creation device 100 according to one aspect of the present disclosure includes a data acquisition unit that acquires first data indicating a factory state, an extraction range acquisition unit that acquires an extraction range for the first data, an extraction rule candidate creation unit that creates a data extraction rule candidate including a condition for extracting the first data based on the extraction range for the first data, and data The data extraction system includes an extraction rule evaluation unit that calculates an evaluation value of an extraction rule candidate, and an extraction rule optimization unit that creates a data extraction rule that optimizes the data extraction rule candidate. A data extraction rule creation method according to one aspect of the present disclosure includes: acquiring first data indicating a factory state; acquiring an extraction range for the first data; and creating a data extraction rule candidate including a condition for extracting the first data based on the extraction range for the first data. data An evaluation value of the candidate extraction rule is calculated, and a data extraction rule is created by optimizing the candidate data extraction rule based on the evaluation value. A computer-readable storage medium according to one aspect of the present disclosure, when executed by one or more processors, acquires first data indicating a factory state, acquires an extraction range for the first data, and creates a data extraction rule candidate including a condition for extracting the first data based on the extraction range for the first data; data The computer-readable instructions are stored to calculate an evaluation value of a candidate data extraction rule, and to create a data extraction rule that optimizes the candidate data extraction rule based on the evaluation value. [Effects of the Invention]

[0009] According to one aspect of the present invention, necessary data can be extracted. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a block diagram showing a configuration of a data extraction system according to the first disclosure. [Figure 2] 10 is an example of a cutout range of first data. [Figure 3] 10 is a flowchart illustrating a process of the data extraction rule creating device. [Figure 4] 10 is a flowchart illustrating a process of the data extraction device. [Figure 5] FIG. 10 is a block diagram showing the configuration of a data extraction system according to the second disclosure. [Figure 6] This is an example of the extraction range of the second data. [Figure 7] 10 is an example of a cutout rule. [Figure 8] FIG. 10 is a diagram illustrating optimization of data extraction rules. [Figure 9] 10 is a flowchart illustrating a process of the data extraction rule creating device. [Figure 10] 10 is a flowchart illustrating a process of the data extraction device. [Figure 11] FIG. 2 is a diagram illustrating a hardware configuration of a data extraction rule creation device. [Figure 12] FIG. 1 is a diagram illustrating a configuration of a data extraction system. [Figure 13] FIG. 1 is a diagram illustrating an application example of a data extraction system. DETAILED DESCRIPTION OF THE INVENTION

[0011] [First Disclosure] FIG. 1 is a block diagram showing the configuration of a data extraction system 1000. As shown in FIG. The data extraction system 1000 includes a data extraction rule creating device 100 that creates data extraction rules, and a data extraction device 200 that extracts data using the data extraction rules.

[0012] The data extraction rule creating device 100 includes a data acquiring unit 11 , an extraction range acquiring unit 13 , an extraction rule candidate creating unit 14 , an extraction rule evaluating unit 15 , an extraction rule optimizing unit 16 , and an output unit 17 .

[0013] The data acquisition unit 11 acquires data indicating the state of the factory. The data indicating the state of the factory is acquired from sensors installed in the factory, programmable logic controllers (PLCs) that collect information from the sensors and control the machines, numerical control devices that control machine tools, data loggers that record data collected from these machines, and the like. The data indicating the state of the factory may include, for example, the position of each motor that operates the machine tools installed in the factory, the speed of each motor, the acceleration of each motor, the torque command of each motor, the temperature of each part, detection values from sound sensors and optical sensors, the state of signals from peripheral devices such as safety doors, the ambient temperature, and the ambient humidity. The data indicating the state of the factory is continuous time-series data. In data analysis, only the valid range of the continuous time-series data is extracted.

[0014] The cut-out range acquisition unit 13 acquires the cut-out range of the time-series data. The cut-out range may be acquired automatically or manually. When the cutout range is automatically acquired, for example, deviation from normal data is used. In this case, the cutout range acquisition unit 13 learns the waveform of normal data in advance and acquires the range that deviates from the normal data as the cutout range. When the extraction range is acquired manually, the extraction range is specified by a person. The manual specification of the extraction range is performed by an experienced person who has knowledge and experience of the factory or machinery. The experienced person visually checks the data and specifies the extraction range. Figure 2 shows an example of the data extraction range. The area between the dotted lines is the extraction range.

[0015] The extraction rule candidate creation unit 14 creates candidates for extraction rules for time series data. The data extraction rule candidates created by the extraction rule candidate creation unit 14 are tentative data extraction rules, which are later evaluated by the extraction rule evaluation unit 15 and optimized by the extraction rule optimization unit 16 based on the evaluation. The extraction rule candidate creation unit 14 creates candidates for data extraction rules based on the values of multiple time series data and their changes in the extraction range acquired by the extraction range acquisition unit 13. The data extraction rule includes conditions such as a threshold for the data to be extracted, a time for extracting the time series data, and a rate of change of the time series data at a certain point. A combination of these conditions becomes an extraction rule. As an example, suppose the torque command of the motor of the moving axis is 6% ± a predetermined margin value M t A rule is created that cuts out a range where the torque command of the motor of the moving axis is 5% to 7%. A rule including such a condition can be created by, for example, calculating the average or median of the values of each piece of time-series data in the specified cut-out range. As another example, a rule is created that cuts out a range where the torque command of the motor of the moving axis is 5% to 7%. A rule including such a condition can be created based on, for example, the maximum or minimum value of each piece of time-series data in the specified cut-out range. The cut-out rule candidate creation unit 14 may create multiple candidates for data cut-out rules.

[0016] The extraction rule evaluation unit 15 calculates an evaluation value for evaluating the data extraction rule candidate created by the extraction rule candidate creation unit 14. The extraction rule evaluation unit 15 calculates, for example, the compatibility (degree of match) between the extraction range based on the data extraction rule candidate and the extraction range acquired by the extraction range acquisition unit 13 as the evaluation value. This compatibility may be, for example, the degree of overlap of the extraction ranges. For example, if the extraction range acquisition unit 13 has acquired one extraction range, the evaluation value may be the rate at which the width of that extraction range overlaps with the width of the extraction range based on the data extraction rule candidate. If the extraction range acquisition unit 13 has acquired multiple extraction ranges (e.g., if extraction ranges have been acquired for multiple pieces of abnormal data), the evaluation value may be the average value of the rates at which the widths of the respective extraction ranges overlap. If the evaluation value is low, the extraction rule optimization unit 16 changes or adjusts the rule of that data extraction rule candidate.

[0017] The extraction rule evaluation unit 15 may further reflect versatility, which indicates the versatility of the conditions included in the data extraction rule, in the evaluation value. In this case, a versatility value (for example, in the range of 0.5 to 1.5) is set in advance for the type of time-series data used as a condition. For example, from the viewpoint of noise resistance and portability, a higher versatility is set for digital signals than for analog signals. Also, from the viewpoint of ease of data acquisition, a higher versatility is set for speed signals than for torque command signals, and a higher versatility is set for control signals than for speed signals. Then, the extraction rule evaluation unit 15 may calculate a final evaluation value by multiplying the evaluation value by the versatility of the time-series data used as a condition.

[0018] The extraction rule optimization unit 16 optimizes the data extraction rule candidate created by the extraction rule candidate creation unit 14 based on the evaluation value calculated by the extraction rule evaluation unit 15. The extraction rule optimization unit 16 creates new candidate data extraction rules by changing the conditions included in the data extraction rule candidate (for example, changing a condition based on shaft position to a condition based on shaft speed) or adjusting the condition value by adding a predetermined difference. Next, the extraction rule evaluation unit 15 calculates an evaluation value for the new data extraction rule candidate after this change or adjustment. Then, if a data extraction rule candidate with a higher evaluation value than the initial data extraction rule candidate is obtained, the new data extraction rule candidate is adopted. The extraction rule optimization unit 16 repeats the above process until a data extraction rule candidate with a higher evaluation value is obtained. Then, when a data extraction rule candidate with the highest evaluation value is obtained, the extraction rule optimization unit 16 regards the data extraction rule candidate as an optimized data extraction rule candidate and outputs it.

[0019] The extraction rule optimized by the extraction rule optimization unit 16 is output by the output unit 17. The output unit 17 may display and output the optimized extraction rule on a display device included in the data extraction rule creation device 100. Alternatively, the output unit 17 may output the optimized extraction rule to an internal storage device or an external storage device included in the data extraction rule creation device 100. Alternatively, the output unit 17 may output the optimized extraction rule to the data extraction device 200 via a wired / wireless network (not shown) or the like.

[0020] Next, the data extraction device 200 will be described. The data extraction device 200 includes an extraction rule storage unit 21, a data acquisition unit 22, and a data extraction unit 24.

[0021] The extraction rule storage unit 21 stores the data extraction rules created by the data extraction rule creation device 100. The data acquisition unit 22 acquires time series data to be extracted. The data extraction unit 24 extracts data from the time series data acquired by the data acquisition unit 22 based on the data extraction rules stored in the extraction rule storage unit 21.

[0022] The data extraction device 200 can automatically extract a valid range of continuous time-series data using the data extraction rules created by the data extraction rule creating device 100.

[0023] The operation of the data extraction system 1000 will be described with reference to FIGS. Fig. 3 is a flowchart showing a schematic example of the operation of the data extraction rule creation device 100. As illustrated in Fig. 3, the data extraction rule creation device 100 acquires time-series data (step S1). Next, it acquires an extraction range for the acquired time-series data (step S2). The extraction range may be determined manually or automatically. The manual extraction range is specified by a veteran operator with experience and knowledge. The automatic extraction range is determined, for example, by the deviation from normal data.

[0024] The data extraction rule creation device 100 creates data extraction rule candidates (step S3). The data extraction rule creation device 100 calculates a data extraction range based on the data extraction rule candidates (step S4). The data extraction rule creation device 100 calculates the compatibility between the calculated extraction range and the extraction range acquired in step S2 as an evaluation value, and judges the evaluation value (step S5).

[0025] If the evaluation value exceeds a predetermined threshold (step S5; appropriate), the data extraction rule creation device 100 considers the data extraction rule candidate to be the optimal data extraction rule and ends the process. If the evaluation value is less than a predetermined threshold (step S5; inappropriate), the data extraction rule creation device 100 adjusts the data extraction rule candidate (step S6), and then proceeds to step S4 to repeat the process. The data extraction rule creation device 100 repeats the adjustment of the data extraction rule candidate and creates a data extraction rule that is more suitable for the extraction range.

[0026] Fig. 4 is a flowchart showing a schematic operation example of the data extraction device 200. As shown in Fig. 4, the data extraction device 200 stores the data extraction rules created by the data extraction rule creating device 100 (step S11). The data extraction device 200 acquires time series data to be extracted (step S12). The data extraction device 200 extracts data from the acquired time series data in accordance with the data extraction rules (step S13).

[0027] As described above, in the data extraction system 1000 of the first disclosure, once a data extraction range is first specified manually or automatically, a data extraction rule that fits this extraction range can be automatically created. The data extraction device 200 can extract data in accordance with the created extraction rule. By extracting the necessary data, it is possible to reduce the storage capacity, the time required to search for the necessary data, and the calculation load required for data processing.

[0028] [Second Disclosure] Next, a data extraction system according to the second disclosure will be described with reference to Fig. 5. Similar to the first data extraction system, the data extraction system according to the second disclosure includes a data extraction rule creation device 100 that creates data extraction rules, and a data extraction device 200 that extracts data using the data extraction rules.

[0029] The data extraction rule creating device 100 includes a data acquiring unit 11 , an extraction range acquiring unit 13 , an extraction rule candidate creating unit 14 , an extraction rule evaluating unit 15 , an extraction rule optimizing unit 16 , and an output unit 17 .

[0030] The data acquisition unit 11 acquires data indicating the state of the factory. The data indicating the state of the factory is acquired from sensors installed in the factory, PLCs that collect information from the sensors and control machines, numerical control devices that control machine tools, data loggers that record collected data, etc.

[0031] The cut-out range acquisition unit 13 acquires a first cut-out range and a second cut-out range. The first cut-out range is a cut-out range of main data (called first data). The second cut-out range is a cut-out range of data (called second data) cut out based on the first cut-out range. The first cutout range may be obtained automatically or manually. When the first cutout range is automatically acquired, for example, normal data is used. The cutout range acquisition unit 13 learns the waveform of the normal data in advance and acquires a range deviating from the normal data as the cutout range. When the first cutout range is manually acquired, the cutout range is specified by a person. The cutout range is specified by an experienced person who has knowledge and experience of the factory or machinery. The experienced person visually checks the data and specifies a valid cutout range.

[0032] Figure 6 is an example of a data extraction range. The range between the dotted lines is the first extraction range for the first data (torque command in this case). The top row of Figure 6 shows the waveform of normal data, and the bottom row shows the waveform of abnormal data. An enlarged view of the extraction range is displayed to the right of the torque command. The waveform of the extraction range differs between normal data and abnormal data. When the extraction range is obtained automatically, the part where the normal data and abnormal data diverge becomes the extraction range. When the extraction range is obtained manually, an experienced operator specifies the extraction range based on the data waveform.

[0033] Once the extraction range for the first data has been determined, the extraction range for the second data is also determined. For example, the extraction range for the second data can be set to the same time period as the first extraction range. In Figure 6, the extraction ranges for "rotation speed" and "Z-axis position" are displayed as the second data. The cutting signal, which indicates the machine status, is also extracted. Because the cutting signal is in the "ON" state, it can be seen that this is data during cutting.

[0034] The extraction rule candidate creation unit 14 creates candidates for data extraction rules for the first data and the second data. The extraction rule candidate creation unit 14 creates candidates for data extraction rules that satisfy the possible range of the data. In Fig. 6, the possible range of "rotation speed" is set to [0 to 10000 rpm], the possible range of "torque command" is set to [0 to 100%], and the possible range of "Z-axis position" is set to [-100 to +100 mm].

[0035] The data extraction rule according to the second disclosure is a combination of conditions such as values of multiple time-series data of the first data and the second data, and their changes. The extraction rule candidate creation unit 14 creates candidates for data extraction rules based on the values of multiple time-series data and their changes in the extraction range acquired by the extraction range acquisition unit 13. The data extraction rule includes conditions such as a threshold value for the data to be extracted, a time for extracting the time-series data, and a rate of change of the time-series data at a certain point. A combination of these conditions becomes a data extraction rule. As an example, suppose the torque command of the motor of the moving axis is 6% ± a predetermined margin value M t The rotation speed of the spindle motor is 3000 rpm ± a predetermined margin value M s , the Z-axis position is -50 mm ± a predetermined margin value M pz A rule is created to extract a range in which the torque command of the motor of the moving axis is 5% to 7%, the rotation speed of the spindle motor is 2950 to 3050 rpm, and the speed of the Z-axis (change in Z-axis position) is -80 mm / s to -120 mm / s. A rule that includes such a condition can be created based on, for example, the maximum value or minimum value of each time series data in the specified extraction range. The extraction rule candidate creation unit 14 may create a plurality of data extraction rule candidates.

[0036] FIG. 7 is an example of a data extraction rule candidate created by the extraction rule candidate creating unit 14. In the data extraction rule candidate illustrated in FIG. 7, the torque command “6±M t %, rotation speed "3000±M s rpm”, Z-axis position “-50±M zp The extraction rule candidate creating unit 14 may create a plurality of extraction rules.

[0037] The extraction rule evaluation unit 15 calculates an evaluation value for evaluating the data extraction rule candidates created by the extraction rule candidate creation unit 14, as in the first disclosure.

[0038] The extraction rule optimization unit 16 optimizes the data extraction rule candidate created by the extraction rule candidate creation unit 14 based on the evaluation value calculated by the extraction rule evaluation unit 15. The extraction rule optimization unit 16 creates a new candidate data extraction rule by changing the conditions included in the data extraction rule candidate (for example, by deleting the shaft position from the conditions and adding a torque command signal to the conditions) or by adjusting the condition values by adding a predetermined difference. Next, the extraction rule evaluation unit 15 calculates an evaluation value for the new data extraction rule candidate after this change or adjustment. Then, if a data extraction rule candidate with a higher evaluation value than the initial data extraction rule candidate is obtained, the new data extraction rule candidate is adopted. The extraction rule optimization unit 16 repeats the above process until a data extraction rule candidate with a higher evaluation value is obtained. Then, when the data extraction rule candidate with the highest evaluation value is obtained, the extraction rule optimization unit 16 considers the data extraction rule candidate as the optimized data extraction rule candidate and outputs it.

[0039] The simplest example of adjusting a data extraction rule is to create new data extraction rule candidates by adjusting each condition included in the data extraction rule candidate by a predetermined amount and then evaluate their suitability. For example, for the data extraction rule candidates illustrated in FIG. 7, a data extraction rule is created in which the torque command is changed in increments of 1%, the spindle rotation speed is changed in increments of 10 rpm, and the Z-axis position is changed in increments of 1 mm. Then, for each data extraction rule, the suitability of the extraction range acquired by the extraction range acquisition unit 13 is determined. As a result, the most suitable extraction rule is adopted as the next data extraction rule candidate. The extraction rule optimization unit 16 repeats this process and finally considers the most suitable data extraction rule candidate as the optimized data extraction rule. FIG. 8 shows an example of data extraction rule optimization. 8, the extraction rule optimization unit 16 starts a search from the data extraction rule candidate A created by the extraction rule candidate creation unit 14, and creates data extraction rule candidates B and C by changing the respective condition values. Then, the compatibility of each is determined, and the most compatible extraction rule is set as the optimized data extraction rule.

[0040] More efficient optimization algorithms for data extraction rules include well-known algorithms such as GP-EI (Bayesian Optimization-Evolutionary Strategy) and CMA-ES (Covariance Matrix Adaptive Evolutionary Strategy). GP-EI (GP-EI) (I) sets a search space containing data to be used as a data extraction rule, randomly selects a rule from the search space, and predicts an objective function using a Gaussian process. (II) Searches for a point where the acquisition function (e.g., the expected improvement (EI)) is maximized. (III) Calculates the similarity between each condition and the selected region and evaluates the obtained point. The extraction rule optimization unit 16 repeats steps (I) to (III) to search for a data extraction rule that best fits the extraction range acquired by the extraction range acquisition unit 13.

[0041] The data extraction rule optimized by the extraction rule optimization unit 16 is output by the output unit 17. The output unit 17 may display and output the optimized extraction rule on a display device included in the data extraction rule creation device 100. Alternatively, the output unit 17 may output the optimized extraction rule to an internal storage device or an external storage device included in the data extraction rule creation device 100. Alternatively, the output unit 17 may output the optimized extraction rule to the data extraction device 200 via a wired / wireless network (not shown) or the like.

[0042] Next, the data extraction device 200 will be described. The data extraction device 200 includes an extraction rule storage unit 21, a data acquisition unit 22, and a data extraction unit 24.

[0043] The extraction rule storage unit 21 stores the data extraction rules created by the data extraction rule creation device 100. The data extraction rule of the second disclosure is a combination of first data and second data. The data acquisition unit 11 acquires the first data and the second data. The data extraction unit 24 determines whether the data acquired by the data acquisition unit 11 conforms to the data extraction rule, and if the time-series data conforms to the data extraction rule, calculates an extraction range and extracts the first data and the second data.

[0044] The data extraction device 200 can automatically extract not only the first data but also the valid range of the second data related to the first data.

[0045] The operation of the data extraction system 1000 of the second disclosure will be described with reference to FIGS. 9 is a flowchart showing a schematic operation example of the data extraction rule creating device 100. The data extraction rule creating device 100 acquires first data and second data (step S21). Next, the data extraction rule creating device 100 acquires an extraction range for the first data (step S22). The extraction range may be determined manually or automatically. The data extraction rule creating device 100 acquires the extraction range of the second data based on the extraction range of the first data (step S23).

[0046] The data extraction rule creation device 100 creates a data extraction rule candidate including conditions for the first data and the second data (step S24). The data extraction rule creation device 100 calculates a data extraction range from the data extraction rule candidate (step S25). The data extraction rule creation device 100 calculates the compatibility between the calculated extraction range and the extraction ranges acquired in steps S22 and S23 as an evaluation value, and judges the evaluation value (step S26).

[0047] If the evaluation value exceeds a predetermined threshold (step S26; appropriate), the data extraction rule creation device 100 considers the data extraction rule candidate to be the optimal data extraction rule and ends the process. If the evaluation value is less than the predetermined threshold (step S26; inappropriate), the data extraction rule creation device 100 adjusts the data extraction rule candidate (step S27), and then proceeds to step S25 to repeat the process. The data extraction rule creation device 100 repeats the adjustment of the data extraction rule candidate and creates a data extraction rule that is more suitable for the extraction range.

[0048] 10 is a flowchart showing a schematic operation example of the data extraction device 200. The data extraction device 200 stores the data extraction rules created by the data extraction rule creating device 100 (step S31). The data extraction device 200 acquires first data and second data, which are time-series data to be extracted (step S32). The data extraction device 200 extracts data from the first data and second data in accordance with the acquired time-series data according to the data extraction rules (step S33).

[0049] As described above, in the data extraction system disclosed in the second disclosure, when a range for extracting first data is first manually or automatically specified, extraction rules for the first data and second data can be automatically created. The data extraction device can extract a range effective for analyzing the first data and second data in accordance with the created data extraction rules.

[0050] The hardware configuration of the data extraction rule creating device 100 will be described with reference to FIG. The data extraction rule creation device 100 is an information device such as a server installed inside or outside a factory, or a PC (personal computer) that monitors the state of the factory. The CPU 111 provided in the data extraction rule creating device 100 is a processor that controls the entire data extraction rule creating device 100. The CPU 111 reads out a system program processed in the ROM 112 via the bus, and controls the entire data extraction rule creating device 100 in accordance with the system program. The RAM 113 temporarily stores temporary calculation data, display data, various data input by the user via the input unit 71, etc.

[0051] The display unit 70 is a monitor or the like attached to the data extraction rule creating device 100. The display unit 70 displays an operation screen, a setting screen, and the like of the data extraction rule creating device 100.

[0052] The input unit 71 is a keyboard, a touch panel, or the like that is integrated with the display unit 70 or is separate from the display unit 70. The user operates the input unit 71 to input data to the screen displayed on the display unit 70. The display unit 70 and the input unit 71 may be mobile terminals.

[0053] The nonvolatile memory 114 is a memory that is backed up by, for example, a battery (not shown) and retains its stored state even when the power to the data extraction rule creating device 100 is turned off. The nonvolatile memory 114 stores programs read from external devices via an interface (not shown), programs input via the input unit 71, and various data acquired from various parts of the data extraction rule creating device 100 and from sensors in the factory. The programs and various data stored in the nonvolatile memory 114 may be expanded into the RAM 113 when executed / used. Furthermore, various system programs are written in the ROM 112 in advance.

[0054] The data extraction device 200 also has the same hardware configuration as the data extraction rule creation device 100. The data extraction device 200 is an information processing device including a CPU, a RAM, and the like.

[0055] [Third Disclosure] The third disclosure will be described. The data extraction system 1000 of the third disclosure is a system in which the components of the data extraction system 1000 of the second disclosure are distributed in a factory system. FIG. 12 shows an example of the factory system. The factory system is composed of an edge, a fog, and a cloud. The edge is an area close to sensors and control devices in terms of the network. Edge computing performs real-time processing of large volumes of data obtained from numerous sensors installed in factories and from the control devices of factory equipment and machinery. The cloud is not an internal hard drive or local server, but an external system connected via the internet. Cloud computing involves collecting large amounts of data at the edge, storing it in the cloud, and analyzing and interpreting the data. The fog is located between the cloud and the edge in terms of the network. The fog may perform data processing that would normally be performed in the cloud. Because the data is processed before being sent over the Internet, it can respond quickly to changes in the environment.

[0056] In the data cutout system 1000 of the third disclosure, a data acquisition unit 11, a cutout range acquisition unit 13, a cutout rule candidate creation unit 14, a cutout rule evaluation unit 15, a cutout rule optimization unit 16, and an output unit 17 are implemented on the cloud. A cutout rule storage unit 21, a data acquisition unit 22, and a data cutout unit 24 are implemented on the edge or fog. Specifically, the cutout rule storage unit 21, the data acquisition unit 22, and the data cutout unit 24 are implemented on an information processing device (e.g., a numerical control device, a PLC) close to the edge in terms of the network, or on an information processing device (e.g., a gateway) close to the fog in terms of the network.

[0057] In the data extraction system 1000 disclosed in the third embodiment, data extraction rules are optimized on the cloud. At the edge or fog, data is extracted using the data extraction rules. In the data extraction system, data extracted at the edge or fog is transmitted to the cloud or recorded in a storage device, thereby reducing the storage capacity and communication load of the storage device.

[0058] Furthermore, as shown in Figure 13, the on-site operator can also specify the extraction range of the first data. The extraction range and data extraction rules of the second data output from the cloud can be confirmed on an edge or fog terminal. By extracting data, it becomes possible to send and receive only the necessary data over the network, allowing high-load processing to be performed on the upper information processing device, while lower-load processing such as specifying the extraction range and checking the calculation results can be performed on the lower terminal. [Explanation of symbols]

[0059] 1000 Data Extraction System 100 Data extraction rule creation device 11 Data Acquisition Section 13 Extraction range acquisition section 14 Extraction rule candidate creation section 15 Extraction rule evaluation part 16 Extraction rule optimization part 17 Output section 200 Data Extraction Device 21 Extraction rule memory 22 Data Acquisition Section 24 Data extraction section 111 CPU 112 ROM 113 RAM 114 Non-volatile memory

Claims

1. a data acquisition unit that acquires first data indicating a state of the factory; an extraction range acquisition unit that acquires an extraction range of the first data; an extraction rule candidate creation unit that creates a data extraction rule candidate including a condition for extracting the first data based on the extraction range of the first data; a cutout rule evaluation unit that calculates an evaluation value of the data cutout rule candidate; a segmentation rule optimization unit that generates a data segmentation rule by optimizing the data segmentation rule candidate based on the evaluation value calculated by the segmentation rule evaluation unit; A data extraction rule creation device comprising:

2. the data acquisition unit acquires, in addition to the first data, second data different from the first data; the cut-out range acquisition unit acquires a cut-out range of the second data based on the cut-out range of the first data; the extraction rule candidate creation unit creates a data extraction rule candidate including a condition for extracting the first data and the second data based on extraction ranges of the first data and the second data; the extraction rule evaluation unit calculates an evaluation value of the data extraction rule candidate, the extraction rule optimization unit creates a data extraction rule by optimizing the data extraction rule candidate based on the evaluation value calculated by the extraction rule evaluation unit. The data extraction rule creating device according to claim 1 .

3. the cut-out range acquisition unit acquires the cut-out range of the first data based on a deviation from normal data; 2. The data extraction rule creating device according to claim 1.

4. the cut-out range acquisition unit accepts manual designation of a cut-out range and acquires the cut-out range of the first data; 2. The data extraction rule creating device according to claim 1.

5. acquiring first data indicative of a factory condition; acquiring a cutout range of the first data; creating a data extraction rule candidate including a condition for extracting the first data based on the extraction range of the first data; calculating an evaluation value of the candidate data segmentation rule; creating a data segmentation rule that optimizes the data segmentation rule candidate based on the evaluation value; A method for creating data extraction rules that execute processing on a computer.

6. When executed by one or more processors, acquiring first data indicative of a factory condition; acquiring a cutout range of the first data; creating a data extraction rule candidate including a condition for extracting the first data based on the extraction range of the first data; calculating an evaluation value of the candidate data segmentation rule; creating a data segmentation rule that optimizes the data segmentation rule candidate based on the evaluation value; A storage medium that stores computer-readable instructions.

Images