Industrial data analysis system and control method thereof

Abstract

Disclosed are an industrial data analysis system and a control method thereof. The control method of the industrial data analysis system of the present invention comprises the steps of: collecting industrial data; and extracting an analysis result according to a preconfigured algorithm by using the collected industrial data, and the method further comprises, before the step of extracting, a step of, when data of an item not included in the industrial data is required to extract the analysis result, automatically adding the corresponding item to the industrial data, and then receiving corresponding item data from an external server by using a correlation between the added item and a pre-existing item to add the item data to the industrial data.

Description

Industrial data analysis system and its control method

[0001] The present invention relates to an industrial data analysis system and a control method thereof, and more particularly, to a system for extracting highly useful analysis results for industrial data by field and a control method thereof.

[0002] Recent changes in the IT environment have accelerated the transition to digital technology in the industrial sector. However, compared to many competing countries, Korea's performance in analyzing and utilizing industrial data is poor due to a lack of relevant specialized personnel and data.

[0003] However, according to the 'SME Strategic Technology Roadmap 2020-2022' published by the Small and Medium Business Information Promotion Agency (TIPA), the number of big data-based decision-making systems in each industry is expected to increase in line with the growth of the big data market.

[0004] However, small and medium-sized enterprises face many obstacles in utilizing industrial data, such as a lack of human resources to collect, refine, and analyze data, and the high cost of building the necessary systems.

[0005] Additionally, industrial data may not always be fully available with all the necessary data, in which case a significant amount of data may go unused and become lost.

[0006] There is a need for a method to maximize the utilization of data that is being discarded like this.

[0007] Accordingly, the present invention has been devised to meet the above-mentioned conventional demands, and its purpose is to provide an industrial data analysis system and a control method thereof that maximizes the utilization of limited industrial data and minimizes wasted industrial data.

[0008] In order to achieve the above-described purpose, a control method of an industrial data analysis system according to the present invention includes a step of collecting industrial data; and a step of extracting analysis results according to a preset algorithm using the collected industrial data, and before the step of, if data of an item not included in the industrial data is required for the extraction of the analysis results, the control method may further include a step of automatically adding a corresponding item to the industrial data, and then receiving corresponding item data from an external server using a correlation between the added item and an existing item and adding the received corresponding item data to the industrial data.

[0009] Here, analysis results can be extracted using an artificial intelligence module built through machine learning.

[0010] Here, the above may include a step of selecting an analysis artificial intelligence module by inputting the item names included in the industrial data into a primary artificial intelligence module; and a step of extracting analysis results by inputting data for each item of the industrial data into the selected artificial intelligence module.

[0011] Here, if there is a missing value in the item data of the industrial data, a step of determining similarity using items other than the item with the missing value and inferring the missing value using the determined similarity may be further included.

[0012] Here, a step of performing machine learning for constructing the artificial intelligence module is further included, and when the data for machine learning is less than a preset number, a change rate can be determined based on the minimum and maximum value ranges for each item, and data for machine learning reflecting the determined change rate can be generated.

[0013] In addition, in order to achieve the above-described purpose, the industrial data analysis system according to the present invention includes an industrial data collection unit that collects industrial data; an analysis result extraction unit that extracts analysis results according to a preset algorithm using the industrial data collected by the industrial data collection unit; and, when data of an item not included in the industrial data is required for extraction of the analysis result of the analysis result extraction unit, the system may further include an item generation unit that automatically adds a corresponding item to the industrial data, and then receives corresponding item data from an external source using a correlation between the added item and an existing item and adds the data to the industrial data.

[0014] Here, the analysis result extraction unit can extract the analysis result using an artificial intelligence module built through machine learning.

[0015] Here, the analysis result extraction unit can input the item names included in the industrial data into the first artificial intelligence module to select an analysis artificial intelligence module, and then input data for each item of the industrial data into the selected artificial intelligence module to extract the analysis results.

[0016] Here, if there is a missing value in the item data of the above industrial data, a missing value correction unit may be further included to determine similarity using items other than the item with the missing value, and to infer the missing value using the determined similarity.

[0017] Here, the artificial intelligence module generation unit further includes an artificial intelligence module generation unit that performs machine learning for building the artificial intelligence module, and the artificial intelligence module generation unit determines a change rate based on a minimum and maximum value range for each item when the number of data for machine learning is less than or equal to a preset number, and can generate data for machine learning reflecting the determined change rate.

[0018] Figure 1 is a schematic diagram of the entire system including an industrial data analysis system according to one embodiment of the present invention.

[0019] Figure 2 is a functional block diagram of the industrial data analysis system of Figure 1.

[0020] Figures 3 and 4 are control flow diagrams of the industrial data analysis system of Figure 1.

[0021] Hereinafter, the present invention will be described in detail with reference to the attached drawings.

[0022] The following embodiments of the present invention are merely examples intended to aid understanding of the present invention, and the present invention is not limited to these embodiments. In particular, the present invention may be configured by a combination of at least one of the individual components, individual functions, or individual steps included in each embodiment.

[0023] In particular, for convenience, some claims in the claims include alphabets such as '(a)', but these alphabets do not specify the order of each step.

[0024] A schematic configuration of the entire system including an industrial data analysis system (100) according to one embodiment of the present invention is as illustrated in FIG. 1.

[0025] As shown in the drawing, it can be configured to include a user terminal (400), an external server (300), and an industrial data analysis system (100).

[0026] Here, the user terminal (400) is a terminal operated by a user who uses industrial data, and may be, for example, a laptop or personal computer provided at home, or a mobile communication terminal.

[0027] The industrial data provision server (200) provides industrial data and performs the function of providing industrial data upon request from the industrial data analysis system (100).

[0028] The external server (300) is distinct from the industrial data provision server (200), and is a server that provides necessary data upon request from the industrial data analysis system (100). For example, it may be a specific server provided on the Internet, such as a website.

[0029] The industrial data analysis system (100) analyzes certain industrial data to derive necessary results, and in particular, can extract such analysis results by having an artificial intelligence module.

[0030] An example of a specific functional block of this industrial data analysis system (100) is as shown in FIG. 2.

[0031] As illustrated in the drawing, an industrial data analysis system (100) according to one embodiment of the present invention may be configured to include an industrial data collection unit (110), an analysis result extraction unit (120), an item generation unit (130), a missing value correction unit (140), and an artificial intelligence module generation unit (150).

[0032] Here, the industrial data collection unit (110) performs the function of collecting industrial data.

[0033] For example, the industrial data collection unit (110) may collect industrial data by requesting and receiving necessary industrial data from the industrial data provision server (200), or may collect industrial data by reading out data stored in a pre-existing database (not shown).

[0034] The item creation unit (130) performs the function of adding a new item within industrial data and receiving data of the item from the outside and reflecting it in the industrial data.

[0035] That is, as described below, the analysis result extraction unit (120) can extract analysis results desired by users, etc., using industrial data. In this case, if the item data required for extracting the analysis results is not included in the collected industrial data, the item creation unit (130) can automatically add corresponding items to the industrial data, and then receive corresponding item data from the outside using the correlation between the added items and existing items, and then add it to the industrial data.

[0036] More specific examples of this will be given later.

[0037] The missing value correction unit (140) performs an imputation processing function for missing values ​​when missing values ​​exist in the industrial data item data. In this embodiment, 'missing value' means that a value does not exist.

[0038] Specifically, the missing value correction unit (140) can determine similarity using items other than items with missing values, and infer missing values ​​using the determined similarity. A more specific example of this will be described later.

[0039] The analysis result extraction unit (120) performs the function of extracting analysis results according to a preset algorithm using industrial data collected by the industrial data collection unit (110).

[0040] The algorithm set here may be an algorithm for the final analysis results, or it may be an algorithm related to preprocessing performed before using the artificial intelligence module, as described later.

[0041] An example of the processing process of the analysis result extraction unit (120) is as follows.

[0042] If the industrial data is industrial data related to waste discharge, including, for example, 'residential apartment address', 'waste discharge amount by resident', 'discharge date', etc., the analysis result extraction unit (120) can extract analysis results such as monthly average discharge amount, seasonal lifetime discharge amount, daily / time-based discharge amount, and discharge cycle according to a preset algorithm.

[0043] However, when an analysis of emissions according to residential area is requested by a user, the analysis result extraction unit (120) extracts the analysis result, and since there is no information on residential area in the previously collected industrial data, in this case, the item creation unit (130) described above adds the necessary items and data.

[0044] For example, after adding an item called 'residential area' to the industrial data, the item creation unit (130) can obtain necessary data from an external server (300) by using the correlation between the existing 'residential apartment address' and 'residential area'.

[0045] For example, if data matching 'apartment address' and 'average area' exists in the external server (300), the item creation unit (130) can obtain the average area corresponding to the residential apartment address in the industrial data from the external server (300) and then reflect it in the industrial data.

[0046] Accordingly, industrial data can be updated with data including ‘residential apartment address’, ‘amount of waste discharged by resident’, ‘discharge date’, and ‘residential area’.

[0047] In addition, referring to these examples, the function of the missing value correction unit (140) mentioned above is as follows.

[0048] If data for the 'waste discharge per resident' item of a specific household is missing among the items of industrial data, similar households can be selected based on the remaining data, and the average value of the 'waste discharge per resident' of those households can be estimated as the waste discharge of the missing household.

[0049] In particular, the analysis result extraction unit (120) can extract analysis results using an artificial intelligence module built through machine learning, and there may be multiple such artificial intelligence modules.

[0050] For example, if an artificial intelligence module is built for each type of industrial data, the analysis result extraction unit (120) can select an appropriate artificial intelligence module for each type of industrial data and extract the necessary analysis results.

[0051] Specifically, the analysis result extraction unit (120) can input the item names included in the industrial data into the first artificial intelligence module to select an analysis artificial intelligence module, and input data for each item of the industrial data into the selected artificial intelligence module to extract the analysis results.

[0052] That is, the analysis result extraction unit (120) can extract the final analysis result using a two-stage artificial intelligence module.

[0053] Industrial data may have different collected items depending on the type, and there may also be artificial intelligence modules specialized in industrial data analysis. In the present invention, by using an appropriate artificial intelligence module for each industrial data, the accuracy of the analysis results can be increased.

[0054] To this end, a process of generating an artificial intelligence module by performing machine learning using industrial data is required in advance, and this function is processed by the artificial intelligence module generation unit (150).

[0055] That is, the artificial intelligence module generation unit (150) performs machine learning for building an artificial intelligence module. At this time, if there is insufficient data for machine learning, for example, if the data for machine learning is less than a preset number, the change rate can be determined based on the minimum and maximum value ranges for each item, and data for machine learning reflecting the determined change rate can be generated.

[0056] For example, if industrial data is composed of 'waste amount' and 'waste disposal time', and the analysis of each data shows that the maximum value of 'waste amount' is 5.5, the minimum value is 0.5, the maximum value of 'waste disposal time' is 22:30, and the minimum value is 05:30, the change rate can be determined based on the difference between the maximum and minimum values ​​of each item.

[0057] For example, for 'waste discharge amount', new data can be created by adding 0.5 to each existing data, and for 'waste discharge time', new data can be created by adding 1 hour to each existing data.

[0058] Here, the machine learning data additionally generated by the artificial intelligence module generation unit (150) is generated by reflecting the change ratio described above based on the existing data. For example, if there is data in which the amount of waste discharged is 1 and the time of waste discharge is 19:50, the artificial intelligence module generation unit (150) can generate new data in which the amount of waste discharged is 1.5 and the time of waste discharge is 20:50 by reflecting the change ratio calculated previously.

[0059] After new machine learning data is generated in this way, the artificial intelligence module generation unit (150) can perform machine learning.

[0060] Hereinafter, the overall control flow of the industrial data analysis system (100) according to one embodiment of the present invention will be described with reference to FIGS. 3 and 4.

[0061] First, referring to Figure 3, the following is provided.

[0062] The industrial data analysis system (100) collects industrial data (step S1).

[0063] For example, industrial data can be collected by requesting it to an industrial data provision server (200).

[0064] The industrial data analysis system (100) adds new items and corresponding data to the industrial data (step S5) when new items and corresponding data are required in addition to the collected industrial data (i.e., this can be confirmed through a preliminary check of items required for extracting analysis results) (step S3).

[0065] Next, the industrial data analysis system (100) performs correlation analysis of industrial data (step S7), and then requests and receives data of newly added items from an external server (300) (step S9).

[0066] The industrial data analysis system (100) reflects item data received from an external server (300) into industrial data (step S11) and extracts analysis results using the industrial data (step S13).

[0067] Next, referring to Fig. 4, the following is provided.

[0068] The industrial data analysis system (100) collects industrial data (step S21) and analyzes the collected industrial data to determine whether there are missing values ​​(step S23).

[0069] Missing values ​​here mean that some data for a specific item does not exist.

[0070] At this time, the industrial data analysis system (100) performs record-by-record similarity determination based on data stored in the industrial data (step S25).

[0071] The industrial data analysis system (100) infers missing values ​​based on this similarity determination (step S27).

[0072] For example, the industrial data analysis system (100) can compare data stored in each item where there are no missing values ​​by record, select other records similar to the records with missing values, and then infer missing values ​​using the values ​​of the other selected records.

[0073] Specifically, if records A, B, C, D, and E exist, and each record has items a, b, and c, and there is a missing value in item b of B, the industrial data analysis system (100) compares the remaining items a and c of each record with each other and selects other records similar to items a and c of B.

[0074] For example, if the selected records are A and E, the industrial data analysis system (100) can infer the b value of B by taking the average of the b item values ​​of A and the b item values ​​of E.

[0075] When data for missing values ​​is inferred in this way, the industrial data analysis system (100) updates the industrial data by replacing the missing values ​​with the inferred values ​​(step S29), and extracts analysis results using the industrial data thus updated (step S31).

[0076] Meanwhile, it goes without saying that the process of performing each of the above-described embodiments can be performed by a program or application stored in a predetermined recording medium (e.g., computer-readable). Here, the recording medium includes electronic recording media such as RAM (Random Access Memory), magnetic recording media such as hard disks, and optical recording media such as CDs (Compact Disks).

[0077] At this time, the program stored in the recording medium can be executed on hardware such as a computer or smartphone to perform each of the embodiments described above. In particular, at least one of the functional blocks of the industrial data analysis system according to the present invention described above can be implemented by such a program or application.

[0078] Furthermore, the present invention is not limited to the specific embodiments described above, and various modifications and variations can be made without departing from the spirit and scope of the invention. It will be apparent that such modifications and variations are included within the scope of the appended claims.

[0079] As described above, according to the present invention, more accurate analysis results can be extracted by analyzing industrial data using an appropriate algorithm or artificial intelligence module.

[0080] Additionally, if an item required for extracting analysis results does not exist in the previously collected industrial data, the item and corresponding data can be automatically collected to provide the user with the analysis results they desire, thereby increasing the usability of industrial data.

[0081] In particular, the amount of wasted industrial data can be minimized by allowing industrial data that cannot be used due to the presence of missing values ​​to be used through the missing value estimation process.

Claims

1. (a) Step of collecting industrial data; (b) a step of extracting analysis results according to a preset algorithm using the industrial data collected in step (a); (c) A control method for an industrial data analysis system, characterized in that the method further comprises a step of automatically adding a corresponding item to the industrial data, if data of an item not included in the industrial data is required for extracting the analysis result of the step (b), and then using the correlation between the added item and an existing item, receiving corresponding item data from an external server and adding the data to the industrial data.

2. In paragraph 1, A control method for an industrial data analysis system, characterized in that in the step (b) above, analysis results are extracted using an artificial intelligence module built by machine learning.

3. In paragraph 2, Step (b) above, (b1) a step of selecting an analysis artificial intelligence module by inputting the item names included in the above industrial data into the first artificial intelligence module; (b2) A control method for an industrial data analysis system, characterized by including a step of inputting data for each item of the industrial data into the artificial intelligence module selected in the step (b1) and extracting analysis results.

4. In paragraph 3, (d) A control method for an industrial data analysis system, characterized in that it further includes a step of determining similarity using items other than the items with missing values ​​when there are missing values ​​in the item data of the industrial data, and inferring the missing values ​​using the determined similarity.

5. In paragraph 2, (e) further comprising a step of performing machine learning for constructing the artificial intelligence module; A control method for an industrial data analysis system, characterized in that in the step (e) above, if the number of data for machine learning is less than a preset number, a change rate is determined based on a minimum and maximum value range for each item, and data for machine learning reflecting the determined change rate is generated.

6. A computer-readable recording medium recording a program for executing any one of the methods of clauses 1 to 5.

7. An application program stored in a computer-readable recording medium for executing any one of the methods of claims 1 to 5 in combination with hardware.

8. Industrial data collection department that collects industrial data; Includes an analysis result extraction unit that extracts analysis results according to a preset algorithm using the industrial data collected from the above industrial data collection unit, An industrial data analysis system characterized in that, when data of an item not included in the industrial data is required for extraction of the analysis result of the above analysis result extraction unit, the system further includes an item generation unit that automatically adds a corresponding item to the industrial data, and then receives corresponding item data from outside using the correlation between the added item and an existing item and adds it to the industrial data.

9. In paragraph 8, An industrial data analysis system characterized in that the above analysis result extraction unit extracts analysis results using an artificial intelligence module built through machine learning.

10. In paragraph 9, An industrial data analysis system characterized in that the above analysis result extraction unit inputs the item names included in the industrial data into the first artificial intelligence module to select an analysis artificial intelligence module, and then inputs data for each item of the industrial data into the selected artificial intelligence module to extract the analysis results.

11. In paragraph 10, An industrial data analysis system characterized by further including a missing value correction unit that, when there is a missing value in the item data of the above industrial data, determines the similarity using items other than the item with the missing value, and infers the missing value using the determined similarity.

12. In paragraph 9, Further comprising an artificial intelligence module generation unit that performs machine learning for constructing the artificial intelligence module; An industrial data analysis system characterized in that the artificial intelligence module generation unit determines a change rate based on a minimum and maximum value range for each item when the number of data for machine learning is less than a preset number, and generates data for machine learning reflecting the determined change rate.

Images