Information processing device and information processing method

Abstract

To propose an information processing device and an information processing method with which it is possible to select an optimum data model from a plurality of data models in accordance with the site of a factory and efficiently acquire site data.SOLUTION: A processor is configured to: receive site identification information for identifying a site from an input / output unit; calculate a first cost representing a cost at the site identified by the site identification information on the basis of site information on the site identified by the site identification information; select a specific data model corresponding to the site having a second cost similar to the first cost from among a plurality of data models on the basis of the first cost and related information; and output the selected specific data model to the input / output unit.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to an information processing apparatus and an information processing method, and is suitably applied to, for example, an information processing apparatus that collects data from a production facility of a factory using a plurality of data collecting devices arranged in the production facility of the factory. [Background technology]

[0002] In recent years, a large number of data collection devices are installed on the production lines of production facilities in factories, and on-site data is collected in real time using these data collection devices. The on-site data collected in this way is stored as a data lake on the platform. The platform is equipped with not only this data lake, but also a data model and an API.

[0003] When a user wants to use data collected at production facilities, the application operated by the user requests on-site data from the platform. In response to a request from the application, the API of the platform retrieves data from the data lake using a data model.

[0004] A publicly known example of a method for collecting data in production facilities is Patent Document 1. In Patent Document 1, the impact of acquiring on-site data on performance is comprehensively assessed according to the access pattern from an application, and then a decision is made as to whether or not to use a data model. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2011-154653 A Summary of the Invention [Problem to be solved by the invention]

[0006] However, in the method disclosed in Patent Document 1, a decision is made as to whether or not to use a data model by comprehensively assessing the impact on the performance of the production equipment based on the access pattern from the application. However, this decision does not take into account the characteristics of the on-site data to be acquired, and the data model used is not necessarily the optimal data model.

[0007] The present invention has been made in consideration of the above points, and aims to propose an information processing device and an information processing method that can select the most appropriate data model from among multiple data models according to the factory site and efficiently acquire site data. [Means for solving the problem]

[0008] In order to solve such problems, the present invention provides a memory having site data, which is information related to a site, costs, which are information representing the characteristics of the site, a plurality of data models defining the connection relationships between the site data, and associated information in which the costs for each site are associated with the plurality of data models, a processor, and an input / output unit, wherein the processor is provided with a cost calculation processing unit that receives site-specific information that identifies a site from the input / output unit and calculates a first cost, which is the cost of the site identified by the site-specific information, based on the site information of the site identified by the site-specific information, and a data model selection unit that selects a specific data model corresponding to a site having a second cost similar to the first cost from among the plurality of data models based on the first cost and the associated information, and outputs the selected specific data model to the input / output unit.

[0009] In addition, the present invention provides an information processing method in an information processing device equipped with a memory having site data, which is information regarding a site, costs, which are information representing the characteristics of the site, a plurality of data models defining the connection relationships between the site data, and associated information in which the costs for each site are associated with the plurality of data models, a processor, and an input / output unit, the method including a cost calculation processing step in which the processor receives site-specific information identifying a site from the input / output unit, and calculates a first cost, which is the cost of the site identified by the site-specific information, based on the site information of the site identified by the site-specific information, and a data model selection step in which the processor selects a specific data model corresponding to the site having a second cost similar to the first cost from among the plurality of data models based on the first cost and the associated information, and outputs the selected specific data model to the input / output unit.

[0010] According to the information processing device and method of the present invention, an optimal specific data model can be selected from among a plurality of data models according to the factory site so that the performance of the data collection equipment is optimized depending on the cost of the characteristics of the site data, and the site data can be efficiently acquired using the optimal specific data model. Effect of the Invention

[0011] According to the present invention, it is possible to realize an information processing device and an information processing method that can select an optimal data model from among a plurality of data models according to a factory site and efficiently acquire site data. [Brief description of the drawings]

[0012] [Figure 1] FIG. 1 is a diagram illustrating an example of a conceptual hardware configuration of a data acquisition system according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a diagram illustrating an example of a software configuration of the data acquisition system according to the present embodiment. [Diagram 3] 13 is a flowchart showing an example of a cost calculation process for a local site. [Figure 4] 13 is a flowchart illustrating an example of a past performance result and cost information acquisition process. [Diagram 5] FIG. 2 is a diagram showing an example of past performance record information in each factory. [Figure 6] FIG. 11 is a diagram showing an example of site cost information in each factory. [Figure 7] FIG. 13 is a diagram showing an example of site cost summary information. [Figure 8] FIG. 11 is a diagram illustrating an example of data model performance information. [Figure 9] FIG. 13 is a diagram illustrating an example of an estimated data model table. [Figure 10] 13 is a flowchart illustrating an example of a data model selection process. [Figure 11] FIG. 11 is a diagram showing an example of data for specifically explaining how a data model is selected in the data acquisition system according to the present embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] An embodiment of the present invention will now be described in detail with reference to the drawings.

[0014] 1 shows an example of a conceptual hardware configuration of a data acquisition system 100 according to this embodiment. The data acquisition system 100 includes a production line of production equipment in a factory 200, a platform 300, an application server 400, a terminal 500, and a network 600. This data acquisition system 100 is configured such that the factory 200, the platform 300, the application server 400, and the terminal 500 are connected to the network 600.

[0015] The factory 200 is equipped with an ERP (Enterprise Resource Planning), a PLM (Product Lifecycle Management), an MES (Manufacturing Execution System), and production equipment, and at least one data collection device such as a sensor (in this embodiment, multiple devices) is placed on the production line of the production equipment in a form known as IoT (Internet of Things).

[0016] In this embodiment, a sensor 201 is mainly exemplified as a data collection device. A plurality of sensors 201 are provided on a production line of production equipment in a factory 200, and have a function of collecting preset numerical values ​​on the production line. The numerical values ​​collected by the sensor 201 are output from the sensor 201 as on-site data.

[0017] The application server 400 runs a group of applications that collect and analyze on-site data from sensors 201 arranged on the production line of the production equipment of the factory 200 via the platform 300 .

[0018] The terminal 500 is a terminal operated by a user, and is a terminal device capable of operating the applications of the application server 400 .

[0019] The platform 300 includes at least one server or database server. The platform 300 is, for example, configured by a computer, and includes at least a memory, a processor, and an input / output unit (not shown). The memory includes site data, which is information about a site, a cost, which is information representing the characteristics of the site, a plurality of data models defining a connection relationship between the site data, and related information in which the cost for each site is associated with the plurality of data models. The site data corresponds to, for example, a production plan and performance record data of the production equipment. The related information corresponds to a site-cost-data model shown in FIG. 11(A) and FIG. 11(B) described later. The input / output unit receives site-specific information for identifying a site, and outputs a specific data model selected as described later. The server group accumulates site data from a sensor 201 installed in a production line of the production equipment of the factory 200 in a format using a data model. The server group is provided with a plurality of types of data models. The site data may be accumulated in a database server. Details of the platform 300 will be described later.

[0020] The platform 300 and the application server 400 may adopt a so-called on-premise format in which they are managed in a company's own environment, or may adopt a so-called cloud format in which the system is used within resources provided by a vendor.

[0021] Fig. 2 shows an example of the software configuration of the data acquisition system 100 according to this embodiment. In Fig. 1, it is assumed that there are multiple factories, but in Fig. 2 and subsequent figures, unless it is necessary to mention multiple factories, the factory 2 will be mainly illustrated and described.

[0022] The data acquisition system 100 mainly comprises a plurality of sensors 201 provided on a production line of a production facility in a factory 200, a platform 300, and an application server 400, all connected to a network 600.

[0023] As described above, a plurality of sensors 201 serving as a plurality of data collecting devices for collecting on-site data of each piece of production equipment are provided in the production line of the production equipment in the factory 200. The sensor 201 collects values ​​relating to preset measurement contents in the production line of the production equipment in operation, and outputs the on-site data.

[0024] The platform 300 is an example of an information processing device, and includes at least a plurality of data model groups 320, a cost calculation processing unit 350, each site cost information 380, and a data model selection unit 330, and preferably further includes a data lake 310, a data acquisition API group 340, production plan information 360, and past performance information 390.

[0025] The data lake 310 accumulates on-site data collected by each sensor 201 installed on the production line of the production facility of the factory 200. Note that the on-site data may be accumulated in a database (not shown) instead of the data lake 310.

[0026] The data model group 320 includes a first data model 321, a second data model 322, a third data model 323, etc. The first data model 321, the second data model 322, the third data model 323, etc. are data models related to performance records of a plurality of sensors 201 provided on a production line of production equipment in a factory 200. In the following embodiment, a description will be given mainly assuming that three data models, the first data model 321, the second data model 322, and the third data model 323, exist in the platform 300.

[0027] Each data model in the data model group 320 may adopt, for example, a structure such as a so-called ER (Entity Relationship) diagram connected by keys, a graph structure, a tree structure, or a network structure. The platform 300 may have a plurality of data models with various structures, and may be in a form capable of selecting a specific data model from among the plurality of data models so as to efficiently provide data in response to a request from the application server 400 by utilizing various costs.

[0028] Here, in this embodiment, "cost" means an index for determining which data model should be used for analysis depending on the scale (amount) and complexity of data that can be obtained from the site, that is, an index that quantifies data characteristics that are determined depending on the scale (amount) and complexity of the data. In this embodiment, examples of such costs include actual cost, planned cost, and site cost. The actual cost includes rework cost, loop cost, and operation order cost, and the planned cost includes bill of materials (BOM) cost and bill of process (BOP) cost.

[0029] Each site cost information 380 is information relating to site costs indicating the scale of production equipment in the production line of the production equipment of the factory 200 and the degree of difficulty in producing the finished product.

[0030] The past performance information 390 is information relating to past performance results that indicates the performance results when on-site data was acquired in the past by each sensor 201 provided on the production line of the production facility of the factory 200 .

[0031] The production plan information 360 is information related to a production plan for finished products and their parts to be produced on the production lines of the production facilities of the factory 200 .

[0032] The cost calculation processing unit 350 calculates a first cost (hereinafter also abbreviated as "cost"), which is the cost of the site specified by the site identification information, based on the site information of the site specified by the site identification information received from the input / output unit. This cost calculation processing unit 350 calculates an actual cost based on the performance actual data of the production equipment as a characteristic of the site data. Specifically, the cost calculation processing unit 350 calculates a cost based on information on the production plan based on the production plan information 360 and the data of each sensor 201 accumulated in the data lake 310, and generates site cost information 370 as a cost representing the scale and production difficulty of the production equipment for each production line of the production equipment of the factory 200. The above processing by the cost calculation processing unit 350 may be executed by the data model selection unit 330, or the data model selection unit 330 may cause the cost calculation processing unit 350 to execute it.

[0033] In this embodiment, "cost" affects performance because on-site data at production facilities differs depending on the process and amount of work at the production facilities, the amount of products handled, and the complexity of the work. For example, in a factory, even if the same data model is used, the work results, etc., will be significantly different between mass-producing one type of product and small-scale manufacturing of many different types of products.

[0034] In this embodiment, since the more complex the production equipment, the greater the impact on performance, the cost that may affect the performance of each piece of production equipment is calculated. That is, this embodiment is characterized in that the data model to be selected is flexibly changed depending on the cost (of the characteristics of the on-site data) due to differences specific to the production equipment.

[0035] The data model selection unit 330 selects a specific data model corresponding to a site having another cost (second cost) similar to the cost (first cost) calculated by the cost calculation processing unit 350 based on the cost (first cost) and related information from among the multiple data models, and outputs the selected specific data model to the input / output unit. The data model selection unit 330 refers to the data model selection table 331 according to the past performance results of the multiple sensors 201 and the cost of the characteristics of the site data of the production equipment, and selects an optimal specific data model from the multiple data models so that the performance of the multiple sensors 201 is optimized. When the specific data model selected in this way is optimal, the performance results of the site data from the sensor 201a arranged on the production line of the production equipment can be good.

[0036] The data model selection unit 330 calculates the planned cost based on the production plan of the production equipment as the characteristics of the site data. The data model selection unit 330 matches the planned cost and actual cost of the new site by combining these planned costs with the actual cost of the production equipment, and determines the data model to be selected for the new production equipment by referring to the performance actual data of the site with the closest combination.

[0037] The data model selection unit 330 calculates the bill of material (BOM) cost and bill of process (BOP) cost as planned costs from the bill of material (BOM) and bill of process (BOP) of the production equipment. For a BOM, the more intermediate products and materials there are for a product, the more complex it is considered to be. It is said that BOM cost = Σ number of terminal leaves. The number of terminal leaves indicates the number of material items of the intermediate products that make up the finished product.

[0038] Regarding BOP, the more processes there are in the production line of the production equipment, and the deeper the process hierarchy, the more complex the cost becomes. BOP cost = Σ hierarchy x number of processes. The number of processes refers to the number of downstream processes that make up the upstream process.

[0039] In this embodiment, not only the production plan is acquired, but also on-site resource information (for example, the base CPU performance of the platform, etc.) may be acquired. This is because the performance regarding on-site data acquisition differs between a site using a high-performance computer and a site using a low-performance computer.

[0040] The data acquisition API group 340 includes a number of APIs (Application Programming Interfaces) that collect field data from the sensor 201 using the data model selected by the data model selection unit 330. The data acquisition API group 340 is an example of an API providing unit, and provides a plurality of applications with APIs (Application Programming Interfaces) for utilizing field data collected by the sensor 201 using a specific data model selected by the data model selection unit 330. The data acquisition API group 340 measures the performance of each sensor 201, for example, from the start time to the completion time of execution using an API included in the data acquisition API group 340, and collects the performance results of each sensor 201. In this way, it can be detected that the amount of data accumulated in the factory 200 has increased and the previous data model is no longer able to provide sufficient performance.

[0041] Furthermore, when sufficient performance cannot be ensured, the data acquisition API group 340 performs a process of selecting a data model capable of ensuring better performance from performance information of other factories.

[0042] The data model selection unit 330 calculates the bill of material (BOM) cost and the bill of process (BOP) cost as planned costs from the bill of material (BOM) and the bill of process (BOP) of the production equipment.

[0043] The data model selection unit 330 extracts rework, loops, and work order changes performed in the production equipment from the past performance records of the data collection equipment, and calculates the rework cost, loop cost, and work order cost as the respective performance costs.

[0044] The application server 400 includes a first application 410, a second application 420, a third application 430, and the like, as a plurality of applications that utilize on-site data acquired by a plurality of sensors 201 using a data model via a data acquisition API group 340. Hereinafter, in this embodiment, the application server 400 will be described as including the first application 410, the second application 420, and the third application 430.

[0045] The data acquisition system 100 according to the present embodiment has the above-mentioned configuration, and an example of the operation of the data acquisition system 100 will be described below. First, to give an overview, the present embodiment relates to a data acquisition method for collecting on-site data by the sensor 201 using a data model related to the performance record of the sensor 201 that collects on-site data of a production facility in a factory, and includes a data model selection step in which a data model selection unit 330 selects a specific data model from a plurality of data models so that the performance of the sensor 201 is optimized according to the past performance record of the sensor 201 and the cost of the characteristics of the on-site data, and an API provision step in which an API (Application Programming Interface) for using the on-site data collected by the sensor 201 using the specific data model selected by the data model selection unit 330 is provided to a plurality of applications 410 via a data acquisition API group 340. Next, cost calculation will be described.

[0046] 3 is a flow chart showing an example of a cost calculation process for the local site. The local production facility referred to here is, for example, the production facility of the factory 200 described above.

[0047] In step S101, the cost calculation processing unit 350 in the platform 300 acquires production plan information 360 of the local site. This production plan information 360 includes BOM information and BOP information. Next, in step S102, the cost calculation processing unit 350 calculates the BOM cost by analyzing the complexity of the number of parts from the BOM information, and calculates the BOP cost by analyzing the complexity of the process from the BOP information.

[0048] Next, in step S103, the cost calculation processing unit 350 analyzes the time-series performance record data of the own site, detects the number of reworks, the number of loops, and the amount of change in the work order from the work order of the same lot No., and calculates the rework cost, the loop cost, and the cost of the amount of change in the work order. Next, in step S104, the cost calculation processing unit 350 records the calculation results as site cost information 370 corresponding to the own site.

[0049] Here, "rework" refers to performing work again on a rejected product that results from a certain operation. Rework cost = Σ number of operations × number of times. The number of analysis lines will increase for the amount of rework. "Loop" refers to the same operation being performed repeatedly in a certain operation. Loop cost = Σ number of operations × number of times. The number of analysis lines will increase for the amount of loop. "Work order change" refers to changing the order of multiple consecutive operations. Work order cost = Σ master difference × number of times. Note that master difference refers to the difference from the standard work order. Performance analysis such as that of a job shop model incurs costs.

[0050] Here, the timing for executing the cost calculation process for the site itself, which is the site that is about to select a certain data model and execute the API, may be periodically, such as once every few months. This timing may also be when there is a change in the work at the site, since it may be better to change the data model if the amount of data increases over time. This is because there is a possibility that the cost will change significantly if the work changes.

[0051] 4 is a flowchart showing an example of a past performance record and cost information acquisition process. First, in step S201, the data model selection unit 330 acquires (1) past performance record information (past performance record information 390) for each site and (2) site cost information 370 corresponding to the site. Next, in step 202, the data model selection unit 330 acquires (1)' performance record information (record (performance) information) for the own site in the own factory 200 and (2)' site cost information 370 corresponding to the site.

[0052] Next, in step S203, the data model selection unit 330 eliminates duplicates, merges (1) past performance actual information (past performance actual information 390) with (1)' performance actual information, and records this in the database as past performance actual collection information, and also merges (2) site cost information 370 with (2)' site cost information 370, and records this in the database as site cost collection information.

[0053] In this embodiment, past performance information and cost information may be acquired from a number of specified factories, or may be acquired from all factories. Also, past performance information and cost information of each site may be collected in one of the servers, and a portion of the information (data of factories with similar costs) may be acquired from that. Furthermore, past performance information and cost information of each site may be acquired periodically, or the presence or absence of updates may be checked and information on sites that have been updated may be acquired, or the information may be acquired at the required timing.

[0054] Fig. 5 is an example of past performance information for each factory. Here, sites 1 and 2 are shown as examples. The past performance information for site 1 manages the data model, performance results, and cost ID for each API No. For example, if the API No. is "A01", the data model is the "second data model", the performance results are "100ms", and the cost ID is "site 1_3".

[0055] Meanwhile, the past performance information for site 2 is managed by data model, performance results, and cost ID for each API No. For example, if the API No. is "A01", the data model is "third data model", the performance results are "500 ms", and the cost ID is "site 2_1".

[0056] Here, the API No. can be the API name such as "A01" or it can be a number that indicates the version such as "A01_0001". Note that even APIs with the same processing may have different performance depending on the version. Also, past performance information and on-site cost information are used as a set for each factory.

[0057] FIG. 6 is an example of site cost information for each factory. Here, sites 1 and 2 are shown as examples. The site cost information for site 1 manages the BOM cost, BOP cost, rework cost, loop cost, sub-operation cost, and date and time for each number (corresponding to the "#" in the figure). For example, if the number is "site 1_1", the BOM cost is "350", the BOP cost is "76", the rework cost is "25", the loop cost is "50", the sub-operation cost is "100", and the date and time is "February 15, 2023".

[0058] Here, the cost ID is assigned an ID that can identify each factory. This may be recalculated every few months, or may be recalculated and changed based on the process or work when the process or work changes.

[0059] Meanwhile, the site cost information for site 2 is managed by number (corresponding to the "#" in the illustration), with the BOM cost, BOP cost, rework cost, loop cost, sub-operation cost, and date and time. For example, if the number is "site 2_1", the BOM cost is "1000", the BOP cost is "320", the rework cost is "152", the loop cost is "410", the sub-operation cost is "780", and the date and time is "January 24, 2023". Here, the product being manufactured has changed from the number "site 2_2", and each cost has changed significantly.

[0060] Figure 7 is an example of site cost summary information. The site cost summary information manages BOM cost, BOP cost, rework cost, loop cost, and operation stage cost for each site. For example, if the number is "site 1_1", the BOM cost is "350", the BOP cost is "76", the rework cost is "25", the loop cost is "50", and the operation stage cost is "100".

[0061] Here, the following can be done using a table that collects the costs of each site. For example, the costs can be tallied at each site, or one of the servers can tallie the costs and each site can download the tallied table. Also, combinations of similar costs can be clustered and grouped together by average performance to reduce the total number of tables and the search load.

[0062] Figure 8 is an example of data model performance information. The data model performance information manages the data model, performance results, and cost ID for each API No. For example, if the API No. is "A01", the data model is "second data model", the performance results are "100ms", and the cost ID is "site 1_3".

[0063] Here, the data may be tabulated at each site, or one of the servers may tabulate the data and each site may download the tabulated table. Furthermore, when performance is measured at the same site and with the same data model, all data may be kept, or the data with the best performance may be recorded as the performance record, or the average value may be recorded.

[0064] Figure 9 shows an example of an estimated data model table. The estimated data model table manages the data model and estimated / measured values ​​for each API No. For example, if the API No. is "A01", the data model is the "second data model" and the estimated / measured value is "100 ms".

[0065] Here, the results of the estimation according to the cost at each site may be stored so that the estimation process does not have to be performed every time. Also, if the actual measurement value is worse than the estimated value by a certain percentage, a search is performed to see if there is a data model that can further improve the performance.

[0066] 10 is a flowchart showing an example of the data model selection process. This data model selection process is executed by the above-mentioned data model selection unit 330. In step S501, the data model selection unit 330 executes the data acquisition API group 340. Next, in step S502, the data model selection unit 330 acquires the site cost information 370 and the past performance result information 390.

[0067] Next, in step S503, the data model selection unit 330 checks whether there is an estimation result in a data model estimation table (not shown). That is, the data model selection unit 330 judges whether there is a data model estimation result corresponding to the usage site of the executed API in the data model estimation table and whether there is "no flag". Here, "no flag" means that the process does not proceed to selecting a data model with similar cost.

[0068] If the result in step S502 is positive, in step S504, the data model selection unit 330 selects a data model corresponding to the site of use and executes the API. Next, in step S505, the data model selection unit 330 assigns a flag if the API execution performance is reduced by a predetermined percentage from the previously recorded performance result. Here, assigning a flag means that when a data model corresponding to the site of use is selected but does not provide sufficient performance, the operation proceeds to selecting a data model with a similar cost.

[0069] Next, in step S509, the data model selection unit 330 records the cost and API execution performance result information.

[0070] On the other hand, if the result in step S503 is negative, in step S506, the data model selection unit 330 judges whether or not there is a cost combination similar to that of the target site. If there is a cost combination similar to that of the target site, in step S507, the data model selection unit 330 selects a specific data model with the best performance from past performance results for the site with a cost similar to that of the site cost aggregation information, and executes the API. Next, the data model selection unit 330 executes the above-mentioned step S509.

[0071] On the other hand, if there is no cost combination similar to that of the target site, in step S508, the data model selection unit 330 selects a data model by default or at random and executes the API. Next, the data model selection unit 330 executes the above-mentioned step S509.

[0072] 11(A) to 11(C) show examples of data (corresponding to "site-cost-data model" as an example of the related information) for specifically explaining how a data model is selected in the data acquisition system 100 according to this embodiment. Here, it is assumed that the platform 300 is equipped with a first data model 321 and a second data model 322, and is applied to a site N.

[0073] [Step A] In the platform 300, the data model selection unit 330 searches for a site cost close to the site N. For example, as shown in FIG. 11A, the site cost aggregation information of the site N is BOM cost "300", BOP cost "55", rework cost "20", loop cost "5", and operation sequence cost "30". In contrast, the site cost aggregation information of the site 1 is BOM cost "350", BOP cost "76", rework cost "25", loop cost "50", and operation sequence cost "100". The site cost aggregation information of the site 3_1 is BOM cost "400", BOP cost "60", rework cost "10", loop cost "5", and operation sequence cost "50".

[0074] The data model selection unit 330 compares the site cost collection information of the site N with the site cost collection information of the site 1, and also compares the site cost collection information of the site N with the site cost collection information of the site 3_1. Based on the comparison results, the data model selection unit 330 selects a specific data model from the multiple data model groups 320 so that the combination of BOM cost, BOP cost, rework cost, loop cost, and operation order cost is close to the site cost collection information of the site N. Specifically, the data model selection unit 330 selects a data model that is close in shape to the graph of each of these costs, for example. The data model selection unit 330 may also evaluate whether they are similar to each other by dividing the production plan and the actual results.

[0075] [Step B] Next, the data model selection unit 330 uses the past performance data to acquire site data of site 3 close to site N. For example, as shown in Fig. 11(B) , the data model selection unit 330 compares past performance result information in which API No. is "A01", the result is "500 ms", the site is "second data model", and the cost ID is "site 3_1", with past performance result information in which the result is "1000 ms", the site is "third data model", and the cost ID is "site 3_1", and since the "third data model" has better performance, it adopts the third data model 323 when executing the API with API No. "A01" at site N.

[0076] [Step C] Next, the data model selection unit 330 creates a data model and an estimated value to be selected for each API. For example, as shown in Fig. 11(C), the data model selection unit 330 creates an estimated data model with API No. "A01", data model "third data model", and estimated value "100 ms", an estimated data model with API No. "A02", data model "first data model", and estimated value "500 ms", and an estimated data model with API No. "A03", data model "second data model", and estimated value "1.2 sec".

[0077] Based on the estimated data model created in the above manner, the data model selection unit 330 can estimate which data model will have high performance when using the API at site N, and select a specific data model from the multiple data model groups 320.

[0078] As described above, in this embodiment, the platform 300 comprises a memory having site data, which is information related to the site, costs, which are information representing the characteristics of the site, a plurality of data models defining the connection relationships between the site data, and associated information in which the costs for each site are associated with the plurality of data models, a processor, and an input / output unit, and the processor has a function of receiving site-specific information that identifies a site from the input / output unit, and calculating a first cost, which is the cost of the site identified by the site-specific information, based on the site information of the site identified by the site-specific information (cost calculation processing unit 350), and a function of selecting a specific data model corresponding to a site having another cost (second cost) similar to the first cost from among the plurality of data models, based on the first cost and the associated information, and outputting the selected specific data model to the input / output unit (data model selection unit 330).

[0079] In addition, the information processing method according to this embodiment is an information processing method of a platform 300 including a memory having site data, which is information about a site, costs, which are information representing the characteristics of the site, a plurality of data models defining the connection relationships between the site data, and associated information in which the costs for each site are associated with the plurality of data models, a processor, and an input / output unit, and includes a cost calculation processing step in which the processor of the platform 300 (cost calculation processing unit 350) receives site-specific information that identifies a site from the input / output unit, and calculates a first cost, which is the cost of the site identified by the site-specific information, based on the site information of the site identified by the site-specific information, and a data model selection step in which the processor of the platform 300 (data model selection unit 330) selects a specific data model corresponding to the site having a second cost similar to the first cost from the plurality of data models based on the first cost and the associated information, and outputs the selected specific data model to the input / output unit.

[0080] With this configuration, an optimal specific data model can be selected from among the multiple data models 321, 322, and 323 depending on the cost of the characteristics of the on-site data at the factory site, and the optimal specific data model can be used to efficiently acquire on-site data from the sensor 201a installed on the production line of the production equipment of factory 2.

[0081] In this embodiment, a platform 300 including a plurality of data model groups 320, a data model selection unit 330, and a data acquisition API group 340 is provided.

[0082] With this configuration, a platform 300 that can be used at various sites can be provided by efficiently acquiring site data by selecting the most appropriate specific data model from among multiple data models 321, 322, 323 in accordance with the scale of each site and the characteristics specific to the production site.

[0083] In this embodiment, the data model selection unit 330 calculates a planned cost based on the production plan of the production facility as a characteristic of the site data. In this way, the calculated planned cost is used to efficiently acquire site data from the sensor 201a installed on the production line of the production facility of the factory 2 using an optimal specific data model from among the multiple data models 321, 322, and 323 according to the site of the factory 2.

[0084] In this embodiment, the cost calculation processing unit 350 of the platform 300 calculates the actual cost based on the performance actual data of the production equipment as a characteristic of the on-site data. In this way, the processing load of the data model selection unit 330 can be reduced.

[0085] In this way, the calculated actual costs can be used to efficiently acquire on-site data from the sensor 201a installed on the production line of the production equipment of factory 2 using a specific data model that is optimal from among the multiple data models 321, 322, and 323 depending on the site of factory 2.

[0086] In this embodiment, the data model selection unit 330 matches the planned cost and actual cost of the new site by combining the planned cost and the actual cost of the production equipment, and refers to the performance actual data of the site with the closest combination to determine the specific data model to be selected for the new production equipment.

[0087] In this way, by using the performance record data of the site that has the closest combination, it is possible to efficiently acquire site data from the sensor 201a installed on the production line of the production equipment of factory 2 using the most appropriate specific data model from among the multiple data models 321, 322, and 323 depending on the site of factory 2.

[0088] In this embodiment, the data model selection unit 330 calculates a bill of material (BOM) cost and a bill of process (BOP) cost as planned costs from the bill of material (BOM) and bill of process (BOP) of the production equipment.

[0089] In this way, by using BOM cost and BOP cost, it is possible to efficiently acquire on-site data from the sensor 201a installed on the production line of the production equipment of the factory 2 by using a specific data model that is optimal from among multiple data models 321, 322, and 323 depending on the site of the factory 2.

[0090] In this embodiment, the data model selection unit 330 extracts rework, loops, and work order changes performed in the production equipment from the past performance result information 390 related to the past performance results of the sensor 201, and calculates them as rework cost, loop cost, and work order cost, respectively, as the actual costs.

[0091] In this way, by using actual costs including rework costs, loop costs, and work order costs, it is possible to efficiently acquire on-site data from the sensor 201a installed on the production line of the production equipment of factory 2 by using a specific data model that is optimal from among the multiple data models 321, 322, and 323 depending on the site of factory 2.

[0092] In this embodiment, the data model selection unit 330 selects a specific data model from among the multiple data models 321, 322, and 323 so as to obtain a similar combination of bill of materials (BOM) cost, bill of process (BOP) cost, rework cost, loop cost, and operation order cost.

[0093] In this way, by using a specific data model that has a similar combination of bill of materials (BOM) cost, bill of process (BOP) cost, rework cost, loop cost, and operation order cost from among the multiple data model groups 320, on-site data can be efficiently acquired from the sensor 201a installed on the production line of the production facility in the factory 2.

[0094] In this embodiment, it has been described that one data model is selected from a plurality of data models, but the present invention is not limited to this, and multiple data models may be selected from a large number of data models.

[0095] In the data acquisition system 100 according to this embodiment, the sensor 201 is illustrated as an example of a data collection device installed on the production line of the production equipment at the factory site, but the present invention is not limited to this, and other data collection devices, ERP, PLM, MES, other production equipment, or any combination of these can also be exemplified. [Industrial Applicability]

[0096] The present invention can be applied to an information processing device that acquires on-site data from a data collection device installed on a production line of a production facility in a factory using a data model. [Explanation of symbols]

[0097] 100...data acquisition system, 200...factory, 201...sensor, 300...platform, 310...data lake, 320...data model group, 330...data model selection unit, 340...data acquisition API group, 350...cost calculation processing unit, 360...production plan information, 370...site cost information, 380...each site cost information, 390...past performance information, 400...application server, 410...application.

Claims

1. A memory having site data, which is information related to a site, costs, which are information representing characteristics of the site, a plurality of data models defining connection relationships between the site data, and association information in which the costs for each site are associated with the plurality of data models; A processor; An input / output unit, The processor, a cost calculation processing unit that receives site identification information that identifies a site from the input / output unit, and calculates a first cost that is a cost of the site identified by the site identification information based on site information of the site identified by the site identification information; a data model selection unit that selects a specific data model corresponding to a site having a second cost similar to the first cost from among the plurality of data models based on the first cost and the related information, and outputs the selected specific data model to an input / output unit; An information processing device comprising:

2. The data model selection unit Selecting the particular data model from the plurality of data models so as to optimize the performance of the data collection device according to past performance records of the data collection device that collects the on-site data and the cost of the characteristics of the on-site data.

2. The information processing apparatus according to claim 1,

3. The data model selection unit calculates a planned cost based on a production plan of the production facility as a characteristic of the on-site data.

3. The information processing apparatus according to claim 2.

4. 4. The information processing apparatus according to claim 3, wherein the cost calculation processing unit calculates an actual cost based on performance actual data of the production facility.

5. The data model selection unit performs matching between the planned cost and the actual cost of the new site by combining the planned cost and the actual cost of the production equipment, refers to the performance actual data of the site with the closest combination, and determines the specific data model to be selected for the new production equipment.

5. The information processing apparatus according to claim 4.

6. The data model selection unit calculates a bill of material (BOM) cost and a bill of process (BOP) cost as the planned cost from a bill of material (BOM) and a bill of process (BOP) of the production equipment.

4. The information processing apparatus according to claim 3.

7. The data model selection unit extracts rework, loops, and work order changes performed in the production equipment from past performance information related to past performance results of the data collection device, and calculates the rework cost, loop cost, and work order cost, respectively, as the actual costs.

5. The information processing apparatus according to claim 4.

8. The data model selection unit selects the specific data model from among the plurality of data models so that a combination of a bill of materials (BOM) cost, a bill of process (BOP) cost, the rework cost, the loop cost, and the operation sequence cost is close to each other.

8. The information processing apparatus according to claim 6, wherein the information processing apparatus is a computer.

9. An information processing method in an information processing device including a memory having site data, which is information related to a site, a cost, which is information representing a characteristic of the site, a plurality of data models defining a connection relationship between the site data, and association information in which the cost for each site is associated with the plurality of data models, a processor, and an input / output unit, a cost calculation process step in which the processor receives site identification information that identifies a site from the input / output unit, and calculates a first cost, which is a cost of the site identified by the site identification information, based on site information of the site identified by the site identification information; a data model selection step in which the processor selects a specific data model corresponding to a site having a second cost similar to the first cost from among the plurality of data models based on the first cost and the related information, and outputs the selected specific data model to an input / output unit; 13. An information processing method comprising:

Images