Discharge management device

Abstract

To provide an emissions management apparatus which can accurately estimate emissions of carbon dioxide generated in a production line, in a design phase.SOLUTION: An emissions management apparatus includes: an emission estimation unit which calculates an estimated value of emissions that are generated during production of a planned product which is to be produced in a production line, using a correspondence relationship between design information of production conditions including specification information of products and setting information of the production line and an actually measured value of the emissions of carbon dioxide generated in the production line; and a determination unit which calculates a difference, for each product, between the calculated estimated value of the emissions and the actually measured value of the emissions generated during production of a product corresponding to the planned product, and issues a warning when the calculated difference is equal to or larger than a predetermined upper limit value, or equal to or smaller than a predetermined lower limit value.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present disclosure relates to emissions management devices. [Background technology]

[0002] For example, Patent Document 1 discloses a CO2 emission reduction support system that acquires the amount of light and heat usage, such as electricity, and the average temperature in a home, and generates a regression equation for estimating the amount of light and heat usage using the acquired average temperature and the amount of light and heat usage. This CO2 emission reduction support system estimates the amount of light and heat usage and daily CO2 emissions in the home using the average temperature and the regression equation for a target day, and also calculates the actual amount of CO2 emissions for that day from the amount of light and heat usage for the target day. The user evaluates their daily actions to reduce CO2 emissions by comparing the estimated CO2 emissions with the actual CO2 emissions. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2011-13755 Summary of the Invention [Problem to be solved by the invention]

[0004] When estimating CO2 emissions generated on a production line, the number of possible contributing factors increases as the number of parts in the product and the number of processes on the production line increases. Conventional techniques may not be able to accurately estimate CO2 emissions on a production line. [Means for solving the problem]

[0005] The present disclosure can be realized in the following forms. A first aspect of the present disclosure is an emission management device for managing carbon dioxide emissions generated by the production of a product, the emission management device including an emission estimation unit that calculates an estimated value of the emission amount generated during the production of a product to be produced on the production line using a correspondence relationship between design information for production conditions including product specification information and production line setting information, and an actual measurement value of the emission amount of carbon dioxide generated on the production line, and calculates, for each product, a difference between the calculated estimated value of the emission amount and the actual measurement value of the emission amount generated during the production of a product corresponding to the product to be produced, and calculates, when the calculated difference is equal to or greater than a predetermined upper limit value or less than a predetermined upper limit value. The emission management device includes a determination unit that notifies of an abnormality when the value is equal to or less than a predetermined lower limit, and a cause analysis unit that identifies the cause of the abnormality, wherein the cause analysis unit inputs the actual measurement information of the production conditions during the production of a product in which the abnormality has occurred into an analytical learning model that has been trained using an analytical dataset including actual measurement information of the production conditions during the production of a normal product, thereby calculating a numerical value indicating whether the actual measurement information of the production conditions is abnormal for each item included in the production conditions, and identifies the item for which the calculated numerical value is higher than a predetermined threshold as the cause of the abnormality.

[0006] (1) According to one aspect of the present disclosure, there is provided an emissions management device for managing carbon dioxide emissions generated in the production of a product. The emissions management device includes an emissions estimation unit that calculates an estimate of emissions generated during the production of a planned product to be produced on the production line by using a correspondence relationship between design information for production conditions, including product specification information and production line setting information, and actual measured values of carbon dioxide emissions generated on the production line, and a determination unit that calculates, for each product, a difference between the calculated estimated value of emissions and the actual measured value of emissions generated during the production of a product corresponding to the planned product, and issues an abnormality notification when the calculated difference is equal to or greater than a predetermined upper limit or equal to or less than a predetermined lower limit. This type of emission management device can accurately estimate the amount of emissions generated on the production line by using the correspondence between the design information of the production conditions and the actual measured values of carbon dioxide emissions. Furthermore, by notifying an abnormality, it becomes easier to analyze the cause of the abnormality. (2) The emission control device of the above aspect may further include a cause analysis unit that identifies the cause of the abnormality. According to the emission control device of this aspect, the cause of the abnormality can be identified. (3) In the emission management device of the above aspect, the cause analysis unit may input the actual measurement information of the production conditions during the production of the product in which the abnormality occurred into an analytical learning model trained using an analytical dataset including the actual measurement information of the production conditions during the production of normal products, and calculate a numerical value indicating whether the actual measurement information of the production conditions is abnormal for each item included in the production conditions. The cause analysis unit may identify the item for which the calculated numerical value is higher than a predetermined threshold as the cause of the abnormality. According to the emission amount management device of this aspect, the cause of an abnormality in the estimated value can be identified with high accuracy by using a learning model. (4) In the emission management device of the above form, the emission estimation unit may calculate the estimated value of the emission amount by inputting the design information of the production conditions corresponding to the product to be produced into an estimation learning model trained using an estimation dataset including a combination of design information of the production conditions and the actual measured value of the emission amount. According to the emission amount management device of this embodiment, by utilizing the learning model, it is possible to accurately estimate the emission amount generated on the production line at the design stage. (5) In the emission management device of the above form, if the item extracted by the cause analysis unit as the cause of the abnormality is not included in the estimation dataset, the emission estimation unit may add the extracted item to the estimation dataset. According to the emission amount management device of this aspect, it is possible to compensate for the lack of an estimation data set for estimating emission amounts, and to improve the estimation accuracy of emission amounts. The present disclosure can also be realized in various forms other than an emission management device, such as an information processing device, an emission management method, a learning method for a learning model, a trained model, a control method for an emission management device, a computer program that realizes the control method, a non-transitory recording medium on which the computer program is recorded, etc. [Brief explanation of the drawings]

[0007] [Figure 1] 1 is an explanatory diagram showing the configuration of a system including an emission management device according to a first embodiment of the present disclosure. [Figure 2] FIG. 2 is a block diagram showing the functional configuration of the emission control device. [Figure 3] 10 is a flowchart showing details of a learning process. [Figure 4] 10 is a flowchart showing details of an estimation process. [Figure 5] 10 is a flowchart showing details of an analysis process. [Figure 6] FIG. 4 is an explanatory diagram showing the results of an emission amount estimation process performed by the emission amount management device. DETAILED DESCRIPTION OF THE INVENTION

[0008] A. First embodiment: FIG. 1 is an explanatory diagram showing the configuration of a system including an emissions management device 60 according to a first embodiment of the present disclosure. The emissions management device 60 estimates the amount of carbon dioxide (CO2) emissions generated by the production of a product WK on a production line Ln at the design stage of the product WK. In the example of FIG. 1, the production line Ln includes processes PR1 and PR2. Process PR1 is a casting process equipped with a die-casting machine that forms vehicle parts such as cylinder heads and engine blocks. Process PR2 is a process subsequent to process PR1 for the product, and is equipped with, for example, a machining center or the like that performs post-processing and post-treatment on the product after casting.

[0009] The emission management device 60 calculates an estimated value of the amount of carbon dioxide emission that will occur on the production line Ln when the scheduled products are produced, using the correspondence relationship between the design information of the production conditions and the actual measured value of the amount of carbon dioxide emission. Hereinafter, the amount of carbon dioxide emission will also be simply referred to as the "emission amount."

[0010] "Design information for production conditions" refers to the production conditions at the design stage before the start of product production. The production conditions during product production are also called "measured information for production conditions." "Production conditions" are the various conditions set for producing product WK on production line Ln. Production conditions include, for example, specification information for product WK and setting information for production line Ln. "Specification information for product WK" includes the name and quality of the material, dimensions, weight, and allowable tolerances for dimensions and weight of product WK. Specification information for product WK can also be said to be the target values for the specifications of product WK set at the design stage of product WK. Specification information for product WK is not limited to finished products, but also includes specification information for so-called intermediate products in the middle of production.

[0011] "Setting information for a production line Ln" refers to the specifications of the production line Ln, the specifications of the production equipment included in the production line Ln, and the specifications of the manufacturing process including the production line Ln. The setting information for a production line Ln includes, for example, the installation location of the production line Ln, the installation location of the production equipment included in the production line Ln, the type of production equipment, the number of production equipment units, processing conditions for the production equipment (e.g., processing time, temperature, pressure, current, and voltage), the type of work indicating whether the product WK is processed manually or automatically, and mold information. The setting information for a production line Ln can also be considered the target values of specifications set for the production line Ln during the design phase of the production of the product WK. "Mold information" refers to the specifications of the molds used for forging and casting, such as the volume and shape of the mold's internal space (cavity). The production line information may also include the operation plan of the factory that includes the production line Ln, the planned production date, time, and season, and the worker's work shifts.

[0012] The emission amount estimation by the emission management device 60 uses data stored in a database DB of an external device. The database DB stores product specification information D1, setting information D2 of the production line Ln, weather conditions D3, actual emission amount measurement values D4, and production information D5. This information is recorded in association with product identification information. The product identification information is, for example, a serial number assigned to each product.

[0013] Product specification information D1 and production line setting information D2 belong to the design information of production conditions. Weather conditions D3 include the temperature, weather, and sunshine hours at the time of production. Weather conditions D3 belong to the measured information of production conditions. Weather forecasts and the like may also be stored in the weather conditions D3 as design information of production conditions.

[0014] The actual measured value D4 of the emission amount and the production information D5 belong to the actual measured information of the production conditions. The actual measured value D4 of the emission amount is, for example, data on the emission amount generated on the production line Ln during the production of the product WK. The production information D5 is the actual measured information of the production conditions, and is information that has traceability in the production process of the product WK. Note that the "actual measured information of the production conditions" is not limited to numerical values, but may also include textual information such as the names of materials used in the product WK during production.

[0015] As shown in FIG. 1, the actual emission amount D4 is calculated by multiplying the amount of power consumption acquired by a sensor 70 installed in processes PR1 and PR2 of the production line Ln by a CO2 emission coefficient. The sensor 70 includes a detection unit 72 and a communication unit 74. The detection unit 72 is a wattmeter and detects the amount of power consumption generated in processes PR1 and PR2. The detection unit 72 may be, for example, a gas meter instead of a watthour meter. The detected amount of power consumption is output to the communication unit 74. The communication unit 74 transmits the amount of power consumption to the emission management device 60 via wireless communication in accordance with an arbitrary communication protocol.

[0016] 2 is a block diagram showing the functional configuration of the emission management device 60. The emission management device 60 is a computer equipped with a CPU 62 as a central processing unit, a storage device 64, a display unit 66 such as a liquid crystal display or a touch panel, an input unit 67 such as a keyboard or a mouse, and a communication unit 68. These are connected to each other via a bus 61 and configured to be able to communicate bidirectionally. The input unit 67 is used to input design values for production conditions, etc.

[0017] In this embodiment, the emission management device 60 calculates an estimated emission amount using a learning model (hereinafter also referred to as a "trained model") that has been trained by machine learning, with the actual measurement of the emission amount as a response variable and the production conditions as explanatory variables, as an example of the correspondence relationship between the design information of the production conditions and the actual measurement of the emission amount of carbon dioxide. The emission management device 60 can estimate the emission amount that will occur when the planned production product is produced by inputting the design information of the production conditions of the planned production product into the trained model. Note that multivariate analysis or regression analysis may be used instead of the learning model.

[0018] The communication unit 68 controls communication to receive machine learning datasets and estimation data via a network. Specifically, the communication unit 68 acquires product specification information D1, production line setting information D2, weather conditions D3, actual emission amount measurement values D4, and production information D5 from a database DB of an external device. The communication unit 68 also acquires the power consumption of processes PR1 and PR2 from a sensor 70.

[0019] The storage device 64 is, for example, a RAM, a ROM, or a hard disk drive (HDD). The HDD or ROM stores various programs for realizing the functions provided in this embodiment. The various programs read from the HDD or ROM are expanded on the RAM and executed by the CPU 62. The readable / writable area of the storage device 64 includes a CO2 emission coefficient storage unit 642, an estimated value storage unit 644 that stores estimated values of emissions, an estimation learning model storage unit 646, and an analysis learning model storage unit 648.

[0020] The CO2 emission coefficient storage unit 642 stores a CO2 emission coefficient for deriving emissions. The "CO2 emission coefficient" refers to the amount of carbon dioxide emitted per unit of activity, i.e., the amount of carbon dioxide emitted relative to the amount of electricity consumed per predetermined unit amount. In this embodiment, the CO2 emission coefficient corresponds to the amount of carbon dioxide emitted to generate 1 kWh of electricity, and its unit is, for example, "g / kWh." In this embodiment, the CO2 emission coefficient storage unit 642 stores emission coefficients for each electric utility published by the Ministry of the Environment and the Ministry of Economy, Trade and Industry in accordance with the Act on Promotion of Global Warming Countermeasures (Global Warming Act). However, the CO2 emission coefficient is not limited to being preset as a fixed value, and may be updated sequentially via a wide area network such as the Internet. This configuration allows CO2 emissions to be derived using the latest CO2 emission coefficient.

[0021] The calculated estimated value of the emission amount is stored in the estimated value storage unit 644. More specifically, the estimated value of the emission amount is recorded in the estimated value storage unit 644 in association with the identification information of the product.

[0022] The estimation learning model storage unit 646 stores an estimation learning model used to calculate an estimated value of emission amounts. The estimation learning model is, for example, a model having an architecture such as a recurrent neural network (RNN), a convolutional neural network (CNN), a general regression neural network, or a random forest. The analysis learning model storage unit 648 stores an analysis learning model used to analyze the cause of an abnormality in the estimated value, which will be described later. The analysis learning model is, for example, an RNN with a recurrent structure such as a long short-term memory (LSTM), or a model having an architecture such as a generative adversarial network (GAN).

[0023] The CPU 62 executes programs stored in the storage device 64 to function as a learning model generation unit 622, an emission amount estimation unit 624, a determination unit 626, and a cause analysis unit 628. The learning model generation unit 622 generates a trained model using machine learning datasets including an estimation dataset and an analysis dataset. The estimation dataset is data including a combination of design information on production conditions and actual measurement values of emissions. Specifically, the estimation dataset is, for example, actual measurement values D4 of emissions generated on the production line Ln obtained from the sensor 70, and specification information D1 and setting information D2 stored in the database DB. The estimation dataset may further include weather conditions D3. The analysis dataset is, for example, data including actual measurement information of production conditions during normal product production from production information D5.

[0024] The emission estimation unit 624 uses the trained learning model for estimation to estimate the emission amount that will occur when the planned production product is produced. The determination unit 626 calculates the difference between the estimated emission amount calculated by the emission estimation unit 624 and the actual measurement value D4 of the emission amount that occurred during the production of the product WK corresponding to the planned production product for each product WK, and determines whether the estimated emission amount is abnormal using the calculated difference. In this disclosure, "each product WK" means each individual product that can be distinguished by a serial number or the like. If there is an abnormality in the estimated emission amount, the determination unit 626 notifies the abnormality. The cause analysis unit 628 uses the trained learning model for analysis to calculate a numerical value indicating whether the actual measurement information of the production conditions during the production of the abnormal product in which the abnormality occurred is abnormal for each item included in the production conditions. The cause analysis unit 628 uses the calculated numerical value to identify the cause of the abnormality from the items included in the production conditions.

[0025] The emission amount estimation process executed by the emission amount management device 60 of the present disclosure will be described with reference to Figures 3 to 5. The emission amount estimation process includes a learning step, an estimation step, and a cause analysis step in this order.

[0026] FIG. 3 is a flowchart showing details of the learning process. In the learning process, the learning model generation unit 622 acquires a machine learning dataset and performs machine learning to generate a trained model. In step S100, the learning model generation unit 622 acquires an estimation dataset. Specifically, the learning model generation unit 622 acquires product specification information D1, production line setting information D2, weather conditions D3, and actual measurement values D4 of emissions stored in the database DB via the communication unit 68. In step S110, the learning model generation unit 622 inputs the product specification information D1, production line setting information D2, weather conditions D3, and actual measurement values D4 of emissions acquired from the database DB into the estimation learning model as training data, and generates a trained estimation learning model.

[0027] In step S120, the learning model generation unit 622 acquires an analysis dataset. In this embodiment, the learning model generation unit 622 acquires production information D5 accumulated in the database DB via the communication unit 68. In step S130, the learning model generation unit 622 inputs, from the acquired production information D5, actual measurement information on production conditions during normal product production into the analytical learning model as learning data, and generates a trained analytical learning model.

[0028] 4 is a flowchart showing details of the estimation process. In the estimation process, the emission amount estimation unit 624 estimates the emission amount that will occur when the planned production products are produced using a trained learning model for estimation. In this embodiment, the estimation process is performed for each operating day of the production line Ln.

[0029] In step S202, the emission amount estimation unit 624 acquires the production conditions of the products scheduled to be produced for one day. The production conditions of the products scheduled to be produced can be acquired, for example, from a higher-level production management device, an information processing device used for product design, or a database DB. The production conditions of the products scheduled to be produced may be input by the user operating the input unit 67.

[0030] In step S204, the emission amount estimation unit 624 inputs the acquired production conditions for the products scheduled for production for one day into the trained learning model for estimation. In step S206, the emission amount estimation unit 624 acquires the output value from the learning model for estimation as an estimated emission amount. In this embodiment, the emission amount estimation unit 624 also estimates the emission amount for each predetermined unit time, such as each hour, and the emission amount for each product serial number by referencing the details of the production schedule for one day. That is, the emission amount estimation unit 624 calculates the total estimated value for one day, as well as a subtotal of the estimated values of the emission amount for each unit time and the estimated value of the emission amount for each product serial number. The emission amount estimation unit 624 stores the calculation results in the estimated value storage unit 644 and terminates processing.

[0031] 5 is a flowchart showing the details of the analysis process. The analysis process starts when production of the product WK on the production line Ln starts.

[0032] In step S302, the determination unit 626 acquires the actual measured value of the amount of emissions during production of the product WK. The determination unit 626 acquires the amount of power consumption during production of the product WK from the sensor 70, and calculates the actual measured value of the amount of emissions by multiplying the acquired amount of power consumption by the CO2 emission coefficient. In this embodiment, the determination unit 626 calculates the actual measured value of the amount of emissions for each product serial number.

[0033] In step S304, the determination unit 626 calculates the difference between the estimated emission amount and the actual emission amount for each product serial number. More specifically, the determination unit 626 compares the estimated emission amount for each product serial number calculated by the emission amount estimation unit 624 in step S206 with the actual emission amount for the product corresponding to that serial number calculated in step S302.

[0034] In step S306, the determination unit 626 determines whether the actual measured value of the discharge amount deviates from the estimated value by a predetermined value or more. More specifically, the determination unit 626 determines whether the calculated difference is equal to or greater than a predetermined upper limit and whether it is equal to or less than a predetermined lower limit. In this embodiment, the upper limit of the difference is set to +20% of the estimated value, and the lower limit is set to -20% of the estimated value. If the difference is greater than the lower limit but less than the upper limit (S306: NO), the determination unit 626 proceeds to step S310. In step S310, the determination unit 626 determines whether a termination condition is satisfied. The termination condition is, for example, whether analysis has been performed on all of the products WK for one day whose discharge amount has been estimated. If the termination condition is satisfied (S310: YES), the determination unit terminates the analysis process. If the termination condition is not satisfied (S310: NO), the determination unit proceeds to step S302.

[0035] In step S306, if the difference is equal to or greater than the upper limit value or equal to or less than the lower limit value (S306: YES), the process proceeds to step S308. In step S308, the determination unit 626 notifies the user or the like of the abnormality in the estimated value, for example, by displaying a screen on the display unit 66. In this embodiment, the determination unit 626 notifies the user or the like of the occurrence of the abnormality in the estimated value, as well as the serial number of the abnormal product in which the abnormality in the estimated value occurred. Note that the abnormality may be notified by voice or the like instead of by the display unit 66.

[0036] In step S312, the cause analysis unit 628 acquires actual measurement information on the production conditions during the production of the abnormal product. In this embodiment, the cause analysis unit 628 acquires production information D5 associated with the serial number of the abnormal product from the database DB. The cause analysis unit 628 inputs the acquired production information D5 corresponding to the abnormal product into the trained analytical learning model. For each item of production conditions included in the production information D5, the analytical learning model calculates a numerical value indicating whether the actual measurement information on the production conditions is abnormal. The numerical value indicating whether the information is abnormal is expressed, for example, as a numerical value between 0 and 1. In this embodiment, if the actual measurement information on the production conditions is abnormal, the numerical value is high.

[0037] In step S314, the cause analysis unit 628 identifies the cause of the abnormality by comparing the calculated numerical value with a predetermined threshold. In this embodiment, the cause analysis unit 628 identifies an item whose numerical value is higher than the threshold as the cause of the abnormality in the estimated value. The threshold can be set arbitrarily, for example, to 0.5 (50%), 0.6 (60%), 0.7 (70%), 0.8 (80%), 0.9 (90%), or the like. If the numerical value is low when the actual measurement information of the production conditions is abnormal, the numerical value lower than a threshold such as 20% may be identified as the cause of the abnormality.

[0038] In step S316, the emission amount estimation unit 624 determines whether the item extracted by the cause analysis unit 628 as the cause of the abnormality is included in the estimation dataset. If it is included in the estimation dataset, the processing ends. If it is not included in the estimation dataset, there is a concern that the cause of the abnormality in the estimated value is, for example, a lack of learning data for the estimation learning model. In this embodiment, if the extracted item is not included in the estimation dataset, the extracted item is added to the estimation dataset and the processing ends. By expanding the range of learning for the estimation learning model, the accuracy of the emission amount estimation by the emission amount estimation unit 624 can be improved.

[0039] FIG. 6 is an explanatory diagram showing the results of the emission amount estimation process performed by the emission management device 60. FIG. 6 schematically shows a screen 66D displayed on the display unit 66. The upper part of FIG. 6 shows data related to the emission amount for each product. More specifically, a bar graph G1 showing the actual emission amount (unit: kg CO2 / unit) for each product serial number and a line graph G2 showing the estimated emission amount (unit: kg CO2 / unit) for each product serial number are shown. Above that, a plot G3 showing whether the estimated emission amount is normal or abnormal is shown for each product serial number as a result of the determination made by the determination unit 626. Note that the bar graph G1 is displayed by stacking bar graphs of different colors for each production facility used during production.

[0040] An example of actual measurement information on production conditions is shown in the lower part of Fig. 6. More specifically, a bar graph G4 showing the actual measurement value of the discharge amount per unit time (one hour in this embodiment) and a line graph G5 showing the temperature are shown. Note that the bar graph G4 is displayed by arranging bar graphs of different colors for each production facility during production. Also, above it, as the determination result by the determination unit 626, a plot G6 showing the presence or absence of an abnormality in the estimated value of the discharge amount and a plot G7 showing the operating and stopped status of the production facility are shown. Note that the stoppage of the production facility includes both an abnormal stoppage and a planned stoppage.

[0041] As shown by plots G3a and G3b in the upper part of Fig. 6, abnormalities in the estimation results occurred in the products with serial numbers Pa and Pb. Note that for the product with serial number Pa, the actual measured value of emissions was +20% or more of the estimated value, as shown by bar graph G1a, and for the product with serial number Pb, the actual measured value of emissions was -20% or less of the estimated value, as shown by bar graph G1b, and thus was determined to be abnormal.

[0042] As shown by plot G6a in the lower part of Figure 6, it can be seen that the production equipment was shut down when an abnormality in the estimated value occurred. The time period during which the product with serial number Pa was produced coincides with the time period to which plot G6a belongs. The cause of the shutdown of the production equipment in question was a breakdown due to deterioration over time, in other words, the cumulative operating time of the production equipment was longer than normal. As a result, the processing time for the product with serial number Pa was longer than normal, as shown by bar graph G1a, and the actual measured value of the amount of emissions was greater than the estimated value.

[0043] The cause analysis unit 628 inputs the actual measured values of the production conditions corresponding to the product with serial number Pa into the analytical learning model, thereby extracting that the value of the item "cumulative operation time of production equipment" is higher than the threshold value, and identifies the cumulative operation time of the production equipment as the cause of the abnormality in the estimated value of the emission amount. Note that if "cumulative operation time of production equipment" is not included in the estimation dataset, the cause analysis unit 628 adds it as an item in the estimation dataset.

[0044] As shown in the lower part of Figure 6, an abnormality in the estimated value also occurs in plot G6b. The time period in which the product with serial number Pb was produced coincides with the time period to which plot G6b belongs. The product with serial number Pb was determined to be defective during production, and production was terminated midway without stopping the production equipment. Therefore, the production equipment for the product with serial number Pb was not stopped, and the processing time was shorter than normal. Note that the production equipment was not stopped. As a result, the actual measured value of the discharge amount for serial number Pb is lower than normal, as shown by bar graph G1b. The cause analysis unit 628 uses the analytical learning model to identify product defects as the cause of the abnormality in the estimated value. Furthermore, if an item indicating whether the product is non-defective is not included in the estimation dataset, the cause analysis unit 628 adds it as an item in the estimation dataset.

[0045] As described above, the emission management device 60 of this embodiment includes an emission estimation unit 624 that calculates an estimated amount of emission generated during the production of a planned product using a correspondence relationship between the design information of the production conditions, including specification information of the product WK and setting information of the production line Ln, and the actual measured amount of carbon dioxide emission. The emission management device 60 also includes a determination unit 626 that calculates the difference between the estimated amount of emission and the actual measured amount of emission generated during the production of the planned product WK for each product WK, and notifies the user of an abnormality in the estimated amount if the calculated difference exceeds a predetermined value. The emission management device 60 of this embodiment can accurately estimate the amount of emission generated on the production line during the design stage by using the correspondence relationship between the design information of the production conditions and the actual measured amount of carbon dioxide emission. Furthermore, by notifying the user of an abnormality in the estimated amount for each product, the cause of the abnormality can be easily analyzed for each product.

[0046] According to the emission management device 60 of this embodiment, the emission estimation unit 624 calculates an estimated value of the emission amount by inputting design information of the production conditions corresponding to the product to be produced into an estimation learning model trained using an estimation dataset including a combination of design information of the production conditions and actual measurement values of the emission amount. By using the learning model, the emission amount generated on the production line can be accurately estimated at the design stage.

[0047] The emission management device 60 of this embodiment further includes a cause analysis unit 628 that identifies the cause of the abnormality using a numerical value indicating whether the actual measurement information of the production conditions is abnormal. Therefore, it is possible to provide an emission management device 60 that can identify the cause of the abnormality in the estimated value of the emission amount.

[0048] According to the emission management device 60 of this embodiment, the cause analysis unit 628 inputs the actual measurement information of the production conditions during the production of the abnormal product into an analytical learning model trained using an analytical data set including the actual measurement information of the production conditions during the production of the normal product WK, and identifies the item whose value is higher than a predetermined threshold as the cause of the abnormality. By using the learning model, the cause of the abnormality in the estimated value can be identified with high accuracy.

[0049] According to the emission amount management device 60 of this embodiment, when the item extracted as the cause of the abnormality by the cause analysis unit 628 is not included in the estimation data set, the emission amount estimation unit 624 adds the extracted production condition to the estimation data set. This makes it possible to supplement the deficiency of the estimation data set for the emission amount estimation, and improves the estimation accuracy of the emission amount.

[0050] B. Other Embodiments: (B1) In the above embodiment, an example was shown in which the emission management device 60 was provided with the cause analysis unit 628. In contrast to this, for example, in cases where the cause analysis is performed by a person, the emission management device 60 does not need to be provided with the cause analysis unit 628. Even with this form of emission management device 60, it is possible to accurately estimate the amount of emissions generated on the production line at the design stage by using the correspondence between the design information of the production conditions and the actual measured values of carbon dioxide emissions.

[0051] (B2) In the above embodiment, the cause analysis unit 628 identifies the cause of an abnormality using an analytical learning model. However, the cause analysis unit 628 may identify the cause of an abnormality using multivariate analysis or regression analysis instead of a learning model. Furthermore, the cause analysis unit 628 may determine whether the performance information of the production conditions is abnormal using, for example, a predetermined setting value such as process capacity.

[0052] (B3) In the above embodiment, an example was shown in which the emission amount estimation unit 624 adds an identified item to the estimation dataset when the item identified as the cause of the abnormality is not included in the estimation dataset. In contrast, an item identified as the cause of the abnormality may not be added to the estimation dataset, for example, when the processing load of the estimation process is to be reduced or when the effect of improving the estimation accuracy relative to the processing load of the estimation process is not sufficient. Furthermore, for example, whether or not to add an item may be determined based on the identified item.

[0053] The present disclosure is not limited to the above-described embodiments and can be realized in various configurations without departing from the spirit thereof. For example, the technical features in the embodiments corresponding to the technical features in each aspect described in the Summary of the Invention section can be appropriately replaced or combined to solve some or all of the above-described problems or achieve some or all of the above-described effects. Furthermore, if a technical feature is not described as essential in this specification, it can be appropriately deleted. [Explanation of symbols]

[0054] 60...emissions management device, 61...bus, 62...CPU, 64...storage device, 66...display unit, 66D...screen, 67...input unit, 68...communication unit, 70...sensor, 72...detection unit, 74...communication unit, 622...learning model generation unit, 624...emissions estimation unit, 626...determination unit, 628...cause analysis unit, 642...CO2 emission coefficient storage unit, 644...estimated value storage unit, 646...estimation learning model storage unit, 648...analysis learning model storage unit, D1...product specification information, D2...production line setting information, D3...weather conditions, D4...actual measurement value of emissions, D5...production information, DB...database, Ln...production line, WK...product

Claims

1. An emission control device for controlling carbon dioxide emissions generated by the production of products, an emission estimation unit that calculates an estimated value of the emission amount that will be generated during the production of a planned product on the production line, using a correspondence relationship between design information on production conditions, including product specification information and production line setting information, and an actual measurement value of the emission amount of carbon dioxide generated on the production line; a determination unit that calculates, for each product, a difference between the calculated estimated value of the emission amount and an actual measured value of the emission amount generated during production of the product corresponding to the planned production product, and notifies of an abnormality when the calculated difference is equal to or greater than a predetermined upper limit value or equal to or less than a predetermined lower limit value; a cause analysis unit that identifies the cause of the abnormality, The cause analysis unit inputting the actual measurement information of the production conditions during the production of the product in which the abnormality has occurred into an analytical learning model trained using an analytical dataset including the actual measurement information of the production conditions during the production of a normal product, thereby calculating a numerical value indicating whether the actual measurement information of the production conditions is abnormal for each item included in the production conditions; Identifying the item for which the calculated value is higher than a predetermined threshold as the cause of the abnormality. Emission control device.

2. The emission control device according to claim 1, the emission amount estimation unit calculates the estimated value of the emission amount by inputting the design information of the production conditions corresponding to the planned production product into an estimation learning model trained using an estimation dataset including a combination of the design information of the production conditions and the actual measurement value of the emission amount; Emission control device.

3. The emission control device according to claim 2, when the item extracted by the cause analysis unit as the cause of the abnormality is not included in the estimation data set, the emission amount estimation unit adds the extracted item to the estimation data set. Emission control device.

Images