WO2022149344A1 - 品質異常解析方法、金属材料の製造方法および品質異常解析装置 - Google Patents
品質異常解析方法、金属材料の製造方法および品質異常解析装置 Download PDFInfo
- Publication number
- WO2022149344A1 WO2022149344A1 PCT/JP2021/040936 JP2021040936W WO2022149344A1 WO 2022149344 A1 WO2022149344 A1 WO 2022149344A1 JP 2021040936 W JP2021040936 W JP 2021040936W WO 2022149344 A1 WO2022149344 A1 WO 2022149344A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- quality
- manufacturing
- quality prediction
- metal material
- manufacturing conditions
- Prior art date
Links
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 247
- 230000005856 abnormality Effects 0.000 title claims abstract description 101
- 238000004458 analytical method Methods 0.000 title claims abstract description 53
- 239000007769 metal material Substances 0.000 title claims description 137
- 238000004364 calculation method Methods 0.000 claims abstract description 25
- 238000000034 method Methods 0.000 claims description 198
- 230000008569 process Effects 0.000 claims description 122
- 239000000047 product Substances 0.000 claims description 67
- 238000013441 quality evaluation Methods 0.000 claims description 26
- 230000002123 temporal effect Effects 0.000 claims description 10
- 230000007704 transition Effects 0.000 claims description 9
- 239000012467 final product Substances 0.000 claims description 6
- 238000010238 partial least squares regression Methods 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 6
- 238000012417 linear regression Methods 0.000 claims description 5
- 238000010801 machine learning Methods 0.000 claims description 5
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 238000012628 principal component regression Methods 0.000 claims description 4
- 238000007637 random forest analysis Methods 0.000 claims description 4
- 238000003908 quality control method Methods 0.000 claims description 2
- 238000001303 quality assessment method Methods 0.000 abstract 3
- 239000000463 material Substances 0.000 description 33
- 238000005520 cutting process Methods 0.000 description 32
- 238000010586 diagram Methods 0.000 description 15
- 229910000831 Steel Inorganic materials 0.000 description 13
- 239000010959 steel Substances 0.000 description 13
- 238000005097 cold rolling Methods 0.000 description 9
- 238000001816 cooling Methods 0.000 description 9
- 238000000137 annealing Methods 0.000 description 8
- 230000007547 defect Effects 0.000 description 8
- 238000012545 processing Methods 0.000 description 8
- 238000005098 hot rolling Methods 0.000 description 7
- 238000005457 optimization Methods 0.000 description 7
- 239000010960 cold rolled steel Substances 0.000 description 6
- 238000005096 rolling process Methods 0.000 description 6
- 238000007796 conventional method Methods 0.000 description 5
- 238000009628 steelmaking Methods 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 238000003723 Smelting Methods 0.000 description 3
- 238000005266 casting Methods 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000010438 heat treatment Methods 0.000 description 3
- 239000000203 mixture Substances 0.000 description 3
- 239000000126 substance Substances 0.000 description 3
- 229910001335 Galvanized steel Inorganic materials 0.000 description 2
- 239000002253 acid Substances 0.000 description 2
- 230000002950 deficient Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 239000008397 galvanized steel Substances 0.000 description 2
- 230000010365 information processing Effects 0.000 description 2
- 239000002184 metal Substances 0.000 description 2
- 238000007747 plating Methods 0.000 description 2
- 238000010791 quenching Methods 0.000 description 2
- 230000000171 quenching effect Effects 0.000 description 2
- 239000011265 semifinished product Substances 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000005275 alloying Methods 0.000 description 1
- FFBHFFJDDLITSX-UHFFFAOYSA-N benzyl N-[2-hydroxy-4-(3-oxomorpholin-4-yl)phenyl]carbamate Chemical compound OC1=C(NC(=O)OCC2=CC=CC=C2)C=CC(=C1)N1CCOCC1=O FFBHFFJDDLITSX-UHFFFAOYSA-N 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000005499 meniscus Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005554 pickling Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
- G05B19/41875—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM] characterised by quality surveillance of production
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
- G05B19/41885—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM] characterised by modeling, simulation of the manufacturing system
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0224—Process history based detection method, e.g. whereby history implies the availability of large amounts of data
- G05B23/0227—Qualitative history assessment, whereby the type of data acted upon, e.g. waveforms, images or patterns, is not relevant, e.g. rule based assessment; if-then decisions
- G05B23/0232—Qualitative history assessment, whereby the type of data acted upon, e.g. waveforms, images or patterns, is not relevant, e.g. rule based assessment; if-then decisions based on qualitative trend analysis, e.g. system evolution
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/32—Operator till task planning
- G05B2219/32187—Correlation between controlling parameters for influence on quality parameters
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/32—Operator till task planning
- G05B2219/32193—Ann, neural base quality management
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/32—Operator till task planning
- G05B2219/32194—Quality prediction
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Definitions
- the present invention relates to a quality abnormality analysis method, a metal material manufacturing method, and a quality abnormality analysis device.
- Patent Documents 1 to 8 disclose the following methods as a method for predicting quality for arbitrary requirements.
- this method for example, the distance between a plurality of past observation conditions stored in the actual database and the desired required condition is calculated, the weight of the observation data (actual data) is calculated from the calculated distance, and the calculated weight is used. Create a function that fits the neighborhood of the requirement condition. Then, the quality for the required condition is predicted using the created function.
- Patent Documents 1 to 8 the quality for any requirement is calculated from the data stored in the performance database.
- This actual database stores actual values of a plurality of manufacturing conditions and actual values of quality of metal materials manufactured under these manufacturing conditions.
- Patent Documents 1 to 8 disclose a method for constructing a prediction model for predicting quality from stored actual data of a plurality of manufacturing conditions.
- Patent Documents 1 to 8 do not mention a technique for estimating what is the cause of an abnormality when a product quality abnormality occurs.
- the present invention has been made in view of the above, and is based on a quality prediction model that predicts quality for arbitrary manufacturing conditions, and when a quality abnormality occurs, a candidate for the cause can be presented. It is an object of the present invention to provide an analysis method, a method for manufacturing a metal material, and a quality abnormality analysis device.
- the quality abnormality analysis method is a quality abnormality analysis method for products manufactured by the manufacturing process, and a plurality of manufacturing conditions of the manufacturing process are input.
- a quality prediction step for predicting the quality of the product by inputting the manufacturing conditions to the quality prediction model generated by using the quality of the product as a variable and the quality of the product as an output variable, and the actual product manufactured by the manufacturing process.
- the quality evaluation step for calculating the quality evaluation value of the product, and the quality prediction error calculation step for calculating the difference between the quality prediction value obtained as the output of the quality prediction step and the quality evaluation value as the quality prediction error.
- a step of presenting a cause of quality abnormality which presents a manufacturing condition that causes the quality abnormality of the product, is included.
- the quality contribution calculation step calculates the quality contribution based on each partial regression coefficient of the quality prediction model and the value of each variable. ..
- the quality abnormality cause presenting step presents the quality prediction error and the temporal integrated value of the quality contribution of each manufacturing condition in time series. Moreover, the time transition of the candidate of the manufacturing condition that causes the quality prediction error and the quality abnormality is visualized and presented.
- the quality abnormality analysis method focuses on the manufacturing conditions in which the quality contribution is large when the quality prediction error exceeds a predetermined value in the quality abnormality cause presentation step in the above invention.
- the manufacturing conditions having a large temporal integral value of the quality contribution are presented in order as candidates for the manufacturing conditions that cause the quality abnormality.
- the quality prediction model performs machine learning including linear regression, local regression, principal component regression, PLS regression, neural network, regression tree, random forest, and XGBoost. Generated using.
- the quality prediction model is a quality prediction model of a metal material manufactured through one or a plurality of steps, and the manufacturing conditions of each step are set in advance.
- the method for manufacturing a metal material according to the present invention is a method for manufacturing a metal material manufactured through a plurality of manufacturing steps, and is before the final manufacturing step is carried out.
- the quality of the final product is predicted by the quality prediction model generated by the above quality abnormality analysis method, and based on the prediction result, it is the manufacturing condition of the subsequent manufacturing process. Therefore, manufacturing conditions that have a high degree of contribution to quality and can be changed are selected, and the selected manufacturing conditions are determined and operated so that the quality of the final product falls within the preset quality control over the entire length.
- the quality abnormality analysis device is a quality abnormality analysis device for products manufactured by the manufacturing process, and inputs a plurality of manufacturing conditions of the manufacturing process.
- a quality prediction means for predicting the quality of the product by inputting the manufacturing conditions to the quality prediction model generated by using the quality of the product as a variable and the quality of the product as an output variable, and the actual product manufactured by the manufacturing process.
- a quality evaluation means for calculating the quality evaluation value of the product, and a quality prediction error calculation means for calculating the difference between the quality prediction value obtained as the output of the quality prediction means and the quality evaluation value as a quality prediction error.
- a means for presenting a cause of quality abnormality which presents manufacturing conditions that cause the quality abnormality of the product, is provided.
- a quality prediction error and a quality contribution of each manufacturing condition are obtained by using a manufacturing condition of each process and a quality prediction model for predicting the quality of a product manufactured under this manufacturing condition. Therefore, candidates for the cause of quality abnormality can be presented. Further, according to the present invention, it is possible to manufacture a metal material having good product quality over the entire length of the product.
- FIG. 1 is a block diagram showing a configuration of a quality prediction model generation device and a quality prediction device according to an embodiment of the present invention.
- FIG. 2 is a flowchart showing a flow of a quality prediction model generation method and a quality prediction method according to an embodiment of the present invention.
- FIG. 3 is a diagram showing an example of performance data collected by the manufacturing performance collection unit 11 in the quality prediction model generation method according to the embodiment of the present invention.
- FIG. 4 is a diagram showing an example of actual data edited by the manufacturing actual result editing unit 12 in the quality prediction model generation method according to the embodiment of the present invention.
- FIG. 5 is a diagram showing an example of a case where a metal material is manufactured through a plurality of steps in the quality prediction model generation method according to the embodiment of the present invention.
- FIG. 1 is a block diagram showing a configuration of a quality prediction model generation device and a quality prediction device according to an embodiment of the present invention.
- FIG. 2 is a flowchart showing a flow of a quality
- FIG. 6 is a diagram showing an example of a metal material in each step in the quality prediction model generation method according to the embodiment of the present invention.
- FIG. 7 is a diagram showing an example of actual data edited by the integrated process actual result editing unit in the quality prediction model generation method according to the embodiment of the present invention.
- FIG. 8 is a block diagram showing a configuration of a quality abnormality analysis device according to an embodiment of the present invention.
- FIG. 9 is a flowchart showing the flow of the quality abnormality analysis method according to the embodiment of the present invention.
- FIG. 10 is a diagram schematically showing the configurations of the conventional and actual database of the present invention.
- FIG. 11 is a diagram showing prediction errors of the conventional method and the method of the present invention in predicting the tensile strength of a highly workable high-strength cold-rolled steel sheet.
- FIG. 12 is a diagram showing prediction errors of the conventional method and the method of the present invention in predicting the hardness of the front and back surfaces of a thick steel plate.
- FIG. 13 is a diagram showing the error rate of the conventional method and the method of the present invention in predicting defects on the front and back surfaces of a hot-dip galvanized steel sheet.
- FIG. 14 is a diagram showing an intensity distribution when the manufacturing conditions of the subsequent steps are changed based on the quality prediction results of the conventional method and the method of the present invention.
- FIG. 15 is a diagram showing an outline of a quality abnormality analysis method according to an embodiment of the present invention.
- FIG. 16 is an example of the quality abnormality analysis method according to the embodiment of the present invention, and shows the transition of the actual value, the predicted value, and the quality prediction error (actual value-predicted value) of the objective variable (strength).
- FIG. 17 is an example of the quality abnormality analysis method according to the embodiment of the present invention, in which the values of the quality contribution of each explanatory variable (each manufacturing condition) time-integrated in the shaded section are shown in the top 20 in descending order. The figure drawn by the histogram is shown.
- FIG. 18 is a diagram showing a transition of the manufacturing condition H in FIG.
- Quality prediction model generation method, quality prediction model, quality prediction method, metal material manufacturing method, quality prediction model generation device, quality prediction device, quality abnormality analysis method, metal material manufacturing method and quality abnormality according to the embodiment of the present invention.
- the analyzer will be described with reference to the drawings.
- the quality prediction device is a device for predicting the quality of a metal material manufactured through one or a plurality of steps (processes).
- the metal material in the present embodiment include semi-finished products such as slabs and products such as steel plates manufactured by rolling the slabs, which are steel products.
- the quality prediction device 1 is specifically realized by a general-purpose information processing device such as a personal computer or a workstation.
- the quality prediction device 1 has, for example, a processor including a CPU (Central Processing Unit) and a memory (main storage unit) including a RAM (Random Access Memory) and a ROM (Read Only Memory) as main components.
- a processor including a CPU (Central Processing Unit) and a memory (main storage unit) including a RAM (Random Access Memory) and a ROM (Read Only Memory) as main components.
- a processor including a CPU (Central Processing Unit) and a memory (main storage unit) including a RAM (Random Access Memory) and a ROM (Read Only Memory) as main components.
- RAM Random Access Memory
- ROM Read Only Memory
- the quality prediction device 1 includes a manufacturing record collecting unit 11, a manufacturing record editing unit 12, a front-end replacement record collecting unit 13, a front-back replacement record collecting unit 14, and a cutting record collecting unit. It is equipped with 15.
- the quality prediction device 1 includes an integrated process performance editing unit 16, a performance database 17, a model generation unit 18, and a quality prediction unit 19.
- the quality prediction model generation device according to the present embodiment is composed of elements of the quality prediction device 1 excluding the quality prediction unit 19. In the following, the quality prediction model generation device will also be described in the description of the quality prediction device 1.
- a sensor (not shown) is connected to the manufacturing record collection unit 11.
- the manufacturing result collecting unit 11 collects the manufacturing results of each process according to the measurement cycle of this sensor, and outputs the manufacturing results to the integrated process result editing unit 16.
- the above-mentioned “manufacturing record” includes the manufacturing conditions of each process and the quality of the metal material manufactured through each process. Further, the above-mentioned “manufacturing conditions” include the components of the metal material in each process, temperature, pressure, plate thickness, plate passing speed, and the like. Further, the above-mentioned “quality of metal material” includes tensile strength, defect mixing rate (number of defects expressed per unit length), and the like.
- the manufacturing conditions of each process collected by the manufacturing record collecting unit 11 include not only the measured values of the manufacturing conditions measured by the sensors but also the set values of the manufacturing conditions set in advance. That is, depending on the process, the sensor may not be installed. In such a case, the set value is collected as the actual manufacturing value instead of the actual value.
- the manufacturing record collection unit 11 collects the manufacturing conditions of each process for each predetermined range of predetermined metal materials. Further, the manufacturing record collecting unit 11 evaluates and collects the quality of the metal material manufactured through each process for each of the predetermined ranges described above.
- the above-mentioned "predetermined range” means, for example, a certain range in the longitudinal direction of the metal material when the metal material is a slab or a steel plate. This predetermined range is determined based on the moving distance (passing speed) of the metal material according to the transport direction in each step. The specific processing contents by the manufacturing record collecting unit 11 will be described later (see FIG. 2).
- the manufacturing record data of each process (hereinafter referred to as “actual data”) is provided by this one manufacturing record collecting unit 11. Is supposed to be collected. However, for example, a plurality of manufacturing record collecting units 11 may be provided according to the number of each process, and the actual data of each process may be collected by different manufacturing record collecting units 11.
- the manufacturing record editing unit 12 edits the actual data of each process input from the manufacturing record collecting unit 11. That is, the manufacturing record editing unit 12 edits the actual data collected by the manufacturing record collecting unit 11 in hourly units into the actual data in length units of the metal material, and outputs it to the integrated process actual result editing unit 16.
- the specific processing contents by the manufacturing record editorial unit 12 will be described later (see FIG. 2).
- a material charging machine (not shown) for charging a metal material into each process is connected to the tip-tail replacement record collecting unit 13. Whether or not the tail end of the metal material has been replaced (reversed) when the metal material is charged from the front-end process to the back-end process through this material charging machine. The actual data is collected for each metal material. Then, the tail end replacement result collecting unit 13 outputs the actual data regarding the presence or absence of the replacement of the front end of the metal material to the integrated process result editing unit 16.
- the material charging machine described above is connected to the front / back replacement record collecting unit 14.
- the front / back surface replacement record collecting unit 14 determines whether the front and back surfaces of the metal material have been replaced (reversed) when the metal material is charged from the front-end process to the back-end process through this material charging machine. Collect actual data for each metal material. Then, the front / back surface replacement result collecting unit 14 outputs the actual data regarding the presence / absence of replacement of the front and back surfaces of the metal material to the integrated process result editing unit 16.
- a cutting machine (not shown) for cutting the tip end portion and the tail end portion of the metal material is connected to the cutting record collecting unit 15.
- the cutting record collecting unit 15 collects actual data such as the cutting position (distance from the tip of the metal material at the time of cutting) and the number of cuttings (hereinafter referred to as "cutting position") of the metal material. Collect each. Then, the cutting result collecting unit 15 outputs the actual data regarding the cutting position and the like of the metal material to the integrated process actual result editing unit 16.
- the front end replacement record collection unit 13 the front and back surface replacement record collection unit 14, and the cutting record collection unit 15, or the number of each process.
- a plurality of them may be provided according to the above.
- the integrated process performance editing unit 16 edits the performance data input from the manufacturing performance editing unit 12, the front and back replacement performance collection unit 13, the front and back replacement performance collection unit 14, and the cutting performance collection unit 15.
- the integrated process result editing unit 16 stores the manufacturing conditions of each process and the quality of the metal material manufactured under these manufacturing conditions in the performance database 17 in association with each predetermined range.
- the integrated process result editing unit 16 specifies a predetermined range in consideration of whether or not the tip and tail ends of the metal material are replaced in each process, whether or not the front and back surfaces are replaced, and the cutting position. Then, the integrated process result editing unit 16 determines the manufacturing conditions of each process and the quality of the metal material manufactured under these manufacturing conditions, whether or not the tail end of the metal material is replaced in each process, and the front and back surfaces. It is stored in the performance database 17 in a form in which the presence / absence of replacement and the cutting position can be distinguished. Further, the integrated process result editing unit 16 stores the manufacturing conditions of each process and the quality of the metal material manufactured under these manufacturing conditions in the performance database 17 in association with each predetermined range.
- the integrated process result editing unit 16 evaluates the volume from the tip of the metal material and specifies a predetermined range when, for example, each process is a rolling process and the shape of the metal material is deformed by passing through each process. .. Then, the manufacturing conditions of each process and the quality of the metal material manufactured under these manufacturing conditions are associated with each predetermined range and stored in the actual database 17. In the actual result database 17, the actual data edited by the integrated process actual result editing unit 16 is accumulated.
- the model generation unit 18 generates a quality prediction model that predicts the quality of the metal material for each predetermined range from the manufacturing conditions for each predetermined range in each process stored in the actual database 17.
- the model generation unit 18 uses, for example, XGBoost as a machine learning method.
- XGBoost a machine learning method
- various methods such as linear regression, local regression, principal component regression, PLS regression, neural network, regression tree, and random forest can be used.
- the quality prediction unit 19 predicts the quality of the metal material manufactured under arbitrary manufacturing conditions for each predetermined range by using the quality prediction model generated by the model generation unit 18. For example, when the metal material to be predicted is a slab, the quality of the entire slab is predicted by the conventional method, but in the present embodiment, the quality within a predetermined range in the length direction of the slab can be predicted.
- the manufacturing record collection unit 11 collects record data regarding the manufacturing conditions and quality of each process (step S1).
- the manufacturing record collection unit 11 collects manufacturing condition and quality record data of each process for each metal material and for each process.
- the actual data collected by the manufacturing actual collection unit 11 is data in which actual values (or installation values) of a plurality of manufacturing conditions are arranged for each time.
- the actual data shown in the figure shows the time t 1 , t 2 ..., the speed of the metal material (plate passing speed) v 1 , v 2 ... at the time, and a plurality of manufacturing conditions measured by the sensor at the time. It has items consisting of x 1 1 , x 1 2 ..., x 2 1 , x 2 2 ....
- the actual data collected in the final process includes items related to the quality of the metal material in addition to the items shown in the figure.
- the front-end replacement record collection unit 13, the front-back replacement record collection unit 14, and the cutting record collection unit 15 collect the record data (step S2).
- This actual data is actual data regarding the presence / absence of replacement of the tail end of the metal material in each process, the presence / absence of replacement of the front and back surfaces of the metal material in each process, the cutting position of the metal material in each process, and the like.
- the manufacturing record editing unit 12 converts the record data collected by the manufacturing record collecting unit 11 into units of length of the metal material (step S3). That is, the manufacturing result editing unit 12 converts the actual data collected in the time unit as shown in FIG. 3 into the actual data in the length unit of the metal material as shown in FIG.
- a method of converting the actual data of FIG. 3 into the actual data of FIG. 4 will be described.
- the manufacturing record editing unit 12 calculates the position of the metal material at each time in FIG. 3 by utilizing the property that the distance is obtained by multiplying the time and the speed (passing speed). Next, the manufacturing record editing unit 12 has the property that the record data is recorded when the metal material passes through the sensors installed in each process, and the missing value is recorded when the metal material does not pass. Utilize to detect the tip of a metallic material. Next, the manufacturing record editing unit 12 creates performance data corresponding to the position from the tip to the tail of the metal material, except when the metal material has not passed through the sensor.
- the data is in the length unit of the metal material as it is, it is not the data in the fixed period. Therefore, for example, by performing linear interpolation or the like, the data is converted into the actual data in the length unit of the metal material and the fixed period. .. That is, in each process, when the plate passing speed of the metal material is slow, the actual data that can be collected becomes fine, and when the plate passing speed of the metal material is high, the actual data that can be collected becomes coarse. Therefore, the above interpolation is performed in order to make the roughness of the actual data uniform.
- the manufacturing record editing unit 12 creates performance data in units of length of the metal material as shown in FIG. 4 by performing the above processing.
- the integrated process performance editing unit 16 aligns and combines the performance data of all processes in units of length of the metal material (step S4).
- the integrated process results editorial unit 16 includes performance data in units of length of the metal material, presence / absence of replacement of the tip and tail ends of the metal material, presence / absence of replacement of the front and back surfaces of the metal material, and actual data regarding the cutting position of the metal material. Based on, the actual data is combined. That is, the integrated process performance editing unit 16 prepares a plurality of manufacturing conditions and quality performance data of the metal material of all processes in units of the length of the metal material on the side of the final process based on the above performance data. And combine. The actual data in units of length of the metal material is created by the manufacturing actual editing unit 12.
- the actual data regarding the presence or absence of the replacement of the tail end of the metal material is collected by the tail end replacement record collecting unit 13.
- the actual data regarding whether or not the front and back surfaces of the metal material are exchanged is collected by the front and back surface exchange results collecting unit 14.
- the actual data regarding the cutting position of the metal material and the like are collected by the cutting actual collecting unit 15.
- the integrated process result editing unit 16 associates the manufacturing conditions of each process with the quality of the metal material manufactured under these manufacturing conditions for each predetermined range in the length direction of the metal material, and the results are obtained. Save in database 17.
- an example of processing by the integrated process result editing unit 16 will be described.
- Steps 1 to 3 are, for example, rolling steps, and the length of the material in the longitudinal direction increases with each step.
- the material A is divided into the material A1 and the material A2 when moving from the step 1 to the step 2, and the material A1 becomes the material A11 and the material A12 when moving from the step 2 to the step 3. It is divided.
- FIG. 6 shows an image of the material of each process, and is a diagram focusing on the part B of FIG.
- the manufacturing record collecting unit 11 collects the actual data of M1 items of, for example, X 1 1 to X 1 M 1 every 50 mm in the range of the length of 5300 mm from the tip to the tail end.
- the cutting actual data collecting unit 15 is collected by the cutting actual data collecting unit 15.
- the tip portion of 0 mm (tip) to 250 mm was truncated, the material A1 was taken at 250 mm to 3300 mm, the material A2 was taken at 3300 mm to 4950 mm, and the tail end portion of 4950 mm to 5300 mm (tail end) was truncated. Actual data is collected.
- the manufacturing record collecting unit 11 collects the actual data of M2 items of, for example, X 2 1 to X 2 M 2 every 100 mm in the range of the length from the tip to the tail end of 68000 mm. Further, for the material A1, the following actual data is collected by the cutting actual data collecting unit 15.
- the tip portion of 0 mm (tip) to 500 mm was truncated, the material A11 was taken at 500 mm to 34500 mm, the material A12 was taken at 34500 mm to 66800 mm, and the tail end portion of 66800 mm to 68000 mm (tail end) was truncated. Actual data is collected.
- the manufacturing record collecting unit 11 collects the actual data of M3 items of , for example, X3 1 to X3 M3 every 500 mm in the range of the length from the tip to the tail end of 65,000 mm. Further, in the material A11, the tip portion of 0 mm (tip) to 2500 mm is truncated by the cutting record collecting unit 15, the material A11 is taken at 2500 mm to 59700 mm, and the tail end portion of 59700 mm to 65000 mm (tail end) is truncated. Actual data is collected.
- the integrated process result editing unit 16 is finely collected in the longitudinal direction by a sensor (not shown) while considering actual data such as the presence / absence of replacement of the tail end of the metal material in each process, the presence / absence of replacement of the front and back surfaces, and the cutting position.
- Process the actual data of all processes That is, in order to combine the actual data of all the processes into the length unit of the metal material in the final process, as shown in FIG. 6, according to the material length of the process 3 which is the final process, the process 2 and the process 1 Scale the material length (see broken line in the figure).
- the integrated process result editing unit 16 specifies the position where each metal material is taken while considering the tip portion and the tail end portion cut off in each process. Then, in each predetermined range of the metal material in the final process, the quality in the predetermined range is associated with the manufacturing conditions of all the processes in the predetermined range, and stored in the actual database 17. For example, in FIG. 6, the shaded portion where the material A11 is taken in the final step 3 is specified retroactively to the material A1 in the step 2 and the material A in the step 1. By repeating such processing for all metal materials, as shown in FIG. 7, actual data of a plurality of manufacturing conditions (and quality) of the metal materials in all processes are arranged in units of length of the metal materials. Create combined performance data. Hereinafter, the description will be continued by returning to FIG.
- the model generation unit 18 generates a quality prediction model that predicts the quality of the metal material from the manufacturing conditions in each process (step S4). Subsequently, the quality prediction unit 19 predicts the quality of the metal material manufactured under arbitrary manufacturing conditions for each predetermined range by using the quality prediction model generated by the model generation unit 18 (step S5).
- the quality prediction model generation method quality prediction model, quality prediction method, quality prediction model generation device, and quality prediction device according to the present embodiment as described above, the following effects are obtained. That is, by generating a quality prediction model in which the manufacturing conditions of each process and the quality of the metal material manufactured under these manufacturing conditions are associated with each predetermined range, the quality of the metal material for any manufacturing condition can be obtained. , It is possible to predict with higher accuracy than before.
- the quality prediction model generation method, quality prediction model, quality prediction method, quality prediction model generation device, and quality prediction device have the following effects. That is, the actual data of a plurality of manufacturing conditions (and quality) of all processes is obtained by considering the replacement of the tail end, the replacement of the front and back surfaces, the cutting position, etc. in each process, and the length of the metal material on the exit side of the final process. Align and combine in units of length. Therefore, since the quality is predicted by effectively utilizing the actual data of the manufacturing conditions collected in detail in the longitudinal direction of the metal material by the sensor, the quality can be predicted with higher accuracy than before.
- the quality prediction method according to the present embodiment is applied to the method for manufacturing a metal material, for example, the following processing is performed. First, the production conditions determined during the production of the metal material are fixed, and then the quality of the metal material produced under the fixed production conditions is predicted for each predetermined range by the quality prediction method according to the present embodiment. Then, based on the prediction result, the manufacturing conditions of the subsequent processes are changed. In addition, the change in manufacturing conditions is such that the quality of all predetermined ranges included over the entire length of the metal material to be manufactured falls within a predetermined control range.
- the quality prediction method according to the present embodiment to the method for manufacturing a metal material in this way, it is possible to predict the final quality of the metal material in the middle of manufacturing, and the manufacturing conditions are changed accordingly. Therefore, the quality of the metal material to be manufactured is improved.
- the quality abnormality analysis device is a device for analyzing the cause of quality abnormality of a product manufactured by a manufacturing process.
- Examples of the product in the present embodiment include semi-finished products such as slabs and products such as steel plates manufactured by rolling the slabs, which are steel products.
- the quality abnormality analysis device 2 is specifically realized by a general-purpose information processing device such as a personal computer or a workstation.
- a processor including a CPU and a memory (main storage unit) including a RAM and a ROM are realized.
- a main storage unit including a RAM and a ROM are realized.
- Etc. are the main components.
- the quality abnormality analysis device 2 includes a quality prediction unit 21, a quality evaluation unit 22, a quality prediction error calculation unit 23, a quality contribution calculation unit 24, and a quality abnormality cause presentation unit 25. It is equipped with.
- the quality prediction unit 21 inputs arbitrary manufacturing conditions to a quality prediction model generated in advance with a plurality of manufacturing conditions of the manufacturing process collected from the actual plant 3 as input variables and product quality as an output variable. Predicts the quality of the product.
- the quality prediction unit 21 outputs a quality prediction value as a result of the quality prediction.
- the quality prediction model used in the quality prediction unit 21 is generated by using machine learning including, for example, linear regression, local regression, principal component regression, PLS regression, neural network, regression tree, random forest, and XGBoost. Further, the quality prediction model may be a model generated by the quality prediction model generation method (see FIG. 2) according to the above-described embodiment.
- the quality prediction model is a quality prediction model of a metallic material manufactured through one or more steps, and goes through a first collection step, a second collection step, a storage step, and a quality prediction model generation step. Generated.
- the manufacturing record collection unit 11 collects the manufacturing conditions of each process for each predetermined range of predetermined metallic materials. Further, in the second collection step, the manufacturing record collecting unit 11 (see the figure) evaluates and collects the quality of the metal material manufactured through each step for each predetermined range. Further, in the storage step, the integrated process performance editing unit 16 (see the figure) stores the manufacturing conditions of each process and the quality of the metal material manufactured under these manufacturing conditions in association with each predetermined range. .. Further, in the quality prediction model generation step, the model generation unit 18 (see the figure) generates a quality prediction model for predicting the quality of the metal material from the manufacturing conditions in each of the stored steps.
- the quality evaluation unit 22 calculates the quality evaluation value of the actual product manufactured by the manufacturing process. Examples of the quality evaluation value include the strength of cold-rolled thin steel sheets. Specifically, the quality evaluation unit 22 is composed of measuring equipment, material test equipment, and the like.
- the quality prediction error calculation unit 23 calculates the difference between the quality prediction value obtained as the output of the quality prediction unit 21 and the quality evaluation value obtained as the output of the quality evaluation unit 22 as the quality prediction error.
- the quality prediction error calculation unit 23 sequentially calculates an error between the actual quality evaluation value calculated by the quality evaluation unit 22 and the quality prediction value each time the quality prediction unit 21 makes a quality prediction using the quality prediction model. evaluate.
- the quality contribution calculation unit 24 calculates the quality contribution of each input manufacturing condition when predicting the quality of the product using the quality prediction model.
- the quality abnormality cause presenting unit 25 presents the manufacturing conditions that cause the quality abnormality of the product to the display unit 4 based on the quality prediction error and the quality contribution.
- the display unit 4 is a data output means processed by the quality abnormality analysis device 2, and is composed of, for example, an LCD (liquid crystal display), an OLED (organic EL display), or the like.
- the quality abnormality cause presenting unit 25 calculates the quality contribution degree based on each partial regression coefficient of the quality prediction model and the value of each variable.
- the quality abnormality cause presenting unit 25 presents the quality prediction error and the temporal integrated value of the quality contribution of each manufacturing condition in chronological order, and is a candidate for the manufacturing condition that causes the quality prediction error and the quality abnormality.
- the temporal transition is visualized and presented to the display unit 4. In this way, by presenting the temporal transition of the candidate manufacturing conditions that cause the quality prediction error and the quality abnormality, it is possible to easily grasp the manufacturing conditions that are presumed to be the cause of the quality abnormality.
- the quality abnormality cause presenting unit 25 pays attention to the manufacturing condition having a large quality contribution, and from the manufacturing condition having a large temporal integral value of the quality contribution, the quality abnormality. They are ordered and presented to the display unit 4 as candidates for the manufacturing conditions that cause the above. In this way, by presenting the manufacturing conditions having a large temporal integral value of the quality contribution side by side, it is possible to easily grasp the manufacturing conditions presumed to be the cause of the quality abnormality.
- the specific processing contents of the quality evaluation unit 22, the quality prediction error calculation unit 23, the quality contribution calculation unit 24, and the quality abnormality cause presentation unit 25 will be described in Examples described later.
- the quality abnormality analysis method according to the present embodiment will be described with reference to FIG. In the quality abnormality analysis method according to the present embodiment, the processes of steps S11 to S15 shown in FIG. 9 are performed.
- the quality prediction unit 21 predicts the quality of the product by inputting the manufacturing conditions to the quality prediction model generated in advance (step S11). Subsequently, the quality evaluation unit 22 calculates the quality evaluation value of the actual product manufactured by the manufacturing process (step S12). Subsequently, the quality prediction error calculation unit 23 calculates the difference between the quality prediction value obtained in step S11 and the quality evaluation value obtained in step S12 as a quality prediction error (step S13).
- the quality contribution calculation unit 24 calculates the quality contribution of each manufacturing condition input to the quality prediction model when predicting the quality (step S14).
- the quality abnormality cause presenting unit 25 presents the manufacturing conditions that cause the quality abnormality of the product to the display unit 3 based on the quality prediction error and the quality contribution (step S15).
- quality abnormality is obtained by obtaining the quality prediction error and the quality contribution of each manufacturing condition by using the manufacturing conditions of each process and the quality prediction model that predicts the quality of the products manufactured under these manufacturing conditions. Can suggest the cause of.
- Example 1 An example of the quality prediction method according to the present embodiment will be described.
- the quality prediction method according to the present embodiment is applied to the prediction of the tensile strength of a high-strength cold-rolled steel sheet having high workability, which is a kind of cold-rolled thin steel sheet.
- the objective variable (quality) for quality prediction in this example is the tensile strength of the product (highly workable, high-strength cold-rolled steel sheet).
- the explanatory variables (manufacturing conditions) are the chemical composition of the metal material in the smelting process, the temperature of the metal material in the casting process, the temperature of the metal material in the heating process, the temperature of the metal material in the hot rolling process, and the metal in the cooling process. The temperature of the material. Further, the explanatory variables (manufacturing conditions) are the temperature of the metal material in the cold rolling process, the temperature of the metal material in the annealing process, and the like.
- each manufacturing condition and quality are predicted from the conventional performance database (see (a) in FIG. 10) in which representative values such as one average value are stored for each product. Further, in this embodiment, each manufacturing condition and quality are predicted from the quality prediction model generated from the actual database of the quality prediction method according to the present embodiment (see (b) in FIG. 10) for one product. .. Then, in this example, these two prediction results were compared. The number of samples in the performance database was 40,000, the number of explanatory variables was 45, and the prediction method used was local regression. As a result of the prediction, the prediction error by the quality prediction method according to the present embodiment (see (b) in FIG. 11) is a root average as compared with the prediction error by the conventional quality prediction method (see (a) in FIG. 11). It was confirmed that the square error (RMSE: Root Mean Square Error) can be reduced by 23%.
- RMSE Root Mean Square Error
- the quality prediction method according to this embodiment was applied to the prediction of the hardness of the front and back surfaces of a thick steel sheet.
- the objective variable is the hardness of the front and back surfaces of the product
- the explanatory variables are the chemical composition of the smelting process, the front and back temperature of the casting process, the front and back temperature of the heating process, the front and back temperature of the rolling process, and the front and back surfaces of the cooling process. Temperature etc.
- each manufacturing condition and quality were predicted from the conventional performance database (see (a) in FIG. 10) in which representative values such as one average value are stored for each product.
- each manufacturing condition and quality were predicted from the quality prediction model generated from the actual database of the quality prediction method according to the present embodiment (see (b) in FIG. 10) for one product.
- the two prediction results were compared.
- the number of samples in the performance database was 10,000, the number of explanatory variables was 30, and the prediction method used was linear regression.
- the prediction error by the quality prediction method according to the present embodiment is the prediction error by the conventional quality prediction method (see (a) in FIG. 12). It was confirmed that the root mean square error (RMSE) can be reduced by 26%.
- RMSE root mean square error
- the quality prediction method according to this embodiment was applied to the prediction of front and back defects of a hot-dip galvanized steel sheet, which is a kind of cold-rolled thin steel sheet.
- the objective variable is the presence or absence of defects on the front and back surfaces of the product.
- the explanatory variables are the chemical composition of the smelting process, the front and back temperature of the casting process, the meniscus flow velocity, the mold molten metal level, the front and back temperature of the heating process, the front and back temperature of the hot rolling process, and the front and back temperature of the cooling process. , The acid concentration in the pickling process, the acid temperature, and the front and back temperature of the cold pressure process. Further, the explanatory variables are the front and back temperature of the annealing process, the amount of plating adhered in the plating process, the degree of alloying, and the like.
- each manufacturing condition and quality were predicted from the conventional performance database (see (a) in FIG. 10) in which representative values such as one average value are stored for each product.
- each manufacturing condition and quality were predicted from the quality prediction model generated from the actual database of the quality prediction method according to the present embodiment (see (b) in FIG. 10) for one product.
- the two prediction results were compared.
- the number of samples in the performance database was 4000, the number of explanatory variables was 250, and the prediction method used a decision tree.
- the wrong answer rate by the quality prediction method according to the present embodiment is the wrong answer rate by the conventional quality prediction method ((a) of FIG. 13). It was confirmed that the reduction was 14% compared to (see).
- the quality prediction method according to the present embodiment is applied to the tensile strength prediction of a high-strength cold-rolled steel sheet, which is a kind of cold-rolled thin steel sheet, and based on the prediction result, the manufacturing conditions of the subsequent processes are changed.
- the cooling temperature after quenching which is the manufacturing condition of the final stage of the cold rolling process, is obtained at the stage during manufacturing where the actual values of the manufacturing conditions before the final stage of the steelmaking process, hot rolling process and rolling process are obtained.
- the cooling temperature after quenching which is the manufacturing condition of the final stage of the cold rolling process
- this embodiment Based on the actual values of the manufacturing conditions before the final stage of the steelmaking process, hot rolling process and cold rolling process, and the standard value of the cooling temperature after quenching, which is the manufacturing condition of the final stage of the cold rolling process, this embodiment is used.
- the predicted tensile strength values at each position of the total length of the product predicted using the quality prediction method are shown below.
- the amount of change in the tensile strength at each position of the total length of the product predicted by using the quality prediction method according to the present embodiment is shown as follows.
- y LL and y UL are the control lower limit and the control upper limit of the tensile strength, respectively, and ⁇ u * is the optimum solution of this optimization problem.
- This optimization problem can be solved by a mathematical programming method such as a branch-and-bound method. By changing the cooling temperature after the annealing temperature by ⁇ u * , it is possible to obtain a cold-rolled steel sheet in which the tensile strength of the entire length does not deviate from the control range, that is, the quality is not defective over the entire length.
- the actual data stored in the actual database of the quality prediction method according to the present embodiment can be as follows. That is, in each predetermined range of the metal material in the final process, the hardness or the presence / absence of defects and the manufacturing conditions of the entire process are determined while considering the actual data such as the presence / absence of replacement of the tail end, the presence / absence of replacement of the front and back surfaces, and the cutting position. It will be possible to retroactively combine precise actual data. Then, since the predicted value under arbitrary manufacturing conditions is calculated based on the quality prediction model generated from the performance database constructed in this way, it is possible to predict the quality of the metal material with high accuracy.
- the quality prediction method according to the present embodiment was applied to the strength prediction of one kind of cold-rolled thin steel sheet, and the manufacturing conditions of the subsequent processes were changed based on the prediction result.
- an example of changing the annealing temperature in the cold rolling process will be described in the middle of manufacturing where the actual values of the manufacturing conditions up to the steelmaking process and the hot rolling process have been obtained.
- the strength of each position of the total length of the product predicted by using the quality prediction method according to the present embodiment based on the actual values of the manufacturing conditions up to the steelmaking process and the hot rolling process, and the standard value of the annealing temperature of the cold rolling process.
- the predicted values are shown as follows.
- the annealing temperature changes by ⁇ u from the reference value
- the amount of change in the strength at each position of the total length of the product predicted by using the quality prediction method according to the present embodiment is shown as follows. Based on the above, the optimization problem expressed by the following equation (1) is solved.
- y LL and y UL are the control lower limit and the control upper limit of the strength, respectively, and ⁇ u * is the optimum solution of this optimization problem.
- This optimization problem can be solved by a mathematical programming method such as a branch-and-bound method. By changing the annealing temperature by ⁇ u * , it is possible to obtain a cold-rolled steel sheet in which the strength of the entire length does not deviate from the control range, that is, the quality is not defective over the entire length.
- FIG. 14A shows the intensity distribution when the manufacturing conditions of the subsequent steps are changed based on the prediction result by the conventional quality prediction method. Further, (b) in the figure shows the intensity distribution when the manufacturing conditions of the subsequent steps are changed based on the prediction result by the quality prediction method according to the present embodiment. As shown in the figure, by using the quality prediction method according to the present embodiment, it is possible to reduce the variation in strength.
- the actual data stored in the actual database of the quality prediction method according to the present embodiment can be as follows. That is, in each predetermined range of the metal material in the final process, the hardness or the presence / absence of defects and the manufacturing conditions of the entire process are determined while considering the actual data such as the presence / absence of replacement of the tail end, the presence / absence of replacement of the front and back surfaces, and the cutting position. It will be possible to retroactively combine precise actual data. Then, since the predicted value under arbitrary manufacturing conditions is calculated based on the quality prediction model generated from the performance database constructed in this way, it is possible to predict the quality of the metal material with high accuracy.
- Example 2 An example of the quality abnormality analysis method according to the present embodiment will be described.
- the quality abnormality analysis method according to the present embodiment is applied to the strength prediction of one kind of cold-rolled thin steel sheet, and the quality abnormality is analyzed based on the quality prediction result.
- FIG. 15 shows an outline of the quality abnormality analysis method according to the present embodiment.
- y indicates the actual value of the strength
- y ⁇ indicates the predicted value of the strength
- x 1 ... x M indicates the actual value of the manufacturing conditions up to the steelmaking process, the hot rolling process, and the cold rolling process.
- a quality prediction model is used to calculate and present the quality contribution Cx 1 ... C x M of each explanatory variable x 1 ... x M for the prediction error y-y ⁇ .
- the reason for calculating the quality prediction error is as follows. It is considered that the reason why the quality prediction error becomes large is that the relationship between the manufacturing conditions and the quality in the conventional manufacturing process has become different. Therefore, when the quality prediction error becomes large, an abnormality occurs in the plant, or the product is manufactured under manufacturing conditions outside the conventional range, and the quality of the product also becomes abnormal.
- the quality contribution of each manufacturing condition input to the quality prediction model can be calculated by, for example, the following equation (2).
- Cx k indicates the quality contribution
- a k indicates the standard partial regression coefficient
- x k with an overline indicates the average value of the explanatory variables x k .
- the explanatory variable (manufacturing condition) having the largest contribution to quality calculated by the above formula (2) is the cause of the quality abnormality of the product.
- the PLS regression model was used as the quality prediction model, and the absolute value of the standard partial regression coefficient was used as the quality contribution.
- the average value of the explanatory variables x k is the average of each explanatory variable calculated based on the normal data used when creating the PLS regression model.
- FIG. 16A shows the transition of the actual value and the predicted value of the objective variable (intensity) in the example of the quality abnormality analysis method according to the present embodiment. Further, (b) in the figure shows the transition of the quality prediction error (actual value-predicted value) of the objective variable (strength).
- the shaded sections (see parts A and B) have quality defects in which the actual strength values are larger than those in the other sections. In this example, the cause of the quality abnormality was analyzed by focusing on this shaded section.
- FIG. 17 shows the time-integrated values of the quality contribution of each explanatory variable (each manufacturing condition) in the shaded section (see parts A and B) of FIG. 16 drawn in a histogram in descending order. Is shown.
- the explanatory variables are arranged in descending order of the temporal integral value of the quality contribution degree, thereby contributing to the quality abnormality.
- the explanatory variables that are presumed to be are clarified.
- the quality contribution of the manufacturing condition H for example, the cooling condition of the cold rolling process
- the manufacturing condition H is very large as compared with the others, and therefore, the manufacturing condition H is focused on.
- FIG. 18 shows the transition of the manufacturing condition H (for example, the cooling condition of the cold rolling process).
- the value of the manufacturing condition H is very large in the shaded section (see part C), that is, in the section where the strength is continuously very large. From this, it can be seen that the cause of the abnormality should be analyzed by focusing on the manufacturing condition H, and that the manufacturing condition H is related to the quality abnormality. As described above, by using the quality abnormality analysis method according to the present embodiment, it is possible to identify the cause of the quality abnormality of the product from, for example, about 1000 kinds of manufacturing conditions.
- the manufacturing conditions that contribute to the quality are clarified. Therefore, the following can be done.
- the quality of the final product is predicted by a quality prediction model at a stage during manufacturing, that is, at a stage where an arbitrary manufacturing process before carrying out the final manufacturing process is completed. Then, based on the prediction result, it is possible to select and change the manufacturing conditions that have a high quality contribution and can be changed, which are the manufacturing conditions of the subsequent manufacturing process.
- the quality of the final product is within the preset control range over the entire length of the product for the selected changeable manufacturing conditions. It will be possible. This makes it possible to manufacture a metal material having good product quality over the entire length of the product in a method for manufacturing a metal material manufactured through a plurality of manufacturing steps.
- the quality prediction model generation method, the quality prediction model, the quality prediction method, the metal material manufacturing method, the quality prediction model generation device, the quality prediction device, the quality abnormality analysis method, the metal material manufacturing method and the quality abnormality analysis according to the present invention has been specifically described with reference to embodiments and examples for carrying out the invention.
- the gist of the present invention is not limited to these statements, and must be broadly interpreted based on the statements of the claims. Needless to say, various changes, modifications, etc. based on these descriptions are also included in the gist of the present invention.
- the above-mentioned integrated process result editing unit 16 specified a predetermined range in consideration of whether or not the tip and tail ends of the metal material were replaced in each process, whether or not the front and back surfaces were replaced, and the cutting position.
- the integrated process result editing unit 16 considers at least one of the actual data of the presence / absence of replacement of the tail end of the metal material, the presence / absence of replacement of the front and back surfaces of the metal material, and the actual data of the cutting position of the metal material.
- a predetermined range may be specified.
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Manufacturing & Machinery (AREA)
- Automation & Control Theory (AREA)
- Quality & Reliability (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- General Engineering & Computer Science (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- General Factory Administration (AREA)
- Investigating And Analyzing Materials By Characteristic Methods (AREA)
- Analysing Materials By The Use Of Radiation (AREA)
Abstract
Description
本実施形態に係る品質予測装置および品質予測モデル生成装置の構成について、図1を参照しながら説明する。品質予測装置は、一つまたは複数の工程(プロセス)を経て製造される金属材料の品質を予測するための装置である。なお、本実施形態における金属材料としては、例えば鉄鋼製品であって、スラブ等の半製品や、このスラブを圧延して製造される鋼板等の製品が挙げられる。
本実施形態に係る品質予測方法および品質予測モデル生成方法について、図2~図7を参照しながら説明する。本実施形態に係る品質予測方法は、図2に示したステップS1~ステップS6の処理を行う。また、本実施形態に係る品質予測モデル生成方法は、同図に示したステップS6を除いたステップS1~ステップS5の処理を行う。
本実施形態に係る品質異常解析装置の構成について、図8を参照しながら説明する。品質異常解析装置は、製造プロセスにより製造された製品の品質異常の原因を解析するための装置である。なお、本実施形態における製品としては、例えば鉄鋼製品であって、スラブ等の半製品や、このスラブを圧延して製造される鋼板等の製品が挙げられる。
本実施形態に係る品質異常解析方法について、図9を参照しながら説明する。本実施形態に係る品質異常解析方法は、図9に示したステップS11~ステップS15の処理を行う。
本実施形態に係る品質予測方法の実施例について説明する。本実施例では、本実施形態に係る品質予測方法を、冷延薄鋼板の一種である高加工性高強度冷延鋼板の引張強度の予測に対して適用した。
本実施形態に係る品質異常解析方法の実施例について説明する。本実施例では、本実施形態に係る品質異常解析方法を、冷延薄鋼板のある一品種の強度予測に対して適用し、品質の予測結果に基づいて、品質異常の解析を行った。
11 製造実績収集部
12 製造実績編集部
13 先尾端入替実績収集部
14 表裏面入替実績収集部
15 切断実績収集部
16 一貫工程実績編集部
17 実績データベース
18 モデル生成部
19 品質予測部
2 品質異常解析装置
21 品質予測部
22 品質評価部
23 品質予測誤差算出部
24 品質寄与度算出部
25 品質異常原因提示部
3 実プラント
4 表示部
Claims (8)
- 製造プロセスにより製造された製品の品質異常解析方法であって、
前記製造プロセスの複数の製造条件を入力変数とし、前記製品の品質を出力変数として生成された品質予測モデルに対して、前記製造条件を入力することにより、前記製品の品質を予測する品質予測ステップと、
前記製造プロセスにより製造された実際の製品の品質評価値を算出する品質評価ステップと、
前記品質予測ステップの出力として得られた品質予測値と、前記品質評価値との差を、品質予測誤差として算出する品質予測誤差算出ステップと、
前記品質予測モデルを用いて前記製品の品質を予測する際に、入力した各製造条件の品質寄与度を算出する品質寄与度算出ステップと、
前記品質予測誤差および前記品質寄与度に基づいて、前記製品の品質異常の原因となる製造条件を提示する品質異常原因提示ステップと、
を含む品質異常解析方法。 - 前記品質寄与度算出ステップは、前記品質予測モデルの各偏回帰係数と各変数の値とに基づいて、前記品質寄与度を算出する請求項1に記載の品質異常解析方法。
- 前記品質異常原因提示ステップは、前記品質予測誤差および前記各製造条件の品質寄与度の時間的な積分値を時系列で提示し、かつ前記品質予測誤差と前記品質異常の原因となる製造条件の候補の時間的推移を可視化して提示する請求項1または請求項2に記載の品質異常解析方法。
- 前記品質異常原因提示ステップは、前記品質予測誤差が所定の値を超えた場合、前記品質寄与度が大きい製造条件に着目し、前記品質寄与度の時間的な積分値が大きな製造条件から、前記品質異常の原因となる製造条件の候補として順序付けて提示する請求項1から請求項3のいずれか一項に記載の品質異常解析方法。
- 前記品質予測モデルは、線形回帰、局所回帰、主成分回帰、PLS回帰、ニューラルネットワーク、回帰木、ランダムフォレスト、XGBoostを含む機械学習を用いて生成される請求項1から請求項4のいずれか一項に記載の品質異常解析方法。
- 前記品質予測モデルは、
一つまたは複数の工程を経て製造される金属材料の品質予測モデルであり、
各工程の製造条件を、予め定めた前記金属材料の所定範囲ごとに収集する第一の収集ステップと、
前記各工程を経て製造される前記金属材料の品質を、前記所定範囲ごとに評価して収集する第二の収集ステップと、
前記各工程の製造条件と、この製造条件の下で製造される前記金属材料の品質とを、前記所定範囲ごとに関連付けて保存する保存ステップと、
保存した前記各工程における前記所定範囲ごとの製造条件から、前記金属材料の前記所定範囲ごとの品質を予測する前記品質予測モデルを生成する品質予測モデル生成ステップと、
を経て生成される請求項1から請求項5のいずれか一項に記載の品質異常解析方法。 - 複数の製造工程を経て製造される金属材料の製造方法であって、
最終製造工程を実施する前の任意の製造工程が終了した段階で、請求項6に記載の品質異常解析方法によって生成された品質予測モデルにより、最終製品の品質を予測し、
その予測結果に基づいて、その後の製造工程の製造条件であって、品質寄与度が高く、かつ変更可能な製造条件を選択し、最終製品の品質が、全長にわたって予め設定した品質管理内に入るように、前記選択された製造条件を決定して操業する、
金属材料の製造方法。 - 製造プロセスにより製造された製品の品質異常解析装置であって、
前記製造プロセスの複数の製造条件を入力変数とし、前記製品の品質を出力変数として生成された品質予測モデルに対して、前記製造条件を入力することにより、前記製品の品質を予測する品質予測手段と、
前記製造プロセスにより製造された実際の製品の品質評価値を算出する品質評価手段と、
前記品質予測手段の出力として得られた品質予測値と、前記品質評価値との差を、品質予測誤差として算出する品質予測誤差算出手段と、
前記品質予測モデルを用いて前記製品の品質を予測する際に、入力した各製造条件の品質寄与度を算出する品質寄与度算出手段と、
前記品質予測誤差および前記品質寄与度に基づいて、前記製品の品質異常の原因となる製造条件を提示する品質異常原因提示手段と、
を備える品質異常解析装置。
Priority Applications (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202180089024.7A CN116724328A (zh) | 2021-01-06 | 2021-11-08 | 品质异常分析方法、金属材料的制造方法以及品质异常分析装置 |
EP21917569.2A EP4258069A4 (en) | 2021-01-06 | 2021-11-08 | METHOD FOR ANALYSIS OF QUALITY ANOMALIES, METHOD FOR PRODUCING METAL MATERIAL AND DEVICE FOR ANALYSIS OF QUALITY ANOMALIES |
KR1020237022030A KR20230116870A (ko) | 2021-01-06 | 2021-11-08 | 품질 이상 해석 방법, 금속 재료의 제조 방법 및 품질 이상 해석 장치 |
JP2022507731A JP7251687B2 (ja) | 2021-01-06 | 2021-11-08 | 品質異常解析方法、金属材料の製造方法および品質異常解析装置 |
MX2023008069A MX2023008069A (es) | 2021-01-06 | 2021-11-08 | Metodo de analisis de anomalias de calidad, metodo de fabricacion de material de metal y dispositivo de analisis de anomalias de calidad. |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2021000695 | 2021-01-06 | ||
JP2021-000695 | 2021-01-06 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022149344A1 true WO2022149344A1 (ja) | 2022-07-14 |
Family
ID=82357412
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2021/040936 WO2022149344A1 (ja) | 2021-01-06 | 2021-11-08 | 品質異常解析方法、金属材料の製造方法および品質異常解析装置 |
Country Status (7)
Country | Link |
---|---|
EP (1) | EP4258069A4 (ja) |
JP (1) | JP7251687B2 (ja) |
KR (1) | KR20230116870A (ja) |
CN (1) | CN116724328A (ja) |
MX (1) | MX2023008069A (ja) |
TW (1) | TWI808540B (ja) |
WO (1) | WO2022149344A1 (ja) |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2004355189A (ja) | 2003-05-28 | 2004-12-16 | Jfe Steel Kk | 結果予測装置 |
JP2006309709A (ja) | 2005-03-30 | 2006-11-09 | Jfe Steel Kk | 結果予測装置、制御装置及び品質設計装置 |
JP2008112288A (ja) | 2006-10-30 | 2008-05-15 | Jfe Steel Kk | 予測式作成装置、結果予測装置、品質設計装置、予測式作成方法及び製品の製造方法 |
JP2009230412A (ja) | 2008-03-21 | 2009-10-08 | Jfe Steel Corp | 結果予測装置、及び、これを用いた製品品質予測方法 |
JP2014013560A (ja) | 2012-06-04 | 2014-01-23 | Jfe Steel Corp | 結果予測装置および結果予測方法 |
JP2014071858A (ja) | 2012-10-02 | 2014-04-21 | Jfe Steel Corp | 結果予測方法及び結果予測装置 |
JP2014071859A (ja) | 2012-10-02 | 2014-04-21 | Jfe Steel Corp | 結果予測方法及び結果予測装置 |
JP2017120638A (ja) | 2015-12-24 | 2017-07-06 | Jfeスチール株式会社 | 結果予測装置及び結果予測方法 |
JP2018113027A (ja) * | 2017-01-10 | 2018-07-19 | Jfeスチール株式会社 | プロセスの異常状態診断方法および異常状態診断装置 |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TWI269990B (en) * | 2005-02-04 | 2007-01-01 | Univ Nat Cheng Kung | Quality prognostics system and method for manufacturing processes with generic embedded devices |
CN102033523B (zh) * | 2009-09-25 | 2014-01-01 | 上海宝钢工业检测公司 | 基于偏最小二乘的带钢质量预测、炉况预警与故障诊断方法 |
TWI625615B (zh) * | 2016-11-29 | 2018-06-01 | 財團法人工業技術研究院 | 預測模型建立方法及其相關預測方法與電腦程式產品 |
JP7020500B2 (ja) * | 2019-02-07 | 2022-02-16 | Jfeスチール株式会社 | 予測モデルの生成方法、金属材料の腐食量予測方法、予測モデルの生成プログラムおよび予測モデルの生成装置 |
CN111506030B (zh) * | 2020-04-16 | 2021-06-25 | 福州泰全工业有限公司 | 汽车生产线品质管理系统及方法 |
-
2021
- 2021-11-08 WO PCT/JP2021/040936 patent/WO2022149344A1/ja active Application Filing
- 2021-11-08 CN CN202180089024.7A patent/CN116724328A/zh active Pending
- 2021-11-08 KR KR1020237022030A patent/KR20230116870A/ko unknown
- 2021-11-08 EP EP21917569.2A patent/EP4258069A4/en active Pending
- 2021-11-08 JP JP2022507731A patent/JP7251687B2/ja active Active
- 2021-11-08 MX MX2023008069A patent/MX2023008069A/es unknown
- 2021-11-19 TW TW110143174A patent/TWI808540B/zh active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2004355189A (ja) | 2003-05-28 | 2004-12-16 | Jfe Steel Kk | 結果予測装置 |
JP2006309709A (ja) | 2005-03-30 | 2006-11-09 | Jfe Steel Kk | 結果予測装置、制御装置及び品質設計装置 |
JP2008112288A (ja) | 2006-10-30 | 2008-05-15 | Jfe Steel Kk | 予測式作成装置、結果予測装置、品質設計装置、予測式作成方法及び製品の製造方法 |
JP2009230412A (ja) | 2008-03-21 | 2009-10-08 | Jfe Steel Corp | 結果予測装置、及び、これを用いた製品品質予測方法 |
JP2014013560A (ja) | 2012-06-04 | 2014-01-23 | Jfe Steel Corp | 結果予測装置および結果予測方法 |
JP2014071858A (ja) | 2012-10-02 | 2014-04-21 | Jfe Steel Corp | 結果予測方法及び結果予測装置 |
JP2014071859A (ja) | 2012-10-02 | 2014-04-21 | Jfe Steel Corp | 結果予測方法及び結果予測装置 |
JP2017120638A (ja) | 2015-12-24 | 2017-07-06 | Jfeスチール株式会社 | 結果予測装置及び結果予測方法 |
JP2018113027A (ja) * | 2017-01-10 | 2018-07-19 | Jfeスチール株式会社 | プロセスの異常状態診断方法および異常状態診断装置 |
Also Published As
Publication number | Publication date |
---|---|
EP4258069A1 (en) | 2023-10-11 |
TWI808540B (zh) | 2023-07-11 |
KR20230116870A (ko) | 2023-08-04 |
JPWO2022149344A1 (ja) | 2022-07-14 |
CN116724328A (zh) | 2023-09-08 |
MX2023008069A (es) | 2023-07-17 |
TW202244646A (zh) | 2022-11-16 |
EP4258069A4 (en) | 2024-06-05 |
JP7251687B2 (ja) | 2023-04-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2021014804A1 (ja) | 品質予測モデル生成方法、品質予測モデル、品質予測方法、金属材料の製造方法、品質予測モデル生成装置および品質予測装置 | |
JP4383493B2 (ja) | 780MPa以上のTSを持つハイテン出荷鋼板の材質情報提供方法及び材質情報利用方法 | |
EP2286935B1 (en) | Steel plate quality assurance system and method | |
CN107179749A (zh) | 热镀锌产品全流程质量控制方法 | |
KR102145465B1 (ko) | 산업 공정의 제품 특성 및 제조 비용의 최적화를 위한 방법 | |
KR20200017506A (ko) | 프로세스의 이상 상태 진단 장치 및 이상 상태 진단 방법 | |
CN110989510A (zh) | 一种热镀锌产品全流程质量控制与等级自动判断系统 | |
Uyan et al. | Industry 4.0 foundry data management and supervised machine learning in low-pressure die casting quality improvement | |
JP2009205554A (ja) | 機器の劣化度算出方法およびリスク評価方法 | |
JP5003362B2 (ja) | 製品品質の制御方法及び制御装置 | |
Kumar et al. | Framework for lean implementation through fuzzy AHP-COPRAS integrated approach | |
WO2022149344A1 (ja) | 品質異常解析方法、金属材料の製造方法および品質異常解析装置 | |
KR102589428B1 (ko) | 중간 금속 제품(들)의 그룹으로부터의 최종 금속 제품(들)의 그룹의 제조를 제어하기 위한 방법 및 전자 장치, 관련 컴퓨터 프로그램, 제조 방법 및 설비 | |
Gao et al. | Experimental methodology and predictive tools in austenitic stainless steel with different ferrite content welds | |
JP2009245141A (ja) | 出力値予測方法、該装置および該方法のプログラム | |
JP4926883B2 (ja) | 出荷鋼板の材質情報提供方法及び出荷鋼板の材質情報利用方法 | |
JP5577583B2 (ja) | 厚鋼板のオンライン材質保証システム | |
JP5577581B2 (ja) | 厚鋼板の温度履歴による材質保証システム | |
Musatova et al. | Operating cycles of automated production systems | |
JP7031713B1 (ja) | 異常診断モデルの構築方法、異常診断方法、異常診断モデルの構築装置および異常診断装置 | |
JP5577580B2 (ja) | 厚鋼板の材質保証システム | |
NOTRANJIH | Determination of the cause of the formation of transverse internal cracks on a continuously cast slab | |
Zhikharev et al. | Artificial Intelligence and Machine Learning In Metallurgy. Part 2. Application Examples | |
CN115130272A (zh) | 变形阻力预测系统、变形阻力预测方法以及变形阻力预测装置 | |
CN117808499A (zh) | 一种基于大数据的工程造价管理方法及系统 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
ENP | Entry into the national phase |
Ref document number: 2022507731 Country of ref document: JP Kind code of ref document: A |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21917569 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 20237022030 Country of ref document: KR Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202180089024.7 Country of ref document: CN |
|
WWE | Wipo information: entry into national phase |
Ref document number: MX/A/2023/008069 Country of ref document: MX |
|
ENP | Entry into the national phase |
Ref document number: 2021917569 Country of ref document: EP Effective date: 20230704 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |