CN115345434A - Improved dynamic data mining method and device for continuous casting quality judgment model - Google Patents
Improved dynamic data mining method and device for continuous casting quality judgment model Download PDFInfo
- Publication number
- CN115345434A CN115345434A CN202210833764.3A CN202210833764A CN115345434A CN 115345434 A CN115345434 A CN 115345434A CN 202210833764 A CN202210833764 A CN 202210833764A CN 115345434 A CN115345434 A CN 115345434A
- Authority
- CN
- China
- Prior art keywords
- quality
- time period
- continuous casting
- index parameter
- quality index
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 188
- 238000009749 continuous casting Methods 0.000 title claims abstract description 107
- 238000007418 data mining Methods 0.000 title claims abstract description 32
- 230000008569 process Effects 0.000 claims abstract description 139
- 238000001514 detection method Methods 0.000 claims abstract description 99
- 238000005259 measurement Methods 0.000 claims abstract description 94
- 238000005266 casting Methods 0.000 claims description 26
- 230000015654 memory Effects 0.000 claims description 19
- 238000004590 computer program Methods 0.000 claims description 17
- 238000012549 training Methods 0.000 claims description 15
- 238000005070 sampling Methods 0.000 claims description 11
- 238000003860 storage Methods 0.000 claims description 8
- 238000010606 normalization Methods 0.000 claims description 5
- 230000002457 bidirectional effect Effects 0.000 claims description 3
- 230000006403 short-term memory Effects 0.000 claims description 3
- 230000007787 long-term memory Effects 0.000 claims description 2
- 238000004519 manufacturing process Methods 0.000 abstract description 28
- 230000007547 defect Effects 0.000 description 19
- 238000013461 design Methods 0.000 description 8
- 229910000831 Steel Inorganic materials 0.000 description 7
- 230000000875 corresponding effect Effects 0.000 description 7
- 238000012545 processing Methods 0.000 description 7
- 239000010959 steel Substances 0.000 description 7
- 238000010924 continuous production Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 4
- 238000007689 inspection Methods 0.000 description 4
- 238000010923 batch production Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 230000001360 synchronised effect Effects 0.000 description 3
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 239000002994 raw material Substances 0.000 description 2
- 238000007711 solidification Methods 0.000 description 2
- 230000008023 solidification Effects 0.000 description 2
- 238000003723 Smelting Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000012942 design verification Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000004512 die casting Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 229910052742 iron Inorganic materials 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 229910052751 metal Inorganic materials 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 239000002893 slag Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Images
Classifications
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/217—Database tuning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2474—Sequence data queries, e.g. querying versioned data
-
- 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
-
- 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
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Databases & Information Systems (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Educational Administration (AREA)
- Entrepreneurship & Innovation (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Manufacturing & Machinery (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- General Factory Administration (AREA)
Abstract
An improved dynamic data mining method and device for a continuous casting quality judgment model comprises the following steps: acquiring the occurrence time of the product quality accident and the type of the product quality accident; determining a quality index parameter associated with the product quality accident according to the type of the product quality accident and preset prior knowledge information; calculating the detection time period of the quality index parameters according to the occurrence time; acquiring an actual measured value of a quality index parameter in a detection time period; determining the numerical value of one or more process parameters related to the quality index parameter in a detection time period according to a pre-constructed continuous casting quality judgment model and an actual measurement value; and determining the process parameters related to the product quality accident according to the numerical value of each process parameter in the detection time period and the preset value range of the process parameter. Through the scheme, mass data generated in the production process of the continuous casting billet can be utilized to analyze the quality accident reason, so that the product quality is improved.
Description
Technical Field
The invention relates to the technical field of data mining, in particular to an improved dynamic data mining method and device for a continuous casting quality judgment model.
Background
The iron and steel industry plays an important role in economic development as a basic raw material industry of national economy. In the process of producing various steel products in a steel plant, two methods are used for solidifying and forming molten steel: conventional die Casting and Continuous Casting (Continuous Casting for short). Compared with the traditional method, the continuous casting technology has the remarkable advantages of greatly improving the metal yield and the casting blank quality, saving energy and the like.
The continuous casting slab is a product obtained by casting molten steel smelted by a steel smelting furnace through a continuous casting machine, and various quality defects are easy to occur in the continuous casting slab in the actual production process. How to utilize mass data generated in the production process of continuous casting billets and analyze the reasons of quality accidents so as to improve the product quality is one of the problems to be solved urgently.
Disclosure of Invention
The embodiment of the application provides an improved dynamic data mining method facing a continuous casting quality judgment model, which can analyze the reasons of quality accidents by utilizing mass data generated in the production process of continuous casting billets, thereby being beneficial to improving the product quality.
In order to solve the above problem, an embodiment of the present invention provides an improved dynamic data mining method for a continuous casting quality determination model, where the method includes: acquiring the occurrence time of a product quality accident and the type of the product quality accident; determining a quality index parameter associated with the product quality accident according to the type of the product quality accident and preset prior knowledge information; calculating the detection time period of the quality index parameter according to the occurrence time; acquiring an actual measured value of the quality index parameter in the detection time period; determining the numerical value of one or more process parameters related to the quality index parameter in the detection time period according to a pre-constructed continuous casting quality judgment model and the actual measurement value, wherein the continuous casting quality judgment model is used for describing the mathematical relationship between the quality index parameter and the process parameter; and determining the process parameters related to the product quality accident according to the numerical value of each process parameter in the detection time period and the preset value range of the process parameter.
Optionally, the a priori knowledge information includes: before determining the value of one or more process parameters associated with the quality index parameter in the detection time period according to a pre-constructed continuous casting quality judgment model and the actual measurement value, the method further comprises the following steps: calculating a characteristic value of the actual measurement value of the quality index parameter in the detection time period according to the type of the characteristic value; if the characteristic value of the actual measurement value is in the preset value range of the characteristic value, determining that the actual measurement value in the detection time period is qualified data, otherwise, determining that the actual measurement value in the detection time period comprises unqualified data; determining the value of one or more process parameters associated with the quality index parameter in the detection time period according to a pre-constructed continuous casting quality judgment model and the actual measurement value comprises the following steps: and determining the numerical value of one or more process parameters related to the quality index parameter according to the continuous casting quality judgment model and the actual measured value under the condition that the actual measured value in the detection time period is qualified data.
Optionally, determining the values of one or more process parameters associated with the quality index parameter in the detection time period according to a pre-constructed continuous casting quality judgment model and the actual measurement value further includes: under the condition that the actual measurement value in the detection time period comprises unqualified data, dividing the detection time period into a plurality of time intervals; respectively judging whether the actual measurement value of the quality index parameter in each time interval comprises unqualified data; actual measurement values in a time interval including unqualified data are removed to obtain the removed actual measurement values; and determining the numerical value of one or more process parameters related to the quality index parameter in the detection time period according to the removed actual measured value and the continuous casting quality judgment model.
Optionally, the method for constructing the continuous casting quality determination model includes: acquiring a sample data sequence of the quality index parameter, wherein the sample data sequence comprises a plurality of historical measurement values of the quality index parameter; acquiring a sample data sequence of the process parameter, wherein the sample data sequence of the process parameter comprises a plurality of historical measurement values of the process parameter, and each historical measurement value has sampling time point information; performing data normalization on the sample data sequence of the quality index parameter according to the sampling time point information to obtain a normalized data sequence, wherein the normalized data sequence and the sample data sequence of the process parameter are time-aligned; and fitting according to the sample data sequence of the process parameters and the structured data sequence to obtain the continuous casting quality judgment model.
Optionally, the product quality accident is: longitudinal cracks of the casting blank occur, and the quality index parameters are as follows: the casting blank pulling speed, and one or more process parameters related to the quality index parameters comprise: thickness of the shell at the exit of the crystallizer.
Optionally, the method further includes: acquiring an actual measurement value of the quality index parameter in a first time period; inputting the actual measurement value of the quality index parameter in the first time period into a quality prediction model obtained by pre-training so as to obtain a predicted quality result output by the quality prediction model, wherein the predicted quality result is used for indicating whether a product quality accident related to the quality index parameter occurs in the first time period; the quality prediction model is obtained by training a preset model in advance by adopting training data, wherein the training data comprises: and the sample measurement value sequence of the quality index parameter and the actual quality result corresponding to the sample measurement value sequence.
Optionally, the preset model is: a bidirectional long and short term memory network.
The embodiment of the invention also provides an improved dynamic data mining device facing the continuous casting quality judgment model, which comprises: the first acquisition module is used for acquiring the occurrence time of a product quality accident and the type of the product quality accident; the quality index parameter determining module is used for determining a quality index parameter related to the product quality accident according to the type of the product quality accident and preset prior knowledge information; the detection time period determination module is used for calculating the detection time period of the quality index parameter according to the occurrence time; the second acquisition module is used for acquiring the actual measurement value of the quality index parameter in the detection time period; a third obtaining module, configured to determine, according to a pre-constructed continuous casting quality determination model and the actual measurement value, a numerical value of one or more process parameters associated with the quality index parameter in the detection time period, where the continuous casting quality determination model is used to describe a mathematical relationship between the quality index parameter and the process parameter; and the quality tracing module is used for determining the process parameters related to the product quality accident according to the numerical value of each process parameter in the detection time period and the preset value range of the process parameter.
Embodiments of the present invention also provide a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, causes the above-described improved continuous casting quality determination model-oriented dynamic data mining method to be performed.
The embodiment of the present invention further provides a terminal, which includes a memory and a processor, where the memory stores a computer program capable of running on the processor, and the processor executes the steps of the above improved continuous casting quality determination model-oriented dynamic data mining method when running the computer program.
Compared with the prior art, the technical scheme of the embodiment of the application has the following beneficial effects:
in the scheme of the embodiment of the invention, the occurrence time and the type of the product quality accident are obtained, on one hand, the quality index parameter associated with the product quality accident is determined according to the type of the product quality accident and the preset prior knowledge information, and on the other hand, the detection time period of the quality index parameter is calculated according to the occurrence time of the product quality accident. Further, the actual measurement value of the quality index parameter in the detection time period is obtained, and the numerical value of the process parameter in the detection time period is obtained by reversely solving according to the pre-constructed continuous casting quality judgment model and the actual measurement value of the quality index parameter, so that whether the process parameter is related to the product quality accident or not can be determined according to the numerical value of the process parameter, and the process parameter which may cause the product quality accident is determined. By adopting the scheme, the possible reasons of the product quality problems can be efficiently and accurately determined based on the actual measured values of the quality index parameters in the production process.
Further, in the solution of the embodiment of the present invention, the prior knowledge information includes: and before determining the values of one or more process parameters associated with the quality index parameter in the detection time period, judging whether the actual measurement value in the detection time period comprises unqualified data, and rejecting the unqualified data. By adopting the scheme, the qualified data is verified based on the prior knowledge information, so that the possible parameter reasons causing the product quality accidents can be analyzed in the state of qualified historical data, and the accuracy of the result is favorably ensured.
Further, in the scheme of the embodiment of the invention, the sample data sequence of the quality index parameters and the sample data sequence of the process parameters are firstly obtained, then the data normalization is carried out on the sample data of the quality index parameters to obtain a normalized data sequence, and finally the fitting is carried out according to the normalized data sequence and the sample data sequence of the process parameters to obtain the continuous casting quality judgment model. Because the time point information is adopted according to each historical measured value in the sample data sequence of the process parameters, the structured data sequence and the sample data sequence of the process parameters are time-aligned, and the time-aligned data sequence is time-aligned with the process parameters by adding a time-sequence factor into the measured values of the discrete quality index parameters, so that the correlated dynamic data mining can be performed on the data in the time space, and the accuracy of a continuous casting quality judgment model is improved.
Drawings
FIG. 1 is a schematic flow chart of an improved continuous casting quality decision model-oriented dynamic data mining method according to an embodiment of the present invention;
fig. 2 is a schematic flowchart of a method for constructing a continuous casting quality determination model according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of an improved dynamic data mining device for a continuous casting quality determination model according to an embodiment of the present invention.
Detailed Description
As described in the background art, how to analyze the cause of quality accidents by using mass data generated in the continuous casting billet production process, so as to improve the product quality is one of the problems to be solved urgently.
The product quality accident is usually caused by Process Parameter (Process Parameter), and the reasons for the poor product quality may include: the design values of the process parameters are not reasonable and the actual values of the process parameters during production are not controlled within reasonable values. The process parameters refer to a series of basic data or indexes of a process for completing a certain work, and more specifically, the process parameters refer to index parameters of machinery and/or equipment related to the process, and can be used for embodying some information such as performance of the process. In the solution of an embodiment of the invention, the process parameters may refer to parameters used or set in a continuous casting process. Therefore, the establishment of the incidence relation between the product quality accident and the process parameters is of great importance for analyzing the reason of the continuous casting billet quality accident.
The inventor of the present invention has found that the cause of the poor product quality can be reflected in the time series data of the process parameters in the production process. That is, changes in certain characteristics of the time series data of the process parameters can cause variations in product quality. However, in the actual production process, the actual values of some process parameters are difficult to measure, and in this case, because the actual measured values of the process parameters cannot be obtained, the characteristics of the process parameter time series data cannot be determined, and there is a technical obstacle to analyzing the product quality problem according to the characteristics of the process parameter time series data.
In order to solve the above problems, embodiments of the present invention provide an improved dynamic data mining method for a continuous casting quality determination model, in a scheme of an embodiment of the present invention, occurrence time and type of a product quality accident are obtained, on one hand, a quality index parameter associated with the product quality accident is determined according to the type of the product quality accident and preset prior knowledge information, and on the other hand, a detection time period of the quality index parameter is calculated according to the occurrence time of the product quality accident. Further, the actual measurement value of the quality index parameter in the detection time period is obtained, and the numerical value of the process parameter in the detection time period is obtained by reversely solving according to the pre-constructed continuous casting quality judgment model and the actual measurement value of the quality index parameter, so that whether the process parameter is related to the product quality accident or not can be determined according to the numerical value of the process parameter, and the process parameter which may cause the product quality accident is determined. By adopting the scheme, the possible reasons of the product quality problems can be efficiently and accurately determined based on the actual measured values of the quality index parameters in the production process.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in detail below.
Referring to fig. 1, fig. 1 is a schematic flow chart of an improved continuous casting quality determination model-oriented dynamic data mining method according to an embodiment of the present invention. The method may be executed by a terminal, and the terminal may be various existing devices with data receiving and data processing capabilities, for example, a mobile phone, a computer, a tablet computer, an internet of things device, a server, and the like, which is not limited in this embodiment of the present invention. The method shown in fig. 1 may include steps S101 to S106:
step S101: acquiring the occurrence time of a product quality accident and the type of the product quality accident;
step S102: determining a quality index parameter associated with the product quality accident according to the type of the product quality accident and preset prior knowledge information;
step S103: calculating the detection time period of the quality index parameter according to the occurrence time;
step S104: acquiring an actual measurement value of the quality index parameter in the detection time period;
step S105: determining the numerical value of one or more process parameters related to the quality index parameter in the detection time period according to a pre-constructed continuous casting quality judgment model and the actual measurement value, wherein the continuous casting quality judgment model is used for describing the mathematical relationship between the quality index parameter and the process parameter;
step S106: and determining the process parameters related to the product quality accident according to the numerical value of each process parameter in the detection time period and the preset value range of the process parameter.
It is understood that in a specific implementation, the method may be implemented by a software program running in a processor integrated within a chip or a chip module; alternatively, the method can be implemented in hardware or a combination of hardware and software.
In the specific implementation of step S101, the occurrence time of the product quality accident and the type of the product quality accident may be acquired. In the embodiment of the present invention, the product refers to a continuous casting slab (also referred to as a "casting slab" for short), the product quality accident refers to a condition that a quality defect occurs in the continuous casting slab, and the type of the product quality accident may refer to a type of the quality defect.
In a specific implementation, the detection may be performed on a continuous casting slab to obtain a detection result, and in the case that the continuous casting slab has a quality defect, the detection result may include: the time of occurrence of the product quality incident and the type of the product quality incident.
In a specific example, the continuous casting slab may be artificially detected to obtain the detection result, and the detection result is input to the terminal.
In another specific example, the terminal may automatically perform detection on the continuous casting slab to obtain the detection result. For example, the continuous casting billet can be shot to obtain a continuous casting billet image, and the continuous casting billet image is subjected to image processing to obtain a detection result.
In an actual production scenario, the type of the product quality accident may be any one of the following: the method is not limited to the method, and the method can also be used for solving the problems that the casting blank has longitudinal cracks, the casting blank has transverse cracks, the casting blank has bubbles, the surface of the casting blank is subjected to slag inclusion and the like, and the quality defects can also be generated in the actual production process. Hereinafter, a non-limiting description will be given by taking the type of "longitudinal crack of a cast slab" as an example.
Further, the occurrence time of the product quality accident may refer to a time when a quality defect of the continuous casting slab is detected.
In other embodiments, the time of occurrence of the quality defect may be estimated according to the detection time of the occurrence of the quality defect of the continuous casting slab, and the estimated time may be used as the occurrence time. For example, the condition of the quality defect of the continuous casting billet can be analyzed to calculate the time when the quality defect occurs. In a specific implementation, the whole continuous casting billet can be detected to obtain a detection result of the continuous casting billet, and the detection result can comprise a plurality of quality defects on the continuous casting billet. For each quality defect, the position of the quality defect on the continuous casting billet can be determined according to the detection time and the quality defect distribution of the continuous casting billet, and then the time for the quality defect to occur is determined according to the position of each quality defect on the continuous casting billet. The quality defect distribution of the continuous casting billet refers to the distribution of the quality defects on the whole continuous casting billet. By adopting the scheme, the occurrence time of the quality problem can be more accurately determined under the condition that the detection of the quality defect has delay, so that the detection time period of the subsequent quality index parameter can be more accurately determined.
In a specific implementation of step S102, preset a-priori knowledge information may be read, where the a-priori knowledge information may be stored in a memory of the terminal in advance, or may also be stored in a database coupled to the terminal, which is not limited in this embodiment.
In particular, the a priori knowledge information may include a quality indicator parameter associated with each type of product quality incident and mode information for the quality indicator parameter. Wherein the mode information may include: the characteristic value type of the quality index parameter and the preset value range of the characteristic value.
In the scheme of the embodiment of the invention, the quality index parameter refers to a parameter for characterizing the quality of the product. More specifically, different quality index parameters may be used to characterize the product quality of different production nodes in the production process.
Further, the type of the feature value refers to a type of a feature value of the time-series data of the quality index parameter, and the type of the feature value may be any one of the following items: variance, mean, covariance, etc., but are not limited thereto.
For example, the quality index parameters related to the occurrence of longitudinal cracks in the casting blank may be: and (5) casting blank drawing speed. In specific implementation, the a priori knowledge information of "longitudinal crack of casting slab" may include: in the continuous casting production process, longitudinal cracks are easy to appear when the casting blank pulling speed is unstable, and the variance of the time sequence of the measured values of the casting blank pulling speed is used as a mode for influencing the quality accident of the longitudinal cracks of the casting blank.
It should be noted that, multiple quality index parameters may be associated with each product quality accident type, and steps S103 to S106 may be continuously performed for each quality index parameter.
Further, after the priori knowledge information is read, the quality index parameter may be determined according to the type of the product quality accident and the priori knowledge information acquired in step S101. Taking the product quality accident as "longitudinal crack of casting blank" as an example, according to the priori knowledge information, the associated quality index parameters may be determined to include: and (3) determining the casting blank pulling rate, wherein the type of the characteristic value of the casting blank pulling rate can be a variance.
In a specific implementation of step S103, the detection time period of the quality index parameter may be estimated according to the occurrence time of the product quality accident. It is understood that the detection period is a period before the occurrence time.
In a specific implementation, the detection time period of the quality index parameter can be calculated according to a preset logic relationship and the occurrence time of the product quality accident. More specifically, the a priori knowledge information may include the preset logical relationship.
In one specific example, the occurrence time of the product quality accident is T 1 At the moment, thenDetermining the detection time period as (T) by calculation 1 Δ T) to T 1 A time period between instants, wherein Δ T is used to indicate a duration of the detection time period, which may be predetermined.
In a specific implementation of step S104, the actual measured value of the quality indicator parameter within the detection period may be obtained.
Further, according to the above-mentioned type of the characteristic value, the characteristic value of the actual measurement value of the quality index parameter in the detection time period may be calculated, and the calculated characteristic value may be compared with a preset value range of the characteristic value, so as to determine whether the actual measurement value in the detection time period includes unqualified data.
It should be noted that the preset value range of the characteristic value refers to an allowable value range. Specifically, when the characteristic value of the actual measurement value of the quality index parameter exceeds the preset value range of the characteristic value in the production process, it is indicated that the actual measurement value itself has an error, that is, the data itself has an error. For example, but not limited to, improper measurement may result in data errors that cannot be used for subsequent data mining and analysis. In other words, when the characteristic value of the actual measurement value of the quality index parameter exceeds the preset value range of the characteristic value in the production process, the actual measurement value does not indicate that a product quality accident occurs. For convenience of description, the preset value range of the characteristic value is referred to as a first value range, and the first value range and the type of the characteristic value are in one-to-one correspondence.
Further, if it is determined that the characteristic value of the actual measurement value in the detection time period is in the first value range, the actual measurement value in the detection time period may be determined to be qualified data, and thus, steps S105 and S106 may be continuously performed according to the actual measurement value in the detection time period.
If it is determined that the characteristic value of the actual measurement value in the detection time period is not within the first value range, it may be determined that the actual measurement value in the detection time period includes the non-qualified data.
Further, if it is determined that the actual measurement value within the detection period includes the fail data, the detection period may be divided into a plurality of time intervals. Further, it may be separately determined whether the actual measured value of the quality indicator parameter in each time interval includes non-conforming data.
Specifically, the characteristic value of the actual measurement value of the quality index parameter in each time interval may be calculated, and whether the characteristic value is within the corresponding first value range is determined. For any time interval, if the characteristic value of the actual measurement value is not within the first value range, it may be determined that the time interval includes the unqualified data.
Further, the actual measurement values in the time interval including the unqualified data may be rejected to obtain the rejected actual measurement values. Further, steps S105 and S106 may be performed on the actual measurement values after the culling.
In one non-limiting example, if each time interval includes non-conforming data, each time interval can be further divided to obtain a plurality of time sub-intervals.
Further, for each time sub-interval, it may be determined whether the actual measured value of the quality indicator parameter within the time sub-interval comprises non-conforming data. Specifically, the characteristic value of the actual measurement value of the quality index parameter in each time subinterval may be calculated, and it may be determined whether the actual measurement value is within the corresponding first value range. For any time subinterval, if the characteristic value of the actual measurement value is not in the first value range, it may be determined that the time subinterval includes the unqualified data.
Further, the actual measurement values in the time subinterval including the unqualified data may be removed from the actual measurement values of the entire detection time period to obtain the removed actual measurement values. If each time subinterval includes unqualified data, it may be determined that the unqualified data exists within the detection time period. In this case, an abnormal reminding message may be sent to the user, where the abnormal reminding message is used to indicate that all the detection time periods are unqualified data.
It should be noted that, in the solution of the embodiment of the present invention, the time lengths of the multiple time intervals may be the same, and the time lengths of the multiple time subintervals may also be the same.
Therefore, in the scheme of the embodiment of the invention, possible parameter reasons causing product quality accidents can be analyzed in the state of qualified historical data.
In the specific implementation of step S105, for each quality index parameter determined in step S102, a continuous casting quality determination model corresponding to the quality index parameter may be obtained. The quality index parameters and the continuous casting quality judgment model can be in one-to-one correspondence.
Wherein the continuous casting quality decision model is used for describing a mathematical relationship (or also called a functional relationship) between the quality index parameter and one or more process parameters. More intuitively, the continuous casting quality determination model may be expressed as y = f (x) 1 ,x 2 ,…,x n ) Wherein y is the quality index parameter, x i I is more than or equal to 1 and less than or equal to n, and both i and n are positive integers.
In a specific implementation, the continuous casting quality index model may be pre-constructed and stored in a memory of the terminal or a database coupled to the terminal, which is not limited to this embodiment. Regarding the process of constructing the continuous casting quality determination model, reference may be made to the following description of fig. 2, which is not repeated herein.
Further, the value of each process parameter associated with the quality index parameter in the detection time period may be obtained by solving in reverse according to the actual measurement value of the quality index parameter in the detection time period obtained in step S104 and the continuous casting quality determination model. It should be noted that, if the operation of eliminating the unqualified data is performed in step S104, the numerical value of each process parameter associated with the quality index parameter may be reversely solved according to the actual measured value and the continuous casting quality determination model after elimination
In the case where the product quality accident is "longitudinal crack of a casting slab", and the quality index parameter is casting slab casting speed, the continuous casting quality determination model may be expressed as:
wherein V is casting blank drawing speed, delta is blank shell thickness at the outlet of the crystallizer, K is solidification coefficient of the crystallizer, and L is m Is the effective strength length of the crystallizer. The thickness delta of the shell at the outlet of the crystallizer is a process parameter related to the casting blank pulling speed V, the solidification coefficient K of the crystallizer and the effective strength length L of the crystallizer m Are all predetermined constant values. In step S105, a value of the shell thickness δ at the exit of the mold during the detection period may be calculated according to the actual measurement value of the casting blank pulling rate V during the detection period.
In the specific implementation of step S106, for each process parameter, it may be determined whether the process parameter is related to the product quality accident according to the value of the process parameter and the preset value range of the process parameter in the detection time period. For convenience of description, the preset value range of the process parameter is hereinafter referred to as a second value range, and the second value range and the type of the process parameter may be in one-to-one correspondence.
In specific implementation, for each process parameter associated with the quality index parameter, if the value of the process parameter exceeds the second value range of the process parameter within the detection time period, it may be determined that the process parameter is associated with the product quality accident. The association of the process parameter with the product quality incident may refer to the process parameter being one of the causes of the product quality incident.
For example, the value of the blank shell thickness δ at the mold outlet obtained in step S105 is compared with the second value range corresponding to the blank shell thickness δ to determine whether the blank shell thickness δ at the mold outlet is likely to be one of the causes of longitudinal cracks in the cast slab.
Therefore, in the scheme of the embodiment of the invention, the dynamic data mining method is adopted to mine the information aiming at the dynamic data and the real-time data, so that the mass data generated in the production process of the continuous casting billet can be utilized to analyze the quality accident reason, the condition that the process parameters are not controlled within the range of reasonable values is found, and the subsequent improvement of the product quality is facilitated.
Referring to fig. 2, fig. 2 is a schematic flowchart of a method for constructing a continuous casting quality determination model according to an embodiment of the present invention. As shown in fig. 2, the construction method may include steps S201 to S204:
step S201: acquiring a sample data sequence of the quality index parameter, wherein the sample data sequence comprises a plurality of historical measurement values of the quality index parameter;
step S202: acquiring a sample data sequence of the process parameter, wherein the sample data sequence of the process parameter comprises a plurality of historical measurement values of the process parameter, and each historical measurement value has sampling time point information;
step S203: performing data normalization on the sample data sequence of the quality index parameter according to the sampling time point information to obtain a normalized data sequence, wherein the normalized data sequence and the sample data sequence of the process parameter are time-aligned;
and S204, fitting according to the sample data sequence of the process parameters and the structured data sequence to obtain the continuous casting quality judgment model.
In the specific implementation of step S201 and step S202, a sample data sequence of the quality index parameter and a sample data sequence of the process parameter are respectively obtained, and for convenience of description, the sample data sequence of the quality index parameter may be denoted as a first sample sequence H1, and the sample data sequence of the process parameter may be denoted as a second sample sequence H2. It will be appreciated that the first sequence of samples comprises a plurality of historical measurements of the quality indicator parameter and the second sequence of samples comprises a plurality of historical measurements of the process parameter.
For the steel production process, the internal production process can be divided into a continuous process and a batch process, the data involved in the continuous process is usually time series data, and the data involved in the batch process is usually batch data. The continuous production process is characterized in that raw materials are continuously processed by each process device of the production line to form products, and the design values of the operating parameters of each process device are constant values. Batch type production processes are characterized by a single batch of product in the same process equipment, which typically experiences multiple processing periods, and therefore, typically, the design values for the various operating parameters will take on different values at different times during the production process. The continuous casting process mainly takes a continuous production process as a main process, but has the characteristic of batch type for each batch of ladle data. Therefore, different preprocessing normalization modes are adopted for time series data and batch data.
In the embodiment of the present invention, the process parameter is continuous process data (also referred to as "time series data"). Specifically, the second sample sequence H1 may be obtained by sampling the measured values of the process parameter according to a preset sampling period, for this purpose, the second sample sequence H1 includes a plurality of historical measured values, and each historical measured value has sampling time point information.
The quality indicator parameter is batch-type process data (also referred to as "batch data"), whereby the above-mentioned first sequence of samples H1 comprises a plurality of historical measurement values of the quality indicator parameter, wherein the time intervals between the plurality of historical measurement values in the first sequence of samples H1 are irregular, for which purpose the first sequence of samples H1 may be data-normalized by performing step S203.
In a specific implementation, the step S201 may be performed first, that is, the first sample sequence H1 is obtained first, and since the historical measurement values in the first sample sequence H1 are obtained by performing quality detection on the product, the historical measurement values of the process parameters in the time period that have an influence on the historical measurement values in the first sample sequence H1 may be extracted. Wherein the duration of the influential time period may be preset. In other words, the continuous process data is structured in such a way that the measured values of different process parameters at different moments on part of products participating in the quality sampling inspection and the quality inspection result form a sample together, that is, a time sequence of the measured values of the process parameters influencing a certain quality inspection result is extracted and forms the sample together with the historical measured values of the quality index parameters.
More specifically, the product quality inspection is completed at time t, i.e., the measured value of the quality index parameter at time t can be denoted as yt, and the measurement point at which each process generates yt may be installed in the same or different area of the production line, in other words, the various process parameters associated with yt may be located in the same or different area of the production line.
In a specific implementation of step S203, data warping is performed on the first sample sequence H1 according to the sampling time point information of each historical measurement value in the second sample sequence H2 to obtain a warped data sequence (i.e., warped first sample sequence H1), where the warped data sequence and the sample data sequence of the process parameter are time-aligned. The time alignment may refer to that the historical measured values in the normalized data sequence and the historical measured values in the sample data sequence of the process parameter are in one-to-one correspondence, and the two historical measured values in the one-to-one correspondence have the same time point information.
In other words, the discrete detection values of the measured values of the quality index parameters are aligned by executing step S203, that is, the data alignment is performed by adding a time-series factor.
In a specific implementation of step S204, fitting may be performed according to the second sample sequence and the normalized data sequence to obtain the continuous casting quality determination model. In specific implementations, the fitting method may be any of various fitting methods, for example: polynomial fitting, least squares fitting, etc., and this embodiment does not limit this.
In the scheme of the embodiment of the invention, the preprocessing data of the continuous data (namely, the historical measured values of the process parameters) and the batch process data (namely, the historical measured values of the quality index parameters) in the continuous casting billet production process are normalized, and then the model is fitted according to the normalized data, so that the continuous casting quality judgment model is obtained.
In other embodiments of the present invention, a quality prediction model may also be constructed, which may be used to characterize the relationship between the quality index parameter and the product quality incident.
Specifically, the preset model may be trained using training data, which may include: and the sample measurement value sequence of the quality index parameter and the actual quality result corresponding to the sample measurement value sequence. In other words, the series of sample measurements is the input data and the actual quality result is the tag data. In one specific example, the pre-set model may be a Bidirectional Long Short-Term Memory network (Bi-LSTM). It should be noted that, the training method in the embodiment of the present invention is not limited, and may be various existing training methods for training a neural network model. And when the preset training stopping condition is met, obtaining the quality prediction model. The quality prediction model can learn the rule or relationship between the product quality accident and the quality index parameter through training. For example, it is possible to learn a rule about the influence of the longitudinal crack of the cast slab by factors such as the actual pulling rate.
Further, in a specific implementation, the actual measurement value of the quality index parameter in the first time period may be input to a quality prediction model obtained through pre-training, so as to obtain a predicted quality result output by the quality prediction model, where the predicted quality result is used to indicate whether a product quality accident associated with the quality index parameter occurs in the first time period. Wherein the first time period may be any time period in the production process. For example, it is possible to predict whether or not the influence of the longitudinal crack quality accident will be caused in the state of the conventional pulling rate, and to achieve the target result finally determined in advance.
In the scheme of the embodiment of the invention, the continuous casting quality judgment model and the quality prediction model can be used on line or off line.
When the method is used on line, the numerical value of the quality index parameter in the production process can be predicted in real time according to the measured value of the process parameter and the continuous casting quality judgment model, the quality index parameter does not need to be measured at the moment, and furthermore, the product quality accident can be predicted in real time according to the quality index parameter and the continuous casting quality judgment model.
When the method is used off line, the process design verification can be carried out according to the design value of the process parameter. Specifically, the design value of the process parameter may be input into the continuous casting quality determination model to obtain the quality index parameter corresponding to the design value of the process parameter, and then the quality index parameter output by the continuous casting quality determination model may be input into the quality prediction model, and obtaining a predicted quality result output by the quality prediction model, so that whether the design value of the process parameter is reasonable or not can be verified, and if the design value is unreasonable, adjustment can be performed, so that quality control is performed.
Referring to fig. 3, fig. 3 is a schematic structural diagram of an improved dynamic data mining apparatus for a continuous casting quality determination model according to an embodiment of the present invention, where the dynamic data mining apparatus shown in fig. 3 may include:
a first obtaining module 31, configured to obtain occurrence time of a product quality accident and a type of the product quality accident;
a quality index parameter determination module 32, configured to determine a quality index parameter associated with the product quality accident according to the type of the product quality accident and preset prior knowledge information;
a detection time period determination module 33, configured to calculate a detection time period of the quality indicator parameter according to the occurrence time;
a second obtaining module 34, configured to obtain an actual measured value of the quality indicator parameter in the detection time period;
a third obtaining module 35, configured to determine, according to a pre-constructed continuous casting quality determination model and the actual measurement value, values of one or more process parameters associated with the quality index parameter in the detection time period, where the continuous casting quality determination model is used to describe a mathematical relationship between the quality index parameter and the process parameters;
and the quality tracing module 36 is configured to determine the process parameter associated with the product quality accident according to the value of each process parameter in the detection time period and the preset value range of the process parameter.
In a specific implementation, the dynamic data mining device for the improved continuous casting quality judgment-oriented model may correspond to a chip having a data processing function in a terminal; or to a chip module having a data processing function in the terminal, or to the terminal.
For more details of the working principle, the working mode, the beneficial effects, and the like of the improved continuous casting quality determination model-oriented dynamic data mining device shown in fig. 3, reference may be made to the description above regarding fig. 1 to 3, and details are not repeated here.
Embodiments of the present invention further provide a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor, and the computer program performs the steps of the above improved continuous casting quality determination model-oriented dynamic data mining method. The storage medium may include ROM, RAM, magnetic or optical disks, etc. The storage medium may further include a non-volatile memory (non-volatile) or a non-transitory memory (non-transient), and the like.
The embodiment of the invention also provides a terminal, which comprises a memory and a processor, wherein the memory is stored with a computer program capable of running on the processor, and the processor executes the steps of the method when running the computer program. The terminal can be a mobile phone, a computer, an internet of things device and the like.
It should be understood that, in the embodiment of the present application, the processor may be a Central Processing Unit (CPU), and the processor may also be other general-purpose processors, digital Signal Processors (DSPs), application Specific Integrated Circuits (ASICs), field Programmable Gate Arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, and the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
It will also be appreciated that the memory in the embodiments of the subject application can be either volatile memory or nonvolatile memory, or can include both volatile and nonvolatile memory. The nonvolatile memory may be a read-only memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an Electrically Erasable PROM (EEPROM), or a flash memory. Volatile memory can be Random Access Memory (RAM), which acts as external cache memory. By way of example and not limitation, many forms of Random Access Memory (RAM) are available, such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (enhanced SDRAM), synchronous DRAM (SLDRAM), synchronous Link DRAM (SLDRAM), and direct bus RAM (DR RAM).
The above embodiments may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, the above-described embodiments may be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer instructions or computer programs. The procedures or functions according to the embodiments of the present application are wholly or partially generated when the computer instructions or the computer program are loaded or executed on a computer. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer program may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another computer readable storage medium, for example, the computer program may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire or wirelessly.
In the several embodiments provided in the present application, it should be understood that the disclosed method, apparatus and system may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative; for example, the division of the unit is only a logic function division, and there may be another division manner in actual implementation; for example, various elements or components may be combined or may be integrated into another system, or some features may be omitted, or not implemented. The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be physically included alone, or two or more units may be integrated into one unit. The integrated unit may be implemented in the form of hardware, or in the form of hardware plus a software functional unit. For example, for each device or product applied to or integrated into a chip, each module/unit included in the device or product may be implemented by hardware such as a circuit, or at least a part of the module/unit may be implemented by a software program running on a processor integrated within the chip, and the rest (if any) part of the module/unit may be implemented by hardware such as a circuit; for each device or product applied to or integrated with the chip module, each module/unit included in the device or product may be implemented by using hardware such as a circuit, and different modules/units may be located in the same component (e.g., a chip, a circuit module, etc.) or different components of the chip module, or at least some of the modules/units may be implemented by using a software program running on a processor integrated within the chip module, and the rest (if any) of the modules/units may be implemented by using hardware such as a circuit; for each device and product applied to or integrated in the terminal, each module/unit included in the device and product may be implemented by using hardware such as a circuit, and different modules/units may be located in the same component (e.g., a chip, a circuit module, etc.) or different components in the terminal, or at least part of the modules/units may be implemented by using a software program running on a processor integrated in the terminal, and the rest (if any) part of the modules/units may be implemented by using hardware such as a circuit.
It should be understood that the term "and/or" herein is merely one type of association relationship that describes an associated object, meaning that three relationships may exist, e.g., a and/or B may mean: a exists alone, A and B exist simultaneously, and B exists alone. In addition, the character "/" in this document indicates that the former and latter related objects are in an "or" relationship.
The "plurality" appearing in the embodiments of the present application means two or more.
The descriptions of the first, second, etc. appearing in the embodiments of the present application are only for illustrating and differentiating the objects, and do not represent the order or the particular limitation of the number of the devices in the embodiments of the present application, and do not constitute any limitation to the embodiments of the present application.
Although the present invention is disclosed above, the present invention is not limited thereto. Various changes and modifications may be effected therein by one skilled in the art without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (10)
1. An improved dynamic data mining method for a continuous casting quality judgment model, which is characterized by comprising the following steps:
acquiring the occurrence time of a product quality accident and the type of the product quality accident;
determining a quality index parameter associated with the product quality accident according to the type of the product quality accident and preset prior knowledge information;
calculating the detection time period of the quality index parameter according to the occurrence time;
acquiring an actual measurement value of the quality index parameter in the detection time period;
determining the numerical values of one or more process parameters related to the quality index parameter in the detection time period according to a pre-constructed continuous casting quality judgment model and the actual measurement value, wherein the continuous casting quality judgment model is used for describing the mathematical relationship between the quality index parameter and the process parameters;
and determining the process parameters related to the product quality accident according to the numerical value of each process parameter in the detection time period and the preset value range of the process parameter.
2. The dynamic data mining method of claim 1, wherein the a priori knowledge information comprises: before determining the value of one or more process parameters associated with the quality index parameter in the detection time period according to a pre-constructed continuous casting quality judgment model and the actual measurement value, the method further comprises the following steps:
calculating the characteristic value of the actual measurement value of the quality index parameter in the detection time period according to the type of the characteristic value;
if the characteristic value of the actual measurement value is in the preset value range of the characteristic value, determining that the actual measurement value in the detection time period is qualified data, otherwise, determining that the actual measurement value in the detection time period comprises unqualified data;
determining the value of one or more process parameters associated with the quality index parameter in the detection time period according to a pre-constructed continuous casting quality judgment model and the actual measurement value comprises the following steps:
and determining the numerical value of one or more process parameters related to the quality index parameter according to the continuous casting quality judgment model and the actual measured value under the condition that the actual measured value in the detection time period is qualified data.
3. The dynamic data mining method of claim 2, wherein determining the values of one or more process parameters associated with the quality index parameter during the detection period based on a pre-constructed continuous casting quality decision model and the actual measurement values further comprises:
under the condition that the actual measurement value in the detection time period comprises unqualified data, dividing the detection time period into a plurality of time intervals;
respectively judging whether the actual measurement value of the quality index parameter in each time interval comprises unqualified data;
actual measurement values in a time interval including unqualified data are removed to obtain the removed actual measurement values;
and determining the numerical value of one or more process parameters related to the quality index parameter in the detection time period according to the removed actual measured value and the continuous casting quality judgment model.
4. The dynamic data mining method according to claim 1, wherein the method of constructing the continuous casting quality determination model includes:
acquiring a sample data sequence of the quality index parameter, wherein the sample data sequence comprises a plurality of historical measurement values of the quality index parameter;
acquiring a sample data sequence of the process parameter, wherein the sample data sequence of the process parameter comprises a plurality of historical measurement values of the process parameter, and each historical measurement value has sampling time point information;
performing data normalization on the sample data sequence of the quality index parameter according to the sampling time point information to obtain a normalized data sequence, wherein the normalized data sequence and the sample data sequence of the process parameter are time-aligned;
and fitting according to the sample data sequence of the process parameters and the normalized data sequence to obtain the continuous casting quality judgment model.
5. The dynamic data mining method of claim 1, wherein the product quality incident is: longitudinal cracks of the casting blank occur, and the quality index parameters are as follows: and the casting blank drawing speed is related to the quality index parameter, and one or more process parameters comprise: thickness of the shell at the exit of the crystallizer.
6. The dynamic data mining method of claim 1, the method further comprising: acquiring an actual measurement value of the quality index parameter in a first time period;
inputting the actual measurement value of the quality index parameter in the first time period into a quality prediction model obtained by pre-training so as to obtain a predicted quality result output by the quality prediction model, wherein the predicted quality result is used for indicating whether a product quality accident related to the quality index parameter occurs in the first time period;
the quality prediction model is obtained by training a preset model in advance by adopting training data, wherein the training data comprises: and the sample measurement value sequence of the quality index parameter and the actual quality result corresponding to the sample measurement value sequence.
7. The dynamic data mining method of claim 6, wherein the predetermined model is: a bidirectional long and short term memory network.
8. An improved dynamic data mining device for a continuous casting quality judgment model, which is characterized by comprising:
the first acquisition module is used for acquiring the occurrence time of a product quality accident and the type of the product quality accident;
the quality index parameter determining module is used for determining a quality index parameter related to the product quality accident according to the type of the product quality accident and preset prior knowledge information;
the detection time period determination module is used for calculating the detection time period of the quality index parameter according to the occurrence time;
the second acquisition module is used for acquiring the actual measurement value of the quality index parameter in the detection time period;
a third obtaining module, configured to determine, according to a pre-constructed continuous casting quality determination model and the actual measurement value, a numerical value of one or more process parameters associated with the quality index parameter in the detection time period, where the continuous casting quality determination model is used to describe a mathematical relationship between the quality index parameter and the process parameter;
and the quality tracing module is used for determining the process parameters related to the product quality accident according to the numerical value of each process parameter in the detection time period and the preset value range of the process parameter.
9. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, causes the improved continuous casting quality decision model-oriented dynamic data mining method of any one of claims 1 to 7 to be performed.
10. A terminal comprising a memory and a processor, the memory having stored thereon a computer program operable on the processor, wherein the processor when executing the computer program performs the steps of the improved continuous casting quality decision model oriented dynamic data mining method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210833764.3A CN115345434A (en) | 2022-07-15 | 2022-07-15 | Improved dynamic data mining method and device for continuous casting quality judgment model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210833764.3A CN115345434A (en) | 2022-07-15 | 2022-07-15 | Improved dynamic data mining method and device for continuous casting quality judgment model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115345434A true CN115345434A (en) | 2022-11-15 |
Family
ID=83948312
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210833764.3A Pending CN115345434A (en) | 2022-07-15 | 2022-07-15 | Improved dynamic data mining method and device for continuous casting quality judgment model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115345434A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116541436A (en) * | 2023-04-13 | 2023-08-04 | 华院计算技术(上海)股份有限公司 | Correlation query method and device for continuous casting production line, medium and terminal |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113420157A (en) * | 2021-05-27 | 2021-09-21 | 冶金自动化研究设计院 | Steel product surface longitudinal crack defect traceability analysis method based on knowledge graph |
| CN113486457A (en) * | 2021-06-04 | 2021-10-08 | 宁波海天金属成型设备有限公司 | Die casting defect prediction and diagnosis system |
| CN114444986A (en) * | 2022-04-11 | 2022-05-06 | 成都数之联科技股份有限公司 | Product analysis method, system, device and medium |
| CN114626673A (en) * | 2022-02-08 | 2022-06-14 | 南京钢铁股份有限公司 | Continuous casting billet quality forecasting method based on Bayesian network |
-
2022
- 2022-07-15 CN CN202210833764.3A patent/CN115345434A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113420157A (en) * | 2021-05-27 | 2021-09-21 | 冶金自动化研究设计院 | Steel product surface longitudinal crack defect traceability analysis method based on knowledge graph |
| CN113486457A (en) * | 2021-06-04 | 2021-10-08 | 宁波海天金属成型设备有限公司 | Die casting defect prediction and diagnosis system |
| CN114626673A (en) * | 2022-02-08 | 2022-06-14 | 南京钢铁股份有限公司 | Continuous casting billet quality forecasting method based on Bayesian network |
| CN114444986A (en) * | 2022-04-11 | 2022-05-06 | 成都数之联科技股份有限公司 | Product analysis method, system, device and medium |
Non-Patent Citations (3)
| Title |
|---|
| 吴以凡等: "面向生产过程质量控制的动态数据挖掘方法", pages 195 - 201 * |
| 吴润等: "钢铁质量及其构件失效分析", vol. 1, 北京航空航天大学出版社, pages: 83 - 86 * |
| 张凯;: "基于多元线性回归的注塑机液压系统工具变量与参数检验", no. 03, pages 59 - 63 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116541436A (en) * | 2023-04-13 | 2023-08-04 | 华院计算技术(上海)股份有限公司 | Correlation query method and device for continuous casting production line, medium and terminal |
| CN116541436B (en) * | 2023-04-13 | 2024-02-27 | 华院计算技术(上海)股份有限公司 | Correlation query method and device for continuous casting production line, medium and terminal |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11105758B2 (en) | Prediction method for mold breakout based on feature vectors and hierarchical clustering | |
| CN115345434A (en) | Improved dynamic data mining method and device for continuous casting quality judgment model | |
| CN108705058B (en) | A method of forecast Crystallizer bleed-out is clustered based on K-Means | |
| Ye et al. | Real-time quality prediction of casting billet based on random forest algorithm | |
| CN112108523B (en) | Method and system for processing surface defects of strip steel in cold continuous rolling production | |
| CN111950577A (en) | Point inspection method and device | |
| JP2017060965A (en) | Surface defect judgment method and apparatus for continuous casting piece and billet manufacturing method using surface defect judgment method | |
| CN108941207A (en) | A kind of hot fine rolling model process control subsystem and method | |
| CN114881425A (en) | Method and device for monitoring production in real time, electronic equipment and medium | |
| CN116228766A (en) | Intelligent regulation and control method and system for plasma processing equipment | |
| CN115423218B (en) | Prediction method and system for forge piece defect elimination process | |
| JP2015160239A (en) | Method and apparatus for surface defect determination for continuously cast slab | |
| CN115582522A (en) | High-precision chromium alloy continuous casting molding monitoring method and system | |
| CN117171936A (en) | Slab quality prediction method for extracting real-time characteristic value of crystallizer based on defect mechanism | |
| CN103390114B (en) | A kind of detection with steel finished product thickness precision judges system and method thereof | |
| CN103273032A (en) | On-line calibration method for opening of wide and thick plate continuous casting machine soft-reduction segment | |
| JP7118690B2 (en) | Manufacturing method, information processing device, relational expression calculation device, and manufacturing system | |
| CN106540970A (en) | A kind of hot-rolled intermediate billet length is monitored and aberration alarm method automatically | |
| CN117339985B (en) | Method for manufacturing sectional material by recycling rail materials | |
| Romashkin et al. | Development of an integral production process for manufacturing machinery billets based on computer simulation | |
| CN113112120B (en) | Online precision evaluation method for hot-rolled strip steel side press | |
| CN116523039B (en) | Continuous casting knowledge graph generation method and device, storage medium and terminal | |
| CN113936744B (en) | Online performance prediction method and system for producing vanadium micro-alloy structural steel by hot continuous rolling process | |
| Tschimpke et al. | Statistical Methods for Predicting of the Quality of Aluminum Ingots | |
| CN113263154B (en) | Method, device, equipment and storage medium for predicting metal dendrite spacing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |