A method and system for judging the quality of a steel coil based on continuous casting pouring high-frequency data

Abstract

The application discloses a method for judging the quality of a steel coil based on continuous casting pouring high-frequency data, which comprises the following steps: 100, obtaining continuous casting pouring high-frequency data samples of a plurality of historical steel coils with quality labels, and establishing a corresponding relationship between the continuous casting pouring high-frequency data samples and slab positions; 200, based on the continuous casting pouring high-frequency data samples, establishing the correlation between various types of high-frequency data characteristic parameters and the quality of the steel coil, and determining the judgment threshold of each type of high-frequency data characteristic parameter based on the correlation; 300, collecting actual continuous casting pouring high-frequency data, and obtaining the actual value of each type of high-frequency data characteristic parameter based on the actual continuous casting pouring high-frequency data; 400, comparing the actual value of each type of high-frequency data characteristic parameter with the corresponding judgment threshold, and predicting the quality of the steel coil based on the comparison result.

Description

Technical Field

[0001] This invention relates to a data processing method and system, and more particularly to a method and system for processing continuous casting process data. Background Technology

[0002] During continuous casting, complex physical processes such as molten metal flow and solidification generate a large amount of high-frequency data, including key parameters such as liquid surface fluctuation, casting speed, and tundish steel temperature. These parameters reflect the dynamic changes in the liquid and solid behavior of the metal during continuous casting.

[0003] Traditional methods for identifying high-frequency data anomalies often rely on manual observation of parameter fluctuations and judgment of parameter reasonableness based on human experience. This method, which depends on manual judgment, fails to quantify what constitutes acceptable and unacceptable parameter fluctuation patterns. Furthermore, it does not incorporate user feedback regarding quality complaints to explore correlations, making it impossible to identify patterns in the data for process control and quality assurance. Summary of the Invention

[0004] One of the objectives of this invention is to provide a method for judging the quality of steel coils based on high-frequency data from continuous casting. This method is based on correlation analysis and mathematical statistics to determine the correlation between high-frequency data from continuous casting and steel coil quality anomalies, and to find the allowable fluctuation range and control standards for high-frequency data from continuous casting, thereby aiming to achieve automatic identification and judgment of abnormal steel coil quality states.

[0005] Based on the above-mentioned objectives, this invention provides a method for judging the quality of steel coils based on high-frequency data from continuous casting, comprising the following steps:

[0006] 100: Obtain several historical steel coils with quality labels for continuous casting high-frequency data samples, and establish the correspondence between the continuous casting high-frequency data samples and the slab position;

[0007] 200: Based on the continuous casting high-frequency data samples, establish the correlation between various types of high-frequency data feature parameters and steel coil quality, and determine the judgment threshold of various types of high-frequency data feature parameters based on the correlation.

[0008] 300: Collect measured high-frequency data of continuous casting and obtain the actual values ​​of various high-frequency data characteristic parameters based on the measured high-frequency data of continuous casting;

[0009] 400: Compare the actual values ​​of various high-frequency data feature parameters with the corresponding judgment thresholds, and predict the quality of steel coils based on the comparison results.

[0010] Furthermore, in the method described in this invention, the quality label includes excellent quality and abnormal quality; step 400 specifically includes: if the actual value of at least one type of high-frequency data feature parameter exceeds the corresponding judgment threshold, then the steel coil is predicted to be of abnormal quality.

[0011] Furthermore, in the method described in this invention, the high-frequency data characteristic parameters include at least one of the following: liquid level fluctuation defects, casting speed fluctuation, standard deviation of molten steel weight in the tundish, and temperature fluctuation of molten steel in the tundish.

[0012] Furthermore, in the method described in this invention, step 200 specifically includes:

[0013] 201: Based on the high-frequency data of continuous casting, find liquid surface fluctuation defects, and obtain the judgment threshold of the number of liquid surface fluctuation defects by means of the correlation between the number of liquid surface fluctuation defects and the quality of steel coil;

[0014] 202: Based on the high-frequency data of continuous casting, the optimal value of casting speed fluctuation is found by examining the correlation between casting speed fluctuation and steel coil quality, and this value is used as the threshold for judging casting speed fluctuation.

[0015] 203: Based on the high-frequency data of continuous casting, the optimal value of the standard deviation of molten steel weight in the tundish is found by the correlation between the standard deviation of molten steel weight in the tundish and the quality of the steel coil, and this value is used as the threshold for judging the standard deviation of molten steel weight in the tundish.

[0016] 204: Based on the high-frequency data of continuous casting, the optimal value of tundish molten steel temperature fluctuation is found by examining the correlation between tundish molten steel temperature fluctuation and coil quality, and this value is used as the threshold for judging tundish molten steel temperature fluctuation.

[0017] Furthermore, the method described in this invention also includes step 500: outputting an alarm signal based on the prediction of steel coil quality anomalies.

[0018] Another objective of this invention is to provide a system for judging the quality of steel coils based on high-frequency data from continuous casting. This system can determine the correlation between high-frequency data from continuous casting and steel coil quality anomalies based on correlation analysis and mathematical statistics, and find the allowable fluctuation range and control standards of high-frequency data from continuous casting, thereby aiming to achieve automatic identification and judgment of abnormal steel coil quality states.

[0019] In view of the above-mentioned objectives, the present invention also provides a system for judging the quality of steel coils based on high-frequency data from continuous casting, comprising:

[0020] The data acquisition module acquires several historical steel coils with quality labels and their high-frequency continuous casting data samples, and establishes the correspondence between the high-frequency continuous casting data samples and the slab position.

[0021] The correlation module establishes the correlation between various high-frequency data feature parameters and steel coil quality based on the continuous casting high-frequency data samples, and determines the judgment threshold of various high-frequency data feature parameters based on the correlation.

[0022] The judgment module obtains the actual values ​​of various high-frequency data characteristic parameters based on the measured high-frequency continuous casting data collected by the data acquisition module; it then compares the actual values ​​of various high-frequency data characteristic parameters with the corresponding judgment thresholds, and predicts the quality of the steel coil based on the comparison results.

[0023] Furthermore, in the system described in this invention, the quality label includes excellent quality and abnormal quality; the judgment module compares the actual values ​​of various high-frequency data feature parameters with the corresponding judgment thresholds, and predicts the quality of the steel coil based on the comparison results, specifically including: if the actual value of at least one type of high-frequency data feature parameter exceeds the corresponding judgment threshold, then the steel coil is predicted to be abnormal in quality.

[0024] Furthermore, in the system described in this invention, the high-frequency data characteristic parameters include at least one of the following: liquid level fluctuation defects, casting speed fluctuation, standard deviation of molten steel weight in the tundish, and temperature fluctuation of molten steel in the tundish.

[0025] Furthermore, in the system described in this invention, the correlation module performs the following steps:

[0026] 201: Based on the high-frequency data of continuous casting, find liquid surface fluctuation defects, and obtain the judgment threshold of the number of liquid surface fluctuation defects by means of the correlation between the number of liquid surface fluctuation defects and the quality of steel coil;

[0027] 202: Based on the high-frequency data of continuous casting, the optimal value of casting speed fluctuation is found by examining the correlation between casting speed fluctuation and steel coil quality, and this value is used as the threshold for judging casting speed fluctuation.

[0028] 203: Based on the high-frequency data of continuous casting, the optimal value of the standard deviation of molten steel weight in the tundish is found by the correlation between the standard deviation of molten steel weight in the tundish and the quality of the steel coil, and this value is used as the threshold for judging the standard deviation of molten steel weight in the tundish.

[0029] 204: Based on the high-frequency data of continuous casting, the optimal value of tundish molten steel temperature fluctuation is found by examining the correlation between tundish molten steel temperature fluctuation and coil quality, and this value is used as the threshold for judging tundish molten steel temperature fluctuation.

[0030] Furthermore, the system described in this invention also includes an alarm module, which outputs an alarm signal based on the prediction of abnormal steel coil quality.

[0031] The method and system for judging the quality of steel coils based on high-frequency data from continuous casting, as described in this invention, can determine the correlation between high-frequency data from continuous casting and abnormal steel coil quality based on correlation analysis and mathematical statistics, and find the allowable fluctuation range and control standards of high-frequency data from continuous casting. This aims to achieve automatic identification and judgment of abnormal steel coil quality, thereby promptly alarming and blocking when abnormal high-frequency data is detected, and avoiding further production losses. Attached Figure Description

[0032] Figure 1 The flowchart illustrates the steps of one embodiment of the method for determining the quality of steel coils based on high-frequency data from continuous casting, as described in this invention.

[0033] Figure 2 An example of a liquid level fluctuation defect is shown.

[0034] Figure 3 The diagram shows the architecture of the system for judging the quality of steel coils based on high-frequency data from continuous casting, as described in this invention, in one embodiment. Detailed Implementation

[0035] The method and system for judging the quality of steel coils based on high-frequency data of continuous casting according to the present invention will be further described below with reference to specific embodiments and accompanying drawings. However, this description does not constitute an improper limitation of the present invention.

[0036] Figure 1 The flowchart illustrates the steps of one embodiment of the method for determining the quality of steel coils based on high-frequency data from continuous casting, as described in this invention.

[0037] like Figure 1 As shown, in some embodiments, the method for determining the quality of steel coils based on high-frequency data from continuous casting according to the present invention may include the following steps:

[0038] 100: Obtain several historical steel coil high-frequency data samples for continuous casting with quality labels, and establish the correspondence between the high-frequency data samples for continuous casting and the slab position.

[0039] In some specific implementations, quality labels may include excellent quality and poor quality (or defective quality).

[0040] In some specific implementations, high-frequency data samples of the continuous casting process may include liquid level height, casting speed, tundish steel weight, and tundish steel temperature.

[0041] In a specific example, the data acquisition frequency can be 100ms. Since the casting process is continuous, the casting machine exit will cut the solidified crystals into slabs, so high-frequency data samples can be correlated with the slab positions during data acquisition.

[0042] 200: Based on the continuous casting high-frequency data samples, establish the correlation between various types of high-frequency data feature parameters and steel coil quality, and determine the judgment threshold of various types of high-frequency data feature parameters based on the correlation.

[0043] In some specific implementations, corresponding to the high-frequency data samples of the continuous casting process described above, the high-frequency data characteristic parameters may include at least one of the following: liquid level fluctuation defects, casting speed fluctuation, standard deviation of molten steel weight in the tundish, and temperature fluctuation of molten steel in the tundish.

[0044] Therefore, in some specific implementations, step 200 specifically includes:

[0045] 201: Based on the high-frequency data of continuous casting, find liquid surface fluctuation defects, and obtain the judgment threshold of the number of liquid surface fluctuation defects by means of the correlation between the number of liquid surface fluctuation defects and the quality of steel coil;

[0046] 202: Based on the high-frequency data of continuous casting, the optimal value of casting speed fluctuation is found by examining the correlation between casting speed fluctuation and steel coil quality, and this value is used as the threshold for judging casting speed fluctuation.

[0047] 203: Based on the high-frequency data of continuous casting, the optimal value of the standard deviation of molten steel weight in the tundish is found by the correlation between the standard deviation of molten steel weight in the tundish and the quality of the steel coil, and this value is used as the threshold for judging the standard deviation of molten steel weight in the tundish.

[0048] 204: Based on the high-frequency data of continuous casting, the optimal value of tundish molten steel temperature fluctuation is found by examining the correlation between tundish molten steel temperature fluctuation and coil quality, and this value is used as the threshold for judging tundish molten steel temperature fluctuation.

[0049] In this way, based on the correlation between various high-frequency data characteristic parameters and steel coil quality, the judgment thresholds for these high-frequency data characteristic parameters can be determined. For example... Figure 1 As shown, in this embodiment, the threshold for judging the number of defects in liquid level fluctuation is a number greater than 0, the threshold for judging the speed fluctuation is 1.25, the threshold for judging the standard deviation of molten steel weight in the tundish is 7, and the threshold for judging the temperature fluctuation of molten steel in the tundish is 10.

[0050] 300: Collect measured high-frequency data of continuous casting and obtain the actual values ​​of various high-frequency data characteristic parameters based on the measured high-frequency data of continuous casting.

[0051] 400: Compare the actual values ​​of various high-frequency data feature parameters with the corresponding judgment thresholds, and predict the quality of steel coils based on the comparison results.

[0052] In some implementations, if the actual value of at least one type of high-frequency data feature parameter exceeds the corresponding judgment threshold, the steel coil is predicted to be of abnormal quality. Only when the actual values ​​of all high-frequency data feature parameters do not exceed the corresponding judgment threshold is the steel coil of excellent quality predicted.

[0053] In some preferred embodiments, an alarm signal is also output when the quality of the steel coil is predicted to be abnormal, so as to immediately shut down the continuous casting process and avoid further production losses.

[0054] In some more specific implementations, the threshold for determining the number of liquid level fluctuation defects can be obtained based on the following steps:

[0055] During continuous casting, the molten steel level fluctuates, a phenomenon known as surface fluctuation. When significant surface fluctuations occur at a certain point or within a nearby range, forming a point of abrupt change, this location can be termed a surface fluctuation defect. For example... Figure 2 An example is shown of a liquid level fluctuation defect. Figure 2 The horizontal axis in the diagram represents the measuring point, and the vertical axis represents the liquid level. Figure 2 The area covered by the red dot represents a defect. Furthermore, when multiple adjacent points exceed the threshold τ, they are merged into a single defect; the diameter of the red dot is proportional to the number of merged defects. When significant liquid level fluctuations lead to defects, it affects the quality of the steel coil. Therefore, to analyze cases where liquid level fluctuations deviate from the mean, the deviation value is first calculated using the following formula.

[0056] X i =X i -μ (1)

[0057] Where i represents the measurement point number on the slab, and is an integer. μ is the average liquid level of all liquid levels on the slab, and X... i Let X' be the liquid level value at the i-th point of the slab. i Let X be the deviation value of the liquid level fluctuation at the i-th point on the slab. i The larger the value, the more it deviates from the overall value.

[0058] To quantitatively characterize the number of defects, X i Set a threshold τ and define:

[0059]

[0060] Then, by analyzing the correlation between liquid surface fluctuation defects and steel coil quality, the threshold τ for judging the number of liquid surface fluctuation defects can be calculated.

[0061] To further analyze and determine the optimal deviation threshold for liquid surface fluctuation defects, a reasonable threshold τ is calculated by taking a point with a step size (e.g., 0.1) set between [0.1, +∞) so that liquid surface fluctuation has the strongest correlation with steel coil quality abnormalities.

[0062] In some more specific implementations, the threshold τ for determining the number of liquid surface fluctuation defects is... k The methods for determining this may include:

[0063] (1) When τ is 0.1, calculate whether there are defects in each sample slab according to formula (1) and formula (2).

[0064] (2) Further analysis was conducted on the correlation between the presence or absence of defects in all sample slabs and the presence or absence of quality abnormalities in the corresponding steel coils when τ was set to 0.1. It was found that when τ was set to 0.1, among all sample slabs, M slabs had defects. τ＝0.1 There are N defect-free slabs. τ＝0.1 Of these, P represents slabs with defects, specifically those exhibiting quality abnormalities (or disputed slabs). τ =0.1, the number of slabs without quality abnormalities is Q. τ =0.1, the proportion of slabs with quality abnormalities and defects is The proportion of slabs without abnormalities and defects is Among the slabs without defects, the slab with quality abnormalities is designated as R. τ＝0.1 One slab without any quality abnormalities is designated as S. τ＝0.1 The proportion of disputed slabs and defects. The ratio of undisputed slabs to defects When K 有异议 and L 无异议 The larger K is, the more 无异议 and L 有异议 The smaller the value, the stronger the correlation between defects and quality abnormalities is considered.

[0065] (3) The value of τ increases by 0.1 each time, and steps (1) and (2) are repeated.

[0066] (4) Continue to increase the value of τ until the correlation index reaches an inflection point (i.e., K). 有异议 and L 无异议 (When it is at its peak). When the inflection point occurs, τ is the optimal value, that is, τ = τ_optimum. k A threshold is set for the number of defects in the liquid surface fluctuation. At this threshold, the number of defects is strongly correlated with the quality of the steel coil.

[0067] In a specific instance, obtained in this way, when τ k When the value is 5, the correlation is considered to be the strongest.

[0068] In some more specific implementations, the method for determining the threshold for judging pulling speed fluctuation may include:

[0069] In continuous casting, the casting speed (or stretching rate) is a critical operating parameter that directly affects the quality of the slab, its crystallization behavior, and the efficiency of continuous casting. Continuous changes in the casting speed will alter the slab quality. Therefore, the method described in this invention uses the casting speed fluctuation value as a characteristic parameter. The casting speed fluctuation value is calculated according to the following formula:

[0070] Y' i =Y i -β (3)

[0071] Where i represents the measurement point number on the slab (integer), β is the average tensioning speed of all points on the slab, and Y... i Y' is the drawing speed value at the i-th point of the slab. i Let be the pulling speed fluctuation value at the i-th point on the slab.

[0072] Then, by analyzing the correlation between casting speed fluctuation and coil quality, the optimal value of casting speed fluctuation can be calculated.

[0073] In some more specific implementations, a point is selected in the range [0.1, +∞) with a set step size (e.g., 0.1) to be identified as the pulling speed fluctuation. Finding the optimal value of pulling speed fluctuation When When the slab quality pass rate is highest and the quality disputes are fewest, then... At that time, the pass rate of slab quality was the lowest, and the number of quality disputes was the highest.

[0074] The methods for determining this may include:

[0075] (1) When the value is 0.1, a correlation analysis was conducted on the quality issues of all sample slabs with drawing speed fluctuations greater than 0.1 and less than 0.1, and the corresponding coils. The results were obtained when... When the value is 0.1, among all sample slabs, the slabs with a drawing speed fluctuation greater than 0.1 are: There are slabs with a casting speed fluctuation of less than 0.1. Of these, the slabs with a drawing speed fluctuation of less than 0.1 mm had quality issues. One, the slab with no quality objection is Of the items, the proportion of those with objections regarding slab and drawing speed fluctuations less than 0.1% is [number missing]. The proportion of undisputed slabs and drawing speed fluctuations less than 0.1% is: Among the slabs with a drawing speed fluctuation greater than 0.1, the slabs with quality disputes are: One, the slab with no quality objection is There are 100 cases where there are objections regarding the proportion of slabs and casting speed fluctuations greater than 0.1%. The proportion of slabs with undisputed fluctuations in casting speed greater than 0.1% When K (拉速)有异议 and L (拉速)无异议 The smaller K is, the better. (拉速)无异议 and L (拉速)有异议 The larger the value, the stronger the correlation between pulling speed fluctuations and quality disputes.

[0076] (2) Each time the value is increased by a step size of 0.1, step (1) is repeated.

[0077] (3) Continue to increase The value is maintained until the correlation indicator reaches an inflection point. When the inflection point occurs, then... The optimal value is, The optimal value for tensile fluctuation is taken, at which point the tensile speed fluctuation is strongly correlated with the quality of the steel coil, and therefore it is used as the threshold for judging tensile speed fluctuation.

[0078] In a specific instance, it is obtained in this way.

[0079] In some more specific implementations, the method for determining the threshold for judging the standard deviation of molten steel weight in the tundish may include:

[0080] First, calculate the standard deviation of the molten steel weight in the tundish using the following formula:

[0081]

[0082] Where i represents the measuring point number on the slab, and N is the measuring point length, which is an integer. Z i Let ε be the weight of molten steel in the tundish at the i-th point on the slab, ε be the average weight of molten steel in the tundish on the slab, and σ be the weight of molten steel in the tundish at the i-th point on the slab. i This represents the standard deviation of the molten steel weight in the tundish.

[0083] Then, through correlation analysis between the standard deviation of molten steel weight in the tundish and the quality of the steel coil, the judgment threshold σ of the standard deviation of molten steel weight in the tundish is calculated. k .

[0084] In some more specific implementations, a point is taken in the range [0.05, +∞) with a set step size (e.g., 0.05) to determine the standard deviation σ of the molten steel weight in the tundish, and the optimal value σ is sought. k , so that when σ i When σ ≤ σ, the slab quality pass rate is the highest and the quality disputes are the fewest. i When the value is ≥σ, the pass rate of slab quality is the lowest, and the number of quality disputes is the highest.

[0085] σk The methods for determining this may include:

[0086] (1) When σ is 0.05, a correlation analysis was conducted on the quality discrepancies between the coils with standard deviations of molten steel weight in the tundish of all sample slabs greater than 0.05 and the corresponding coils when σ is 0.05. The results showed that when σ is 0.05, among all sample slabs, M slabs with standard deviations of molten steel weight in the tundish greater than 0.05 are... σ＝0.05 There are N slabs with a standard deviation of molten steel weight less than 0.05 in the tundish. σ＝0.05 Of the slabs with a standard deviation of molten steel weight less than 0.05 in the tundish, slabs with quality disputes numbered P. σ＝0.05 One slab with no quality objections is designated as Q. σ＝0.05 The proportion of cases where the standard deviation of the weight of slabs and molten steel in tundishes is less than 0.05 is [number missing]. The proportion of undisputed slab and tundish steel weight standard deviations less than 0.05 is: Among the slabs with a standard deviation of molten steel weight greater than 0.05 in the tundish, the slabs with quality disputes are designated as R. σ＝0.05 One slab with no quality objections is designated as S. σ＝0.05 There are 100 cases where the standard deviation of the weight of slabs and molten steel in tundishes is greater than 0.05. The proportion of slab and tundish steel weight standard deviations greater than 0.05 without objection When K (钢水重量)有异议 and L (钢水重量)无异议 The smaller K is, the better. (钢水重量)无异议 and L (钢水重量)有异议 The larger the value, the stronger the correlation between the standard deviation of molten steel weight in the tundish and quality disputes.

[0087] (2) The value of σ is increased by 0.05 each time, and step (1) is repeated.

[0088] (3) Continue to increase the value of σ until the correlation index reaches an inflection point. When the inflection point appears, σ is the optimal value, i.e., σ = σ k A threshold is set for the standard deviation of the molten steel weight in the tundish, at which point there is a strong correlation between the standard deviation of the molten steel weight in the tundish and the quality of the steel coil.

[0089] In a specific instance, obtained in this way, when σ k The correlation is strongest when the value is 7.

[0090] In some more specific embodiments, the method for determining the threshold for judging the temperature fluctuation of molten steel in the tundish may include:

[0091] In continuous casting, the temperature of the molten steel in the tundish is a critical parameter that directly affects the quality of the cast billet and production efficiency. Maintaining an appropriate temperature for the molten steel in the tundish ensures the smooth progress of the continuous casting process and minimizes surface and internal quality problems in the cast billet. Therefore, in some implementations, the temperature fluctuation of the molten steel in the tundish is used as a characteristic parameter.

[0092] First, calculate the temperature fluctuation of the molten steel in the tundish using the following formula:

[0093] W' i =W i -α (5)

[0094] Where i represents the measurement point number on the slab, and is an integer. W i Let α be the temperature of the molten steel in the tundish at the i-th point on the slab, and let W be the average temperature of the molten steel in the tundish on the slab. i 'This refers to the temperature fluctuation of molten steel in the tundish.'

[0095] Then, the optimal value θ of the molten steel temperature fluctuation in the tundish was calculated through correlation analysis between the molten steel temperature fluctuation in the tundish and the quality discrepancy of the steel coil. k This serves as a threshold for judging temperature fluctuations in molten steel in the tundish.

[0096] In some more specific implementations, a point is selected in the range [0.5, +∞) with a set step size (e.g., 0.5) to identify the temperature fluctuation of the molten steel in the tundish, and the optimal value θ is sought. k , so that when W i When W ≤ θ, the slab quality pass rate is the highest and the quality disputes are the fewest. i When the value is ≥θ, the pass rate of slab quality is the lowest, and the number of quality disputes is the highest.

[0097] θ k The methods for determining this may include:

[0098] (1) When θ is 0.05, a correlation analysis was conducted on the quality disputes between the coils with tundish molten steel temperature fluctuations greater than 0.5 and less than 0.5 in all sample slabs. The results showed that when θ is 0.5, among all sample slabs, there were Mθ = 0.5 slabs with tundish molten steel temperature fluctuations greater than 0.5, and N slabs with tundish molten steel temperature fluctuations less than 0.5. θ＝0.5 Of the slabs with molten steel temperature fluctuations of less than 0.5°C in the tundish, slabs with quality disputes were P. θ＝0.5 One slab with no quality objections is designated as Q. θ＝0. Of the 5 cases, the proportion where the temperature fluctuation of slab and molten steel in the tundish was less than 0.5 mm was disputed. The proportion of slab and tundish molten steel temperature fluctuations less than 0.5% without objection is: Among the slabs with molten steel temperature fluctuations greater than 0.5 in the tundish, 0.5 slabs had quality disputes (Rθ = 0.5), and 0.5 slabs had no quality disputes (Sθ = 0.5). The proportion of disputed slabs and slabs with molten steel temperature fluctuations greater than 0.5 in the tundish is... The proportion of slab and tundish molten steel temperature fluctuations exceeding 0.5% is undisputed. When K (钢水温度)有异议 and L (钢水温度)无异议 The smaller K is, the better. (钢水温度)无异议 and L (钢水温度)有异议 The larger the value, the stronger the correlation between tundish molten steel temperature fluctuations and quality disputes.

[0099] (2) The value of θ is increased by 0.5 each time, and step (1) is repeated.

[0100] (3) Continue to increase the value of θ until the correlation index reaches an inflection point. When the inflection point appears, σ is the optimal value, i.e., θ = θ k The threshold for judging the temperature fluctuation of molten steel in the tundish is set, at which point there is a strong correlation between the temperature of molten steel in the tundish and the quality of the steel coil.

[0101] In a specific instance, this method is used to obtain the result when θ k When the value is 10, the correlation is considered to be the strongest.

[0102] Table 1 lists the judgment thresholds for each high-frequency data feature parameter in a specific instance.

[0103] Table 1.

[0104] High-frequency data characteristic parameters High-quality steel coils Poor quality steel coils Liquid surface fluctuation defects 0 [1,+∞) Pull speed fluctuation [0,1.25) [1.25,+∞) Standard deviation of molten steel weight in tundish [0,7) [7,+∞) Temperature fluctuations in molten steel in the tundish [0,10) [10,+∞)

[0105] In another embodiment of the present invention, a system for judging the quality of steel coils based on high-frequency data from continuous casting is also provided.

[0106] Figure 3 The diagram shows the architecture of the system for judging the quality of steel coils based on high-frequency data from continuous casting, as described in this invention, in one embodiment.

[0107] like Figure 3 As shown, in some implementations, the system may include:

[0108] The data acquisition module 500 acquires several historical steel coil high-frequency data samples with quality labels for continuous casting, and establishes the correspondence between the high-frequency data samples for continuous casting and the slab position.

[0109] The correlation module 600 establishes the correlation between various types of high-frequency data feature parameters and steel coil quality based on the continuous casting high-frequency data samples, and determines the judgment threshold of various types of high-frequency data feature parameters based on the correlation.

[0110] The judgment module 700 obtains the actual values ​​of various high-frequency data characteristic parameters based on the measured high-frequency continuous casting data collected by the data acquisition module 500; it compares the actual values ​​of various high-frequency data characteristic parameters with the corresponding judgment thresholds, and predicts the quality of the steel coil based on the comparison results.

[0111] In some more specific embodiments, the quality labels include excellent quality and abnormal quality. In this case, the judgment module 700 compares the actual values ​​of various high-frequency data feature parameters collected by the data acquisition module 500 with the corresponding judgment thresholds, and predicts the quality of the steel coil based on the comparison results. Specifically, if the actual value of at least one type of high-frequency data feature parameter exceeds the corresponding judgment threshold, the steel coil is predicted to be of abnormal quality.

[0112] In some more specific embodiments, the high-frequency data characteristic parameters include at least one of the following: liquid level fluctuation defects, casting speed fluctuation, standard deviation of molten steel weight in the tundish, and temperature fluctuation of molten steel in the tundish.

[0113] In some more specific implementations, the correlation module 600 performs the following steps:

[0114] 201: Based on the high-frequency data of continuous casting, find liquid surface fluctuation defects, and obtain the judgment threshold of the number of liquid surface fluctuation defects by means of the correlation between the number of liquid surface fluctuation defects and the quality of steel coil;

[0115] 202: Based on the high-frequency data of continuous casting, the optimal value of casting speed fluctuation is found by examining the correlation between casting speed fluctuation and steel coil quality, and this value is used as the threshold for judging casting speed fluctuation.

[0116] 203: Based on the high-frequency data of continuous casting, the optimal value of the standard deviation of molten steel weight in the tundish is found by the correlation between the standard deviation of molten steel weight in the tundish and the quality of the steel coil, and this value is used as the threshold for judging the standard deviation of molten steel weight in the tundish.

[0117] 204: Based on the high-frequency data of continuous casting, the optimal value of tundish molten steel temperature fluctuation is found by examining the correlation between tundish molten steel temperature fluctuation and coil quality, and this value is used as the threshold for judging tundish molten steel temperature fluctuation.

[0118] Furthermore, in some preferred embodiments, the system of the present invention may also include an alarm module that outputs an alarm signal based on a prediction of abnormal steel coil quality.

[0119] In this invention, the data acquisition module 500, the correlation module 600, the judgment module 700, and the alarm module can be implemented in any suitable manner. For example, they can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicon Labs C8051F320.

[0120] Furthermore, those skilled in the art will recognize that, besides implementing the controller using purely computer-readable program code, the method steps can be logically programmed to enable the modules to perform the same function in the form of logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers (PLCs), and embedded microcontrollers. Therefore, such a module can be considered a hardware component, and the devices included within it for implementing various functions can also be considered structures within that hardware component. Alternatively, the devices for implementing various functions can be considered as both software modules implementing the method and structures within a hardware component.

[0121] The present invention will be further illustrated by a specific example below.

[0122] We obtained a batch of 2,880 historical steel coils with high-frequency continuous casting data samples. Among them, 340 steel coils were marked with quality labels indicating quality issues (i.e., abnormal or poor quality), and 2,540 steel coils were marked with quality labels indicating no quality issues (excellent quality).

[0123] Perform correlation analysis as described above, when τ k When taking 5 samples, 2545 steel coils were defect-free, and 355 coils had defects. Of the defective coils, 330 had quality objections, accounting for 93%. Of the defect-free coils, 2515 had no quality objections, accounting for 98%. At that time, 320 coils had a drawing speed fluctuation greater than 1.25, of which 285 coils had quality issues, accounting for 89%. 2560 coils had a drawing speed fluctuation less than or equal to 1.25, of which 2505 coils had no quality issues, accounting for 97%. When σ... kWhen θ = 7, the number of coils with a standard deviation greater than 7 in the weight of molten steel in the tundish is 405, of which 305 coils have quality disputes, accounting for 75%. The number of coils with a standard deviation less than or equal to 7 in the weight of molten steel in the tundish is 2475, of which 2440 coils have no quality disputes, accounting for 98%. k When the temperature fluctuation is 10, there are 380 coils with a temperature fluctuation greater than 10 in the tundish, of which 310 coils have quality objections, accounting for 81%. There are 2400 coils with a temperature fluctuation of less than or equal to 10 in the tundish, of which 2370 coils have no quality objections, accounting for 98%.

[0124] During actual testing, measured high-frequency data of continuous casting of steel coil with coil number 14192375000 were obtained.

[0125] Based on the liquid level data, the actual liquid level fluctuation defect rate was 0, which did not exceed the judgment threshold. Based on the casting speed data, the casting speed fluctuation was calculated to be between 0 and 1.25, which also did not exceed the judgment threshold. Based on the molten steel weight data in the ladle, the standard deviation of the molten steel weight was calculated to be between 0 and 7, which did not exceed the judgment threshold. Based on the molten steel temperature data in the ladle, the temperature fluctuation was calculated to be greater than 10, which exceeded the judgment threshold. Therefore, the final prediction for the quality of this steel coil is poor.

[0126] Therefore, this invention can determine the correlation between high-frequency data of continuous casting and steel coil quality anomalies based on correlation analysis and mathematical statistics, and find the allowable fluctuation range and control standards of high-frequency data of continuous casting. It aims to achieve automatic identification and judgment of abnormal steel coil quality states, so as to promptly alarm and block when abnormal high-frequency data is detected, and avoid further production losses.

[0127] It should be noted that the scope of protection of the prior art in this invention is not limited to the embodiments given in this application. All prior art that does not contradict the solution of this invention, including but not limited to prior patent documents, prior publications, prior public uses, etc., can be included in the scope of protection of this invention.

[0128] Furthermore, the combination of the technical features in this case is not limited to the combination methods described in the claims of this case or the combination methods described in the specific embodiments. All technical features described in this case can be freely combined or combined in any way, unless they contradict each other.

[0129] It should also be noted that the embodiments listed above are merely specific embodiments of the present invention. Obviously, the present invention is not limited to the above embodiments, and similar changes or modifications made thereto are those that can be directly derived or easily conceived by those skilled in the art from the content disclosed in the present invention, and should all fall within the protection scope of the present invention.

Claims

1. A method for judging the quality of steel coils based on high-frequency data from continuous casting, characterized in that, The process includes the following steps: 100: acquiring several historical steel coils with quality labels for continuous casting high-frequency data samples, and establishing the correspondence between the continuous casting high-frequency data samples and the slab position; 200: Based on the continuous casting high-frequency data samples, establish the correlation between various types of high-frequency data feature parameters and steel coil quality, and determine the judgment threshold of various types of high-frequency data feature parameters based on the correlation. 300: Collect measured high-frequency data of continuous casting and obtain the actual values ​​of various high-frequency data characteristic parameters based on the measured high-frequency data of continuous casting; 400: Compare the actual values ​​of various high-frequency data feature parameters with the corresponding judgment thresholds, and predict the quality of steel coils based on the comparison results.

2. The method as described in claim 1, characterized in that, The quality labels include excellent quality and abnormal quality; step 400 specifically includes: if the actual value of at least one type of high-frequency data feature parameter exceeds the corresponding judgment threshold, then the steel coil is predicted to be of abnormal quality.

3. The method as described in claim 2, characterized in that, The high-frequency data characteristic parameters include at least one of the following: liquid level fluctuation defects, casting speed fluctuation, standard deviation of molten steel weight in the tundish, and temperature fluctuation of molten steel in the tundish.

4. The method as described in claim 3, characterized in that, Step 200 specifically includes: 201: Based on the high-frequency data of continuous casting, find liquid surface fluctuation defects, and obtain the judgment threshold of the number of liquid surface fluctuation defects by means of the correlation between the number of liquid surface fluctuation defects and the quality of steel coil; 202: Based on the high-frequency data of continuous casting, the optimal value of casting speed fluctuation is found by examining the correlation between casting speed fluctuation and steel coil quality, and this value is used as the threshold for judging casting speed fluctuation. 203: Based on the high-frequency data of continuous casting, the optimal value of the standard deviation of molten steel weight in the tundish is found by the correlation between the standard deviation of molten steel weight in the tundish and the quality of the steel coil, and this value is used as the threshold for judging the standard deviation of molten steel weight in the tundish. 204: Based on the high-frequency data of continuous casting, the optimal value of tundish molten steel temperature fluctuation is found by examining the correlation between tundish molten steel temperature fluctuation and coil quality, and this value is used as the threshold for judging tundish molten steel temperature fluctuation.

5. The method as described in claim 1, characterized in that, It also includes step 500: predicting and outputting an alarm signal based on steel coil quality anomalies.

6. A system for judging the quality of steel coils based on high-frequency data from continuous casting, characterized in that, include: The data acquisition module acquires several historical steel coils with quality labels and their high-frequency continuous casting data samples, and establishes the correspondence between the high-frequency continuous casting data samples and the slab position. The correlation module establishes the correlation between various high-frequency data feature parameters and steel coil quality based on the continuous casting high-frequency data samples, and determines the judgment threshold of various high-frequency data feature parameters based on the correlation. The judgment module obtains the actual values ​​of various high-frequency data characteristic parameters based on the measured high-frequency continuous casting data collected by the data acquisition module; it then compares the actual values ​​of various high-frequency data characteristic parameters with the corresponding judgment thresholds, and predicts the quality of the steel coil based on the comparison results.

7. The system as described in claim 6, characterized in that, The quality labels include excellent quality and abnormal quality; the judgment module compares the actual values ​​of various high-frequency data feature parameters with the corresponding judgment thresholds, and judges the quality of the steel coil based on the comparison results, specifically including: if the actual value of at least one type of high-frequency data feature parameter exceeds the corresponding judgment threshold, then the steel coil is predicted to be abnormal in quality.

8. The system as described in claim 6, characterized in that, The high-frequency data characteristic parameters include at least one of the following: liquid level fluctuation defects, casting speed fluctuation, standard deviation of molten steel weight in the tundish, and temperature fluctuation of molten steel in the tundish.

9. The system as described in claim 6, characterized in that, The correlation module performs the following steps: 201: Based on the high-frequency data of continuous casting, find liquid surface fluctuation defects, and obtain the liquid surface fluctuation defect quantity judgment threshold by the correlation between the number of liquid surface fluctuation defects and the quality of steel coil. 202: Based on the high-frequency data of continuous casting, the optimal value of casting speed fluctuation is found by examining the correlation between casting speed fluctuation and steel coil quality, and this value is used as the threshold for judging casting speed fluctuation. 203: Based on the high-frequency data of continuous casting, the optimal value of the standard deviation of molten steel weight in the tundish is found by the correlation between the standard deviation of molten steel weight in the tundish and the quality of the steel coil, and this value is used as the threshold for judging the standard deviation of molten steel weight in the tundish. 204: Based on the high-frequency data of continuous casting, the optimal value of tundish molten steel temperature fluctuation is found by examining the correlation between tundish molten steel temperature fluctuation and coil quality, and this value is used as the threshold for judging tundish molten steel temperature fluctuation.

10. The system as described in claim 6, characterized in that, It also includes an alarm module, which outputs alarm signals based on predictions of abnormal steel coil quality.

Images