KR20200000699A - method of estimating at least one of coating weight and alloy degree of zinc alloy plated sheet - Google Patents

Abstract

According to an embodiment of the present invention, a method of predicting at least one of the galvanizing amount and alloying degree of a galvanized steel sheet includes the following steps of: (a) preparing a plurality of galvanized steel sheets having different galvanizing amounts and alloying degrees; (b) projecting an X-ray into the galvanized steel sheets, and measuring the intensity of a plurality of fluorescent X-rays detected from the respective galvanized steel sheets; (c) conducting a wet analysis method on each of the galvanized steel sheets, thereby measuring the galvanizing amount and alloying degree of each of the galvanized steel sheets; (d) calculating a correlation between the intensity of the fluorescent X-rays measured individually and at least one of the galvanizing amount and alloying degree according to the wet analysis method with respect to the galvanized steel sheets, thereby determining at least one of a first fluorescent X-ray, which is at least one of the fluorescent X-rays with a high degree of relation between the galvanizing amount and the correlation, and a second fluorescent X-ray, which is at least one of the fluorescent X-rays with a high degree of relation between the alloying degree and the correlation; (e) performing machine learning on at least one of a first correlation between the first fluorescent X-ray and the galvanizing amount and a second correlation between the second fluorescent X-ray and the alloying degree, thereby deriving a prediction model for at least one of the first and second correlations; and (f) measuring the intensity of at least one of the first and second fluorescent X-rays from a targeted galvanized steel sheet through the prediction model, and then, predicting at least one of the galvanizing amount and the alloying degree from the measured intensity.

Description

Prediction method of at least one of coating weight and alloy degree of zinc alloy plated sheet

The present invention relates to at least one prediction method of plating amount and alloying degree of galvanized steel sheet.

At present, galvanized steel sheet is widely applied in the fields of home appliances, automobiles and dry materials. The galvanized steel sheet may vary in its use and corrosion resistance depending on the plating amount and alloying degree. Therefore, there is a need to precisely measure and manage the plating amount and alloying degree of the galvanized steel sheet.

In general, a wet analysis method, a fluorescence X-ray analysis method, or the like is applied as a method for measuring the plating amount and the degree of alloying. The wet analysis method may be performed by measuring the weight of the galvanized steel sheet by separating only the plating layer by wet, and the fluorescent X-ray analysis irradiates X-rays on the galvanized steel sheet sample and measures the intensity of the fluorescent X-rays emitted from the sample. Can be done in a way.

Related prior art documents include Korean Patent Laid-Open Publication No. 2000-0025344 (published on May 06, 2000), which discloses a method for measuring plating amount and alloying degree using fluorescent X-rays.

The problem to be solved by the present invention provides a method for reducing the error between the predicted value and the measured value when predicting at least one of the coating amount and the alloying degree of the galvanized steel sheet by X-ray fluorescence analysis.

According to an aspect of the present invention, a method for predicting at least one of the plating amount and alloying degree of a galvanized steel sheet may include: (a) preparing a plurality of galvanized steel sheets having at least one of plating amount and alloying degree different from each other; (b) injecting X-rays into the plurality of galvanized steel sheets and measuring the intensity of a plurality of fluorescent X-rays detected from each of the galvanized steel sheets; (c) performing a wet analysis method on the plurality of galvanized steel sheets to measure at least one of plating amount and alloying degree of each of the galvanized steel sheets; (d) calculating a correlation between the plurality of fluorescence X-ray intensities measured at least one of the plating amount and the alloying degree according to the wet analysis method and the plating amount and the correlation with respect to the plurality of galvanized steel sheets, respectively. At least one of a first fluorescent X-ray that is at least one of the plurality of highly related fluorescent X-rays and a second fluorescent X-ray that is at least one of the plurality of fluorescent X-rays that are highly related to the alloying degree Determining; (e) performing machine learning on at least one of a first correlation between the first fluorescence X-rays and the plating amount and a second correlation between the second fluorescence X-rays and the degree of alloying, wherein Deriving a prediction model for at least one of the first and second correlations; And (e) measuring the intensity of at least one of the first and second fluorescent X-rays from a desired galvanized steel sheet using the predictive model, and predicting at least one of plating amount and alloying degree from the measured intensity. Steps.

In one embodiment, the galvanized steel sheet may be an alloyed hot dip galvanized steel sheet.

In another embodiment, in step (b), the plurality of fluorescent X-rays may include Kα, Kβ and Lα lines for iron (Fe), and Kα, Kβ and Lα lines for zinc (Zn). .

In another embodiment, the first fluorescent X-ray includes any one of Kα, Kβ, and Lα lines for iron (Fe), and any one of Kα, Kβ, and Lα lines for zinc (Zn). The second fluorescent X-ray may include any one of Kα, Kβ, and Lα rays for the iron (Fe), and may include any one of Kα, Kβ, and Lα rays for the zinc (Zn). have. In this case, the first fluorescent X-rays and the second fluorescent X-rays may be of different types.

In another embodiment, the method for predicting the plating amount and alloying degree of the galvanized steel sheet may include at least one of the first and second fluorescent X-rays corresponding to at least one of the plating amount and the alloying degree predicted in step (f). The method may further include storing the strength of the data in a database for modeling optimization of the machine learning.

According to an embodiment of the present invention, by measuring the fluorescence X-ray intensity from a predetermined galvanized steel sheet and substituting the measured fluorescence X-ray intensity into a predictive model derived in advance according to machine learning, the plating amount and alloying of the galvanized steel sheet At least one of the figures may be predicted. In particular, the method of the present embodiment overcomes the disadvantages of the conventional prediction method through the regression analysis or the basic variable method when estimating at least one of the plating amount and the alloying degree of the alloyed hot-dip galvanized steel sheet. Or, reduce the prediction error between the fluorescence X-ray intensity and the degree of alloying.

1 is a flowchart schematically showing a method for predicting at least one of plating amount and alloying degree of a galvanized steel sheet according to an embodiment of the present invention.
2 is a graph comparing the plating amount prediction value and the plating amount measurement value derived by the prediction method according to an embodiment of the present invention.
3 is a graph comparing the alloying degree prediction value and the alloying degree quantitative analysis value derived by the prediction method according to an embodiment of the present invention.

Hereinafter, with reference to the accompanying drawings will be described in detail to be easily carried out by those of ordinary skill in the art. As those skilled in the art would realize, the described embodiments may be modified in various different ways, all without departing from the spirit or scope of the present invention. Like reference numerals designate like or similar components throughout the specification. In addition, detailed descriptions of well-known functions and configurations that may unnecessarily obscure the subject matter of the present invention will be omitted.

An embodiment of the present invention provides a prediction method of at least one of the plating amount and the alloying degree of the galvanized steel sheet. In particular, the embodiment of the present invention can be effectively applied to at least one prediction method of the coating amount and the alloying degree of galvanized galvanized steel sheet (Galvannealed steel sheet, GA steel sheet).

Recently, the galvannealed steel sheet (GA) has a higher surface quality such as weldability and paintability than general hot dip galvanized steel sheet, and thus its demand is increasing. The alloyed hot-dip galvanized steel sheet is produced by hot-dip galvanizing zinc on the steel sheet and then forming a Fe-Zn alloy phase in the plating layer through heat treatment. Since the surface quality (processability, powdering property, paintability, corrosion resistance) of the alloyed hot-dip galvanized steel sheet is closely related to the alloying degree (Fe%) in the plating layer, the alloying degree is improved to improve the surface quality of the alloyed hot-dip galvanized steel sheet. Accurate measurement and control are required.

On the other hand, when fluorescence X-ray analysis is applied to an alloyed hot-dip galvanized steel sheet, since the iron of the steel sheet and the iron of the plating layer are present at the same time, it is essential to distinguish the fluorescent X-ray intensity emitted from the steel sheet and the plating layer, respectively. Is requested.

In addition, when the conventional linear regression equation is used to obtain the degree of alloying from the intensity of the fluorescent X-ray, an error occurs when the alloying degree is predicted because the intensity of the fluorescent X-ray does not increase linearly with the plating amount. In addition, when the conventional basic variable method is used to determine the degree of alloying from the intensity of fluorescent X-rays, an alloying degree measurement error also occurs because the plating layer of the alloyed hot dip galvanized steel sheet is not made of a homogeneous alloy phase. The embodiment of the present invention provides a new prediction method for deriving a prediction model between the results of the wet analysis method and the X-ray fluorescence analysis through machine learning, away from the conventional linear regression and the basic variable method.

1 is a flowchart schematically showing a method for predicting at least one of plating amount and alloying degree of a galvanized steel sheet according to an embodiment of the present invention.

First, in step S110 of FIG. 1, at least one of the plating amount and the alloying degree is prepared a plurality of galvanized steel sheet different from each other. The plurality of galvanized steel sheets may be test specimens for providing a plurality of experimental examples for deriving a prediction model for at least one of plating amount and alloying degree. Accordingly, the plurality of galvanized steel sheets may be selected from at least one of the plating amount and the alloying degree are different from each other. In this case, the plating amount may mean an amount of the iron (Fe) -zinc (Zn) alloy plating layer formed on the substrate. The alloying degree may mean an amount of iron present in the iron (Fe) -zinc (Zn) alloy plating layer.

Next, in step S120 of FIG. 1, X-rays are incident on the plurality of galvanized steel sheets, and the intensity of the plurality of fluorescent X-rays detected from each of the galvanized steel sheets is measured. In this case, the plurality of detected fluorescent X-rays may include, for example, Kα, Kβ and Lα lines for iron (Fe), and Kα, Kβ and Lα lines for zinc (Zn).

Next, by performing a wet analysis method on the plurality of galvanized steel sheets in step S130 of FIG. 1, at least one of the plating amount and the alloying degree of each of the galvanized steel sheets may be measured. The wet analysis method may apply a conventional method of chemically analyzing the iron (Fe) -zinc (Zn) alloy plating layer from the substrate.

Next, in step S140 of FIG. 1, first, for each of the plurality of galvanized steel sheets, at least one correlation between the measured fluorescence X-ray intensities and the plating amount and alloying degree according to the wet analysis method is calculated. do. And at least one of a first fluorescent X-ray that is at least one of the plurality of fluorescent X-rays highly related to the plating amount, and a second fluorescent X-ray that is at least one of the plurality of fluorescent X-rays highly related to the alloying degree. Decide on one. That is, in one embodiment, when the plating amount is to be predicted, the first fluorescent X-rays may be determined among the plurality of fluorescent X-rays. In another embodiment, when the degree of alloying is to be predicted, the second fluorescent X-rays may be determined among the plurality of fluorescent X-rays. In another embodiment, when the amount of plating and the degree of alloying are to be predicted, the first and second fluorescent X-rays may be determined among the plurality of fluorescent X-rays.

In this embodiment, as the first fluorescent X-ray, any one of the Kα, Kβ and Lα lines for the iron (Fe) is selected and any one of the Kα, Kβ and Lα lines for the zinc (Zn). Can be screened. In another embodiment, as the second fluorescent X-ray, any one of Kα, Kβ and Lα lines for iron (Fe) is selected, and any one of Kα, Kβ and Lα lines for zinc (Zn) is selected. Can be screened. In another embodiment, the first fluorescent X-rays and the second fluorescent X-rays may be selected together in the same manner as described above. In this case, the first fluorescent X-rays and the second fluorescent X-rays may be of different types.

The manner of performing step S140 may be described in more detail below. First, the method of calculating the correlation may proceed as follows. Through Table 1 below, at least the magnitude of the correlation between the measured fluorescence X-ray intensity and the coating amount according to the wet method, and the magnitude of the correlation between the measured fluorescence X-ray intensity and the alloying degree calculated according to the wet method. You can analyze one.

Fluorescence X-ray Types Detected Correlation with Plating Amount by Wet Method Correlation with Alloying Degree Calculated by Wet Method Fe Kα 0.79 0.42 Fe Kβ 0.80 0.39 Fe Lα 0.00 0.72 Zn Kα 0.80 0.43 Zn Kβ 0.80 0.42 Zn Lα 0.37 0.79

According to the results of Table 1, it can be seen that the Fe Kα, Fe Kβ, Zn Kα and Zn Kβ rays among the plurality of fluorescent X-rays have a high correlation with the plating amount, and among the plurality of fluorescent X-rays, Fe Lα and Zn Lα It can be seen that the line has a high correlation with the alloying degree.

In this manner, after the correlation is calculated, at least one of Fe Kα, Fe Kβ, Zn Kα and Zn Kβ rays having a high correlation between the plating amount and the correlation may be determined as the first fluorescent X-ray. Similarly, at least one of Fe Lα and Zn Lα rays having high correlation between the alloying degree and the correlation may be determined as the second fluorescent X-ray.

Next, referring to S150 of FIG. 1, the machine learning about at least one of the first correlation between the first fluorescent X-rays and the plating amount and the second correlation between the second fluorescent X-rays and the alloying degree ( machine learning) to derive the prediction model for at least one of the first and second correlations.

As an example, the neural network may be used to perform machine learning on at least one of the first correlation and the second correlation. The machine learning may use a variety of commercially available algorithms.

Next, referring to S160 of FIG. 1, by using the prediction model, the intensity of at least one of the first and second fluorescent X-rays measured from the desired galvanized steel sheet is measured, and the plating amount is measured from the measured intensity. And predict at least one of degree of alloying. That is, specifically, the X-rays are irradiated to the target galvanized steel sheet having no information about the amount of plating and the degree of alloying by the wet analysis method, from which at least one of the first and second fluorescent X-rays is irradiated. Detect. And, by substituting the intensity of at least one of the first and second fluorescent X-rays into the prediction model, at least one of the plating amount and the alloying degree of the target galvanized steel sheet can be predicted.

Meanwhile, at least one of the plating amount and the alloying degree predicted in step S160 and the intensity of at least one of the corresponding first and second fluorescent X-rays may be stored in a database for modeling optimization of the machine learning. .

By performing the steps including the above-described S110 to S160, it is possible to perform at least one prediction method of the plating amount and the alloying degree of the galvanized steel sheet according to an embodiment of the present invention.

2 is a graph comparing the plating amount prediction value and the plating amount measurement value derived by the prediction method according to an embodiment of the present invention. 3 is a graph comparing the alloying degree prediction value and the alloying degree quantitative analysis value derived by the prediction method according to an embodiment of the present invention.

First, referring to FIG. 2, the X axis represents a plating amount predicted through the predictive model of the present embodiment, and the Y axis represents a plating amount quantitatively analyzed by a wet analysis method. The mean square root error (RMES) of 3.69% is found to be consistent with each other. Next, referring to FIG. 3, the X axis represents the alloying degree predicted through the predictive model of the present embodiment, and the Y axis represents the alloying degree quantitatively analyzed by the wet analysis method. The mean square root error (RMES) of 0.68% is found to be consistent with each other.

As described above, according to an embodiment of the present invention, by measuring the fluorescence X-ray intensity from a predetermined galvanized steel sheet, and substituting the measured fluorescence X-ray intensity into a predictive model derived in advance according to machine learning, the zinc plating At least one of the plating amount and the alloying degree of the steel sheet can be predicted. In particular, the method of the present embodiment overcomes the disadvantages of the conventional prediction method through the regression analysis or the basic variable method when estimating at least one of the plating amount and the alloying degree of the alloyed hot-dip galvanized steel sheet. Or, reduce the prediction error between the fluorescence X-ray intensity and the degree of alloying.

In the above description, the embodiment of the present invention has been described, but various changes and modifications can be made at the level of those skilled in the art. Such changes and modifications may belong to the present invention without departing from the scope of the present invention. Therefore, the scope of the present invention will be determined by the claims described below.

Claims (5)

(a) preparing a plurality of galvanized steel sheets having at least one of plating amount and alloying degree different from each other;
(b) injecting X-rays into the plurality of galvanized steel sheets and measuring the intensity of a plurality of fluorescent X-rays detected from each of the galvanized steel sheets;
(c) performing wet analysis on the plurality of galvanized steel sheets to measure at least one of plating amount and alloying degree of each of the galvanized steel sheets;
(d) a correlation between at least one of the plurality of fluorescence X-ray intensities measured respectively and the plating amount and alloying degree according to the wet analysis method is calculated for the plurality of galvanized steel sheets, At least one of a first fluorescent X-ray that is at least one of the plurality of highly related fluorescent X-rays and a second fluorescent X-ray that is at least one of the plurality of fluorescent X-rays having a high correlation with the alloying degree Determining;
(e) performing machine learning on at least one of a first correlation between the first fluorescence X-rays and the plating amount and a second correlation between the second fluorescence X-rays and the degree of alloying, wherein Deriving a prediction model for at least one of the first and second correlations; And
(f) measuring the intensity of at least one of the first and second fluorescent X-rays from the desired galvanized steel sheet using the predictive model, and predicting at least one of plating amount and alloying degree from the measured intensity Containing
Prediction method of at least one of plating amount and alloying degree of galvanized steel sheet.
The method of claim 1,
The galvanized steel sheet is an alloyed hot dip galvanized steel sheet
Prediction method of at least one of plating amount and alloying degree of galvanized steel sheet.
The method of claim 1,
in step (b),
The plurality of fluorescent X-rays
Kα, Kβ and Lα lines for iron (Fe) and Kα, Kβ and Lα lines for zinc (Zn)
Prediction method of at least one of plating amount and alloying degree of galvanized steel sheet.
The method of claim 3, wherein
As the first fluorescent X-rays, any one of Kα, Kβ and Lα rays for iron (Fe) is selected and any one of Kα, Kβ and Lα rays for zinc (Zn) is selected.
As the second fluorescent X-ray, any one of Kα, Kβ and Lα rays for iron (Fe) is selected, and any one of Kα, Kβ and Lα rays for zinc (Zn) is selected,
The first fluorescent X-rays and the second fluorescent X-rays are different kinds
Prediction method of at least one of plating amount and alloying degree of galvanized steel sheet.
The method of claim 1,
storing the intensity of at least one of the first and second fluorescent X-rays corresponding to at least one of the plating amount and the alloying degree predicted in step (f) in a database for modeling optimization of the machine learning. More containing
Prediction method of at least one of plating amount and alloying degree of galvanized steel sheet.

Images