CN113868782A - Method for correcting artificial error/deviation in spring steel decarburized layer thickness measurement - Google Patents

Abstract

The invention discloses a method for correcting artificial errors/deviations in the measurement of the thickness of a decarburized layer of spring steel, which relates to the technical field of steel production, continuously calibrates observation points or test points of measurement errors or deviations caused by artificial factors after fitting correction of a large amount of data is adopted, gradually defines the boundary of a part of the decarburized layer, realizes the calibration of experimental data under the condition of accurate judgment boundary, simultaneously obtains a multiple linear regression equation according with the experimental result, and has guiding significance for controlling the thickness of the part of the decarburized layer of a product, ensuring the reliable delivery of rolled products and having excellent performance.

Description

Method for correcting artificial error/deviation in spring steel decarburized layer thickness measurement

Technical Field

The invention relates to the technical field of steel production, in particular to a method for correcting artificial errors/deviations in the measurement of the thickness of a decarburized layer of spring steel.

Background

Before hot rolling or hot forging, spring steel blanks generally need to be heated to a high temperature continuously in a non-vacuum or protective atmosphere heating furnace and then kept for a certain time (the high temperature generally needs to be heated to about 1200 ℃ according to different steel types). In the actual heating process, because the heating furnace is in a non-vacuum or protective atmosphere state, oxidation and decarburization phenomena are generated in the process of heating the steel billet to high temperature and the subsequent heat preservation process. The scale formed by oxidation can be removed by subsequent high-pressure water descaling, and the decarburized layer formed by decarburization can be remained on the surface of the billet. The decarburized layer remaining on the surface of the steel slab is generally thinned in the subsequent hot rolling or hot forging process with the reduction in the size of the rolled material, and a thin decarburized layer is finally formed on the surface of the finished rolled material. The presence of a decarburized layer on the surface of the rolled spring steel material has an extremely detrimental effect on the fatigue properties thereof. Studies have shown that spring steel bar products generally require a decarburized layer on the surface of the spring steel rolling stock of less than 0.5% of the effective thickness of the rolling stock in order to provide high fatigue life and resistance to elastic degradation. Therefore, accurate measurement of the decarburized layer is a prerequisite to ensure reliable delivery of the rolled stock and to have good properties.

The decarburized layer is generally measured by a metallographic method. The method comprises the steps of firstly, manually preparing a sample containing the decarburized layer, then observing the decarburized layer under a metallographic microscope, and finally, manually measuring the thickness of the decarburized layer by taking a metallographic photograph containing the decarburized layer. The decarburized layer can be observed under a metallographic microscope, and the thickness of the decarburized layer can be directly measured on line by using the equipped software. Regardless of the approach taken to measure the decarburized layer thickness, the boundary between the decarburized layer and the non-decarburized layer needs to be finally determined by the measurer or observer. Thus, accurate definition of this limit is a prerequisite to ensure accurate measurement of the decarburized layer thickness.

In most cases, there is a clear boundary between the decarburized layer and the non-decarburized layer, and therefore the exact definition of this boundary is generally less influenced by human factors. However, sometimes the boundary between the decarburized layer and the non-decarburized layer is not sharp, and in this case the exact definition of this boundary is greatly influenced by human factors, which leads to human errors and even deviations in the measurement of the thickness of the decarburized layer.

Since the decarburized layer is formed during the continuous heating and holding of the billet, studies have shown that the thickness of the decarburized layer is related to factors such as heating rate, heating temperature, holding time and oxygen content (heating atmosphere). Since the formation of the decarburized layer is diffusion-controlled, there must be some inherent relationship between the thickness of the decarburized layer and the heating rate, heating temperature, holding time and oxygen content. Through the establishment of this internal relationship, corrections can be made to the decarburized layer thickness measurements for those cases where the boundary definition between the decarburized layer and the non-decarburized layer is significantly affected by human factors.

The invention discloses a method for measuring depth of a decarburized layer of quenched steel, which is disclosed by the patent CN 105387824B. The method uses a wavelength dispersion spectrometer to measure the carbon element characteristic X-ray K alpha line of a cross-section metallographic specimen, and a measuring point extends from the surface layer of the cross-section metallographic specimen to a matrix. The edge distance of the sample corresponding to the transition node where the background signal and the signal intensity gradually rise is the total decarburized layer thickness, the edge distance corresponding to the transition node where the signal intensity gradually rises and the signal intensity gradually becomes stable is the total decarburized layer depth, and the difference between the total decarburized layer and the total decarburized layer is the partial decarburized layer depth. The method is suitable for measuring the depth of the decarburized layer of the quenched steel by adopting a metallographic method, and has higher precision than methods for measuring the carbon content by adopting a hardness method, a chemical method and a spectral analysis method. Although some methods are adopted in the scheme to reduce the influence of problems such as sample preparation pollution and pollution in the measurement process on the accuracy and the measurement efficiency of the measured carbon element characteristic X-ray, the scheme lacks the correction of experimental data and does not reduce the error to the maximum extent, so the accuracy is relatively poor.

The invention provides a method for calculating the thickness of a decarburized layer on the surface of decarburized bearing steel, belonging to the technical field of high-speed wire steel rolling. The method comprises the steps of firstly, calculating the theoretical equilibrium carbon concentration of the surface of a steel piece under different decarburization atmosphere conditions; then calculating the carbon content in the decarburization layer and the diffusion coefficient of carbon in ferrite under different heating temperatures and heating time conditions; and finally, calculating the thickness of the decarburized layer on the surface of the steel part according to a decarburizing model calculation formula and by combining Gaussian error function analysis. The online prediction of the thickness of the decarburized layer in the hot continuous rolling process can be carried out in real time through the furnace atmosphere parameters, the temperature parameters and the time parameters in the real-time database of the hot rolling production line, and the hot rolling process parameters are adjusted according to the online prediction, so that the aims of reducing the thickness of the decarburized layer and improving the surface quality of the bearing steel are fulfilled. Although the decarburized layer thickness is theoretically calculated, the influence of the oxide layer on the decarburized layer thickness is not considered, and a model of the oxide layer on the decarburized layer thickness is not known.

Disclosure of Invention

The invention aims at the technical problems, overcomes the defects of the prior art, provides a method for correcting the artificial error/deviation in the measurement of the thickness of a decarburized layer of spring steel,

firstly, researching the influence of heating temperature, heating speed, heat preservation time and oxygen content on the thickness of the steel decarburization layer by adopting a four-factor three-level orthogonal test design method and a metallographic method, and primarily constructing a multivariate linear regression fitting relation among the thickness of the steel decarburization layer, the heating temperature, the heating speed, the heat preservation time and the oxygen content;

and then according to the goodness of fit through correction Adj-R²Determining observation points or test points with measurement errors or deviations possibly caused by human factors, adjusting the thickness values of the decarburized layers corresponding to the observation points or the test points according to the fitting relation, and then performing multiple linear regression fitting again until the Adj-R²The thickness of the decarburized layer is more than 0.90, and the qualitative relation among the thickness of the decarburized layer, the heating speed, the heating temperature, the heat preservation time and the oxygen content does not violate the basic diffusion rule;

finally, comparing the thickness of the decarburized layer of the adjusted observation point or test point with the actual decarburized layer, redefining the boundary between the decarburized layer and the non-decarburized layer, and judging the accuracy of the redefined boundary; if accurate, the correction of errors or deviations caused by human factors can be realized; if the relationship is not accurate, the fitting regression relationship is not proper, and other relationships are needed to be used for re-correction.

The technical scheme of the invention is further defined as follows:

the method for correcting the artificial error/deviation in the measurement of the thickness of the spring steel decarburized layer specifically comprises the following steps:

(1) designing an orthogonal table comprising heat temperature, heating speed, heating and holding time and oxygen content by adopting a four-factor three-level orthogonal test design method, and carrying out an experiment on a sample;

(2) processing the samples after the experiment by adopting a metallographic method, measuring the thickness of part of the decarburized layer of each sample, and preliminarily constructing a multiple linear regression fitting relation among the thickness of the steel decarburized layer, the heating temperature, the heating speed, the heat preservation time and the oxygen content by combining the orthogonal table in the step (1) to obtain a multiple linear regression fitting equation:

in the formula: y is the predicted decarburized layer thickness, mum; a. the₀、A₁、A₂、A₃、A₄Is the coefficient of the fitting equation; x₁、X₂、X₃、X₄Respectively heating temperature, heating speed, heat preservation time and oxygen content;

(3) goodness of fit Adj-R looking at the correction²Whether 0.9 is reached and the qualitative relation between the decarburized layer thickness and the four factors does not violate the basic diffusion law,

if Adj-R²When the qualitative relation between the decarburized layer thickness and the four factors does not violate the basic diffusion rule when the decarburized layer thickness reaches 0.9, the fact that the measured partial decarburized thickness by the multi-element linear fitting equation is accurate is shown; if Adj-R after preliminary fitting²Not reaching 0.9, and observing Adj-R by adjusting the thickness data of all partial decarburized layers for multiple times²Determining observation points or test points of measurement errors or deviations possibly caused by human factors according to the gold corresponding to the sampleThe phase diagram measures the thickness of partial decarburized layers corresponding to the observation points or the test points again, and the partial decarburized layers are brought into an orthogonal table for fitting to obtain a multivariate linear regression equation after quadratic fitting;

(4) looking at Adj-R after quadratic fitting²If the qualitative relation between the decarburized layer thickness and the four factors is not against the basic diffusion rule or not, repeating the step (3) for multiple times until the Adj-R²The thickness of the decarburized layer reaches more than 0.90, and the qualitative relation between the thickness of the decarburized layer and the four factors does not violate the basic diffusion rule;

(5) finally, comparing the thickness of the decarburized layer of the adjusted observation point or test point with the actual decarburized layer, redefining the boundary between the decarburized layer and the non-decarburized layer, and judging the accuracy of the redefined boundary; if accurate, the correction of errors or deviations caused by human factors can be realized; if the relationship is not accurate, the fitting regression relationship is not proper, and other relationships are needed to be used for re-correction.

The invention has the beneficial effects that:

(1) according to the invention, through an equation of a linear regression fitting relation between the thickness of the decarburized layer of the silicon-manganese spring steel and four factors, the thickness of the decarburized layer of a rolled material part can be predicted and controlled, so that guidance is improved for process optimization, and meanwhile, the losses such as rework and the like caused by the overproof decarburized layer are reduced;

(2) the invention eliminates the problem that the boundary of a transition region is difficult to be accurately defined by artificial subjective judgment in the measurement process of a partial decarburized layer to cause measurement error/deviation, continuously calibrates observation points or test points of the measurement error or deviation caused by human factors after fitting correction of a large amount of data is adopted, gradually defines the boundary of the partial decarburized layer, realizes calibration of experimental data under the condition of accurate judgment limit, simultaneously obtains a multiple linear regression equation which accords with the experimental result, and has guiding significance for controlling the thickness of the partial decarburized layer of a product, ensuring reliable delivery of rolled products and having excellent performance.

Drawings

FIG. 1 is a data processing flow diagram of the present invention;

FIGS. 2-4 are schematic diagrams of the metallographic structures of the examples.

Detailed Description

Example 1

In the method for correcting the human error/deviation in the measurement of the thickness of the decarburized layer of spring steel provided in this example, as shown in fig. 1, the experimental steel is 60Si2Mn spring steel, and the dimension is 10 × 10 × 12 mm. A round hole with the diameter of 5mm is taken in the middle of the sample, a nickel-chromium wire penetrates through the round hole, the round hole is hung in a tubular heating furnace, and the sample is heated from room temperature by adopting different heating speeds, heat preservation temperatures, heat preservation time and decarburization atmosphere. The surface of the heat-treated test steel was polished, and the thickness of a part of the decarburized layer was measured by a metallographic method to obtain initial test data as shown in table 1:

TABLE 1 decarburized layer thickness

	X1	X2	X3	X4	Y
						Experimental group	Heating temperature (. degree.C.)	Heating speed (. degree. C./min)	Incubation time (min)	Oxygen concentration (%)	Part decarburized layer mum
1	1080	5	75	2	355.8
						2	1080	6.5	85	5	315.3
3	1080	8	95	7	224.5
						4	1100	5	85	7	285.9
5	1100	6.5	95	2	256.7
						6	1100	8	75	5	242.2
7	1120	5	95	5	350.7
						8	1120	6.5	75	7	290.6
9	1120	8	85	2	295.3

Preliminarily constructing a multi-element linear regression fitting relation among the thickness of the steel decarburized layer, the heating temperature, the heating speed, the heat preservation time and the oxygen content, and correcting the fitting goodness (Adj-R)²) =0.37939 < 0.9, and the qualitative relationship between the decarburized layer thickness and the four factors violates the basic diffusion law.

Adj-R was observed by adjusting the thickness data of all partial decarburized layers several times²Determining observation points or test points which are probably all measurement errors or deviations caused by human factors, re-measuring the thicknesses of partial decarburization layers corresponding to the observation points or the test points according to a metallographic diagram corresponding to the sample, and fitting the measurement points or the test points in an orthogonal table to obtain the thickness of the decarburization layer corresponding to the observation points or the test pointsAnd (5) performing quadratic fitting on the multiple linear regression equation. Look at Adj-R again²Still less than 0.9, after repeating the above steps several times until Adj-R²Above 0.90 and the qualitative relationship between the decarburized layer thickness and the four factors does not violate the basic diffusion law. Finally, the following fitting formula is obtained:

the corrected data are shown in table 2:

TABLE 2 decarburized layer thickness and four factors

	X1	X2	X3	X4	Y
						Experimental group	Heating temperature (. degree.C.)	Heating speed (. degree. C./min)	Incubation time (min)	Oxygen concentration (%)	Part decarburized layer mum
1	1080	5	75	2	345.8
						2	1080	6.5	85	5	275.3
3	1080	8	95	7	214.5
						4	1100	5	85	7	315.9
5	1100	6.5	95	2	286.7
						6	1100	8	75	5	212.2
7	1120	5	95	5	350.7
						8	1120	6.5	75	7	290.6
9	1120	8	85	2	265.3

(iv) fourth group of samples

The heat treatment is carried out at the temperature of 1100 ℃, the heating speed of 5 ℃/min, the heat preservation time of 85min and the oxygen concentration of 7 percent.

The initial experimental data are 285.9 mu m, the predicted value of the first fitting formula is 313.8842 mu m, 27.98421 mu m is increased compared with the initial experimental data, the actual decarburized layer thickness of the metallographic image is compared, the initial data are increased by 10 mu m to correct data, meanwhile, other data are adjusted, the obtained fitting formula still cannot meet the expected value, the correction data are selected to be reduced or continuously increased, and the step is repeated for multiple times until the Adj-R is carried out²Above 0.90 and the qualitative relationship between the decarburized layer thickness and the four factors does not violate the basic diffusion law. Finally, the following fitting formula is obtained:

goodness of fit (Adj-R) corrected by the fitting equation²) And the qualitative relation between the decarburized layer thickness and the four factors does not violate the basic diffusion rule, and the corrected thickness of the fourth group of experimental oxidation decarburized layers is improved by 30 mu m compared with that before adjustment and is similar to the actual thickness of the decarburized layer in the metallographic image.

The experimental data were successfully calibrated as shown in figure 2.

② sixth group of samples

The heat treatment is carried out at the temperature of 1100 ℃, the heating speed of 8 ℃/min, the heat preservation time of 75min and the oxygen concentration of 5 percent.

The initial experimental data are 242.2 mu m, the predicted value of the first fitting formula is 259.64298 mu m, 17.44298 mu m is increased compared with the initial experimental data, the actual decarburized layer thickness of the metallographic image is compared, the initial data are increased by 10 mu m to correct data, meanwhile, other data are adjusted, the obtained fitting formula still cannot meet the expected value, the correction data are selected to be reduced or continuously increased, and the step is repeated for multiple times until the Adj-R is carried out²Above 0.90 and the qualitative relationship between the decarburized layer thickness and the four factors does not violate the basic diffusion law. Finally, the following fitting formula is obtained:

goodness of fit (Adj-R) corrected by the fitting equation²) And the qualitative relation between the decarburized layer thickness and the four factors does not violate the basic diffusion rule, and the corrected thickness of the fourth group of experimental oxidation decarburized layers is reduced by 30 mu m compared with that before adjustment and is close to the actual thickness of the decarburized layer in the metallographic image.

The data was successfully corrected as shown in fig. 3.

Thirdly, carrying out heat treatment on the seventh group of samples at 1120 ℃, at a heating speed of 5 ℃/min, for a heat preservation time of 95min and at an oxygen concentration of 5%.

Initial compactionThe test data is 350.7μm, the predicted value of the first fitting formula is 324.37632μm, 26.3237μm is reduced compared with the initial experiment data, the actual decarburized layer thickness of the metallographic graph is compared, the initial data is improved by 10μm to correct data, meanwhile, other data are adjusted, the obtained fitting formula still cannot meet the expected value, the correction data is selected to be reduced or continuously improved, and the step is repeated for multiple times until the step is added to the step of adding to the step of R²Above 0.90 and the qualitative relationship between the decarburized layer thickness and the four factors does not violate the basic diffusion law. Finally, the following fitting formula is obtained:

goodness of fit (Adj-R) corrected by the fitting equation²) The qualitative relation between the decarburized layer thickness and the four factors does not violate the basic diffusion rule, and the corrected thickness of the fourth group of experimental oxidation decarburized layers is not changed compared with that before the adjustment and is similar to the actual thickness of the decarburized layer in the gold phase diagram.

Data were successfully corrected as shown in fig. 4.

In addition to the above embodiments, the present invention may have other embodiments. All technical solutions formed by adopting equivalent substitutions or equivalent transformations fall within the protection scope of the claims of the present invention.

Claims (2)

1. A method for correcting artificial errors/deviations in the measurement of the thickness of a decarburized layer of spring steel is characterized by comprising the following steps:

firstly, researching the influence of heating temperature, heating speed, heat preservation time and oxygen content on the thickness of the steel decarburization layer by adopting a four-factor three-level orthogonal test design method and a metallographic method, and primarily constructing a multivariate linear regression fitting relation among the thickness of the steel decarburization layer, the heating temperature, the heating speed, the heat preservation time and the oxygen content;

and then according to the goodness of fit through correction Adj-R²Determining observation points or test points of measurement errors or deviations possibly caused by human factors, and fitting the observation points or test points according to the fitting relationAdjusting the decarburized layer thickness value corresponding to the test point, and then performing multiple linear regression fitting until the Adj-R²The thickness of the decarburized layer is more than 0.90, and the qualitative relation among the thickness of the decarburized layer, the heating speed, the heating temperature, the heat preservation time and the oxygen content does not violate the basic diffusion rule;

finally, comparing the thickness of the decarburized layer of the adjusted observation point or test point with the actual decarburized layer, redefining the boundary between the decarburized layer and the non-decarburized layer, and judging the accuracy of the redefined boundary; if accurate, the correction of errors or deviations caused by human factors can be realized; if the relationship is not accurate, the fitting regression relationship is not proper, and other relationships are needed to be used for re-correction.

2. A method of correcting for human error/bias in the measurement of the thickness of a decarburized layer of spring steel according to claim 1, wherein: the method specifically comprises the following steps:

(1) designing an orthogonal table comprising heat temperature, heating speed, heating and holding time and oxygen content by adopting a four-factor three-level orthogonal test design method, and carrying out an experiment on a sample;

(2) processing the samples after the experiment by adopting a metallographic method, measuring the thickness of part of the decarburized layer of each sample, and preliminarily constructing a multiple linear regression fitting relation among the thickness of the steel decarburized layer, the heating temperature, the heating speed, the heat preservation time and the oxygen content by combining the orthogonal table in the step (1) to obtain a multiple linear regression fitting equation:

in the formula: y is the predicted decarburized layer thickness, mum; a. the₀、A₁、A₂、A₃、A₄Is the coefficient of the fitting equation; x₁、X₂、X₃、X₄Respectively heating temperature, heating speed, heat preservation time and oxygen content;

(3) goodness of fit Adj-R looking at the correction²Whether 0.9 is reached and the qualitative relation between the decarburized layer thickness and the four factors does not violate the basic diffusion law,

if Adj-R²When the qualitative relation between the decarburized layer thickness and the four factors does not violate the basic diffusion rule when the decarburized layer thickness reaches 0.9, the fact that the measured partial decarburized thickness by the multi-element linear fitting equation is accurate is shown; if Adj-R after preliminary fitting²Not reaching 0.9, and observing Adj-R by adjusting the thickness data of all partial decarburized layers for multiple times²Determining observation points or test points of measurement errors or deviations possibly caused by human factors, re-measuring the thicknesses of parts of decarburized layers corresponding to the observation points or the test points according to a metallographic diagram corresponding to the sample, and fitting the thicknesses in an orthogonal table to obtain a quadratic fitted multivariate linear regression equation;

(4) looking at Adj-R after quadratic fitting²If the qualitative relation between the decarburized layer thickness and the four factors is not against the basic diffusion rule or not, repeating the step (3) for multiple times until the Adj-R²The thickness of the decarburized layer reaches more than 0.90, and the qualitative relation between the thickness of the decarburized layer and the four factors does not violate the basic diffusion rule;

(5) finally, comparing the thickness of the decarburized layer of the adjusted observation point or test point with the actual decarburized layer, redefining the boundary between the decarburized layer and the non-decarburized layer, and judging the accuracy of the redefined boundary; if accurate, the correction of errors or deviations caused by human factors can be realized; if the relationship is not accurate, the fitting regression relationship is not proper, and other relationships are needed to be used for re-correction.

Images