JP2001065560A - Method for forecasting life of material for rolling bearing and rolling bearing having long life discriminated by forecasting life - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method to precisely forecast the life of a material for a rolling bearing in a short time and a rolling bearing to ensure discrimination of a long life through forecasting. SOLUTION: Frequency distribution of each of inclusion sizes into which a nonmetallic inclusion is classified by a kind is employed as a parameter of the degree of an internal defect of a material by which unevenness in the life of a rolling bearing is widely influenced. By using a life forecasting formula incorporating frequency distribution quantified by mathematical expression, a life of a material for a rolling bearing is high-precisely forecasted in a short time without executing a rolling life test.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for estimating the life of a material for a rolling bearing, and a rolling bearing in which a long life is identified and assured by estimating the life.

[0002]

2. Description of the Related Art In many cases, the life of a rolling bearing is determined by a peeling phenomenon occurring on the surface of a bearing material. This peeling phenomenon is a kind of fatigue failure due to rolling contact of the bearing material, and involves a strength factor that depends on the chemical composition and hardness of the material and an internal defect factor that depends on nonmetallic inclusions contained in the material. I do. The former strength factor mainly affects the level of the peel life, and the latter internal defect factor largely affects the variation of the peel life.

[0003] The internal defect factor is considered to be affected by the number of unavoidable non-metallic inclusions and the distribution of their size. Even if the same steel type, that is, the strength factor is at the same level, the rolling defect of the rolling bearing is reduced. In general, the life varied at a life ratio of 10 times or more. For this reason, when manufacturing rolling bearings, on the premise that the bearing life varies, ten or more test pieces are prepared for each lot and a rolling life test is performed.

In the rolling life test, a test piece is rolled with a mating test piece at a maximum contact stress of about several GPa, and the number of rolling cycles until the test piece is damaged by peeling or the like is investigated.
For long life products, the number of rolling cycles before breaking is 10 ⁸
There are some that exceed the order. Usually, a cylindrical or disk-shaped test piece is used, and a cylindrical or spherical test piece is used as a counterpart test piece. In this rolling life test, the life (the number of rolling cycles) at which 10% of all test pieces are damaged is defined as L10, and the bearing life of each lot is evaluated using this L10.

[0005] On the other hand, in recent years, various cleaning treatment techniques in the steel manufacturing process have been developed, and nonmetallic inclusions in the steel have been greatly reduced. For this reason, as a material for rolling bearings,
Steel materials with less nonmetallic inclusions are supplied, and long-life rolling bearings are manufactured.

[0006]

In the above-described conventional method for evaluating the life of a bearing, it is necessary to perform a rolling life test on at least 10 test pieces for each lot until each test piece is broken. There is a problem that takes a lot of time. Also,
As described above, as a rolling bearing material, a steel material having a small amount of non-metallic inclusions is used, so that the number of rolling cycles before each test piece is broken often exceeds 10 ⁸ , The time required for the rolling life test is even longer.

In order to shorten the rolling life test time, it is conceivable to increase the applied load (maximum contact stress) in the test to perform an accelerated test. In some cases, a failure phenomenon different from the failure mode in the above appears, and it may not be possible to evaluate the bearing life according to actual use.

It is an object of the present invention to provide a method for accurately estimating the life of a rolling bearing material in a short time, and to provide a rolling bearing in which a long life is identified and guaranteed by life estimation.

[0009]

In order to solve the above-mentioned problems, a method for estimating the life of a rolling bearing material according to the present invention is described below.
For each type of non-metallic inclusions, a formula Y is created that uses the frequency distribution of the size of the inclusions in a predetermined test area of the rolling bearing material as a parameter of the degree of internal defect of the material, and a prediction that includes the formula Y This adopts a method of estimating the life of the rolling bearing material by the equation.

That is, as a parameter of the degree of internal defect of a material that greatly affects the variation in the life of a rolling bearing, a frequency distribution of each inclusion size obtained by classifying nonmetallic inclusions by type is adopted, and this frequency distribution is expressed by a mathematical formula. By incorporating the quantified value into the life prediction formula, the life of the rolling bearing material can be predicted with high accuracy without a rolling life test.

The formula Y can be expressed by the following formula.

Y = exp {α ₁ · (index of influence of non-metallic inclusions of A series) + α ₂ · (index of influence of non-metallic inclusions of B + C) + β} (1) What is done can also be adopted.

Y = exp {α ₃ · (B + C-based non-metallic inclusion index) + β ₂ ｝ (2) where the classification of non-metallic inclusions is based on the JIS method k: Inclusion size (discrete variable; μm) f (k): Frequency distribution of inclusion size k (discrete distribution)
Α ₁ , α ₂ , α ₃ , β ₁ , β ₂ , γ: constant m: any integer The above-mentioned formula Y is an exponential function because of the distribution of the variation of the life of a rolling bearing due to internal defect factors. Is close to an exponential distribution represented by the Weibull distribution. In the influence index, the approximation formula f (k) of the frequency distribution is multiplied by the weight coefficient k ^{r because} inclusions having a larger size k act more as the internal defect factor.

[0014] Further, the reduction amount X of the half width in X-ray diffraction after a predetermined rolling test is used as the strength parameter of the material, and the reduction amount X of the half width is included in the above-mentioned prediction formula, whereby the rolling is performed. The influence of the strength factor is taken into the life prediction of the bearing material, and the life prediction accuracy can be further improved. Note that the rolling test for examining the reduction X of the half-value width does not need to be performed until the test piece is broken, and therefore can be completed in a much shorter time than the rolling life test.

The present inventors believe that the strength factor of the rolling bearing material depends on the structural change and the difficulty of the material change accompanying the rolling of the material. A rolling life test and a rolling test to measure the half-width of X-ray diffraction for each rolling over a certain period of time were performed on a total of six types of steel, with the conditions changed to two types of standard quenching and carbonitriding. . Table 2 shows the test conditions for the rolling life test and rolling test.
As shown in FIG.

[0016]

[Table 1]

[0017]

[Table 2]

The half width is measured at a plurality of depth positions in the vicinity of the material surface, and the measured half width at each depth position is plotted against the rolling time by approximating a power distribution. The reduction amount X of the half-value width after rolling for 2,000 minutes and 2,000 minutes was determined.

FIG. 1 shows an example of a graph in which the relationship between the reduction amount X of the half width and the life L10 obtained from the rolling life test is plotted. FIG. 1A shows the relationship between the reduction amount X of the half width at 2000 minutes after rolling on the material surface and the life L10.
(B) is a relationship between the reduction amount X of the half-value width after rolling for 1000 minutes at a position 0.1 mm deep from the material surface and the life L10. In both cases, the two show a high linear correlation, and the life L10 decreases linearly with an increase in the half value width X. Based on this result, the half value width reduction X was used as a strength parameter expressing the strength factor of the rolling bearing material.

As a form in which the reduction amount X of the half width is included in the prediction formula, the following expression incorporating the reduction amount X in a linear expression can be adopted.

Predicted life = Y · (a · X + b) (4) where a and b are constants. Further, the life prediction method of the rolling bearing material of the present invention is based on B + C non-metallic inclusions according to JIS classification. The maximum size of the inclusions in a cross-sectional area or volume corresponding to a predetermined dimension of the rolling bearing material is determined by an extreme value statistical method, based on the distribution of the inclusion sizes in the partial test area of the rolling bearing material. A method of estimating the life of the rolling bearing material based on the estimated maximum size of the inclusion can also be adopted.

In this life prediction method, the size distribution of inclusions is measured with a small area to be inspected, the maximum size of inclusions in dimensions equivalent to a bearing component is estimated using an extreme value statistical method, and the size of the inclusions is estimated. This is a simple method for estimating the life of the rolling bearing material at the estimated maximum size described above, and the prediction accuracy is slightly lower than each of the life estimation methods described above. An image analysis method or the like can be used for measuring inclusions in the area to be inspected.

The B + C non-metallic inclusions were selected as inclusions for the following reasons.

The present inventors conducted a rolling life test on about 20 steel materials having different forms and sizes of nonmetallic inclusions under the test conditions shown in Table 2 and a nonmetallic inclusion test using an image analyzer. The thing was investigated. Investigation of non-metallic inclusions by image analysis was conducted by classifying the A-type nonmetallic inclusions and B + C-type nonmetallic inclusions by binarization and fractionation processing with the sample area of the sample being 300 mm ^2. The area of each inclusion was measured. Based on this measurement result, the area Smax of the largest inclusion in the 30,000 mm ² section of each sample was estimated by the extreme value statistical method, and for the nonmetallic inclusions of each system, the Smax and the life L1 obtained from the rolling life test were determined.
The correlation with 0 was examined.

FIG. 2 (a) shows the S + of B + C-based nonmetallic inclusions.
Relationship between the square root of max (B + C) and the life L10, FIG. 2 (b)
Is the square root of Smax (A) and the life L10 of the A-based nonmetallic inclusion.
The correlation with is shown. ＋ (Sma of B + C non-metallic inclusions
x (B + C)) and the life L10 are in a proportional relationship shown in the following equation.
Show good correlation.

L10 {(Smax (A))} ^-2.16 (5) On the other hand, {(Smax (A)) of the A-based nonmetallic inclusion indicates the life L10
And a multiple regression analysis with √ (Smax (A)) added to the explanatory variable, the multiple correlation coefficient with the life L10 is high, but the degree of freedom double adjustment contribution rate is On the contrary, it gets lower. Based on this finding, a method using the estimated maximum size of B + C-based nonmetallic inclusions was adopted as a simple method for estimating the life of a rolling bearing material.

Further, the rolling bearing according to the present invention is characterized in that, for A type nonmetallic inclusions and B + C type nonmetallic inclusions according to the JIS method classification, the square root of the area of each inclusion in a predetermined test area of the rolling bearing material. The discrete frequency distributions of n (μm) are each approximated by an exponential function f (n), and the expected life L10 (lifetime at which 10% is damaged) expressed by the following formula Y ₀ is 9 × 10 ⁷ or more. A configuration using the rolling bearing material identified as above is employed.

Y₀= Exp ｛−9.52 × 10^-6・ (Influence index of A-type nonmetallic inclusions)-6.94 x 10^-Five・ (B + C-based non-metallic inclusion index) + 18.63 ＋ (6) where f (n) = d₁Exp (-d_Two・ N) (8) d₁, D_Two: Regression constant Fig. 3 shows the test area 300mm^TwoB + C non-metals in Japan
Image analysis of frequency distribution of square root n (μm) of inclusion area
The example which measured with the apparatus is shown. Square root n every 1 μm
It is a discretized variable. For example, the frequency of n = 4 μm is
The frequency of inclusions in the range of 3.5 μm ≦ n <4.5 μm is shown.
You. Measure frequency distribution of several other steel materials
However, all frequency distributions have exponential components as shown in FIG.
Since it was cloth, the frequency distribution of the square root n of the area of the inclusion was
It was approximated by an exponential function f (n). Note that the frequency of n ≧ 31 is ten
Since it is a small value per minute, there is no
I watched. In addition, the life of the conventional typical bearing steel SUJ2
The average value of L10 is 6 × 10 ⁷Degree, so the prediction
The identification value of the life L10 is set to 9 × 10 which is 1.5 times longer⁷
And

According to the rolling bearing of the present invention, for the B + C-based nonmetallic inclusions according to the JIS method classification, the square root n of the area of the inclusions in the predetermined test area of the material for the rolling bearings
(Μm) is represented by an exponential function f
(n), and the expected life L expressed by the following formula Y ₀
A configuration using a rolling bearing material identified so that 10 (lifetime at which 10% is damaged) is 9 × 10 ⁷ or more can also be adopted.

Y ₀ = exp ｛−8.94 × 10 ⁻⁵ · (influence index of B + C-based nonmetallic inclusion) +18.13｝ (9) f (n) = d ₁ · exp (−d ₂ · n) (8) d ₁ , d ₂ : regression constant The rolling bearing of the present invention is an A type nonmetallic inclusion and a B + C type nonmetallic inclusion according to the JIS classification. With respect to the object, the discretized frequency distribution of the square root n (μm) of the area of each inclusion at a predetermined test area of the rolling bearing material is approximated by an exponential function f (n), and the 2000
Minutes of the decrease in the half width of the X-ray diffraction of the material surface after the rolling test and X (°), the following equation Y ₁ expected life L10 represented by (lifetime 10% damage) of 9 × 10
A configuration using a rolling bearing material identified to be ⁷ or more can also be adopted.

Y ₁ = Y ₀ · (−2.79X + 4.01) (10) Here, Y ₀ = exp ｛−9.52 × 10 ^−6. (Influence index of A-type nonmetallic inclusion) −6.94 × 10 ^{− 5.} (B + C-based nonmetallic inclusion index) + 18.63｝ (11) f (n) = d ₁ · exp (−d ₂ · n) (13) d ₁ , d ₂ : regression constant Further, the rolling bearing of the present invention rolls for each B + C non-metallic inclusion according to the JIS classification. The maximum value Smax of the area S in the cross-sectional area or volume corresponding to the predetermined dimension of the material is determined by the extreme value statistical method based on the distribution of the area S of the inclusion in the predetermined partial test area of the bearing material. And the square root of the estimated maximum value Smax is 12.2 .mu.m.
A configuration using a rolling bearing material identified as m or less can also be employed.

11. An identification value of the square root of the maximum value Smax.
2 μm is determined based on the correlation between the square root of Smax (B + C) of the B + C-based nonmetallic inclusion and the life L10 shown in FIG. 2A, and corresponds to L10 = 9 × 10 ⁷ . I do.

[0033]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described below based on examples.

[0034]

EXAMPLES (Example 1) Bearing steel SUJ2 and SUJ2
For each of the three steels, three charges were prepared, and for the C steel shown in Table 1, one charge of the steel material was melted. Standard quenched steel material melted at each charge, SUJ2, SUJ3
For the steel materials produced by one charge (A1, B1 in Table 3) and one charge of C steel (C in Table 3), carbonitrided steels were also prepared, and rolling life tests were performed from these steel materials. In addition, a test piece for a rolling test and a sample for measuring nonmetallic inclusions were produced.

[0035]

[Table 3]

For each steel material shown in Table 3, measurement of the half-value width of X-ray diffraction on the surface of the test piece before and after the rolling test for 2000 minutes, and the area of the sample 3
A frequency distribution measurement of the nonmetallic inclusion size at 00 mm ² was performed. In addition, in order to verify the prediction accuracy of the method for estimating the life of the rolling bearing material of the present invention, a rolling life test was performed on ten test pieces for each steel material, and an actual measured value of the life L10 was obtained. The test conditions of the rolling life test and the rolling test are the same as those shown in Table 2.

The frequency distribution of the inclusion size was determined by dividing the A-based non-metallic inclusions and the B + C-based non-metallic inclusions and discretizing the square root n (μm) of the inclusion area every 1 μm.
Each of these frequency distribution of (13) is approximated to an exponential function with, (11) the formula and (12), and calculates the value of the equation Y ₀ is a parameter of an internal defect degree of the material.

On the other hand, from the measured values of the half width before and after the rolling test, the reduction amount X (°) of the half width after rolling for 2000 minutes was obtained, and the predicted value of the life L10 was calculated by the equation (10). .

Table 3 shows a comparison between the predicted value and the actually measured value of the life L10 of each steel material. The ratio between the predicted value and the measured value is 0.73 ~
1.77, and even 10
Life L1 of rolling bearing material that fluctuates at more than twice the life ratio
It can be seen that 0 can be accurately predicted.

In this embodiment, A1, B1, C in Table 3
A material obtained by carbo-nitriding each of the charged materials and a material obtained by standard quenching of the charged B1 were charged with a life L10 of 9 × 1
Identified as 0 ⁷ or used as a material for bearings and so as to produce a rolling bearing of long lifetime. (Example 2) A sample for measuring non-metallic inclusions was prepared from a steel material obtained by standard quenching of about 20 charge bearing steel SUJ2, and the size of the non-metallic inclusions at a test area of 300 mm ² of each sample was prepared by an image analyzer. Was measured. Further, as in Example 1, a rolling life test was performed under the test conditions shown in Table 2, and an actually measured life L10 was obtained in order to verify the prediction accuracy of the life prediction method.

The frequency distribution of the inclusion sizes is shown in Example 1.
Similarly to the above, it was divided into A-based nonmetallic inclusions and B + C-based nonmetallic inclusions, and the square root n (μm) of the inclusion area was discretized every 1 μm. Each measured frequency distribution is approximated to an exponential function using equation (8), and the influence index of each non-metallic inclusion determined by equation (7) is substituted into equations (6) and (9) to obtain a life span. L
The value of each formula Y ₀ was calculated as 10 predicted values.

FIG. 4 shows the life L10 obtained for each steel material.
The correlation between the predicted value and the actually measured value is shown. FIG.
FIG. 4 (b) shows the frequency of only the B + C non-metallic inclusions when the predicted value Y _{0 of the} equation (6) is adopted, which employs the frequency distributions of both the system nonmetallic inclusions and the B + C nonmetallic inclusions. The distribution is adopted.

The predicted value Y ₀ of the life L10 is in the range of 0.44 to 2.41 times the actually measured value in FIG.
In (b), the measured value is in the range of 0.39 to 2.62 times the measured value, and high prediction accuracy is obtained in each case. In addition,
When the degree of freedom double adjustment contribution rate was examined for each predicted value Y ₀ obtained by the equations (6) and (9), Y _{0 in} the equation (6) including the A-based nonmetallic inclusions contributed more. The rate was confirmed to be high and statistically significant. (Embodiment 3) For the same steel material as in Embodiment ² , B was measured by using an image analyzer on each of the test areas of 300 mm ² .
The frequency distribution of the size of the + C-based nonmetallic inclusion was measured, and the area Smax of the largest inclusion in the 30,000 mm ² section of each sample was estimated by the extreme value statistical method. The graph shown in FIG. 2A shows the square root of the area Smax of the largest inclusion,
9 is a correlation with the measured value of the life L10 obtained in the second embodiment.

In this graph, the life L10 = 9 × 10
A square root of 12.2 μm of the area Smax corresponding to ⁷ was used as a discrimination value, and a material having an estimated maximum inclusion size of 12.2 μm or less was used as a bearing material to produce a long-life rolling bearing.

The prediction accuracy of this life prediction method is described in the first embodiment.
Although it is inferior to the method shown in 2 and 2, it can be adopted as a simple method.

[0046]

As described above, the method for estimating the life of a rolling bearing material according to the present invention classifies nonmetallic inclusions by type as a parameter of the degree of internal defects of the material which largely affects the variation in the life of the rolling bearing. The frequency distribution of each inclusion size was adopted, and the life prediction formula incorporating the quantified frequency distribution was used.Therefore, rolling bearings were completed in a short time without conducting rolling life tests. The life of the application material can be accurately predicted.

Further, the reduction amount X of the half width at X-ray diffraction after the predetermined rolling test is used as the strength parameter of the material,
By including the reduced amount X of the half width in the life prediction formula, the influence of the strength factor can be taken into the life prediction of the rolling bearing material, and the life prediction accuracy can be further improved.

In the rolling bearing of the present invention, the life L10 calculated by using the above-described life prediction formulas is 9 × 10 ^7.
Since the above-mentioned materials are identified and manufactured using the identified materials, long life can be guaranteed with high accuracy.

[Brief description of the drawings]

FIGS. 1a and 1b are graphs each showing a correlation between a reduction amount of a half width at X-ray diffraction and a life L10.

FIGS. 2A and 2B are graphs each showing a correlation between an estimated maximum size of a nonmetallic inclusion and a life L10.

FIG. 3 is a graph showing an example of a frequency distribution of sizes of nonmetallic inclusions.

4A and 4B are graphs each showing a correlation between a predicted value of the life L10 obtained by the life prediction method of the second embodiment and an actually measured value.

Claims (10)

[Claims] For each type of non-metallic inclusions, a formula Y is created which uses the frequency distribution of the size of the inclusions in a predetermined test area of the rolling bearing material as a parameter of the degree of internal defect of the material. By the prediction formula including Y,
A method for estimating the life of a rolling bearing material, wherein the method predicts the life of the rolling bearing material. 2. The method for estimating the life of a rolling bearing material according to claim 1, wherein the equation Y is expressed as follows. Y = exp ｛α ₁ · (Influence index of A-based nonmetallic inclusion) + α
₂・ (B + C effect index of non-metallic inclusions) + β ₁ ｝ Here, classification of non-metallic inclusions: JIS method k: Inclusion size (discrete variable; μm) f (k): Frequency distribution of inclusion size k (discrete distribution)
Α ₁ , α ₂ , β ₁ , γ: constant m: arbitrary integer 3. The method for estimating the life of a rolling bearing material according to claim 1, wherein the formula Y is expressed as follows. Y = exp ｛α ₃ · (B + C effect index of nonmetallic inclusions)
+ Β ₂ ｝ Here, classification of non-metallic inclusions: JIS method k: Inclusion size (discrete variable; μm) f (k): Frequency distribution of inclusion size k (discrete distribution)
Α ₃ , β ₂ , γ: constant m: arbitrary integer 4. The method according to claim 1, wherein a reduction amount X of the half width in X-ray diffraction after a predetermined rolling test is used as a strength parameter of the material, and the reduction amount X of the half width is included in the prediction formula. 4. The method for predicting the life of a rolling bearing material according to any one of the above-mentioned items. 5. The method for predicting the life of a rolling bearing material according to claim 4, wherein the reduction amount X of the half width is included in the prediction formula as follows. Expected life = Y · (a · X + b) where a and b are constants 6. A cross-sectional area or volume corresponding to a predetermined dimension of a rolling bearing material based on a distribution of inclusion sizes in a partial test area of the rolling bearing material for B + C-based nonmetallic inclusions according to JIS classification. The maximum size of the inclusions in the, the extreme value statistical method is used to estimate, the estimated maximum size of the inclusions, the life of the rolling bearing material to predict the life of the rolling bearing material, Forecasting method. 7. A square root n of an area of each inclusion in a predetermined test area of a rolling bearing material for an A-based nonmetallic inclusion and a B + C-based nonmetallic inclusion according to JIS classification.
(Μm) is represented by an exponential function f
(n), and the expected life L expressed by the following formula Y ₀
A rolling bearing using a rolling bearing material identified so that 10 (lifetime at which 10% is damaged) is 9 × 10 ⁷ or more. Y ₀ = exp ｛−9.52 × 10 ⁻⁶ (influence index of A-based nonmetallic inclusions) −6.94 × 10 ⁻⁵ ((B + C-based nonmetallic inclusions) +18.63｝ f (n) = d ₁ · exp (−d ₂ · n) d ₁ , d ₂ : regression constant 8. For the B + C non-metallic inclusions classified by the JIS method, the discretized frequency distribution of the square root n (μm) of the area of the inclusions in a predetermined test area of the rolling bearing material is represented by an exponential function f. (n) and the following formula Y
A rolling bearing using a rolling bearing material identified such that a predicted life L10 (life at which 10% is broken) represented by ₀ is 9 × 10 ⁷ or more. Y ₀ = exp ｛−8.94 × 10 ⁻⁵ (influence index of B + C-based nonmetallic inclusion) +18.13 に where f (n) = d ₁ · exp (−d ₂ · n) d ₁ , d ₂ : regression constant 9. A square root n of an area of each inclusion in a predetermined test area of a rolling bearing material for an A-based nonmetallic inclusion and a B + C-based nonmetallic inclusion according to JIS classification.
(Μm) is represented by an exponential function f
with approximated by (n) predicted, the amount of decrease in half-width of X-ray diffraction of the material surface after the rolling test 2000 minutes of this material as X (°), which is represented by formula Y ₁ below A rolling bearing using a rolling bearing material identified so that its life L10 (life at which 10% is damaged) is 9 × 10 ⁷ or more. Y ₁ = exp ｛−9.52 × 10 ^−6. (Influence index of A type non-metallic inclusions) −6.94 × 10 ^−5. (Influence index of B + C type non-metallic inclusions) +18.63 Δ ・ (−2.79 × + 4 .01) where f (n) = d ₁ · exp (−d ₂ · n) d ₁ , d ₂ : regression constant 10. For each B + C-based non-metallic inclusion according to the JIS method classification, based on the distribution of the area S of the inclusion in a predetermined partial test area of the rolling bearing material, a cut corresponding to a predetermined size of the material. The area S in area or volume
The maximum value Smax of the rolling bearing is estimated by an extreme value statistical method, and the square root of the estimated maximum value Smax is identified as 12.2 μm or less.