The AISI Bar Steel Fatigue Database includes fatigue data for several medium and high carbon steels in the normalized condition. As was mentioned in an earlier posting, normalizing consists of heating the steel to a temperature near 900°C followed by air-cooling. This usually results in a microstructure consisting of ferrite and pearlite.
The normalized steels in the database exhibited hardness values from 163-259 BHN. Comparison of the fatigue properties for these steels shows that the strain-life curves fall within a rather narrow band, especially in the long life regime. Three of the steels in the database, SAE 1038, SAE 10V45 and SAE 1090, can be used to demonstrate this.
The mechanical properties and hardness values obtained for the three steels were as follows:
Steel Yield Str. Tensile Str. Red. in Area Hardness
Grade MPa Mpa % BHN
1038 330.7 582.2 54.0 163
10V45 465.2 764.6 48.0 212
1090 728.8 1090.0 14.0 259
Figure 1 shows the strain-life curves for the three normalized steels. The strain life curve for Iteration No. 18 shows the fatigue properties for SAE 1038, the strain-life curve for Iteration No.21 shows the properties for SAE 10V45, and the properties for SAE 1090 are given by the strain-life curve for Iteration No. 6. A comparison of the three curves in the long life regime shows that SAE 1090 exhibits the best performance, followed by SAE 10V45 and SAE 1038. This would be expected from the comparative hardness values. However, the curves lie in a narrow band, and the actual data points are closely clustered. Under a circumstance such as this, it becomes advantageous to consider using strain-life data to obtain a value of fatigue strength at a particular life level in order to compare fatigue performance. This has the added advantage of determining a maximum allowable stress level to achieve a given life.
Fatigue strength at a particular life level is determined by the following equation, which is a part of the strain-life equation:
Δσ/2 = σf‘(2Nf )b
Where Δσ/2 is the fatigue strength, 2Nf is life in reversals (Nf is life in cycles), σf‘ is the fatigue strength coefficient, and b is the fatigue strength exponent. For each of the steels listed above, the fatigue strength coefficient and the fatigue strength exponent are determined from the fit of the strain-life equation to the fatigue data. For these steels the values of σf‘ and b are given in the AISI Bar Steel Fatigue Database.
At a life of one million cycles (two million reversals), the fatigue strengths for the three steels shown in Figure 1 are as follows:
Steel Grade Fatigue Strength @ 106 cycles, MPa
As can be seen, the differences in fatigue performance among the three steel grades can be readily distinguished, and a value of maximum allowable stress can be established.
A study conducted by AISI showed that the fatigue strengths of normalized medium and high carbon steels can be related to hardness and chemical composition. A simple linear regression equation to predict fatigue strength was developed from the data in the AISI Bar Steel Fatigue Database, and is shown below:
Δσ/2 @ 106 cycles = 0.18BHN+53.48%C+218.27%Mn-2052.98%S-46.4%Si
Figure 2 shows a plot of predicted versus measured fatigue strength. The agreement between predicted values and measured values is quite good.
It should be noted that use of the regression equation above should be restricted to normalized steels with ferrite-pearlite microstructures, and chemical compositions and hardness values within the range of those in the AISI Bar Steel Fatigue Database. The equation should not be extrapolated to high hardness steels with martensite microstructures.
In summary, fatigue strength is a useful tool to distinguish fatigue performance among various steel grades and to establish a maximum allowable stress level to achieve a given fatigue life.