A kind of casting aluminium alloy fatigue performance study

The test material is al-SI cast aluminum alloy, the German brand is AlSi9Cu3, and its chemical composition is shown in Table 1.

Table 1: Chemical composition % of cast aluminum alloy AlSi9Cu3

According to GB/T228 2002, the static tensile test of the sample was carried out on the Zwick Electronic Universal Material testing machine, and the basic mechanical properties of the material were obtained (see Table 2).

Table 2: Mechanical properties of cast aluminum alloy AlSi9Cu3

疲劳试样从某转向器管柱经过线切割得到，转向器管柱以及试样截取位置见图1，试样形状及相关尺寸见图2。为了消除试样试验区的机械加工痕迹，得到光滑试样，采用400、800、1000、2000号砂纸对试验区进行打磨抛光，并对试样的棱角边进行磨圆处理。

The P-S-N curve of cast aluminum alloy AlSi9Cu3 was determined by "group method". The test instrument was RUMUL high frequency resonance fatigue testing machine. The test was carried out at room temperature under symmetrical cyclic tension and compression load, stress ratio R=-1, test frequency 65 Hz. The stress level was divided into 5 grades, which were 170, 160, 145, 135 and 127 MPa respectively. The number of cycles and related anomalies during the fracture of each sample were recorded in the test.

1. Fatigue test results

Due to the great dispersion of the fatigue data, the fatigue life of the sample is closely related to the survival rate P under the same stress level. Therefore, it is necessary to know the relationship between fatigue life stress of materials under certain reliability, that is, the P-S-N curve of materials used. It is generally considered that the fatigue life conforms to the lognormal distribution, i.e. By calculating the mean logarithmic life and standard deviation of the sample, the normal probability density function representing the distribution property of the parent population can be estimated, as shown in Equation (1). According to Equation (2), the logarithmic fatigue life values corresponding to different reliability levels are calculated. The logarithmic fatigue life values corresponding to the same reliability and different stress levels can be obtained by performing the same calculation process for the test data at each stress level, and the P-S-N curve can be plotted by fitting these values. The life of R99, R90 and R50 of the cast aluminum alloy AlSi9Cu3 is shown in Table 3, and the P-S-N curve is shown in Figure 3.

Table 3: Life of cast aluminum alloy AlSi9Cu3 under different reliability

Basquin equation is used to fit each data point in the semi-logarithmic coordinate system, as shown in Equation (3). Among them, 127 MPa stress level data is only used when fitting R50. Because the sample life standard deviation is large at 135 MPa stress level, the correlation coefficient of R99 fitting curve is low.

In the type, b for fatigue strength index; Is the fatigue strength coefficient.

Figure 3: P-S-N curve of cast aluminum alloy AlSi9Cu3

2. Observation of fatigue fracture morphology

Typical samples and abnormal life samples were selected for scanning electron microscope observation at each stress level, and the fracture of the samples was cleaned with ultrasonic cleaning machine and anhydrous ethanol. It was observed that the fatigue fracture of the sample consisted of the fatigue source region, the fatigue crack growth region and the transient fracture region, as shown in Figure 4. The fatigue source region is the region where fatigue crack initiation occurs, which is generally located on the surface or subsurface of the sample, because the surface stress is usually high and is greatly affected by the external environment. If there is a serious discontinuity defect within the material, the fatigue source may also arise within the material.

In the stability of crack extension stage, microscopic feature of fracture surface is the most important is "fatigue stripe", as shown in figure 5. Fatigue strip is a series of basically parallel stripes which are perpendicular to the direction of local crack propagation. With the further crack propagation, the bearing area of the sample decreases gradually. When the crack expands to the critical size, the sample suddenly breaks and forms a transient fault zone, which is relatively rough and similar to the fracture morphology of the static failure of the material. Compared with the low stress level, when the sample is subjected to the high stress level, the fatigue spreading area is smaller and the transient fault area is larger, as shown in FIG. 6.

FIG. 6:16 MPa and 127 MPa fracture fatigue zone and transient fracture zone size comparison

3. Influence of casting defects on fatigue life

It was found in the test that with the decrease of stress level, the dispersion of fatigue life data increased. Under 160 MPa stress, the dispersion of fatigue life was the lowest, with a life standard deviation of 0.238; under 135 MPa, the dispersion of fatigue life was the highest, with a life standard deviation of 0.436. Under high stress level, the fatigue source of the sample is easy to be generated, and the fatigue life is mainly the crack propagation life, so the influence of casting defect on the life dispersion is relatively small. At low stress level, the fatigue life of the sample is mainly reflected in the initiation life of the crack source. If there are casting defects inside the sample, these defects will immediately become the crack source, resulting in abnormal fatigue life, thus increasing the dispersion of the life. Various types of casting defects were found in the observation of fatigue fracture of samples, as shown in FIG. 7. Through observation, the fatigue cracks of most samples originate from the surface or subsurface. The fatigue crack of only one sample originated from the inside and expanded outwards, showing the "fish-eye" feature, as shown in FIG. 7A. The fatigue life of the sample is not only related to the type of defect, but also related to the distance from the surface. Generally speaking, the farther the sample is from the surface, the higher its fatigue life will be. The fatigue life of this sample is 508,000 times, which is higher than the life of most samples in this group. Under the stress level of 135 MPa, the sample life is only about 3000, which belongs to the abnormal data. Combined with the scanning electron microscope image (see Figure 7b), it is confirmed that this is caused by a large area of oxide inclusion. Such a large area of oxide inclusion would not be generated directly from the inside of the liquid aluminum, but should be formed by the oxide film on the surface of the liquid aluminum being sucked into the inside of the casting during the pouring process. FIG. 7C and FIG. 7D are samples of the same grade. The longest diagonal dimension of the defect in FIG. 7C is 0.365 mm, and the fatigue life is 105,000; the longest diagonal dimension of the defect in FIG. 7D is 0.09 mm, and the fatigue life is 240,000, 2.29 times that of the sample in FIG. 7C. This indicates that the larger the defect size, the lower the fatigue life.

Figure 7: Multiple types of crack sources

Because the aluminum alloy is easy to absorb hydrogen and oxidized in the melting process, the oxide film on the surface is easy to be involved in the aluminum liquid, which leads to the casting defects in the casting process. Common casting defects include oxidized inclusions and voids. Pores are divided into two types: pores and shrinkage cavity. Due to irregular shape of shrinkage cavity, the microscopic stress concentration caused by shrinkage cavity is greater, which has more significant influence on fatigue life. The oxide inclusions inside the casting parts are not closely bound to the matrix, which breaks the continuity of the matrix and does not match the elastic modulus and thermal shrinkage coefficient between the matrix. Under the action of alternating stress, these defects act as a gap and produce a large local stress concentration around them, which becomes the starting point of crack initiation. The existence of these defects greatly reduces the initiation time of fatigue cracks and leads to a significant reduction in fatigue life. The uncertain casting defects in each sample lead to an increase in fatigue life dispersion. Miao Guolei et al. studied the dispersion characteristics of high-temperature fatigue life of powder metallurgy nickel-based superalloy FGH96, and drew similar conclusions. They believed that the inclusion leading to surface crack initiation was the leading factor of the worst fatigue life, making the fatigue life dispersion greater.

(1) through the test the redirector AlSi9Cu3 casting aluminum alloy material of high cycle fatigue performance, according to the test results of fitting out of the P - s-n curve, 99% reliability fatigue life under stated one million times of fatigue strength is 120.7 MPa, test results meet the requirements.

(2) The main cause of fatigue failure of the sample is casting defects such as holes and oxidized inclusions. The existence of defects greatly reduces the initiation time of crack sources, leading to a significant reduction in fatigue life.

(3) Casting defects cause an increase in fatigue life dispersion, and the type, size and position of defects will have a great impact on sample life and dispersion. At high stress level, defects have little influence on life dispersion. At low stress level, defects have the greatest influence on life dispersion.

From: Special Casting and Nonferrous Alloys, vol. 39, No. 11, 2019