Repair welding is an important remediation process for castings with slight defects. In this paper, the tensile behaviors of the QT400-18 nodular cast iron with different repair welding sizes were experimentally analyzed. Specimens with different diameters of the filler region were prepared by the same welding process. The fracture initiated in the filler region under uniaxial tensile loading. The modulus, strength, and ductility decreased with the weld diameter increase. The postyield hardening phenomenon was not observed in the repaired specimen. The repair region ratio was defined as the proportion of the repair welding area to the cross-sectional area of the structure. The effective modulus of the repaired specimens decreased with the repair region ratio increase, and the relationship between them was fitted by a negative exponential function. The repair welding region was treated as an inclusion in the matrix of castings, and the volume fraction of inclusion was applied to characterize the repair welding size. Based on the theories of Eshelby tensor and Mori–Tanaka equivalent method, a method for estimating macroscopic effective modulus of repair welding castings was established. The theoretical solutions were in good agreement with the experimental results. The method will be helpful in estimating the safe service limit of repair welded castings.
Some foundry minor defects, such as gas pores, shrinkage porosity, and misrun, exist in large castings more or less [
The equivalent mechanical behaviors of repair welds have always been a research focus in foundry industry. For some materials, such as cast Ti-6Al-4V and Titanium castings, minor weld repair was demonstrated acceptable for creep [
However, the situation also depends on the material, loading conditions, and welding process. The weld defects were introduced in the vicinity of the weldment by an unqualified welding process accidentally, and fatigue cracks initiated from these defects. The microcracks at the welded junctions evolved into subsequent macrocracks in the tensile or cyclic loading [
In this paper, the influence of repair welding on the effective stiffness performance was investigated. Tensile tests for QT400-18 nodular cast iron specimens with different repair welding sizes were carried out. The relations between the effective modulus and the repaired size were discussed. A prediction method of effective modulus for repair welded casting was proposed. This work is helpful to validate repair welded casting structures.
QT400-18 (ISO: 400-18) nodular cast iron was applied as the base material in this study. This material is often made for the cylinder block, cover, crankshaft, etc. in heavy diesel engines. Z308 (corresponding to the AWS standard ENi-CI, alkaline-coated electrode, and HRC : HB150) was used for repair welding on QT400-18 nodular cast iron. The diameter of the Z308 electrode cored wires was 3.2 mm. The chemical compositions (wt.%) of the two materials are presented in Table
Chemical compositions of QT400-18 and Z308 (wt.%).
C | Si | Mn | S | P | Mg | Ni | Fe | |
---|---|---|---|---|---|---|---|---|
QT400-18 | 3.61 | 2.56 | 0.47 | 0.018 | 0.047 | 0.045 | ||
Z308 | ≤2.0 | ≤2.5 | ≤1.0 | ≤0.03 | ≥90 | ≤8 |
For the QT400-18 material, foundry defects often existed in the terminal stage of solidification, and some of these defects in casting were exposed after some machining processes. In practice, the whole portion of shrinkage porosity (if any) was hollowed out from the base material manually, and then the repair welding process was carried out to fill in the hollowed parts. Usually, the perforated defect was more harmful to the structure; special care should be paid to the perforated forms of foundry defects. Therefore, in this study, the perforated welding specimen was manufactured.
Furthermore, in order to investigate the influence of repairing size on effective mechanical performance, a certain amount of repair welding specimens in the same process batch were required, but the repairing sizes were different.
The specimens with different repair welding sizes were prepared by the following procedures: As shown in Figure The rectangular castings were hollowed by two symmetric frustum cones at the center site (in the top and the bottom surface, respectively). The angle of the cone was 60°, and the maximal diameter of cones was 40 mm. The two cones should not penetrate the ingot, as shown in Figure The hollowed cone was polished to eliminate the influence of impurities, to ensure that there were no cracks or other defects. Also, residual stress was relieved through the method of hammering in the sides of the hollowed region. Temperature gradient may result in cold cracks; therefore, local preheating was carried out on the hollowed parts of the ingot before welding, and the welding electrode was also preheated. The welding processes of “layers and intervals” were carried out; that is, a small amount of new consumables were added upon the previous layer, and then the new filled metal was solidified with hammering. The above processes were repeated until the hollowed cone was fulfilled. Manual argon-arc welding was carried out to fill the hollowed regions. The mode of welding power supply was DC, and the welding current was set to 60 A. Considering that residual stress may be introduced in the case of the large volume welding, the short arc welding process was carried out. In this method, the two symmetric frustum cones were fulfilled by several overlaps rather than once entirely welded, and the thickness of each overlap was about 3 mm. After the previous overlap was entirely solidified, in this pause, the residual stress was released by hammering the weld bond. The repaired rectangular castings were cut into slices in the transverse direction after the repair region completely solidified in room temperature, and then the repaired welding area was located in the center of every plate (Figure Each plate was then made into the flat specimen. The geometry of the specimen with a repaired weld is shown in Figure
Schematic diagram of repair welding specimen preparation.
Shape and dimensions of the repaired welding specimen (unit: mm).
The size of the repaired filler region was measured by the mean diameter as follows:
Measurement of the repaired filler region on the specimen (side view).
In order to compare the performance before and after repair welding, intact specimens (without defect or welding repair) of the QT400-18 material were prepared with the same dimensions. Uniaxial tensile tests were carried out on the Instron 8802 servo hydraulic universal machine at room temperature. The displacement loading rate was 2 mm/min. The strain data of the gauge length were obtained by an MTS extensometer. The nominal stress was defined as the loading divided by the original cross-sectional area of the specimen.
The stress-strain responses of the intact specimen and specimen with the welding diameter of 30 mm under uniaxial tensile loading are shown in Figure
Stress-strain responses of the intact specimen and specimen with the welding diameter of 30 mm.
The phenomenon of the fracture initiating in the filler region was observed in all specimens (Figure
Fracture initiation from the filler region in uniaxial tensile loading.
Figure
Fracture behaviors of the intact specimen and the specimens with different weld diameters.
The trends of elastic modulus, tensile stress, and failure strain with weld diameters are shown in Figure
Effective modulus (a), failure strain (b), and failure stress (c) of specimens with different welding diameters.
In this paper, the welding repair region ratio
The fitted model (
The experimental data and predictions by formulas (
Relation of effective modulus of the welded specimen with the repair region ratio.
The effective modulus of repaired cast iron decreased with the repaired region ratio according to the rule of negative exponent.
In order to study the equivalent model of the effective modulus for repair welding casting, a representative volume element (RVE) model was abstracted from the welding-repaired casting, as shown in Figure
Schematic diagram of the RVE: the castings contain repair weld.
The RVE model was deemed as an infinite elastic body containing inclusions, and its elastic field was quantified by the Eshelby equivalent inclusion theory [
As shown in Figure
The linear elastic relation was satisfied in the condition of small deformation. Since the RVE undertook a volume average stress
For the assumption of the homogeneous stress boundary condition, the relation between the volume average stress and the strain can be generalized mathematically as
An elastic solution can be obtained for a single inclusion volume of
The stress and strain of the inclusion were different from those of the base material. The differences were
Simplifying equation (
In the representative volume shown in Figure
Solving equations (
Comprehensively solving equations (
The relationship between the equivalent eigenstrain
The effective properties which were represented by the effective stiffness matrix
According to the Mori–Tanaka method, equation (
Substituting formula (
The macro-stress-strain properties of nodular cast iron can be deduced by the equivalent inclusion method. The dimensions of
The volume fraction of inclusion was one of the important factors affecting the equivalent mechanical properties of castings. The repair welding size can be respected by the welding volume fraction. Here, the material properties of matrix and inclusion in the RVE element were the same as those of the test. The volume fractions of the specimens in the tension tests were 0.02, 0.05, 0.1, 0.2, 0.4, 0.6, 0.7, and 0.7854. Substituting these values into the tensor matrix
Comparison of experimental data and the prediction results of the RVE model.
Besides the volume fraction, the filler material property was another factor that influenced the effective mechanical property. So, for the same model, while the property of the filler material varied, the effective mechanical property was different. The elastic modulus of the filler material with five grades was analyzed, the modulus of the base metal remained unchanged, and the modulus of the filler metal was selected as 100 GPa, 120 GPa, 140 GPa, 160 GPa, and 169 GPa (its value depends on the filler material). Figure
Influence of the filler material property and weld volume fraction on equivalent elastic modulus.
The higher material property of the filler material results in a slower decrease rate of equivalent elastic modulus in a specified volume fraction. Therefore, when selecting electrode materials for repair welding in engineering, in order to reduce the influence of the material property and welding volume fraction on the filler material, the filler material with parameters similar to the base material parameters should be selected.
Fracture initiated in the filler region in uniaxial tensile loading, and the modulus, strength, and ductility tended to decrease with the weld diameter increase. The postyield hardening phenomenon of nondefected cast iron disappeared after repair welding. The effective modulus of repaired cast iron decreased with the repaired region ratio according to the rule of negative exponent. Based on Eshelby equivalent inclusion theory, a model for estimating macroscopic effective modulus of repair welding castings was established. Comparing with the experimental results, the predictions were in good agreement with the experimental results. The volume fraction of repairing and the filler material property were the two major factors for the effective modulus of repaired casting. The effective modulus was closer to the modulus of the base material when the welding volume fraction was smaller. The filler material with properties similar to base material properties should be selected as far as possible to restore the effective modulus of repaired casting.
The data used to support the findings of this study are available from the corresponding author upon request.
The authors declare that they have no conflicts of interest.
This project was supported by the National Natural Science Foundation of China (Grant no. 51875460) and the Aviation Power Foundation (Grant no. 6141B090320).