1. Introduction
Resistance spot welding is a process that offers high productivity owing to its short processing time and is used in more than 70 % of car bodies1-3). To assess the weld quality of resistance spot welds, performance measures such as maximum load, failure mode, and nugget diameter are inspected. While destructive testing is essential to evaluate these parameters, performing destructive testing on automotive body structures is not only difficult but also costly and time-consuming.
In recent years, there has been an increasing demand for environmental regulations on reducing exhaust gas emissions. In the automotive industry, reduction of exhaust gas emissions and increase in fuel efficiency can be achieved by reducing the weight of vehicles4). Modern automotive bodies use materials such as aluminum, high strength steel, and carbon-fiber reinforced plastics (CFRP)5). High strength steel, which exhibits high strength relative to its small thickness, has recently been widely applied to the automotive industry due to the demand for stability and weight reduction of vehicles6). However, problems such as poor weldability arise in high strength steel as various alloying elements are added to the material to provide enhanced mechanical properties7,8). Unlike conventional steel, high-strength steel has a narrow range of welding conditions due to expulsion that occurs even at relatively low currents. Moreover, even when the nugget size becomes sufficiently large, interfacial failure occurs more frequently than pull out failure9-11).
Studies investigating the mechanical properties of newly developed steel as well as low-strength steel have been carried out continuously. Zhao et al. conducted tensile shear tests on resistance spot-welded joints of DP600, examined the effect of electrode force on the tensile shear load and failure energy absorption, and investigated the failure mode and microstructure12). The study concluded that the weld nugget size was affected by the failure mode and failure energy. Zhang et al. performed resistance spot welding on DP780 and DP600 and investigated the microstructure and mechanical properties of single-lap joints between the two materials13). The failure mode was shown to be primarily dependent on the nugget size, and a critical nugget diameter associated with the transition from interfacial failure to pull out failure was identified. Nikoosohbat et al. carried out resistance spot welding on DP980, a material with higher strength compared to that of materials used in the aforementioned studies, and investigated the mechanical properties, microstructure, and failure behavior14). The study reported that, in pull out failure mode, failure generally occurred at the interface of the base material (BM) and the heat-affected zone (HAZ) where softening took place. Furthermore, the study suggested that a new weld quality assurance was necessary for high strength steel, as the weld size recommendation of 4√t was not sufficient to ensure the pull out failure for DP980. Many other studies have been conducted on high strength steels that exhibit higher sensitivity to interfacial failure than pull out failure15-17).
Research on predicting the weld quality of resistance spot welds without involving destructive testing has also continued. Yang et al. developed a computational model to predict the performance of DP590 resistance spot welds18). This model, which consisted of a welding process model and a local model, predicted the residual stress and plastic deformation of the spot welds. Radakovic and Tumuluru conducted resistance spot welding on DP590, DP780, and TRIP780, and finite element modeling (FEM) was performed to predict the failure modes19). The plastic strain and predicted failure loads associated with the failure modes were calculated based on the analysis, and the equations derived from FEM showed that the force required to result in pull out failure is proportional to not only tensile strength but also the steel thickness and nugget diameter. In addition to predicting the weld quality of resistance spot welds using analysis modeling software, the tensile shear strength, nugget diameter, failure mode, and expulsion occurrence were predicted using the external image of the spot weld and a convolution neural network (CNN), which is one of the deep learning algorithms20). However, analysis predicted through computer simulations and methods using artificial intelligence require long computational times. Moreover, predictive methods using artificial intelligence require large volumes of input data, such that a full factorial experiment must be preceded. The maximum strength of the material used in the work described earlier was 980 MPa, and only one type of material was studied in general. Furthermore, the model for weld quality prediction was also performed on one type of material or did not include materials ranging from low strength steel to high strength steel.
In this work, resistance spot welding was conducted on materials of varying strength from low strength steel to advanced high strength steel to investigate the mechanical properties of spot welds. In addition, the nugget size and hardness of spot welds were predicted by using the chemical composition of materials to evaluate the weld quality without destroying the weld joints. The calculated values were subsequently used to predict the failure modes, which were compared to actual failure modes.
2. Materials and experimental methods
2.1 Materials
In this study, resistance spot welding was conducted on 440, 590, 780, 980, and 1180 MPa-grade steels with a thickness of 1 mm. The steel grade was used to denote each material, and the chemical composition is listed in Table 1.
Table 1
Chemical composition (wt. %) of the material used in this study
2.2 Experimental methods
A medium frequency direct current (MFDC) resistance spot welder was used as the welding power source, and a schematic of the shape of the electrode used in this work is displayed in Fig. 1(a). A dome-shaped welding electrode was used, and the electrode tip diameter and tip radius were 6 mm and 40 mm, respectively (Fig. 1(a)). Welding current ranging from 4 kA to 12 kA was used, while the electrode force and welding time were respectively fixed at 3 kN and 333 ms (Table 2). To evaluate the welding quality of the joint upon resistance spot welding, the tensile shear strength, nugget size, and hardness were measured. Tensile shear testing was conducted using AG-300kNXPlus (Shimazu, Japan) at a strain rate of 10 mm/min. The tensile test specimen was prepared in accordance with KS B 0851 (Specimen Dimensions and Procedure for Shear Testing Resistance Spot and Embossed Projection Welded Joints) by overlapping 100 mm-long and 30 mm-wide materials by 30 mm and performing spot welding the joint (Fig. 1(b)). Hardness measurements were conducted with HM-101 (Mitutoyo, Japan) using an indentation load of 200 g. The weld nugget size was measured using an optical microscope after cross-sectioning and etching the center of the nugget. Failure modes were determined from the results of the tensile shear tests.
3. Results and discussion
3.1 Mechanical properties of resistance spot welds
Fig. 2 and 3 respectively show the tensile shear strength and nugget size with respect to the change in welding current. Both tensile shear strength and nugget size exhibited increasing trends with increasing current, and the tensile shear strength generally leveled off at currents higher than 9 kA. Except in the 1180 MPa-grade steel, the nugget size also continued to increase up to a current of 9 kA, beyond which the nugget size showed a small increase or approached a steady value. The correlation between nugget size and tensile shear strength is presented in Fig. 4, which demonstrates that the slopes are different in all steel grades, but the tensile shear strength increases linearly as the nugget size increases.
Tensile shear tests showed that failure modes of interfacial failure (IF) and pull out failure (PF) occurred, and the failure modes of the spot welds of each material with respect to current are summarized in Table 3. PF occurred at a lower current range of 5 kA in the 440 MPa- grade steel, at 6 kA in 590, 780, and 1180 MPa-grade steels, and 7 kA at the 980 MPa-grade steel.
Table 3
Failure mode of welds with steel grade according to welding current
Fig. 5 shows the hardness profile of each steel grade. To accurately capture the representative hardness value of the heat-affected zone (HAZ), the average hardness of the HAZ was used when HAZ softening did not occur, while the minimum hardness was used when HAZ softening was present in the HAZ. The hardness values of the base material (BM) of each steel grade were 157.1, 212.2, 263.8, 326.2, and 395.3 HV, and the hardness increased as the BM strength increased. The hardness values of each steel grade at the fusion zone (FZ) were 407.4, 424.2, 414.2, 421.9, and 451.9 HV, where the hardness was the lowest in the 440 MPa-grade steel, highest in 1180 MPa-grade steel, and similar across 590, 780, and 980 MPa-grade steels. HAZ softening, in which the hardness of the HAZ is lower than that of the BM, occurred in 780, 980, and 1180 MPa-grade steels, where the lowest hardness values were 254.8, 316.2, and 311.6 HV, respectively. The difference between the hardness of the BM and the minimum hardness of the HAZ in 780, 980, and 1180 MPa-grade steels was 9, 10, and 84 HV, respectively, which was the greatest in the 1180 MPa-steel.
3.2 Microstructure characteristics of resistance spot welds
Fig. 6 shows the microstructure of resistance spot welds. The microstructure of resistance spot welds can be divided into three types: BM, HAZ, and FZ. Here, the HAZ can be subdivided into three zones based on the distance between the HAZ and FZ: upper critical HAZ (UCHAZ) that is closest to the FZ, sub critical HAZ (SCHAZ) that is closest to the BM, and inter critical HAZ (ICHAZ) that is between the UCHAZ and SCHAZ. HAZ and FZ from the 440 MPa-grade steel are presented as representative images of the microstructure of HAZ and FZ.
Investigation on the BM region of each steel grade revealed that the proportion of martensite increased and the size of grains decreased with increasing steel strength. During resistance spot welding, the FZ is fully melted first and transforms into austenite. After this process, the cooling water flowing through the electrode leads to rapid cooling to form martensite with high hardness values21). The microstructure of the HAZ changes based on the highest temperature reached during resistance spot welding. The UCHAZ, which is subjected to a maximum temperature above Ac3 and undergoes austenite transformation, consists of coarsened grains formed during rapid cooling and is composed of martensite. The ICHAZ, which experiences a maximum temperature between Ac1 and Ac3 where partial austenite transformation occurs, is comprised of martensite and ferrite with fine grains formed during quenching22). Finally, the SCHAZ undergoes a maximum temperature below Ac1, and martensite tempering occurs if martensite is present in the BM22). Furthermore, martensite tempering in the SCHAZ leads to the decrease in hardness in the HAZ. Based on this phenomenon, the reason for the decrease in hardness in the HAZ of 780, 980, and 1180 MPa-grade steels can be confirmed.
3.3 Hardness prediction using carbon equivalent (CE) equations
As described in the introduction section, the aim of this work is to predict the failure modes and hardness of resistance spot welds using the chemical composition of the materials. To achieve this objective, the relationship between the hardness and strength of the material was established, as shown in equation (1).
Here, σ is the ultimate strength of the material, HV is the hardness of the material, and c is a constant. For metals, a c value of 1/3 is commonly used23). The strength calculated through equation (1) from the hardness values of steel grades studied in this work is listed in Table 4. As shown in Table 4, since the relative error between the strength values calculated from the hardness of the materials and actual strength values of the steel grades was all within 10 %, the strength of the BM was estimated using equation (1).
Table 4
Calculated strength using equation (1) and relative error
| Steel grade (MPa) | 440 | 590 | 780 | 980 | 1180 |
|---|---|---|---|---|---|
| Calculated strength (MPa) | 471.2 | 636.7 | 791.3 | 978.5 | 1185.9 |
| Relative error (%) | 7.08 | 7.92 | 1.45 | 0.15 | 0.50 |
To estimate the hardness of the FZ, carbon equivalent (CE) equations were used. Equations (2)-(8) show the different CE equations that are commonly known. The element in each equation varies, and the coefficient associated with each element also differs. In all equations, 1 was universally used as the coefficient of the carbon element.
When CE was calculated using the equations above, the strength of the material and CE exhibited a linear relationship, as shown in Fig. 7. A regression equation was derived from the calculated CE to predict the hardness of the FZ and the minimum hardness of the HAZ. The regression equation and coefficient of determination are presented in Table 5.
Fig. 7
Carbon equivalent (CE) according to various CE equations of materials with various steel grades
Table 5
Regression equation and coefficient of determination (R2) for predicting hardness of the FZ and HAZ using CE
CE equations with the highest coefficient of determination for predicting the hardness of the FZ and HAZ were kaizu and dearden equations, respectively. Therefore, kaizu and dearden equations were respectively used to predict the hardness of the FZ and HAZ, and the estimated hardness values and relative error are presented in Table 6. In the FZ, the relative error was the lowest in the 440 MPa-grade steel and was within 1.7 % in all other steel grades, indicating that the predicted and measured hardness values were similar. In the HAZ, the relative error was the lowest in the 980 MPa-grade steel at 1.82 % and the largest in the 440 MPa-grade steel at 13.44 %, and the overall relative error in the HAZ was greater than that in the FZ hardness prediction.
Table 6
Measured and predicted hardness of FZ and HAZ with relative error
3.4 Failure mode prediction
The load applied to resistance spot welds during tensile loading was calculated to predict the failure modes of the spot welds subjected to tensile tests. As shown in Fig. 8(a), the nugget and HAZ regions of the resistance spot weld were assumed as a cylindrical shape. When PF occurred in this experiment, it was assumed that normal stress was applied to half of the cylindrical region of a material because the weld joint became separated at one of the two materials (Fig. 8(b)). When IF occurred, it was assumed that shear stress was applied to the cross-sectional area of the spot weld due to fracture across the FZ (Fig. 8(c)). Considering the relevant region and stress during PF and IF, the load in each failure mode was calculated using equations (9-10) and indicated as PPF and PIF, respectively.
Fig. 8
Schematic images of (a) resistance spot welds, (b) pull out failure mode and (c) interfacial failure mode with tensile load direction
Here, d is the nugget diameter and t is the thickness of the material. When HAZ softening occurred, the HAZ diameter as opposed to the nugget diameter was used as d. HAZ diameter measurements on the spot welds of all materials were larger than the nugget diameter by approximately 2 mm, such that the HAZ diameter was assumed as d + 2. σPFL and τFZ are the normal stress at the failure location and shear stress in the FZ, respectively. The hardness at the failure location, hardness of the FZ, and equation (1) were used to calculate the stress. In general, PF occurred at the BM or HAZ region. In this experiment, assuming that the failure location of 780, 980, and 1180 MPa-grade steels in which HAZ softening occurred was the HAZ and that the fracture of 440 and 590 MPa-grade steels without HAZ softening occurred at the BM, the minimum hardness of the HAZ and the hardness of the BM were respectively used as the hardness for calculating σPFL. On the other hand, only shear stress was present at spot welds when IF occurred, such that the hardness of the FZ was used to calculate the normal stress, which was converted to shear stress using the von Mises criterion shown in equation (11).
Using σ of the BM and the estimated hardness of the FZ and HAZ in equations (9-10), PPF and PIF at each welding current value were calculated and compared (Table 7). When PPF is greater than PIF, it is assumed that IF occurs due to the insufficient driving force for the fracture to propagate from the failure location where PF would take place (gray). In the opposite case, it is assumed that PF occurs (dark gray). Comparison with the actual experimental results shown in Table 3 showed that the experimental and predicted results were consistent across all regions except at 5 kA in the 590 MPa-grade steel, 5 kA in the 780 MPa-grade steel, and 6 kA in the 980 MPa-grade steel. The difference between the loads PPF and PIF at 5 kA in the 590 MPa-, 5 kA in the 780 MPa-, and 6 kA in the 980 MPa-grade steels was 2.36, 0.25, and 0.79 kN, respectively. Since the same hardness is used when calculating PPF and PIF at each welding current, the nugget diameter plays a dominant role in determining PPF and PIF. During this step, it is possible that an error arises during the process of measuring the nugget diameter, and this may be the source of discrepancy between the predicted and experimental results. The measurement error associated with the nugget diameter can be minimized by processing the center of the nugget during cross-sectioning of the weld nugget or by increasing the number of samples measured.
Table 7
Calculated load of pull-out and interfacial failure mode according to welding currents
4. Conclusion
In this work, resistance spot welding was conducted on 440 MPa-grade steel (low strength steel) to 1180 MPa-grade steel (high strength steel), followed by investigation on the mechanical properties and prediction on the failure modes. To evaluate the mechanical properties, tensile testing was performed, and the nugget size and hardness were measured. The hardness of the FZ and HAZ were estimated using the chemical composition of the materials, and failure modes were predicted by comparing the calculated load occurring at the spot welds.
1) As the welding current increased, the tensile shear strength increased and eventually approached a steady value beyond a certain current value. Similar to tensile strength, the nugget diameter also increased and leveled off with increasing welding current. Based on these findings, it was confirmed that the nugget diameter has a significant effect on the strength of spot welds.
2) Hardness measurement results showed that the hardness of the BM increased in the order of 440, 590, 780, 980, and1180 MPa-grade steel, while that of the FZ was the lowest in the 440 MPa-grade steel, highest in the 1180 MPa-grade steel, and similar across 590, 780, and 980 MPa-grade steels. In addition, HAZ softening occurred in 780, 980, and 1180 MPa-grade steels.
3) Various CE equations were used to calculate the CE of each material. Kaizu and Dearden equations were respectively used to estimate the hardness of the FZ and HAZ. The predicted hardness results were shown to be similar to the experimental values.
4) Using the estimated hardness values, the load occurring at the region surrounding the spot welds during the application of tensile load was calculated for each failure mode. The calculated PPF and PIF were compared to predict the failure mode, and these results were similar to those from actual experiments.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
