Experimental and numerical study on local mechanical properties and failure analysis of laser welded DP980 steels

Abstract

DP980 steel was welded using Nd:YAG laser with low heat input (4 m/min) and high heat input (1 m/min). The ultimate tensile strength (UTS) of the low heat input welded joint reached 99.7% of base metal, while the joint with high heat input reached 95.6%. The sub-critical heat affected zone (HAZ) was softened with obvious yielding platform, and higher heat input caused lower strength of local HAZs. The low heat input welded joint fractured at base metal, while the joint with high heat input failed at sub-critical HAZ with significantly lower fracture strain. The simulation results revealed that UTS of joint decreased with the increase of HAZ width and HAZ softening degree. Moreover, there were critical values of HAZ width and softening degree to predict the failure location of the joint. With the increase of fusion zone width, the UTS of joint increased gradually closing to that of base metal, and the joint tended to fail at base metal due to biaxial stress intensification effect.

Introduction

The environment friendly vehicles demand less fuel consumption as well as lower CO₂ emissions. The ever-increasing use of advanced high strength steel (AHSS) is a promising way of lightening car body in the automotive industry. Ferrite-martensite dual-phase (DP) steel is one of the AHSS families widely adopted for body-in-white [1], [2]. DP steels consist of martensite islands embedded in ferritic matrix. These martensite islands contribute to the strength of DP steel, while the ductility arises from ferrite [3]. The combination of martensite and ferrite offers higher initial work hardening rate along with considerably uniform elongation compared to conventional steels [4]. Welding is the mostly used joining technique in the automobile industry. Among various welding methods, laser welding is playing a vital role in the joining of AHSS due to its flexibility and high energy density [5]. The microstructures of joint are transformed locally under welding thermal cycles, and then the mechanical properties change correspondingly. The safety of vehicles is closely related to the mechanical behavior of welded joint. Thus, characterizing and understanding of local microstructures and mechanical properties of welded joints is important for modeling and predicting the overall mechanical behavior of DP welded joints [6].

Extensive studies have been conducted on the mechanical performance of fusion zone (FZ) and HAZ in terms of AHSS welds. Conventional experiment methods to characterize local properties have been reported using mini tensile samples, which are machined directly from FZ and HAZ of the joint, respectively [6], [7]. However, laser welding provides much narrower FZ and HAZ compared with conventional arc welding, making such mini tensile samples are difficult to machine because of insufficient FZ and HAZ in the joints. On the other hand, it is inaccurate for assuming a homogeneous property of HAZ using mini tensile sample method, as the HAZ microstructure varies along with the thermal gradient in real situation [8]. The tensile test based on a “rule of mixture” is another method commonly adopted to extract properties of the FZ and HAZ based on the assumption of iso-strain, of which the weld metal is parallel to the loading direction [9]. Lee et al. extracted the average mechanical properties of the weld bead and the HAZ of tailed-welded blanks via the subsize samples [10]. Abdullah et al. obtained the weld properties from tailed-welded blank using a similar approach [11]. In their study, four different sized samples were tested, and more accurate results were found using smaller samples owing to the larger proportion of weld in the cross section of sample. The tensile test combined with the “rule of mixture” is an appealing method for obtaining local properties due to its simplicity, but only the average properties of the weld or the HAZ could be acquired. As a result, the property variation of the HAZ was ignored, whereas it is important for numerical simulation.

Local constitutive behavior of HAZ could be obtained by scaling stress-strain curve according to the hardness profile, where the variation of the HAZ properties was considered [12], [13], [14]. For example, Pavlina and Van Tyne found that the tensile strengths of steels over the range of 450–2350 MPa presented a linear correlation with hardness [15]. However, Rojek et al. pointed out that this method underestimates the yield stress in the range of small plastic deformation and overestimates the yield stress in the range of large plastic deformation [16]. The newly developed digital image correlation (DIC) has been applied in measuring local strain during tensile test [17], [18], [19]. The corresponding local stress is based on the iso-stress assumption of the whole sample. Only part of the stress-strain curve can be obtained for the hardened zone, as little plastic deformation would occur there.

Thermal simulation is an effective method with precise temperature control that can reproduce a relatively large HAZ samples with homogeneous microstructure. This feature allows for conventional mechanical properties test of HAZ [20]. Goodall et al. have examined the toughness of thermal simulated HAZ of arc-welded X80 line pipe steel by Charpy impact test [21]. Dancette et al. have investigated the local microstructure and constitutive behavior of spot welds of DP steels experimentally with a Gleeble 3500 thermo-mechanical simulator [22]. However, the welding thermal cycles are difficult to measure for the thermal simulation, as the HAZ is too narrow to locate several thermocouples at different positions accurately. In this work, therefore, the thermal cycles experienced during laser welding of DP980 steel were identified using finite element (FE) analysis. The thermal cycles were then reproduced using thermo-mechanical simulator to investigate the local constitutive behaviors of the joint. The overall tensile behavior of the joint was evaluated by experiment and FE analysis. Particular attention was paid to the influence of heat input on the failure location of the joint.