Control of defects in deep drawing of tailor-welded blanks for complex shape automotive panel

With the development of lightweight vehicles, tailor welded blanks (TWBs) are increasingly used in the automotive industry. Splitting and wrinkling are the main defects during the deep drawing of TWBs. A new method to control the forming defects was introduced in the forming process of TWBs in this study. The microstructure and mechanical properties of TWBs were characterized through metallography and tensile tests. Finite element modelling of an automobile rear door inner panel made of TWBs was built to analyse deep drawing. Edge cutting and notch cut were introduced in the drawing to deal with forming defects and reduce the number of stamping tools. The minimum distance between the material draw-in and trimming lines, thinning index and thickening index were defined as the measurable index to analyse the numerical results. Orthogonal experiment, numerical simulation and multiobjective experiment were utilised to optimize the forming parameters. The proposed method and optimised parameters were verified through experiments. The experimental results are basically consistent with the numerical simulation. Results demonstrate that the proposed method can provide some guidance for controlling the defects in deep drawing of TWBs for complex shape automotive panel.


Introduction
With the development of industrial technology, an increasing number of families have acquired private cars, causing heavy pressure to transportation and environment [1]. Therefore, energy saving and environmental protection attract increasing attention from the public.
In recent years, lightweight automobiles made of lightweight materials haves become the development trend, because they can save energy and reduce carbon emission [2]. The forming of Tailor-welded blanks (TWBs) is a valid method used to actualize automotive lightweight [3].
TWBs are welded from several different plates with different thickness, mechanical properties and surface coatings [4]. The use of TWBs forming technique has many advantages in the automotive industry [5,6]. They can reduce the number of parts and stamping tools. In addition, they can enhance the safety performance of vehicles and reduce vehicle weight at the same time [7].
The advantage of the stamping forming of TWBs is different mechanical properties can be obtained at different areas of the part by using the blank with different materials or thickness [8]. Compared with the conventional stamping process, the stamping forming of TWBs can effectively reduce the number of tools, welding work and assembly costs [9].
Studies of the TWBs have been a hot topic in recent years [10]. Overall, the existing reports on TWBs mainly concentrated on themicrostructure, mechanical properties, formability, forming limit diagram, failure, residual stresses, forming and applications of these materials [11][12][13][14]. Mechanical properties of TWBs rest with many factors, such as material, welding method, weld line orientation and thickness ratio of the blanks [15,16]. The uniaxial tensile test is widely used to research the mechanical properties of TWBs. Zadpoor et al. [17] utilized the monaxial tension test to obtain the mechanical properties of TWBs. Ciubotariu et al. [18] investigate the behavior and mechanical properties of the weld line in a TWBs during and after its tensile testing by using parallel tensile tests, micro-hardness tests, thermography, EDX, and microscopy. Xu et al. [19] studied the mechanical properties of TWBs with different weld line orientations by three-point bending tests. Song et al. [5] studied the influences of the thickness ratio of the base materials on the formability of TWBs. Miles et al. [20] found the welding method will affect the formability of TWBs. Rossini et al. [21] researched the mechanical properties, microstructure and failure modes of TWBs with dissimilar materials utilizing metallography, microhardness, and tensile tests. Liu et al. [22] investigated the deformation behavior and failure features of TWBs in hot forming.
Challenges have increased in the development of stamping tools by increasing part complexity and usage of TWBs to reduce the weight of a vehicle [23]. Accurate forming simulations are the driving force during the engineering phase. Accurate simulation can reduce the cost of tools manufacturing and reject rates of parts [24]. Several studies have found that the numerical simulation is an effective tool for analyzing the stamping forming process of TWBs [25][26][27]. The finite element (FE) modelling of TWBs is more complex than that of the single material plate because of the existence of the weld [28].
Two approaches are developed to deal with the weld seam for the numerical simulation of TWBs in practice [8]. The first approach is to consider the geometry and property of the weld seam in the FE modelling. The second approach is to simplify or ignore the weld seam in the FE modelling when it has higher strength than that of the base materials. In this case, failure is more likely to occur in the base materials than in the weld seam. Determining the better method depends on the types, geometry and mechanical properties of the base metal and the weld seam.
Raymond et al. [29] introduced FE modelling of TWBs using solid elements, including weld properties and geometry. Buste et al. [30] considered the weld seam of TWBs as rigid links in the FE modelling. However, few investigations have been done on the defects control in deep drawing of TWBs.
This paper aimed to study the deep drawing of TWBs for complex shape automotive panel.
A typical rear door inner panel was studied as a case in the research. The microstructure and mechanical properties of the base material and TWBs were characterised through metallography and tensile tests to achieve accurate FE modelling. Edge cutting and notch cut were introduced in the drawing process to deal with the forming defects and reduce the number of tools. Orthogonal experiment, numerical simulation and multiobjective experiment were utilised to optimise the forming parameters. The proposed method and optimised parameters were verified through experiments.

Microstructure
As shown in Fig. 1, an automobile rear door inner panel made of TWBs was researched in this article. The blanks of this part were welded by two plates which the thickness were 0.7 mm and 1.2 mm, respectively. Considering the self-weight and assembly position of the part, the area close to B-pillar and hinge needs high strength and stiffness. Thus, the thickness of the part close to B-pillar and hinge was thicker than that close to C-pillar. This feature can enhance the strength and stiffness of the automobile rear door inner panel whilst maintaining its weight.

Fig. 1 Automobile rear door inner panel
The base material of the TWBs used for this part is DC05 and belongs to cold-rolled low carbon steel for deep drawing. The chemical composition of DC05 is presented in Table 1. As shown in Fig. 2, the microstructure of TWBs was determined through metallographic microscopy. The cross-section of the welded joint is illustrated in Fig. 2a. The microstructure of the base material was mainly composed of ferrite (Fig. 2b). The heat-affected zone (HAZ) was made up primarily of bainite and ferrite (Fig. 2c), and the weld zone was made up primarily of bainite (Fig. 2d). The microstructure of the base materials, HAZ and weld seam were different. As shown in Fig. 3, the fracture surfaces of the base material and weld seam were characterised through TESCAN VEGA3 LMH scanning electron microscopy (SEM) after tensile tests. As shown in Fig. 3a, the fracture surface of the base material is numerous isometric dimples, which is ductile dimple fracture pattern. In addition, the fracture morphology of the weld seam made up of a great deal of river pattern and cleavage facets, which is typical ductile and cleavage fractures.

Mechanical properties
Accurate numerical simulation is based on reasonable material model and FE modelling.
Tensile tests were performed before the numerical simulation to determine the yield strength, The yield stress of the TWBs is 325. 16

FE models
Automotive rear doors consist of left and right doors, and they are symmetrical in shape.
Considering the uniqueness of automobile rear door inner panels, they are usually formed in a large tool simultaneously, which is called two-cavity in one tool. Based on the structure analysis of the automobile rear door inner panel, the forming process of the automobile rear door inner panel includes drawing, trimming, piercing, reshaping and separating. The first step, drawing, is performed to form the part shape, followed by trimming and piercing operations and a final step that is needed to reshape and separate the left and right door inner panels. The design of die addendum and binder surfaces of the automobile rear door inner panel is illustrated in Fig.   5. The forming defect of the similar automobile panel usually occurs during drawing. Thus, the investigation of defect control for this part focuses on the drawing stage. Numerical simulation is a useful tool to predict various forming defects in the development of stamping tools [33,34]. The initial blank was set as a triangular element with a size of 1800 mm×1270 mm and thickness of 0.7 and 1.2 mm. The weld seam was taken as the rigid links in the finite element modelling because the weld area is extremely narrow and the strength and hardness of the weld seam are larger than that of the base materials in laser welding [30].

Improvement of process design
The preliminary FE analysis of the part was conducted on the FE software. Since the part is symmetrical, the half of the part on one side of the symmetry line can show the full result.
The formability diagram of the formed part is shown in Fig. 7. As shown in Fig. 7, few defects, such as splitting and wrinkling, were found on the formed part. These defects were caused by many reasons, such as uneven material draw-in, material thinning and complex shape. For similar problems, Ouyang [37] proposed a secondary drawing process to improve the defects.
However, this method increases the drawing process time and leads to a long part production cycle. (1) Edge cutting was introduced at the beginning of the drawing process to cut the excess material of the blank. As shown in Figs. 7 and 8, edge cutting, K1a was introduced to cut the excess material of the sheet for avoiding the splitting in the area of P1. Compared with traditional process, blanking was integrated in the drawing die.
(2) Considering the characteristics of large irregular holes on the part, notch cut was introduced to solve the splitting defects near irregular holes during the drawing process. As shown in Figs  The cut-in time is extremely crucial in obtaining the qualified parts. Thus, the cut-in time of edge cutting and notch cut needs to be accurately controlled.

Orthogonal design and results
Taguchi method is a powerful and commonly used technique for the design of experiments [38,39]. Xu et al. [39] investigated the discrete optimisation design of TWB thin-walled structures using the Taguchi method. Thus, to achieve the best results of the cut-in time of edge cutting and notch cut, orthogonal experimental design and numerical simulation were used to optimise the deep drawing of TWBs. In the orthogonal experiment, the edge cutting of K1a and K1b were defined as experimental factors A, the notch cut of K2a and K2b was defined as experimental factors B, the notch cuts of K3a and K3b were defined as experimental factors C, the notch cut of K4 was defined as experimental factors D, and the blank-holder force (BHF) was defined as experimental factors E. To determine the optimal parameter combination, the experiments with five four-level factors were established using the orthogonal experimental design. The range of those factors was determined through a single-factor experiment. As presented in Table 2, the orthogonal array table can be expressed as L16 (4 5 ).  The calculation formulas of L, T1, and T2 were defined as follows: Ki L is the minimum distance between the material draw-in and trimming lines in notch cut i K . t is the thinning rate. t S refers to the area of the part that meets the requirement of the thinning rate. 0 S is the total area of the part.
The minimum distance between the material draw-in and trimming lines can be measured by showing the trimming and piercing lines in the post processing of the software. The thinning and thickening indices are calculated by highlighting the area where the thinning rate is more than 25% and the thinning rate is less than −5% in different colours. The results are summarised in Table 3.

Multiobjective optimisation
The multiobjective experiment was introduced to determine the optimal combination of the cut-in time of edge cutting, notch cut and BHF. The established multiobjective function and constraint equation is shown in function 4. The optimisation objective of L was defined as the maximum objective function, and the optimisation objective of T1 and T2 was defined as the minimum objective function. The limit inferior for L was 5 mm.  Table 5. Five group results that satisfy the multiobjective function are list in Table 5. The first solution has the highest desirability value and is determined as the optimal parameter combination. The optimal parameter combination obtained by multiobjective optimisation is A=20 mm, B=140 mm, C=145 mm, D=146 mm, and E=120×10 4 N.  As shown in Figs. 11 and 12, the optimal result was verified using the FE software. The forming limit diagram of the part after optimisation is shown in Fig. 11 and the thinning distribution diagram of the part after optimisation is shown in Fig. 12 in drawing,. The forming limit diagram clearly shows that the optimised part was fully formed, the safe area covered most of the area of the drawing part, and the wrinkling area of the part was outside the trimming line.

Fig. 11
Forming limit diagram of the part at the end of the drawing after optimization Fig. 12 Thinning distribution diagram of the part after optimization

Experimental verification
The drawing tools used in the experiment are shown in Fig. 13. The die tryout was conducted in a single-action sheet metal stamping hydraulic press using the optimised

Conclusions
1) The microstructure of the different regions of TWBs were different, which is the main factor that causes the difference in the mechanical properties of the base materials and TWBs. The tensile strength, yield strength and elongation to failure of the thin and thick materials are similar. The yield and tensile strengths of the TWBs are larger than that of the base materials, and the elongation to failure of the TWBs is lower than that of the base materials.
2) The cut-in time of edge cutting and notch cut was investigated in this study. Splitting and wrinkling may occur on the formed part when the cut-in time of edge cutting and notch cut is extremely late. The material boundary line will flow into the trimming line or the piercing line when the cut-in time of edge cutting and notch cut is extremely early.
3) The minimum distance between the material draw-in and trimming lines, thinning index and thickening index were defined as the measurable indices to analyse the numerical results, making it extremely convenient for the statistical results and subsequent optimisation analysis.
4) The orthogonal and multiobjective experiments are valuable tools for analysing and optimising multiple process parameters. The optimal parameter combination was verified through experiments. Numerical simulation and experimental results indicate that the proposed method is feasible to deal with the forming defects and reduce the number of tools and can guide actual production.

Ethics declarations
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Consent to participate
The authors declares that they consent to participate this paper.

Consent to Publish
This manuscript is approved by all authors for publication.

Authors Contributions
H. Wang designed the study, performed the research, analysed data, and wrote the paper. LZ.