A Study On The Lightweight Method Of Automotive Door With Tailor-welded Blank Structure

The lightweight technology of automotive body is not only the emphasis of today’s research in auto industry, but also assumes further social issues ranging from energy reservation to environment protection. Presently, the technique of tailor-welded blank (TWB) applying to the structural design of automotive body is flourishing. However, the research level of this technique is tending to focus on pressing manufacture and not enough to use practically and reliably in engineering design.The thesis takes an automotive door with tailor-welded blank structure as the research object. Through the procedures of simulating the welding line, partitioning automotive door panel, and fitting to finite element analysis (FEA) with surrogate model, the door with tailor-welded blank structure achieves a favorable lightweight result. Pertinent works and conclusions are as follows:1. The thesis mainly discusses on the finite element model of welding line in the condition of front impact, analyzes the characteristics of coincident node model, rigid connection model and beam model, and compares the behaviors of energy absorption and deformation. The results indicate that different finite element model of welding line has a fairly subtle influence on the behaviors of energy absorption and deformation. Further, compared to coincident node model and rigid connection model, beam model has a more reliable and credible effect.2. A method of applying approximation model technology to simulate the static and dynamic FEA of automotive door is proposed in the thesis. From the perspective of generalization capability, the approximate efficiency of two-order response surface method (RSM), three-order RSM, four order RSM, Kriging model and artificial neural network (ANN) are explicitly compared. The research indicates that the two-order RSM is more suitable for the static response approximation like weight, stiffness and natural frequency, while the ANN method gives more advantages in the aspect of dynamic response approximation.3. The last part of the thesis puts emphasis on the weight optimization of the automotive door under the requirements of static and dynamic performance. Using the two-order RSM to approximate static requirements and the ANN method to approximate the dynamic requirements, the door structure is optimized via genetic algorithm (GA). The result shows that the research method obtains a satisfactory weight.The method proposed in the thesis has solved certain problems and difficulties from the structural design and computer aided engineering (CAE) of automotive body, put forward some innovative thinking and laid a tentative foundation of the application of TWB structure in automotive industry.