Research And Application On Sheet Metal Of BIW Welding Assembly Variation Analysis

The Automotive covering parts is composed of thin sheet metal parts and theassembly way is resistance spot welding. The error accumulation of vehiclemanufacture is mainly concentrated in the body welding assembly stage. The weldingassembly process of Body-in-white needs to go through the complex engineeringprocess from the component to the sub-assembly and then to assembly. Due to thecomplexity of the structure and hierarchical relationships between the parts, weldingassembly process will inevitably produce certain deviation in size, cumulative withother error in the subsequent station, and finally have a direct impact on themanufacturing process, production quality and vehicle performance. Currentinternational market environment pay more attention to the body quality, cost andother integrated capabilities and competitiveness. In order to improve the autocompanies from time to time and establish efficient, precise manufacturing qualitymanagement system, the way that manage the welding size deviation to achieve finequantified body welding can be used in the design phase.However, this method isneither able to consider the flexibility of the sheet metal parts nor consider therespective deviation of the welding source. So it is can not meet the requirements ofhighly efficient and precise. Thus, it is necessary to carry out a more in-depth researchwork for this technique.This paper improves random sampling of direct Monte Carlo method tosimulate the assembly error and uses uniform design sampling instead of simplerandom sampling during the analysis of sheet metal assembly error. This samplingmethod is not only to ensure the uniformity of sampling but also significantly reducethe sampling frequency, increasing the efficiency of the convergence, and thus be ableto be calculated in a more complex model. This paper puts forward the samplinguniformity coefficient on the basis of uniform design sampling. This sampling methodis used to control the sheet metal parts with boundary would not be out of boundsbecause of random sampling in selecting an locating point, as well as can effectivelyavoid the distance between locating points and fixture points is too close inconsidering the location of the fixture points.In order to get the deformation error of non-computing point, the paperconstructs the response surface model of the parts assembly error. In order to improve the accuracy and efficiency of the response surface, this paper uses a uniform designand selects the deviation of deviation sources according to the test table. So that it notonly reduces the number of tests but also ensure the accuracy of the model. Thismethod has a high degree of universality and practicality which can be widely used inthe error analysis of sheet metal parts assembly.On the basis of finite element analysis theory, this paper establish sheet metalassembly model by considering the parts under external force, as well as the effectsthat position of the locating point takes on assembly errors. This research use theinfluence coefficient method to construct the deviation of the source assembly errorcoefficient matrix to affect the absolute value of the coefficient of the minimum as theobjective function as a variable to locate the point to create a multi-objectivemathematical optimization model. It put forward new sampling methods whichconsider the new border with clamping point to improve the initial point of theuniform and adopt genetic algorithm to optimize the influence coefficient. At last useone of the car body sheet metal parts assemblies to prove the validity of the method.