Sensitivity Analysis And Optimization Of Static And Dynamic Performances Of Car’s Body-in-White

Bodywork holds the most weight shares of the four basic car assemblies; its performanceis closely related to the overall quality of car, such as safety, fuel economy and drivingcomfort. Furthermore, new styling, which makes the structure change, now is the highlightwhen launching a brand new car model. In this sense, automotive enterprise often takesbodywork as a focus of car develop project. The structural quality is acceptable or not, tomost extent, depends on the design of Body-in-White (BIW). Usually domestic product isheavier because strength is the criterion to consider first when it is designed. This goes againstthe demand for lightweight in car development. The application of ultra high strength steel(UHSS), can help us to balance cutting weight against assuring functions, which makes iteasier to achieve lightweight design. Static and dynamic performances are prior functions toconsider when designing BIW. Thus, optimize it without adding weight, which taken as maintopic of this paper, is a main way of lightweight design.Optimization of static and dynamic performances based on sensitivity analysis, for itsevident effect of refining the car, now is widely applied in actual developing project.Althought the method is quite acceptable, disadvantages can be found in detail process of it.The aim of this paper is to propose an improved process, which makes the method moreefficient. First, refine the FEM model of BIW which has not been frozen yet, use NASTRAN,an integrated CAE software, to estimate present performances of BIW and calculate therelative sensitivity data. Second, define3rules to classify parts of BIW. They are directsensitivity, relative sensitivity and ratio sensitivity. Sort100parts to5groups by every rule.Then parts which contribute most or least to performances can be found out. Also the featureof distribution of these parts can be summarized. Third, take research on regularity aboutchanging of sensitivity. According to this, quantify the prediction of performance changewhen any part’s thickness is modified. In this way, various solutions of optimization areconcluded and verified.The keypoint gotten in the research are listed as follows. First, the concept of ratiosensitivity is defined, which quantifies the differences between performance change andweight change caused by modification of part’s thickness. Second, a law is found out that asthickness of parts changes within a tiny range, diagram of corresponding change of sensitivityappears nearly quadric curve. Then the formula to predict performance change is deduced.Error of stiffness prediction with it is less than1%. Third, an improved process ofperformance optimization which based on variables seletion is concluded. Compared to themethod used now, it needs not iteration computation by software, can help us to concludeacceptable solution. Moreover, it becomes more flexible to select suitable parts to modify.