摘要
针对车身结构耐撞性能响应具有强非线性的特点,以某车身前部结构轻量化为例,在利用支持向量回归建立近似模型阶段,分析比较了3种常用的试验设计方法对模型精度的影响。结果表明,将均匀设计与支持向量回归相结合最适合耐撞性能近似模型的构造。基于所得最优近似模型进行车身前部结构的优化设计,在保证耐撞性能的前提下,达到减重2.61kg、实现6.51%轻量化的效果。
Based on the strong nonlinear characteristics of body structure crashworthiness response, the paper takes lightweight of an automotive front body structure as example and analyzes and compares the effect of three frequently-used experimental design methods on model accuracy in the approximation model phase established with support vector regression （SVR）. The results show that the combination of homogeneous design with SVR is most suitable for the building of crashworthiness approximation model. With the optimized front -body structure design based on the optimal approximation model obtained, a reduction of 2.61 kg or 6.51% of the body structure weight can be realized while crashworthiness is not compromised.
出处
《汽车技术》
北大核心
2010年第2期46-50,共5页
Automobile Technology
基金
国家自然科学基金资助项目(50875164)
关键词
车身
前部结构
轻量化
试验设计方法
支持向量回归
Body
Front -body structure
Automotive lightweight
Experimental design method
Support vector regression.
