在某SUV车型的背门模态分析中,背门1阶固有频率与发动机及路面的激振频率过于接近,这种情况会导致在日常行驶工况中车门产生共振,影响车门的NVH性能。本文中以对车门质量及1阶固有频率的灵敏度较高的几个构件的材料厚度为设计变量,采用最优拉丁方试验设计方法获得样本数据,并用响应面法构建了背门质量及1阶固有频率的近似模型,利用序列二次规划法对该两个近似模型进行多学科并行优化,在保证背门质量不增加的情况下,显著提高了背门的1阶固有频率。
The first natural frequency of rear door is found too close to the excited frequencies of the engine and pavement in the model analysis of a SUV's rear door. Resonance will be produced in daily driving under the cir- cumstances, and the NVH of the rear door will be weakened. In order to conquer the problem, the thickness of some components which have higher sensibility on rear door's first natural frequency and weight is treated as the de- sign variables. Trial sample is got through the experiment design method of optimal Latin hypercube and the multi- disciplinary approximate models are established by using the response surface method. The surrogate model is then optimized by sequential quadratic programming. As a result, the first natural frequency of the rear door is obviously improved in such conditions that the weight of the rear door don't increase.