Abstract: To improve the accuracy and efficiency of fatigue durability simulation for automotive components, it is necessary to obtain high-fidelity load boundary conditions. This study takes an sport utility vehicle as the object and conducts multi-channel load spectrum acquisition on the reinforced pavement of a fatigue test site. Aiming at the baseline drift, singular points, and invalid small load cycles existing in the acquired signals, a systematic preprocessing procedure is proposed: first, a third-order polynomial fitting is employed to eliminate the trend term; then, singular points are identified and repaired by interpolation based on the Pauta criterion (3σ); finally, invalid small load cycles are removed according to the principle of fatigue damage equivalence (with a threshold value of k=10%). To verify the fidelity of the processed load spectrum, it is used as input to drive a multi-body dynamic model for simulation. The results show that the peak errors between the simulated outputs (including wheel center Z-direction acceleration, tie rod force, and spring displacement) and the bench test results are all less than 8%, and the correlation coefficients (CC) are all above 0.95. This study establishes a standardized load spectrum processing procedure, which can provide reliable load input for fatigue durability simulation and has clear engineering value for shortening the research and development cycle.

Key words: load spectrum acquisition; reinforced pavement; signal processing; load spectrum processing

摘要： 为提高汽车零部件疲劳耐久性仿真的精度与效率，需获取高保真的载荷边界条件，文 章以某运动型多用途汽车为对象，在疲劳试验场强化路面上进行了多通道载荷谱采集。针对 采集信号中存在的基线漂移、奇异点及无效小载荷循环，提出了一套系统性的预处理流程： 首先采用三阶多项式拟合消除趋势项，继而基于莱茵达准则（3σ）识别并插值修补奇异点， 最后依据疲劳损伤等效原则（阈值 k=10%）删除无效小载荷。为验证所处理载荷谱的保真度， 将其作为输入驱动多体动力学模型进行仿真。结果表明，仿真输出的轮心 Z 向加速度、转向 横拉杆力及弹簧位移与台架试验结果的峰值误差均不超过 8%，相关系数（CC）均在 0.95 以 上。研究形成了一套标准化的载荷谱处理流程，可为疲劳耐久性仿真提供可靠的载荷输入， 对缩短研发周期具有明确的工程价值。

关键词: 载荷谱采集；强化路面；信号处理；载荷谱处理