首页 / 年会论文 / 正文
基于PCA的操稳客观指标降维分析及调校应用
Dimensional Reduction Analysis and Tuning Application of Objective Indicators of Handling Stability Based on PCA
2023-03-01 09:37
鞠道杰 荣兵 石向南 (Ju Daojie,Rong Bing,Shi Xiangnan)
为了研究各零部件参数对整车操稳性能的影响及指导实车调校,针对能反映整车操稳性能的客观评价指标,建立与主观评价相关的客观评价指标体系,搭建整车多体动力学模型,并与测试数据进行对比验证,参考样车调校清单开展所有调校方案仿真分析,采用主成分分析法,对所有样本仿真结果进行对比分析。研究结果表明,稳态性能、瞬态性能及转向响应特性是整车操稳第一主成分性能,占比达59.70%,应用该技术手段可将原始数据由36维降至5维,同时通过降维后的客观评价指标完成所有调校方案的主成分综合评价,确定方案40等前10组最优调校方案,从而指导后期实车调校,降低整个调校周期,提升工作效率。
To study the influence of various component parameters on vehicle handling performance and guide the tuning of the actual vehicle,aimed at objective evaluation indicators that could reflect the handing stability of the whole vehicle, establishing the objective evaluation index system related to the subjective evaluation. A multi-body dynamics model of the vehicle was built, comparing and verifying with the test data, and all schemes were carried out with reference to the tuning list of the prototype. Using the principal component analysis (PCA) method,the simulation results of all samples were compared and analyzed. Results show that the steady-state performance, transient performance and steering response characteristics are the first main component performance of the vehicle's handing stability, accounting for 59.70%. The application of this technology can reduce the original data from 36 dimensions to 5 dimensions, and at the same time, the principal component comprehensive evaluation of all tuning schemes can be completed through the objective evaluation indexes after dimension reduction, and the top 10 groups of optimal tuning schemes such as scheme 40 can be determined, so as to guide the later actual vehicle tuning, reducing the entire tuning cycle and improving work efficiency.
来源:数字仿真论坛