基于数据分类的思想,通过获取仿真数据进行高速铁路轨道不平顺峰值安全域估计,得到不同速度下的不平顺峰值安全域边界.首先,利用Simpack软件建立高速客车车辆动力学仿真模型,以2种不同水平轨道谱的不平顺作为激励,获取高低和轨向不平顺幅值数据以及脱轨系数、轮重减载率和轮轨横向力等安全性指标数据,并依据一定的安全性评判准则将幅值数据标记为“安全”或“危险”2类;其次,鉴于直接利用SVM进行不平顺幅值数据分类困难的情况,提出一种基于危险点分布比率和SVM相结合的分类方法,极大地降低了分类难度,提高了分类效率,获得了最佳分类面;最后,将试验结果与国内外轨道不平顺峰值管理标准进行比较.结果≯ 月:本文提出方法有效,可为高速铁路轨道平顺状态的管理及标准制定提供参考.%Based on the theory of data classification, the security regions of peak amplitudes of track irregularities for high-speed railways were estimated by the simulation data, and the security regions boundaries with different speeds were solved. First, by a high-speed passenger vehicle dynamics simulation model using Siropack, the inputs including vertical and lateral irregularities and outputs including three safety indicators data (derailment coefficient, wheel load reduction rate and wheel-rail lateral force) were collected at two different level track regularity states, and the irregularities data were marked as 'safe' or 'danger' according to certain safety evaluation criteria. And then, considering the difficulty of direct classification using SVM for irregularities data, a new classification method based on danger points distribution ratio and SVM was proposed which can reduce the difficulty and improve the efficiency greatly. At last, the experimental results were compared with the domestic and international management standards of track geometry. The results show that the proposed approach is effective and practicable. This method can provide references for the management and developing standards of track geometry on high-speed railways.
展开▼
