中文摘要：

利用较为完备的青藏铁路轨道不平顺实测数据,将概率论方法较为系统的用于青藏铁路轨道不平顺谱研究,对不 同波长的功率谱密度样本进行广义极值分布函数拟合、参数怙计及假设检验分析.在此基础上,对青藏铁路的线路几何状态及轨道不平顺谱线特征进行分析：采用Levenberg-Marquardt法及铁科院7参数公式对青藏铁路轨道不平顺统计平均值 谱进行公式拟合及参数提取：结合广义极值分布函数的参数怙计值,提出由轨道不平顺统计平均值谱向任意百分位谱转换 的方法.研究结果表明：青藏铁路轨道不平顺不同波长下的谱密度值均服从广义极值II型分布 （ Frechet分布）：轨道高低 和轨向不平顺谱线呈现复杂的窄带和周期波谱特征,其轨距不平顺状态基本处于德国高干扰谱和美国六级谱之间,应该作为重点关注的不平顺类型：结合不同波长谱密度的概率分布规律,建议在统计平均值谱的基础上,融合尖峰谱线特征,以便提供更为合理的动力仿真激励输入谱.

英文摘要：

Combining with relatively complete measured track irregularities on Qinghai-Tibet Railway, this paper systematically introduced probability theory into the study of track irregularity spectrum, and made an detail research on generalized extreme value distribution function fitting, parameter estimation and hypothesis testing for power spectrum density samples of different wavelengths. On this basis, the railway＇s track geometry status and characteristics of track irregularity spectrum were analyzed, and then formula fitting and parameter extraction were conducted for track irregularity statistical average spectrum using Levenberg-Marquardt algorithm and 7 pa-rameters formula of Chinese trunk railways, last we proposed a method that any percentile spectrum could be transformed by track irregularity statistical average spectrum through which. The research results show that; the spectrum density values of different wavelengths obey the generalized extremum II type distribution ; the track profile vertical and alignment irregularity spectrum present complicated narrow-bands and periodic spectral lines, and its track gauge irregularity is mostly located in German high speed high disturbance line and American sixth grade track line; combined with the probability distribution laws of different wavelength spectrum densities, thispaper suggests that we should consider and blend the peak spectrum line features on the basis of the statistical average spectrum for providing more reasonable dynamic simulation spectrum.