The draft gear is a key component in a railway train system. The stochastic feature of dry friction in the draft gear has been observed by many researchers, however, characterisation studies in this regard have not been published. This paper used a state-of-the-art laboratory testing rig to measured draft gear forces and deflection responses to sinusoidal force inputs. Measured data were then inversed into values for Coefficient of Friction (CoF). The inverse method was developed based on a friction-wedge-spring draft gear model. The results show that CoF values have a range from 0.22 to 0.47; however, the majority of values are located above 0.34. Normal distribution fitting of the CoF values show that they have a mean of 0.42 and a deviation of 0.02. Weibull distribution fitting shows that they have a scale parameter of 0.43 and a shape parameter of 22.48. The normal distribution stochastic friction was used to model a friction draft gear. A simulation was conducted, and the results show that simulations using the stochastic model can better match the measured data than using conventional deterministic friction.

牵引装置是铁路列车系统中的关键部件。许多研究人员已经观察到牵引装置中干摩擦的随机特征，然而，这方面的表征研究尚未发表。本文使用最先进的实验室测试台来测量牵引齿轮力和对正弦力输入的偏转响应。然后将测量数据反转为摩擦系数 (CoF) 的值。反演方法是基于摩擦-楔形-弹簧缓冲器模型开发的。结果表明，CoF 值在 0.22 到 0.47 之间；但是，大多数值都位于 0.34 以上。CoF 值的正态分布拟合表明它们具有 0.42 的平均值和 0.02 的偏差。Weibull 分布拟合表明它们的尺度参数为 0。43 和 22.48 的形状参数。正态分布随机摩擦用于模拟摩擦缓冲器。进行了模拟，结果表明使用随机模型的模拟比使用常规确定性摩擦可以更好地匹配测量数据。