机械

通过车轮多边形轴箱振动响应提出一种基于形态学滤波的车轮多边形故障诊断方法。核心是通过形态学滤波算法对轴箱振动加速度信号进行降噪，通过降噪信号频谱分析确认是否存在多边形故障，并根据多边形主频计算多边形阶次。首先，根据车轮多边形仿真信号研究形态学滤波器类型、结构元素类型、结构元素尺寸对车轮多边形信号降噪的影响，并给出上述参数的选择建议；然后，通过车轮多边形仿真信号验证本文多边形故障诊断方法的有效性；最后，通过线路试验数据和轮对多边形测试验证该方法的有效性。验证结果表明，该方法可实现车轮多边形故障信号降噪，并能有效诊断出车轮多边形故障。

According to the vibration response of axle box, a fault diagnosis method based on morphological filtering is proposed to identify the polygonal wheel in this paper. The core is to denoise the vibration signal of the axle box through the morphological filtering method, then distinguish the polygon fault and calculate the polygon order through spectrum analysis. Firstly, the influence of the morphological filter type, the structure element type and the structure element size on the denoise of wheel polygon signal are discussed according to the wheel polygon simulation signal, and the suggestion on the selection of the above parameters are given.Secondly, the simulation test is conducted to verify the effectiveness of the proposed method. Finally, the line test and the wheel set polygon test are conducted to verify the effectiveness of the proposed method. The results show that the method based on morphological filtering can denoise the wheel polygon signal and diagnose the wheel polygon fault effectively.