摘要
通过现有车轮型面测量仪器获得的实测型面数据中包含大量锯齿形信号和部分奇异值,为提高车轮磨耗情况分析的准确性,需对这些实测数据进行平滑处理。采用了基于二乘原理和卷积概念的3种不同平滑方法,包括平均滑动法、样条滑动法和高斯尺度变换法。对比了独立运用3种方法对车轮实测型面数据平滑处理的结果,在此基础上提出了综合运用3种方法对车轮实测型面按一定规则分段采用不同方法进行平滑处理,再将各段平滑曲线连接成完整型面。结果表明,考虑3种方法的优缺点,综合运用3种方法分段处理车轮实测型面,能使平滑误差和误差的标准差减至较小范围内,平滑速度较快,平滑效果较好,能满足车轮磨耗参数计算与运用状况了解的有关要求。
The wheel profile data got from an existing measuring instrument consists of many saw-tooth signals and some singular values.To improve the analysis accuracy on worn situation of a wheel,smoothing treatment must be done for the measured data.Three methods of data smoothing based on square principle and convolution concept are adopted,including average smoothing,spline smoothing and Gaussian scale transformation.On the basis of the results comparison from these three methods respectively,it can be used to smooth the profile in the curve segmentation form,according to certain rules.Then,the smoothed segmentations would be connected to a whole profile. The results show that the synthesized application of three methods to deal with the data in the segmentation form can decrease smoothing deviations and their standard values to a small range,with rapidly smoothing and better effect.It meets certain requirements of the calculation of wheel worn parameters and the working state of the wheel,when considering the advantages and disadvantages of three methods.
出处
《振动．测试与诊断》
EI
CSCD
北大核心
2013年第6期943-949,1090,共7页
Journal of Vibration,Measurement & Diagnosis
基金
国家自然科学基金资助项目(51175381)
关键词
车轮型面
磨耗
平滑处理
平均
样条
高斯尺度
误差
wheel profile,wear,smoothing treatment,average,spline,Gaussian scale,deviation
