异常载荷剔除中幅值与梯度门限的联合应用

Combined Application of Amplitude Threshold and Gradient Threshold on Abnormal Load Omitting

幅值门限法对于大误差及超数量级的异常数据具有良好的识别性,而梯度门限法对于小幅值的奇异点检测能力强。针对这一特点,分析比较了幅值门限法与梯度门限法在剔除异常载荷数据过程中的优劣特性,提出联合应用这两种方式来实现对奇异信号的处理,并以装载机传动系为例给出了异常载荷数据剔除的过程及结果,避免了基于假定样本数据服从某种典型概率分布,利用标准差剔除异常数据的方法存在的固有局限性,可以快速、准确地剔除不同概率分布下的样本异常数据。

Amplitude threshold method has good recognition of large errors and ultra magnitude of the abnormal data in load history,while the gradient threshold method is effective in detecting the abnormal signals with small amplitude.Considering this characteristic,a method is provided to process the abnormal signals with the combined application of amplitude threshold method and gradient threshold method,which is followed by the case of processing the load history of the transmission.Being efficient in eliminating the abnormal signals in load history with various probability contributions,this method avoids the inherent limitations of assuming the sample data subjecting to a typical probability distribution in existed method.Hence,this method can eliminate the abnormal signals accurately.