针对轴向柱塞泵实际故障诊断中采集到的故障类数据远少于正常类数据的情况,为提升故障分类精确率,提出了一种基于平衡随机森林(Balanced Random Forest,BRF)的轴向柱塞泵故障诊断方法。BRF算法是随机森林(Random Forest,RF)的改进算法,...针对轴向柱塞泵实际故障诊断中采集到的故障类数据远少于正常类数据的情况,为提升故障分类精确率,提出了一种基于平衡随机森林(Balanced Random Forest,BRF)的轴向柱塞泵故障诊断方法。BRF算法是随机森林(Random Forest,RF)的改进算法,将欠采样方法与RF结合,强化了RF处理非均衡数据的能力。通过开源的UCI数据集对该算法的性能进行了测试,相较于RF以及合成少数类过采样(Synthetic Minority Over-sampling Technique,SMOTE)与RF的组合算法SMOTE-RF,BRF算法在少数类分类精确率方面有所提升。最后,将BRF算法应用于轴向柱塞泵的故障诊断中。结果表明,在类间数据不均衡的条件下,相较于RF及SMOTE-RF算法,BRF算法能够取得更高的故障分类精确率。展开更多
The chaotic motion characteristics are expounded by taking the Duffing equation system as an example.The frequency band segmentation ability and the frequency resolution of the orthogonal multiresolution analysis and ...The chaotic motion characteristics are expounded by taking the Duffing equation system as an example.The frequency band segmentation ability and the frequency resolution of the orthogonal multiresolution analysis and the orthogonal wavelet packet analysis are compared.A new orthogonal wavelet packet analysis-based chaos recognition method for chaotic motion characteristics is put forward.The chaotic,random,and periodic motions are identified effectively by use of the subfrequency band energy distribution in the signal spectrum.The characteristic frequency of chaotic motion is thus extracted.展开更多
基金study was supported by the 7th Younger Teacher Fund of Fok Ying Tung Education Foundation (No.71061).
文摘The chaotic motion characteristics are expounded by taking the Duffing equation system as an example.The frequency band segmentation ability and the frequency resolution of the orthogonal multiresolution analysis and the orthogonal wavelet packet analysis are compared.A new orthogonal wavelet packet analysis-based chaos recognition method for chaotic motion characteristics is put forward.The chaotic,random,and periodic motions are identified effectively by use of the subfrequency band energy distribution in the signal spectrum.The characteristic frequency of chaotic motion is thus extracted.