基于LMD近似熵和PSO-ELM的齿轮箱故障诊断

发布时间：2018-03-08 19:23

本文选题：齿轮箱　切入点：局域均值分解　出处：《机械传动》2017年08期 　论文类型：期刊论文

【摘要】：针对齿轮箱使用中常见的故障检测与识别问题,考虑到齿轮箱振动响应信号非线性、非平稳的特性,提出基于局域均值分解(LMD)的近似熵和粒子群优化的极限学习机(PSO-ELM)结合的齿轮箱故障诊断方法。首先,使用LMD分解方法对齿轮箱各工况的振动信号进行分解,结合相关系数选取反映主要故障信息的前4个PF分量。利用近似熵进行定量描述,组成特征向量。最后用粒子群算法对ELM的输入权值与隐含层神经元阈值进行优化,建立PSO-ELM模型,并将近似熵特征值输入到ELM和PSO-ELM模型中,对齿轮箱不同工况进行故障识别与分类。结果表明,基于LMD近似熵和粒子群优化的ELM有更高的分类正确率,验证了该方法的可行性。
[Abstract]:In view of the common problems of fault detection and identification in the use of gearbox, considering the nonlinear and non-stationary characteristics of the vibration response signal of the gearbox, An approximate entropy method based on local mean decomposition (LMD) and particle swarm optimization (PSO -ELM) based gearbox fault diagnosis method is proposed. Firstly, LMD decomposition method is used to decompose the vibration signals of the gearbox under different working conditions. The first four PF components reflecting the main fault information are selected according to the correlation coefficient. The approximate entropy is used to quantitatively describe and form the eigenvector. Finally, particle swarm optimization is used to optimize the input weights of ELM and the threshold of hidden layer neurons. The PSO-ELM model is established, and the approximate entropy eigenvalue is input into the ELM and PSO-ELM models to identify and classify the gearbox faults under different working conditions. The results show that the ELM based on LMD approximate entropy and particle swarm optimization has higher classification accuracy. The feasibility of the method is verified.
【作者单位】：中北大学机械与动力工程学院;
【基金】：国家自然科学基金(51175480,50875247)
【分类号】：TH132.41

【相似文献】