基于自适应轻量化模型的轴承故障识别方法研究

Research on Bearing Fault Fdentification Method Based on Adaptive Lightweight Model

提出一种基于自适应轻量化卷积神经网络的轴承故障识别模型。基于凯斯西储大学轴承故障数据集,使用k折交叉验证法划分训练集与测试集,用深度可分离卷积替换第一层为宽卷积核的卷积神经网络中的标准卷积,最后使用优化后的麻雀搜索算法对模型参数进行自适应调整,以提高对轴承故障识别的准确率。实验表明,该模型对轴承故障识别的准确率和损失值优于其他一些现有方法,具有一定的实际应用价值。

Bearing fault identification model based on adaptive lightweight convolutional neural network is proposed.Based on the bearing fault data set of Case Western Reserve University,the training set and the test set are divided by k-fold cross-validation method.Replace the standard convolution with depthwise separable convolution in deep convolutional neural networks with wide first-layer kernels.Finally,the optimized Sparrow Search Algorithm was used to adapt the parameters of the model to improve the accuracy rate of bearing fault identification.Experiments show that the accuracy rate and loss value of the model for bearing fault identification are better than other existing methods,and it has practical application value.