基于HO-VMD优化的TCN轴承故障特征诊断模型
Fault Characteristic Diagnosis Model of TCN Bearing Based on HO-VMD Optimization

现代工业中的大中小型机械设备基本都含有旋转机械，其中轴承连接是一种非常常见的连接方式。但目前针对轴承故障的诊断方式有限，诊断的模型类别较为传统，为此特地提出一种基于HO-VMD优化的TCN轴承故障特征诊断模型，在传统频谱分析以及模态分解方法下无法正确有效的识别轴承故障的背景下，本文利用优化算法领域中较为新锐的HO优化算法(河马优化算法)解决了传统VMD分解(变分模态分解)难以确定模态分解个数与惩罚因子参数的问题，提高了故障特征的提取效率与正确率，此外通过现阶段较为新颖前沿的TCN神经网络模型算法对轴承故障类别进行识别，极大地提高了轴承故障诊断的识别效率与准确率，经过对凯斯西储大学开源轴承故障实验数据的特征提取与故障类型识别，结果表明，基于HO-VMD优化的TCN轴承故障特征诊断模型对于轴承故障诊断有着预测精度高，泛化性能高，稳定性好的特点，为轴承故障诊断提供了一种高效、精准、可靠的方法。
Modern industrial machinery of various sizes typically contains rotating components, with bearings being one of the most common connection methods. However, current bearing fault diagnosis approaches remain limited and rely mainly on traditional model types. In light of the shortcomings of conventional spectrum analysis and modal decomposition methods—which often fail to accurately detect bearing faults-this paper proposes a TCN-based bearing fault feature diagnosis model optimized by HO-VMD. Specifically, we leverage the novel HO (Hippopotamus Optimization) algorithm to tackle the challenge of determining the number of decomposed modes and penalty factors in traditional VMD (Variational Mode Decomposition), thereby improving both the efficiency and accuracy of fault feature extraction. Furthermore, by utilizing the cutting-edge TCN (Temporal Convolutional Network) to classify bearing fault types, the method significantly enhances fault diagnosis accuracy and efficiency. Experimental results on Case Western Reserve University’s open-source bearing fault dataset demonstrate that the HO-VMD-optimized TCN model achieves high prediction accuracy, strong generalization performance, and excellent stability, providing an efficient, precise, and reliable solution for bearing fault diagnosis.