Yuan Qiang, Peng Jinzhi, Wen Xiaofei, Liu Zhihong, Zhou Ruiping, Ye Jun
School of Naval Architecture and Maritime, Zhejiang Ocean University, Zhoushan 316022, China.
School of Naval Architecture, Ocean and Energy Power Engineering, Wuhan University of Technology, Wuhan 430070, China.
Sensors (Basel). 2025 Sep 1;25(17):5400. doi: 10.3390/s25175400.
With the continuous advancement of intelligent, integrated, and sophisticated modern marine equipment, bearing fault diagnosis faces increasingly severe technical challenges. Compared with traditional industrial environments, marine propulsion systems are characterized by multi-bearing coupled vibrations and complex operating conditions. To address these characteristics, this paper proposes a fault diagnosis method that combines a least squares support vector machine (LSSVM) with multi-domain feature extraction based on an improved hippopotamus optimization algorithm (LCM-HO). This method directly extracts time, spectral, and time-frequency domain features from the raw signal, effectively avoiding complex preprocessing and enhancing its potential for field engineering applications. Experimental verification using the Paderborn bearing dataset and a self-built marine bearing test bench demonstrates that the LCM-HO-LSSVM method achieves diagnostic accuracy rates of 99.11% and 98.00%, respectively, demonstrating significant performance improvements. This research provides a reliable, efficient, and robust technical solution for bearing fault diagnosis in complex marine environments.
随着智能、集成和精密的现代船舶设备不断进步，轴承故障诊断面临着日益严峻的技术挑战。与传统工业环境相比，船舶推进系统具有多轴承耦合振动和复杂运行工况的特点。针对这些特点，本文提出了一种基于改进的河马优化算法（LCM-HO）的最小二乘支持向量机（LSSVM）与多域特征提取相结合的故障诊断方法。该方法直接从原始信号中提取时域、频域和时频域特征，有效避免了复杂的预处理，增强了其在现场工程应用中的潜力。使用帕德博恩轴承数据集和自建的船舶轴承试验台进行的实验验证表明，LCM-HO-LSSVM方法的诊断准确率分别达到99.11%和98.00%，显示出显著的性能提升。本研究为复杂海洋环境下的轴承故障诊断提供了一种可靠、高效且稳健的技术解决方案。
