基于改进浣熊算法并行极限学习机机械装备故障预测
Fault Prognosis of Mechanical Equipment Using Improved Raccoon-Optimized Parallel Extreme Learning Machine

本研究提出了一种基于改进浣熊优化算法的并行极限学习机机械装备故障诊断与预测方法。针对并行极限学习机在故障诊断中存在的收敛速度慢、易陷入局部最优等问题，提出一种改进浣熊算法方案，并对并行极限学习机参数进行优化。通过多组对比实验，从诊断准确率、预测精度、收敛速度和鲁棒性等方面验证了所提出方法的优越性。实验结果表明，ICOA优化后的并行极限学习机模型在CWRU轴承数据集上的诊断准确率达到99.6%，在加入不同程度的干扰的情况下，ICOA优化后的并行极限学习具备较强的鲁棒性，显著优于传统优化方法。
This study proposes a mechanical equipment fault diagnosis and prediction method based on an improved Coati Optimization Algorithm (ICOA)-enhanced parallel extreme learning machine (PELM). To address issues such as slow convergence and susceptibility to local optima in fault diagnosis using parallel ELM, an improved Coati Optimization Algorithm scheme is introduced to optimize the parameters of the parallel ELM. Through multiple comparative experiments, the superiority of the proposed method is validated in terms of diagnostic accuracy, prediction precision, convergence speed, and robustness. Experimental results demonstrate that the ICOA-optimized parallel ELM model achieves a diagnostic accuracy of 99.6% on the CWRU bearing dataset. Furthermore, under varying levels of interference, the ICOA-optimized parallel ELM exhibits strong robustness, significantly outperforming traditional optimization methods.