基于AR和PSO_SVR的故障趋势预测

Fault trend prediction based on AR and PSO_SVR

摘要:

针对机械设备产生的非线性、非平稳时间序列,首先使用自回归模型对非平稳数据进行平稳化处理并确定模型的阶数,再使用支持向量回归算法对平稳后的数据进行拟合,并使用粒子群算法优化支持向量回归算法参数。最后,将该模型用于滚动轴承的退化趋势预测,通过提取滚动轴承的时域和频域特征,以经过主成分析降维后的数据为基础进行趋势预测。将该模型预测的结果与单独使用自回归模型和支持向量机模型预测的结果进行对比,实验结果表明该模型预测的效果较好。

Abstract:

Aiming at the nonlinear and non-stationary time series generated by mechanical equipment, the AR algorithm is used to smooth the non-stationary data and determine the order of the model .Then, the SVR algorithm is used to fit the stationary data, and the PSO Algorithm to optimize SVR algorithm parameters. The model is used to predict the degradation trend of rolling bearings. Firstly, the time domain and frequency domain characteristics of the bearings are extracted, and then the data are predicted based on the data after PCA dimension reduction.. Finally, the results of this model are compared with those of AR and PSO_SVR alone. The experimental results show that the model is better.