论文部分内容阅读
针对液压泵压力信号呈现的非线性、非平稳的特性,提出一种将小波包分析、相空间重构理论与支持向量机(SVM)相结合的预测方法,实现液压泵压力信号监测数据的建模及预测。首先将采集到的压力信号通过小波包进行分解,将分解得到的各个分量进行重构,其次对重构后的每一个分量通过混沌支持向量机预测模型进行预测,最后对各预测值进行合成。试验数据表明,该方法能够有效地预测液压泵压力信号的变化趋势,具有较高的预测精度,可有效地应用于系统的状态监测和故障预测。
Aiming at the non-linear and non-stationary characteristics of pressure signal of hydraulic pump, a prediction method combining wavelet packet analysis, phase space reconstruction theory and support vector machine (SVM) is proposed to realize the construction of hydraulic pump pressure signal monitoring data Model and forecast. First, the collected pressure signals are decomposed by wavelet packet to reconstruct the components which are decomposed. Secondly, each component reconstructed is predicted by chaos support vector machine prediction model. Finally, the predicted values are synthesized. The experimental data show that this method can effectively forecast the trend of pressure signal of hydraulic pump and has high prediction accuracy, which can be effectively applied to the condition monitoring and fault prediction of the system.