一种多尺度球磨机筒体振动频谱分析与建模方法

针对基于传统快速傅里叶变换获得的单尺度筒体振动频谱难以有效揭示磨机研磨机理和筒体振动信号组成,以及现有文献中经验模态分解(EMD)技术预测精度低的问题,提出了基于偏最小二乘算法的多尺度筒体振动频谱分析与建模方法.该方法首先采用经验模态分解技术将筒体振动信号分解为具有不同时间尺度的内禀模态函数(IMF),接着通过傅里叶变换获得多尺度频谱,最后采用基于偏最小二乘算法的潜变量贡献率分析和选择不同尺度频谱,并建立融合不同尺度频谱的磨机负荷参数软测量模型.采用实验球磨机的实验数据仿真验证了所提方法的有效性.

Multi-scale Shell Vibration Frequency Spectrum Analysis and Modeling Approach of Ball Mill

Single-scale shell vibration frequency spectrum cannot reflect grinding mechanism of ball mill and analyze the composition of shell vibration signal. The empirical mode decomposition (EMD) based on soft sensor methods in present literature have poor prediction accuracy. Aiming at these problems， a new multi-scale shell vibration frequency spectrum analysis and modeling approach based on the partial least squares (PLS) algorithm was proposed. Firstly， the shell vibration acceleration signal was decomposed into different time-scale intrinsic mode functions (IMF) adaptively. Then， multi-scale frequency spectrum was obtained by using fast Fourier transform to different IMFs. Finally， different scale frequency spectrum was analyzed and selected by the latent variables contribution of the PLS algorithm. In addition， mill load parameters’ soft sensor model was constructed by fusing these selected multi-scale frequency spectrum. Simulation results based on experimental data of the laboratory ball mill validate the effectiveness of the proposed method.