多传感器数据融合的车辆旋转部件振动仿真研究

Research on vehicle rotary parts vibration by multi-sensor data fusion

研究了多传感器、多测试系统采用卷积分离技术提取车辆旋转部件震动特征值的可行性。针对车辆旋转部件振动信号的特点生成仿真信号,采用限值交叉残余误差控制准则的数据融合分离方法,对混合信号进行分离,实验结果验证了准则的有效性。之后,将该方法应用于车辆旋转部件振动信号瞬态成分的提取中,实现了混合震动信号的分离。

This paper points out the feasibility of convolutive separation in the extraction of transients from vehi-cle vibration signal.Furthermore,this paper also proposes a signal separation algorithm for convolutive mixtures adopting residual cross-talking error（RCTE） threshold control criterion.According to the algorithm,a simulation is conducted to prove the efficiency of this criterion.Finally,this algorithm is successfully implemented in the extraction of transients from a gearbox vibration signal.