基于改进小波阈值的自行火炮信号降噪方法研究

Research on Signal Denoising Method of Self-propelled Gun Based on Improved Wavelet Threshold

为有效滤除自行火炮柴油发动机振动信号中的噪声,提出基于改进小波阈值的振动信号降噪方法。运用改进的自适应噪声完备集成经验模态分解处理原始振动信号得到各个本征模态函数分量,通过多尺度排列熵检测分量的随机性,筛选出需要降噪的分量,使用改进的小波阈值降噪方法对筛选出的分量降噪,重构降噪后的分量与无需降噪的分量,获得所需的振动信号。同时,针对人工选取多尺度排列熵中各参数对计算结果影响较大的问题,提出一种改进麻雀搜索算法对多尺度排列熵中各参数进行寻优。分别通过仿真信号和实验室实测数据验证所提方法的可行性和有效性,结果表明:与小波阈值降噪、多小波相邻系数降噪和ICEEMDAN-MPE-小波阈值降噪方法相比,所提方法应用于仿真信号时信噪比分别提升5.989 4 dB、6.078 7 dB和1.565 3 dB;应用于实验室实测数据时,降噪误差比分别降低22.143 3、6.834 9和0.722 7,为自行火炮振动信号降噪提供一种新的思路。

To effectively filter out the noise in the vibration signal of self-propelled gun diesel engine,a vibration signal denoising method based on improved wavelet threshold was proposed.Firstly,the original vibration signals were processed by an improved complete ensemble empirical mode decomposition with adaptive noise to obtain the intrinsic mode function components.Secondly,the randomness of the components was measured by multiscale permutation entropy to screen out components that need to be denoised.These selected components were denoised based on an improved wavelet threshold denoi-sing method.Finally,the required vibration signals were obtained by reconstructing denoised components and components without denoising.Meanwhile,in order to solve the problem that parameters in the multiscale permutation entropy manually selected have a significant influence on the calculation results,an improved sparrow search algorithm was proposed to optimize the parameters in the multiscale permutation entropy.The feasibility and effectiveness of the proposed method was verified by simulated signals and measured data respectively.The results show that the proposed method can improve the signal-to-noise ratio by 5.9894 dB,6.0787 dB and 1.5653 dB respectively compared with wavelet threshold denoising,multiwavelet neighboring coefficient denoising and ICEEMDAN-MPE-wavelet threshold denoising when applied to simulated signal,and reduces the denoising error ratio by 22.1433,6.8349 and 0.7227 respectively when applied to measured data.This method provides a new vibration signal denoising idea for self-propelled gun.