基于总体经验模态分解和连续均方误差的侵彻过载信号分析方法

Penetration Deceleration Signal Processing Method with Ensemble Empirical Mode Decomposition and Consecutive Mean Square Error

侵彻过载是攻坚武器及相关研究的重要参量。针对实测弹载侵彻过载曲线分析处理方法开展了研究,提出采用总体经验模态分解(EEMD)结合连续均方误差(CMSE)理论获取弹体刚体过载信号的方法。通过EEMD获得测试信号的本征模态函数分量,再运用CMSE理论判别高频干扰与侵彻信号的分界点,对不含分界点分量的高频分量进行抛弃处理,将其余低频信号进行重构获得弹体刚体过载信号。积分结果表明,重构信号在有效去除高频干扰的同时,完整保留了侵彻过载中弹体刚体的加速度信号。此外,整个分析过程所具有的信号自驱动特性避免了不同弹靶工况下滤波频率选择困难。

The extraction of rigid-body deceleration characteristic plays a significant role in the research of anti-hard-target weapons and related areas.In this paper,we investigated the methods of processing on-board recorded penetration deceleration data.The intrinsic mode functions were separated from the raw signals by ensemble empirical mode decomposition(EEMD),and a demarcation point between high-frequency interference functions and the projectile rigid-body acceleration signal functions was distinguished by the consecutive mean square error(CMSE)theory.By discarding the first few high-frequency components without demarcation points,the rigid-body acceleration of projectile was reconstructed with the remaining low-frequency components.The consistency of the integral results between the final curve and the original data shows that,the high-frequency interference is removed effectively and the rigid-body penetration over-load is kept completely.In addition,the difficulty of selecting the filter frequency under different target conditions in the traditional filter method is avoided with the characteristic of signal adaptive in the analysis process.