An improved low-frequency noise reduction method in shock wave pressure measurement based on mode classification and recursion extraction.

摘要: Abstract The ill-posed problem of shock wave pressure (SWP) measurement comes from the influence low-frequency noise components to measured responses and leads an inaccurate result. To address this problem, improved method, referred as recursive empirical mode decomposition (REMD), is proposed for filtering noises SWP signals. By means (EMD), signal adaptively decomposed into several intrinsic functions (IMFs) without any prior information. A classification scheme firstly developed select two indexes separating useful, mixed noisy IMFs based on energy gradient ringing amplitude ratio. Then, adaptive diminishing white noise-assisted technology presented iteratively extract remaining useful damping ratio signal. final denoised result achieved by a partial reconstruction with obtained each extraction process. effectiveness method verified through series simulated real Simulated results show that REMD always produces largest SNR smallest RMSE PRD in both single situations. Furthermore, experiments PR PE sensors under different conditions also demonstrate superiority over existing approaches denoising ability integrity.