A Wavelet-based Filtering Algorithm for Enhancing Signal Processing in Coriolis Flow Meters

Applying signal processing method effectively for a Coriolis flow meter (CFM) requires robust filtering strategies. This is because in actual bunkering processes, various noise components can be generated resulting in unreliable mass flow rate measurements. This study introduces a wavelet transform-based filtering algorithm to denoise and extract relevant features from non-stationary signals. It can be observed that the db6 and sym6 wavelets, with SURE and FDR thresholding, achieve the highest SNR values and lowest RMSE. The sym6 wavelet & SURE threshold exhibited good noise reduction effect and were used for further analysis. It was also found that FFT with Zero Padding paired with Wavelet Denoising & Cross Correlation yielded a percentage error of < 1% when comparing with the original simulated signal.