Robustness of the digital filter to differing calibration flows

Abstract

The use of a tubing system is a common means of measuring surface pressures for wind engineering experiments. However, it introduces signal distortions on the pressure fluctuations. This is primarily addressed before data acquisition by altering the physical tube to produce a more suitable behaviour, or post-acquisition by applying a spectral filter. More recently, digital filters have been suggested for correcting the pressure distortions, providing faster correction of data by working in the time-domain. However, the reproducibility, or stability, of the filter coefficients to differing calibration flows has not been assessed. This includes the effect of different dominant frequencies and mean pressures in the calibration signal.Digital filters were derived for a 1500 ​mm flexible tube, sampled at 400 ​Hz with a variety of input signals. These included a controlled signal from a speaker and a turbulent wind tunnel environment. It was found that the digital filter coefficients were stable to the change in input signal, indicating the tube behaviour has been properly characterised by the filter. It was also seen to reject signal noise more effectively than the equivalent spectral filter, while being more computationally-efficient in derivation and application than the latter. The effectiveness of the filter is also discussed with an assessment on the resulting mean, peak and fluctuating error.