Detection and characterization of blade–vortex interaction noise with wavelet analysis

A discrete implementation of the wavelet transform (WT) was used to analyze blade–vortex interaction (BVI) noise in acoustic helicopter noise signals. A BVI detection algorithm was developed which takes advantage of the prominence of BVI noise in certain subbands. Isolated-BVI signals were constructed by detecting impulsive BVI events and extracting them from the subbands. Four metrics were tested for suitability as BVI estimators. Two of these, the rms levels of the isolated-BVI signals and a metric computed from the A-weighted frequency spectra of the isolated-BVI signals, were seen to characterize BVI noise severity, but were computationally expensive. The other two metrics were computed from the amplitudes of the BVI events in the subbands. The first, based on the exponential behavior of the amplitudes across subbands, failed to characterize BVI noise severity for certain signals and was highly sensitive to noise. The second, computed from the amplitudes of BVI events in a single subband, was an effective BVI estimator and was relatively insensitive to noise. This estimator is especially convenient because it is easily used as a decision threshold in the detection algorithm, thereby making possible simultaneous BVI detection and characterization.