Towards an Improved Knock Detection and Quantification using Wavelets and Entropy-based Noise Compensation

This paper deals with an application of wavelets for noise reduction for improved knock detection and quantification results in stationary industrial S.I. engines. In the first part of the paper, a two stage wavelet-based noise reduction algorithm is described in which a novel approach is considered for noise compensation using entropy of the wavelet coefficients in knock bands. In the second part, application results of the proposed method as applied to knock detection and quantification in stationary industrial engines used in pipeline gas transport industry is presented.