Frequency bands selection of the Portevin-LeChatelier effect

Portevin-LeChatelier effect (serration) exemplifies as a series of transient signals on a stress-strain curve. There are several types of serration characterized by various spectral patterns. On the other hand these types of serration occur in different, partially overlapping strain regions, what makes it impossible to separate them by a traditional Fourier analysis. In this paper the authors present the potentialities of a discrete wavelet transformation applied to time-dependent filtering of particular serration types on the basis of stress-strain curves for aluminium alloys. The effectiveness of proposed separation technique can be verified by an application of time-frequency projections of energy of analyzed characteristics obtained using a cone-shaped distribution. (C) 2007 Elsevier B.V. All rights reserved.