Falling sheet envelope method for non-destructive testing time-dependent signals

In this work, an algorithm to compute the envelope of non-destructive testing (NDT) signals is proposed. This method allows increasing the speed and reducing the memory in extensive data processing. Also, this procedure presents advantage of preserving the data information for physical modeling applications of time-dependent measurements. The algorithm is conceived to be applied for analyze data from non-destructive testing. The comparison between different envelope methods and the proposed method, applied to Magnetic Bark Signal (MBN), is studied.

Research highlights
► The use of a flexible envelope method increase the Neural Network classification rate.
► The Falling sheet envelope method allows to preserve the physical information contained in the signal.
► The Falling sheet envelope method presents good results when is used with a PNN for the classfication NDE signals.