VIBROTOOL — Software tool for change detection and diagnosis in vibration signals

The problem of change detection and diagnosis (CDD) is a key point that frequently arises in many application areas, where analysis and modeling of nonstationary signals are required. The paper considers a set of programs, included in a Matlab Toolbox, VIBROTOOL, implementing functions for change detection and diagnosis in vibration engineering. Classical methods based on statistical signal processing and advanced methods based on time-frequency and information processing are considered. Compared to other toolboxes in the same domain, VIBROTOOL Toolbox promotes blind source separation and advanced processing of information from time-frequency analysis, before the change detection and segmentation. These techniques are especially necessary for features extraction of the vibration signals making the object of the analysis, as well as to increase signal to noise ratio or locate certain components in signals to assist detection of machine faults.