Fractional derivative models are widely used to easily characterise more complex damping behaviour than the viscous one, although the underlying properties are not trivial. Several studies about the mathematical properties can be found, but are usually far from the most daily applications. Thus, this paper studies the properties of structural systems whose damping is represented by a fractional model from the point of view of a mechanical engineer. First, a single-degree-of-freedom system with fractional damping is analysed. Specifically, the distribution of the poles and the dynamic response to several excitations is studied for different model parameter values highlighting dissimilarities from systems with conventional viscous damping. In fact, thanks to fractional models, particular behaviours are observed that cannot be reproduced by classical ones. Finally, the dynamics of a machine shaft supported by two bearings presenting fractional damping is analysed. The study is carried out by the Finite Element method, deriving in a system with symmetric matrices. Eigenvalues and eigenvectors are obtained by means of an iterative method, and the effect of damping is visualised on the mode shapes. In addition, the response to a perturbation is computed, revealing the influence of the model parameters on the resulting vibration.

This paper discusses the increasing amount of data handled by companies and the need to use Big Data and Data Analytics to extract value from this data. However, due to the large amount of data collected, challenges related to the computational capacity of machines often arise when performing this analysis to acquire relevant information for the organization, especially when we are using edge computing. The paper aims to train machine learning models using compressed data, with two compression techniques applied to the original data. The results show that models trained with compressed data achieved similar accuracy to those trained with uncompressed data, and different compression techniques were compared. The research extended a previous study by analyzing the use of autoencoders for compression and reducing both instances and dimensionality of the dataset. The accuracy rate of the models when trained with compressed data instead of original data was maintained.