Wireless Sensor Node for Wheelset Defects Identification

Objective of this research is the development of a smart wheelset for high-speed trains, by means of an integrated Wireless Sensor Node (WSN) able to perform 3-axial acceleration measurements, to process the raw data by means of a microcontroller and to transmit wirelessly synthetic data to the on-board control unit. The information on the health of the wheelset allows to perform predictive maintenance, leading to great enhance of high-speed trains from both economic and safety point of view. Algorithms are developed starting from numerical simulations performed by a Multi-Body model of the high-speed train. Modelling of the rail-wheel interaction by means of a deformable railway has been considered. The basic idea is to provide extremely simple identification methods that require very low computational effort, low power consumption and low memory data storage, to be easily implemented on an electronic board. Identification methods able to detect the presence of a defect on the wheelset are based on a statistical approach: the k-NN (Nearest Neighbour) method.