Distributed Signal Processing in Wireless Sensor Networks for Diagnostics

Abstract

This work introduces the on-board vibrational condition monitoring of aircrafts powerplants by its original novel approach of distributed WSN based vibrational condition monitoring. Proposed approach allows to employ computationally intensive methods of vibra- tional signal processing and methods of condition monitoring in computationally weak wireless sensor network. It introduces the fault detection method based on novelty detection which is executed directly in wireless sensor nodes. The novel framework of WSN condition monitoring system with its key attributes Adaptivity, Recon gurability and Three phases of operation enables the immediate fault detection capability while providing long-term monitoring. The proposed methods and framework were evaluated by the means of experiment on the small turbojet engine.