A Small Footprint, Low Power, and Light Weight Sensor Node and Dedicated Processing for Modal Analysis

Abstract

Structural Health Monitoring functionalities are aimed at constantly assessing the health of a building in order to prevent dramatic consequence of a damage. This work describes a well-defined wireless sensor network system installed over a steel beam capable to perform modal parameters estimation, such as natural vibration frequencies and modal shapes. Signal Processing Techniques were aimed at computing Power Spectral Density of the acceleration signals acquired, dealing with parametric and non parametric approaches. Algorithms in frequency domain, together with the Second Order Blind Identification method were implemented for modal shapes reconstruction. Beside a satisfactory agreement between the theoretical model and the output response of the algorithms implemented, versatility, easiness of reconfiguration, scalability and compatibility with long term installation are among the most powerful advantages of the architecture proposed. Light weight, low power consumption also enhance the capabilities of the system to provide real-time information in a relatively cheap way.