Cluster-Based Vibration Analysis of Structures With GSP

Abstract

This work describes a divide-and-conquer strategy suited for vibration monitoring applications. Based on a low-cost embedded network of Micro-ElectroMechanical (MEMS) accelerometers, the proposed architecture strives to reduce both power consumption and computational resources. Moreover, it eases the sensor deployment on large structures by exploiting a novel clustering scheme which consists of unconventional and non-overlapped sensing configurations. Signal processing techniques for inter and intra-cluster data assembly are introduced to allow for a full-scale assessment of the structural integrity. More specifically, the capability of graph signal processing is adopted for the first time in vibrationbased monitoring scenarios to capture the spatial relationship between acceleration data. The experimental validation, conducted on a steel beam perturbed with additive mass, revealed high accuracy in damage detection tasks. Deviations in spectral content and mode shape envelopes were correctly revealed regardless of environmental factors and operational uncertainties. Furthermore, an additional key advantage of the implemented architecture relies on its compliance with blind modal investigations, an approach which favors the implementation of autonomous smart monitoring systems.

Improving the Event-Based Classification Accuracy in Pit-Drilling Operations: An Application by Neural Networks and Median Filtering of the Acceleration Input Signal Data

“…In relation to operational monitoring by the means of automated time studies, recent research on the topic in forestry have proven that high classification accuracies may be achieved by the use as inputs in the machine-learning algorithms of raw signals outputted by various type of sensors such as accelerometers, gyroscopes and sound-pressure level meters [ 8 , 9 , 10 , 11 , 12 , 13 ]. In addition, signals outputted by accelerometers coupled with ML techniques have proven very useful not only in the forestry but also in other engineering disciplines such as those dealing with infrastructure and its monitoring [ 36 , 37 , 38 ]. While the approach of using the raw data as inputs may prove to be useful for fully automated, real-time applications, because it may ease the computational effort, in many ways its use in a processed form for offline modeling and improvement is important for the science.…”

Machine Learning Meets Compressed Sensing in Vibration-Based Monitoring

“…The modal identification process implies long time series are to be collected, stored and processed for each sensing device. If the structure under inspection has large dimensions and complex geometries, which demand the deployment of very dense sensor networks basing on low-cost sensors [ 15 ], the risk of having unacceptable data flooding and network congestion is high. Prompted by these issues, data compression techniques were investigated as viable solutions to alleviate the communication and memory burden caused by such large datasets.…”

Structural Health Monitoring System With Narrowband IoT and MEMS Sensors

“…In particular, civil engineers and inspectors need periodic data to assess the mechanical performance over long time intervals, typically many years, and to analyze the structure's long-term response to environmental and daily stresses (i.e., weather, wind, weight load, corrosion, or act of vandalism). Both static and dynamic measurements are required to achieve this goal: corrosion and crack monitoring are typical static measurements, while modal analysis [4] needs dynamic (velocity, acceleration) measurements. A SHM system comprises four blocks: i) sensors and transducers, ii) remote communication, iii) data storage, and iv) feature extraction and data processing.…”