CANETIA
Canetia Analytics was born in the United States as a startup dedicated to increasing the structural safety of buildings and large projects. Its distinction lies in an exclusive technology equipped with cutting-edge accelerometers, designed to capture the natural vibrations of any structure, generated by its daily use. Canetia's challenge was to transform this vast collection of data into an intelligent system capable of identifying any irregularity in real time based on a comparative analysis of the structure's vibrational history, to ensure its integrity.
At WhiteBox, we accepted the challenge posed to us by Canetia Analytics and created a comprehensive system that unifies and processes data from various IoT devices distributed in the structures under Canetia's supervision. This information is centralized in an analytical database, on which we develop deep learning models, specifically autoencoders, using open-source tools such as Tensorflow. These models are able to discern structural anomalies with great precision, by evaluating discrepancies in the reconstructed signal. This model has been rigorously verified and endorsed by academic specialists from UMass Lowell, marking a milestone in the industry.
The model proved to be outstanding in laboratory tests and excelled in its ability to identify unexpected additions to structures, accurately determining both the size and location of structural anomalies. This achievement not only highlights the commercial viability of the system but it is already in a pilot phase of testing in the Los Angeles subway, promising to revolutionize the way in which structural safety is measured and guaranteed.
Predictive Maintenance
