Structure Monitoring

Structure Monitoring

Comprehensive System for Structure and Bridge Monitoring

Complete system to monitor structural data from ingestion to alert generation and visualization.

Our structure monitoring solution enables data ingestion, advanced analysis, machine learning model training, anomaly alert generation, and visualization of results, based on data from sensors installed on structures.

Main Features

  • Data Ingestion Module

    The data ingestion module allows connection with sensors installed on the structure, whether accelerometers, strain gauges, temperature sensors, etc. The data is acquired and stored in the cloud for further processing.

  • Advanced Machine Learning Algorithms

    We offer a suite of advanced machine learning algorithms for creating relevant features to detect anomalies in structures and bridges. These models are trained on historical data to learn the normal behavior of the structure and then detect anomalies.

  • Alert Generation

    The alert generation module keeps you informed via email of detected anomalous events. These are accompanied by their severity based on their statistical likelihood of occurring.

  • Results Visualization

    To facilitate action, a visualization platform is provided for anomaly models, where you can see the context of the results, such as weather and related sensors, as well as the calculated features used by the models, for example, accelerations from a train passing over a bridge.

    • Autonomous Monitoring and Alert Generation System

    The downloading, signal processing, machine learning model execution, and alert generation will happen automatically on a daily basis, simplifying monitoring to those moments when models consider measurements to be anomalous. Notifications will be sent via email or any other required messaging service.
    structure-4

    Signal Processing

    • The system processes sensor signals taking into account the frequency range of interest and the resolution required for various calculation algorithms such as OMA, train passages, etc.
    • Filtering and decimation techniques are applied to aid in the selection of relevant modes for structural safety.
    • Waveforms from multiple sensors are synchronized and adjusted to be treated together, if the acquisition system used does not provide it natively.
    structure-7

    Advanced Algorithm: Automatic OMA

    We offer an advanced algorithm for calculating frequency modes under zero external excitation, known as Operational Modal Analysis (OMA). This is adjusted to the structure based on historical data to avoid detection of mathematical artifacts. Filtering and decimation algorithms are also applied to aid in the selection of relevant modes for structural safety.
    structure-2

    Cloud Deployment

    The system is designed to be easily deployed in the cloud, allowing the benefits of reduced costs and reliability that it provides.
    structure-2

    Data Storage

    We offer the option to store raw data in the cloud to allow model retraining, research of new uses, and reuse. For this, AWS S3 is used, which, along with compression during storage, provides efficient, scalable, and low-cost storage. Processed signals and model results are stored in an InfluxDB database that enables easy and fast data visualization.

    Request a demo