Real-time Structural Health Monitoring System Using Internet of Things and Cloud Computing

Real-time monitoring of various structural behaviors, particularly displacement and acceleration, serves important and valuable information for people; for example, they can be used for active control or damage warning. With recent advancement of the Internet of Things and client-side web technologies, wireless integrated sensor devices nowadays can process real-time raw sensor signal data into target measurements, such as displacement, and then send the results through a standard protocol to the servers on the Internet. The monitoring results are further processed for visualization purpose in the servers and the computed results are pushed to connected clients like browsers or mobile applications in real-time. We build a real-time cloud-based system that can receive heterogeneous IoT data, allow users to create a three-dimensional model online according to the real world structure, and the monitoring results can be visualized in that model. In this paper, we illustrate the software architecture of the proposed system and focus on the technologies that are used, like client-side scripting, NoSql database, and socket communication. We also present the challenges of displaying the overall movement and shape transformation of the 3D structural model. Thus, each internal-connected element’s rotations and translations are obtained by converting the monitoring results of each sensor device measured in the global coordinate system. To overcome this, we create an inverted movement calculation method. A simple 3D two-level structural model and simulated sensor displacements are used to demonstrate system function and validate the inverted movement calculation method.

💡 Research Summary

This paper presents a comprehensive real‑time structural health monitoring (SHM) platform that leverages Internet of Things (IoT) sensor nodes, cloud computing, and web‑based 3D visualization. The authors address three major shortcomings of traditional SHM systems: delayed data transmission, difficulty integrating heterogeneous sensors, and limited visualization capabilities. Their solution consists of four tightly coupled components: (1) a low‑cost wireless sensor node built around an ESP8266 microcontroller and an ADXL345 accelerometer, which locally processes raw acceleration into displacement, assigns a unique device ID via an embedded web UI, and pushes data to the cloud using standard HTTP; (2) a dual‑server cloud backend, where a conventional HTTP server handles RESTful requests, stores sensor metadata, structural models, and time‑stamped measurements in a MongoDB NoSQL database, and performs health‑evaluation logic such as threshold alerts; a separate real‑time message server built with Node.js and Socket.io maintains bi‑directional WebSocket connections to browsers and mobile clients, instantly broadcasting processed results; (3) a client‑side 3D engine powered by Three.js that lets users construct a virtual replica of the physical structure, map each sensor’s location to a model element, and render dynamic movements; and (4) an “inverted movement calculation method” that converts global‑coordinate displacements measured at sensor nodes into rotation and translation parameters for each structural element.

The core algorithm treats each column as a rigid body. By using the measured top and bottom coordinates of a column, the method derives the y‑axis rotation (Ry) and x‑axis rotation (Tx) through inverse sine functions (Equations 1‑8). Because Three.js applies rotations about an object’s geometric centre, the authors introduce additional correction terms (Equations 9‑11) to align the centre‑based rotation with the real‑world bottom‑anchored motion. This approach scales to multi‑story frames, allowing each connected column to be updated independently while preserving overall kinematic consistency.

Sensor deployment guidelines are discussed: at a minimum, a sensor must be placed at every joint to capture the six‑degree‑of‑freedom motion of each element; additional sensors along the length of a column improve the fidelity of bending representation. The paper demonstrates the system with a two‑level structural model, feeding simulated displacement data into the pipeline. The workflow proceeds as follows: users create the 3D model and input sensor coordinates via a web interface; the sensor nodes broadcast measurements at a predefined sampling rate; the HTTP server stores the raw data, computes rotation/translation using the inverted method, and forwards the results to the Socket.io server; the client receives the data, updates the Three.js scene in real time, and displays any threshold‑exceeding warnings.

Implementation details emphasize the use of open‑source, platform‑agnostic technologies: HTML5/CSS3/JavaScript for the front end, Node.js for server logic, MongoDB for scalable storage, and Socket.io for low‑latency messaging. This stack enables horizontal scaling in cloud environments, making the solution suitable for large‑scale infrastructures such as bridges, dams, or nuclear facilities.

The authors validate the end‑to‑end functionality by comparing the visualized motion against the known simulated inputs, confirming that the inverted calculation reproduces both rotations and translations accurately. They also verify that the alert mechanism triggers when any axis displacement exceeds user‑defined limits.

In conclusion, the paper delivers a fully integrated, real‑time SHM system that bridges the gap between raw IoT sensor streams and intuitive 3D visual feedback. By combining a lightweight wireless node, a dual‑server cloud architecture, and a mathematically sound transformation algorithm, the work provides a practical blueprint for deploying digital twins of civil structures, enabling continuous condition monitoring, rapid damage detection, and informed maintenance decision‑making.