A sensor network is a system of multiple sensors distributed across a structure or site, connected to data acquisition and communication infrastructure, that collectively monitor structural behavior, environmental conditions, or seismic activity. It is the physical foundation of any structural health monitoring system.
A sensor network is a distributed system of sensors β accelerometers, strain gauges, displacement transducers, temperature sensors, tiltmeters, and others β deployed across a structure or site, connected through data acquisition hardware and communication infrastructure. In structural health monitoring (SHM), the sensor network is the physical layer that captures the raw data from which structural behavior is inferred. The design of the sensor network β which sensors, how many, and where they are placed β directly determines what can be observed, and therefore what can be detected, localized, and quantified.
Sensor networks are classified by their architecture and deployment scale. Wired networks use physical cables to connect sensors to a central data acquisition system; they are reliable and high-bandwidth but expensive to install and maintain, especially on large structures. Wireless sensor networks (WSNs) use radio communication (Wi-Fi, ZigBee, LoRa, or custom protocols) to transmit data, eliminating cabling and enabling dense deployment at lower cost; they face challenges in power management, time synchronization, and data reliability. Distributed sensing uses a single fiber optic cable or similar medium to measure strain or temperature continuously along its length β a fundamentally different approach that provides unprecedented spatial resolution. Hybrid networks combine wired and wireless elements for optimal performance in specific applications. Modern SHM increasingly uses MEMS-based wireless sensors, which are low-cost, small, and energy-efficient.
The most critical design decision in a sensor network is sensor placement. Optimal sensor placement (OSP) aims to maximize the information obtained from a limited number of sensors β typically by maximizing the observability of target modes, the sensitivity to expected damage, or the ability to localize damage. The choice of placement depends on the structure type, the expected damage mechanisms, the identification method to be used, and the operational constraints. For modal identification, sensors are typically placed at locations of maximum modal displacement; for damage detection, sensors are placed where damage is most likely to occur and where its signature is strongest. Modern approaches use optimization algorithms (genetic algorithms, simulated annealing, convex optimization) and information-theoretic criteria (Fisher information matrix, entropy) to guide sensor placement. Beyond placement, sensor network design involves decisions about sampling rate, resolution, synchronization, data transmission, power supply, and environmental protection. In earthquake engineering, sensor networks are deployed on buildings, bridges, dams, and lifeline infrastructure for post-earthquake assessment and long-term monitoring. In Iran, sensor networks are increasingly deployed on critical structures β including major bridges and dams β with growing interest in low-cost wireless networks for widespread adoption.