Embedded Edge Ai-Based Smart Dam and Bridge Structural Health Monitoring with Predictive Flood Analytics and Iot System

Abstract

Dams and bridges require continuous monitoring to ensure safety and long-term performance, yet conventional inspection relies on periodic manual measurements that can delay the detection of abnormal behavior. This paper presents an embedded Edge AI-based smart dam and bridge structural health monitoring system with predictive flood analytics and IoT connectivity. ESP32-based monitoring units acquire temperature, humidity, rainfall, water level, load, vibration, acceleration, gyroscope, strain and tilt-related parameters from prototype dam and bridge models, while an ESP32-CAM provides live visual surveillance. Feed-forward artificial neural networks developed with TensorFlow/Keras, using 17 input features for the dam and 15 for the bridge with StandardScaler normalization, classify each structure as NORMAL, WARNING or CRITICAL; the dam model is converted to TensorFlow Lite for edge deployment. A Flask backend connects the units to a web dashboard, and critical conditions trigger LED, buzzer and siren alerts, an automated motor/gate response and e-mail notification. Dashboard outputs, logged sensor records and time-wise event analyses from the prototype demonstrate continuous, low-cost and real-time condition assessment that reduces dependence on manual inspection.