A Self-Calibrating, Energy-Autonomous Sensor Node for Long-Term Structural Health Monitoring Using Hybrid Piezo-Thermoelectric Harvesting

Abstract Structural Health Monitoring (SHM) of infrastructure such as bridges, pipelines, and buildings is essential for safety, but long-term operation of sensor nodes is limited by power constraints and sensor drift. This paper presents the design and implementation of a fully energy-autonomous wireless sensor node using hybrid piezo-thermoelectric energy harvesting. The node combines vibration-based and thermal-gradient harvesting to ensure energy reliability under varying environmental conditions. An onboard microcontroller executes a lightweight machine-learning-based drift-detection algorithm that triggers self-calibration using an internal reference, maintaining long-term accuracy. Experimental results demonstrate that the system sustains continuous operation at a 5-minute sampling rate using harvested energy alone, with a net positive energy margin of 92 mWh/day. The self-calibration reduces drift error by over 90 % compared with static calibration. The prototype thus demonstrates a practical path toward decade-long, maintenance-free SHM deployments. Index Terms Structural Health Monitoring (SHM), Energy Harvesting, Piezoelectric, Thermoelectric, Self-Calibration, Sensor Drift, LoRaWAN, Low-Power IoT.