Energy harvesting using piezoelectric sensors for extremely low-power Internet of Things use cases

This article explores the feasibility of utilizing vibrations as energy sources to develop energy-independent wireless sensing platforms within the Industrial Internet of Things (IIoT) architecture. It proposes deploying piezoelectric sensors on vibrating machinery to enable self-powered sensing in remote or hard-to-access environments. To analyze the energy-harvesting potential of commercial piezoelectric sensors, both preliminary measurements and controlled laboratory experiments are conducted. A vibration-powered LoRaWAN-based sensor node design is introduced, followed by experiments aimed at determining the balance between available energy and optimal sensor sampling rates to ensure continuous operation while maintaining stable energy storage patterns. Ultra-Low-Power Energy-Harvesting Integrated Circuits (EH-ICs) play a key role in efficiently managing harvested energy and regulating device performance. Since the vibration-based energy source is intermittent, system operation—continuous or discontinuous—depends on the balance between harvested and consumed power. Proper transducer selection and configuration of the power generation module are essential for maximizing energy efficiency. To enhance performance evaluation, this study further introduces two hybrid machine learning architectures: one employing Lasso Regression for classification and another using the Self K-Means algorithm for clustering, enabling adaptive energy management and intelligent sensor control in IIoT environments.