Self-tuning triboelectric nanogenerator for omnidirectional broadband (1.2–13.8 m/s) wind energy collection and wind vector detection with deep learning

Wind speed and wind direction are not only important meteorological factors in environmental dynamic monitoring but also crucial parameters for harvesting considerable renewable wind energy. Although the multi-uniform unit arrangement of triboelectric nanogenerator (TENG) is a strategy for extending the responsive wind speed and direction, interunit crosstalk, low wind direction resolution, narrow working wind speed range, and unclear complex interaction between wake and wind capture component hinder further applications of wind-driven TENG. Here, a self-tuning triboelectric nanogenerator (S-TENG) is developed to effectively harvest broadband omnidirectional wind energy and simultaneously meet the high-resolution detection of wind vector. The gas-solid interaction model of this device is proposed as a two-step process. Meanwhile, owning to the self-tuning ability, the device greatly expands its working range from 2.4 to 6.4 m/s to 1.2–13.8 m/s, giving an excellent output power of 47.43 W/m3. Additionally, a real-time wind vector sensing system based on LabVIEW software exhibits wind direction resolution of up to 5° with only three channels. The recognition accuracy of wind direction relying on deep learning model reaches 98.98%. This work provides a guidance for designing effective high entropy wind energy harvesters and a deep understanding of gas-solid interaction processes.