RESUMEN
Continuous information on the suspended sediment in the water system is critical in various areas of industry and hydrological studies. However, because of the high variation of suspended sediment flow, challenges still remain in developing new techniques implementing simple, reliable, and real-time sediment monitoring. Herein, we report a potential method to realize real-time sediment monitoring by introducing a particle-laden droplet-driven triboelectric nanogenerator (PLDD-TENG) combined with a deep learning method. The PLDD-TENG was operated under the single-electrode mode with a triboelectric layer of polytetrafluoroethylene (PTFE) thin film. The working mechanism of the PLDD-TENG was proved to be induced by liquid-PTFE contact electrification and sand particle-electrode electrostatic induction. Then, its performance was explored under various particle parameters, and the results indicated that the output signal of the PLDD-TENG was very sensitive to the sand particle size and mass fraction. A convolutional neural network-based deep learning method was finally adopted to identify the particle parameters based on the output signal. High identifying accuracies over 90% were achieved in most of the cases by the proposed method, which sheds light on the application of the PLDD-TENG in real-time sediment monitoring.
