Artificial neural network reinforced topological optimization for bionics-based tridimensional stereoscopic hydrogen sensor design and manufacture

Abstract

With green energy advancing, the demand of high-sensitive hydrogen sensor is strongly urgent for wide demanding applications, such as energy storage, leak detection, and chemical production. The utilization of resistive hydrogen sensors has brought revolutionary developments for the field of high-performance devices. Regrettably, design and optimization of the hydrogen sensor structure for response enhancement seems to have been overlooked. In this study, we design and manufacture a bionic tridimensional stereoscopic hydrogen sensor (TSHS) based on topology optimization reinforced by artificial neural network, subjected to gas flow. Measurement and observation of TSHS at room temperature demonstrates a 3-times higher flow velocity when compared with the Spiral-serpentine array and the tolerance between simulation and experiment is less than 1 % under the same conditions. Base on theoretical investigation and test validation, the TSHS, which causes itself to favorably stable and exceptional electrical performances at ultralow hydrogen concentration, opening up plenty of opportunities for high-efficiency detection sensors as well as allows for topology optimization in gas detection, early warning equipment and biosensor.