Search

In order to ensure the production quality of high-speed laser welding, it is necessary to simultaneously monitor multiple state properties. Monitoring methods combining vision sensing and deep learning models are popular but most models used can only make predictions on single welding state property. In this contribution, we propose a multi-output model based on a lightweight convolutional neural network (CNN) architecture and introduce the particle swarm optimization (PSO) technique to optimize the loss function of the model, to simultaneously monitor multiple state properties of high-speed laser welding of AISI 304 austenitic stainless steel. High-speed imaging is performed to capture images of the melt pool and the dataset is built. Test results of different models show that the proposed model can achieve monitoring of multiple welding state properties accurately and efficiently. In addition, we make an interpretation and discussion on the prediction of the model through a visualization method, which can help to deepen our understanding of the relationship between the melt pool appearance and welding state. The proposed method can not only be applied to the monitoring of high-speed laser welding but also has the potential to be used in other procedures of welding state monitoring. [ABSTRACT FROM AUTHOR]

Liquid-phase reduced metal electrodes are cleverly used in the fabrication of flexible piezoresistive sensors. Coupled with the integration of porous elastomers and composites, sensors offer permeability, flexibility, outstanding conductivity, and advanced sensing capabilities. They find applications in human physiological signal monitoring, pressure array recognition and handwriting recognition. Furthermore, sensors assisted by deep learning enable the precise recognition and correction of fitness movements. This study provides new ideas for the preparation of piezoresistive sensors and fully demonstrates their application potential in healthcare, exercise monitoring, encrypted transmission of information, and the development of intelligent wearable fitness devices. [Display omitted] • Patterned and reliable electrodes prepared based on a simple and efficient liquid-phase reduction method. • Controlled preparation of porous elastomers by the salt-molded sacrificial method. • Flexible devices with fast response and high sensitivity for multi-state sensing applications. • Deep learning assisted sensors for highly accurate recognition and correction of fitness movements. Flexibility and wearability in electronic devices gain prominence with the rise of national fitness campaigns. Among them, piezoresistive sensors stand out for their ability to accurately monitor health signals due to their high sensitivity. However, conventional metal nanoparticle ink electrodes face issues like peeling, chemical instability, and substrate limitations. This study introduces a novel method for fabricating high-performance flexible piezoresistive sensors using liquid-phase reduced metal electrodes. Integration of porous polydimethylsiloxane (PDMS) substrates with highly conductive interdigital silver electrodes (1.6 × 10−6 Ω·m) addresses conventional electrode shortcomings, offering permeability, flexibility, and outstanding conductivity. Incorporation of a graphene (GR)/carbon nanotube (CNT)/Ecoflex composite enhances sensor piezoresistive sensing capabilities, with features including high sensitivity (3.57 KPa−1), rapid response time (58 ms/72 ms), and excellent cycle stability (>10000 cycles). The sensor finds utility in various applications, including human physiological signal monitoring, pressure array recognition, and handwriting recognition. Additionally, with deep learning techniques, the system achieves accurate recognition (99.25 %) and correction (98.75 %) of diverse fitness movements, aiming to promote safer exercise practices, enhance training efficiency, and advance intelligent wearable fitness devices. [ABSTRACT FROM AUTHOR]