Real Time Self-Monitoring of Adhesion State via Machine Learning-Assisted Traffic Light Color-Coding in...

Despite significant advancements in robust adhesive materials, convenient monitoring adhesion strength under service conditions before adhesion failure is essential yet challenging. In this study, structurally novel silicone-based conductive adhesive from comb-shaped supramolecular elastomers is reported, featuring self-monitoring of adhesion state in real time manner via machine learning-assisted traffic light color-coding approach. Comb-shaped supramolecular ion-conducting polysiloxane P(Ba-co-Apy-co-DMS) are first constructed via effective hydrosilylation, yielding pendant H-bonding barbiturate (Ba) and ionic liquids (Apy) moieties. Adhesion-sensitive electrical properties and experimental database are therefore constructed from P(Ba-co-Apy-co-DMS) via digitally transforming of adhesion strength into capacitance signals. Such database is further used to train a hybrid deep learning architecture that integrates 1D convolutional neural networks (1D-CNN) with long short-term memory (LSTM) units. This model learns the capacitance-adhesion mapping and translates the encoded signal features into States 1, 2, and 3. Moreover, a traffic light color-coding approach on the perceptions of adhesion strength is developed, green, orange, and red LED light symbols are respectively associated with States 1, 2, and 3 leveraged from deep learning architecture, reflecting tight, stretched, and fractured adhesion. Our conductive adhesives illustrate the Internet of Things-empowered adhesion state and structural health monitoring, offering promising opportunity to boost intelligentization and informatization of classical materials.