Water Absorption Behavior of Dual-Sponge Structure Sealing Elastomers Assisted by Machine Learning

Water-absorbing expanded elastomers hold significant importance in the fields of engineering and construction. However, traditional expanded elastomers exhibit common characteristics such as slow swelling rates, leakage after water absorption, and low strength. This research report proposed an approach for developing high-strength water-absorbing expanded elastomers with a dual-sponge structure. The elastomers were prepared by incorporating a composite water-absorbing resin with a porous structure into a fluoroelastomer matrix. Additionally, this research validates this research under the background of machine learning using a random forest model. The water absorption rate of this research material can reach 30 times its own weight with an extremely rapid water absorption response. Its strength can reach 17.37 MPa, retaining more than 50% of moisture and maintaining environmental humidity between 50 and 60%. The R2 value of the machine learning model reaches 0.998, proving the strong guidance significance of the random forest model. Furthermore, the simplicity of the treatment method employed in this research ensures low economic costs and ease of industrial application. The aim of this study is to improve sealing in water-related environments in infrastructure with strengths up to 3–4 times higher than those of seals commonly used at this stage and with greatly improved water absorption response rates. This makes it possible to completely replace the seals commonly used in shield machines today, and medical devices in the field of hemostatic dressings can be developed by using the strategy of this research as a blueprint.