Artificial neural networks (ANNs) and response surface methodology (RSM) for optimizing wetting and anti-bacterial properties of woven fabric

This study aims to enhance woven fabric's antibacterial properties by coating with betel leaf extract using immersion and plasma pre-treatment techniques. It explores nano surface modification and develops mathematical models for inhibition zone values using response surface methodology (RSM) and artificial neural networks (ANNs). We divided the RSM into two groups based on the inputs, model A and model H, and then used artificial neural networks to compare the models. We analyzed fabric treatments using SEM, wetting tests, FT-IR spectrometry, and disc diffusion (Kirby-Bauer) with Staphylococcus aureus to assess nano surface modification and antibacterial properties of fabric-treated with discharged plasma. We compared the optimal inhibition zone values for polyester fabrics with plasma treatment, anti-bacterial coatings, and anti-bacterial and plasma treatment. Fabrics treated with antibacterial coating, plasma treatment, or both exhibited inhibition zones of 2.30 +/- 0.50 mm, 3.00 +/- 0.50 mm, and 6.40 +/- 0.50 mm, respectively. Using RSM model A, wetting properties were predicted with SSE, R-squared, and RMSE values of 0.2402, 0.9999, and 0.2451, respectively. Model A also forecasted inhibition zones for different fabric treatments with R-squared values of 0.899 and 0.9579, and SSE values of 0.8038 and 0.0782, respectively. Additionally, inhibition zones were modeled using RSM model H and ANNs, yielding R-squared values of 0.9128 and 0.97724, respectively. We found that the inhibition zone value using the ANNs model had a better predictive ability than RSM. The new technique can, therefore, be used to enhance the anti-bacterial properties and absorbency of 100% polyester fabrics.