Predicting the effect of thread density on the physical and thermal properties of plain-woven fabric by using a soft computing system

Abstract

This study’s primary goal is to use a fuzzy logic expert system (FLES) to anticipate and create a model for the impact of thread density on plain-woven fabric’s thermal characteristics. Every plain-woven fabric utilized in this study has 100% cotton fiber in it. The input variables for this fuzzy logic expert system are PPI and EPI, while the output variables are thermal transmittance, thermal conductivity, and CLO. It is quite difficult to give the engineers a statistically or mathematically based forecasting model. On the other hand, a large volume of trial data is needed for artificial neural networks (ANN) and adaptive neuro-fuzzy inference systems (ANFIS), which will take a lot of time and effort to maintain. Fuzzy expert systems, on the other hand, may effectively map nonlinear domains and provide a viable model even with a small quantity of experimental data. New experimental results have validated the model in the current study. Predicting production might be a more beneficial approach for industrial practitioners than using the current trial-and-error method. The root mean square, mean absolute error percentage, and coefficient of determination (R2) of CLO, thermal conductivity, and thermal transmittance between the expected and experimental values were determined to be, respectively, 0.002, 0.001, 0.802; 0.857%, 2.642%, 2.359%; and 0.994, 0.969, and 0.955. The results validate the model’s predictions for the CLO, thermal conductivity, and thermal transmittance of plain-woven fabric in the textile sector, and it was discovered that FLES is capable of accurately predicting the result.