Pin-point effect determination using a rigorous approach

Abstract

A new method for evaluating the pin-point effect of pile yarn of carpets before weaving has been introduced. The method has been initially accomplished by presenting a standard method for bundle preparation and consequently the pin-point index is presented by image analysis technique. To this end, yarns with different twists are heat set at various times and temperatures. Comparison of the results shows that increasing the twist, time and temperature positively contribute to the pin-point index. In the last section, an adaptive neuro fuzzy model (ANFIS) and an artificial neural network model (ANN) have been designed to predict the pin-point index of the heat set yarns based on training with the experimental data. The input parameters are twist, time and temperature, and the output is the pin-point index. The results illustrate that the learning capability of the ANFIS model is superior and its generalization ability is slightly better than that of a standalone ANN model.