Aplicación de un modelo de redes neuronales artificiales (RNA) para la dosificación adecuada de los colores en textiles de la Empresa PROLAVTEX

Abstract

In the textile laundry industry, the closer the color achieved is to the color requested by the customer, the higher the degree of acceptance and, thus, user satisfaction. Given the crucial importance of precision in the coloring of garments in this industry, it becomes essential to ensure adequate consistency in the characteristics of the final product. Due to the variability that is usually present in the color and the fact that the relationship between the operating parameters cannot be predicted by statistical models, computer modeling and simulation tools are an interesting option for the constant improvement of these processes, since they avoid the need to repeat experimental tests, saving time and money. In the present work, an artificial neural network approach will be developed based on the most important operational variables of the textile garment dyeing process carried out in the textile laundry company PROLAVTEX. The proposed ANN consists of 3 input variables, represented by the 3 dyes corresponding to the primary colors (red, yellow and blue), which were determined by experimental tests as the most influential parameters in the resulting color in the garment. In addition, 3 output variables were established that objectively determined the color by means of the CIELAB color space and its three coordinates: L (luminosity), a (red/green) and b (yellow/blue). An important point when considering an ANN as a prediction model is to validate the accuracy of the data provided by the network. For this reason, graphical and statistical analyses and comparisons are carried out in order to test the predictive capacity of the model. In this way, it can be affirmed that the proposed ANN can be used for the adequate dosage of colors in textile dyeing.