A critical review on prediction of functional & performance attributes of
textiles by artificial neural network


Jhanji, Y ; Kothari, V K; Gupta, Deepti


Prediction of functional and performance properties of textiles before the actual commencement of fabric production and testing can serve as an effective tool in characterization and designing of fabrics for any desired application. The thermo-physiological properties of textile materials can be predicted by a variety of models, such as statistical, mechanistic and artificial neural network models. Statistical models can give good prediction performance, provided a large data set is presented to make the model and relationship exists between input parameters and response variables. The effect of input parameters on thermo-physiological properties of fabrics cannot be studied in isolation, owing to interdependence and nonlinear relationship of parameters with each other. Statistical models fail to present satisfactory analysis of relationship in such cases. Mechanistic models are useful tools in understanding the fundamentals and physics involved in heat, moisture and liquid transfer through textiles. However, the assumptions considered in the simplification of mechanistic models may not be valid in all conditions and can lead to high prediction errors in real conditions, owing to inherent variability in the textile structures. Moreover the model becomes more complicated as the number of parameters and assumptions increase, thereby limiting the model’s accuracy of prediction. Artificial neural network is a powerful and potent modelling tool which can understand any complex relationship between input and response variables and predicts the thermo-physiological properties of fabrics by considering all fabric parameters at a time. The network exhibits the ability of simulating the functioning of a biological neuron and, in turn, each network component poses analogy to the actual constituents or operations of a biological neuron. In this study, an attempt has been made to highlight the significance of artificial neural network in prediction of comfort properties of textiles.



Artificial neural network;Comfort properties;Neurons;Thermo-physiological properties;Textiles

Full Text: PDF (downloaded 451 times)


  • There are currently no refbacks.
This abstract viewed 863 times