SciELO - Scientific Electronic Library Online

 
vol.20 número2An efficient and highly scalable listless SPIHT image compression frameworkOrange II removal by a horseradish peroxidase immobilized onto chemically modified diatomites is a combination of adsorption and oxidation processes índice de autoresíndice de materiabúsqueda de artículos
Home Pagelista alfabética de revistas  

Servicios Personalizados

Revista

Articulo

Indicadores

Links relacionados

  • No hay artículos similaresSimilares en SciELO

Compartir


Journal of applied research and technology

versión On-line ISSN 2448-6736versión impresa ISSN 1665-6423

Resumen

SONAWANE, Shriram S.; CHARDE, Sarita J.; MALIKA, Manjakuppam  y  THAKUR, Parag. Artificial neural network model for prediction of viscoelastic behaviour of polycarbonate composites. J. appl. res. technol [online]. 2022, vol.20, n.2, pp.188-202.  Epub 27-Ene-2023. ISSN 2448-6736.  https://doi.org/10.22201/icat.24486736e.2022.20.2.1101.

Polymer composites are created by incorporating nanoparticles into polymers and can result in significant gains even with a very tiny amount of reinforcement that can be tailored to specific purposes. To have a better understanding of the behavior of these polymer composites, a variety of characterizations and analysis must be conducted, which demands financial and time investment. Thus, computational techniques can be beneficial in reducing the number of characterizations and studies required to produce polymer composites. Prediction of thermomechanical characteristics has been made possible using a computational technology known as an artificial neural network (ANN). The present study used dynamic mechanical analysis (DMA) to characterise polycarbonate / calcium carbonate-SiO2 core shell composites (polycarbonate composites). The chosen ANN model comprised a network of [2-4-1] (Inputs to the input layer - Neural network count in the hidden layer - Output from the output layer) based on the dataset. Prediction accuracy was approximately 90% when utilising the ANN approach. The applicability and performance of ANN were also confirmed using mean squared error (MSE), which is in the range of 10-5 in this scenario. Correlation coefficient of 0.999 was found between the output predicted by ANN and the actual output. Additionally, sensitivity analysis established the importance of various input variables in terms of output. Optimizing the variables enabled maximization of the circumstances, hence anticipating the glass transition temperature.

Palabras llave : Artificial neural network (ANN); polycarbonate composites; dynamic mechanical analyzer; glass transition temperature; storage modulus.

        · texto en Inglés     · Inglés ( pdf )