Prediction of Mechanical Properties of LDPE-TPS Nanocomposites Using Adaptive Neuro-Fuzzy Inference System

Document Type : Research Paper

Authors

Abstract

The changes in the behaviour of mechanical properties of low density
polyethylene-thermoplastic corn starch (LDPE-TPCS) nanocomposites
were studied by an adaptive neuro-fuzzy interference system. LDPE-TPCS
composites containing different quantities of nanoclay (Cloisite®15A, 0.5-3wt. %) were prepared by extrusion process. In practice, it is difficult to carry out several experiments to identify the relationship between the extrusion process parameters and mechanical properties of the nanocomposites. In this paper, an adaptive neuro-fuzzy inference system (ANFIS) was used for non-linear mapping between the processing
parameters and the mechanical properties of LDPE-TPCS nanocomposites. ANFIS model due to possessing inference ability of fuzzy systems and also the learning feature of neural networks, could be used as a multiple inputs-multiple outputs to predict mechanical properties (such as ultimate tensile strength, elongation-at-break, Young’s modulus and relative impact strength) of the nanocomposites. The proposed ANFIS model utilizes temperature, torque and Cloisite®15A contents as input parameters to predict the desired mechanical properties. The results obtained in this work indicated
that ANFIS is an effective and intelligent method for prediction of the mechanical properties of the LDPE-TPCS nanocomposites with a good accuracy. The statistical quality of the ANFIS model was significant due to its acceptable mean square error criterion and good correlation coefficient (values > 0.8) between the experimental and simulated outputs.

Keywords

Iranian Journal of Polymer Science and Technology
Volume 25, Issue 3 - Serial Number 3
September and October 2012
Pages 171-180

Files

Share

How to cite

Statistics