Tensile Strength Prediction of Polyethylene Terephthalate Bottle Ropes using Neuro - fuzzy

With the huge amounts of plastic bottles being consumed and disposed daily, people have been using different techniques to recycle. One way is to convert them into strings that will be made into plastic ropes. This paper aims to predict the tensile strength of polyethylene terephthalate (PET) bottle ropes using neuro-fuzzy approach given the following inputs: average thickness of strand, strand width, and the number of strands used. Actual data parameters from a previous research, applying design of experiment (DOE), will be utilized as the training and testing data of the network. Moreover, a comparison between the outputs of the neuro-fuzzy model and the statistical model will be shown. The generated fuzzy rules will be analyzed at certain conditions and compared with the optimization results of previous research. Results showed that although the network has very low training errors, testing errors were higher than acceptable values.