Modelling and forecasting the stretch blow moulding temperature based on the weather and surrounding temperature for reducing product defect

Abstract

Machine temperature is an important factor for the defect of the plastic bottle. Mostly defect that found in the plastic industries which using stretch blow moulding is opalescence and pearlesence. Therefore, the temperature of the machine should be controlled during the production process. However, machine temperature of the stretch blow moulding is occasionally influenced by surrounding temperature. This paper aims to reduce the defect by using temperature setting. Moreover, the paper provides machine temperature forecasting using Artificial Neural Network based on surrounding temperature data. The result shows that the machine temperature setting can reduce the defect. Furthermore, the result shows that Artificial Neural Network by using 10 layers, logsig and purelin function can predict the stretch blow moulding machine temperature accurately. This paper recommends for future research to consider machine temperature prediction as an important factor for product scheduling.

Keyword: PET bottle, Machine Temperature Forecasting, Artificial Neural Network