Optimization of Warpage Defects of Base Pencil Box by Using Backpropagation Neural Network and Genetic Algorithm

  • Eko Ari Wibowo Polman Astra
  • Edi Sofyan Swiss German University
  • Ary Syahriar Al Azhar Indonesia University
  • Albertus Aan Dian Nugroho Polman Astra
  • Fuad Widiatmoko Infinitigroup
  • Paulus Gagat Charisma Arwidhiatma PT. Trias Indra Saputra
Keywords: Plastic injection molding, Warpage defect, Backpropagation, Neural Network, Genetic Algorithm

Abstract

The use of plastic products is increasing rapidly nowadays, starting from automotive components, electronics, to office equipment. Injection molding process is a method of making plastic products by injecting the material into the mold. One of the products is a pencil box, but this product has a warpage defect. Defect is indicated by a deflection in the wall, causing misassembles. This study aims to eliminate these defects with parameter optimization. The L27 (34) orthogonal array was used to make the data input design. Data that has been designed is simulated by using MoldFlow to get the value of deflection. Results of the experiment were analyzed by using Backpropagation Neural Network to determine the pattern of relationship between process parameters and response, while Genetic Algorithm method was used for parameter optimization. The composition of the recommended parameters were mold temperature of 15°C, melt temperature of 200°C, packing pressure of 120% and injection time of 6 seconds. As a result, the optimization of deflection reached 44%. The previous maximum deflection of 2.779 mm has decreased to 1.554 mm.

Published
2021-08-03
Section
Articles