Basic study on automatic determination of injection conditions based on automatic recognition of forming states

Masaki Ogawa, Hideki Aoyama, Noriaki Sano

Research output: Contribution to journalArticlepeer-review


Defects may occur when manufacturing plastic products by injection molding if the injection conditions are not appropriate. Thus, it is extremely difficult to produce products with high-dimensional accuracy and low defects. In addition, injection conditions are determined by experience including trial and error which may involve significant time and costs. This is because the relationship between each injection condition and forming defect is not clear. Injection conditions are interdependent; thus, it is difficult to obtain a quantitative correlation with respect to the forming defects. This study proposes a method to automatically recognize forming defects and determine injection conditions to mold non-defective products, thereby creating a basic system. Focusing on shape defects such as burrs, short shots, uneven color, weld lines, and transfer defects, the system photographs the formed product with a camera, recognizes the forming defects by image data processing, and digitizes the forming state. Then, it determines the appropriate injection conditions based on digitized forming states using a neural network. The usefulness of the proposed method is confirmed through experiments conducted under the injection conditions determined by the proposed method, and optimum injection conditions were determined.

Original languageEnglish
Article numberJAMDSM0115
JournalJournal of Advanced Mechanical Design, Systems and Manufacturing
Issue number6
Publication statusPublished - 2018


  • Forming defect
  • Forming state
  • Injection condition
  • Injection molding
  • Neural network
  • Recognition

ASJC Scopus subject areas

  • Mechanical Engineering
  • Industrial and Manufacturing Engineering


Dive into the research topics of 'Basic study on automatic determination of injection conditions based on automatic recognition of forming states'. Together they form a unique fingerprint.

Cite this