Optimisation of micro W-bending process parameters using I-optimal design-based response surface methodology

Xiaoyu Liu, Xiao Han, Shiping Zhao, Yi Qin, Wan Adlan Wan-Nawang, Tianen Yang

Research output: Contribution to journalArticlepeer-review

1 Citation (Scopus)
21 Downloads (Pure)


There is an increasingly recognised requirement for high dimensional accuracy in micro-bent parts. Springback has an important influence on dimensional accuracy and it is significantly influenced by various process parameters. In order to optimise process parameters and improve dimensional accuracy, an approach to quantify the influence of these parameters is proposed in this study. Experiments were conducted on a micro W-bending process by using an I-optimal design method, breaking through the limitations of the traditional methods of design of experiment (DOE). The mathematical model was established by response surface methodology (RSM). Statistical analysis indicated that the developed model was adequate to describe the relationship between process parameters and springback. It was also revealed that the foil thickness was the most significant parameter affecting the springback. Moreover, the foil thickness and grain size not only affected the dimensional accuracy, but also had noteworthy influence on the springback behaviour in the micro W-bending process. By applying the proposed model, the optimum process parameters to minimize springback and improve the dimensional accuracy were obtained. It is evident from this study that the I-optimal design-based RSM is a promising method for parameter optimisation and dimensional accuracy improvement in the micro-bending process.

Original languageEnglish
Article number7
Number of pages10
JournalManufacturing Review
Publication statusPublished - 2 Mar 2021


  • I-optimal design
  • micro-bending
  • micro-forming
  • optimisation
  • Response surface methodology
  • springback


Dive into the research topics of 'Optimisation of micro W-bending process parameters using I-optimal design-based response surface methodology'. Together they form a unique fingerprint.

Cite this