Skip to main navigation Skip to search Skip to main content

Warpage Deformation Prediction and Optimization in Fused Deposition Modeling Using a Hybrid FFNN and PSO Approach

  • Amirpouya Masoumi
  • , Ramen Ghosh
  • , Anto Antony Samy
  • , Atefeh Golbang
  • , Gavin Walker
  • , Rabah Mouras
  • , Marion McAfee
  • Atlantic Technological University
  • Ulster University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Fused deposition modelling (FDM) is a versatile additive manufacturing technique involving extruding thermoplastic polymers layer by layer. FDM enables the fabrication of complex geometries, however prediction and minimization of warpage deformation is critical for high-precision applications such as medical implants. Simulations based on finite element analysis (FEA) have been shown to accurately predict warpage deformation in FDM; however, this approach is computationally intensive and has limitations for process optimisation and monitoring. In this study, a feed-forward neural network (FFNN) model was trained using deformation values from FEA simulations. Input features included nozzle temperature, material deposition rate, bed temperature, layer height, layer number, and spatial coordinates, while warpage deformation served as the output. The trained FFNN was then combined with the particle swarm optimization (PSO) algorithm to find the optimal FDM parameters to minimize the warpage deformation. The combination of a trained FFNN based on FEA with PSO identified optimal FDM parameters, producing a minimum warpage deformation of 0.062 mm. The proposed PSO-FFNN approach demonstrates the feasibility of using machine learning and metaheuristic algorithms to minimize warpage defects in additive manufacturing while reducing computational cost and computation time.

Original languageEnglish
Title of host publication12th 2026 International Conference on Control, Decision and Information Technologies, CoDIT 2026
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages622-627
Number of pages6
ISBN (Electronic)9798319520777
DOIs
Publication statusPublished - 2026
Event12th International Conference on Control, Decision and Information Technologies, CoDIT 2026 - Bari, Italy
Duration: 13 Jul 202616 Jul 2026

Publication series

Name12th 2026 International Conference on Control, Decision and Information Technologies, CoDIT 2026

Conference

Conference12th International Conference on Control, Decision and Information Technologies, CoDIT 2026
Country/TerritoryItaly
CityBari
Period13/07/2616/07/26

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

  1. SDG 2 - Zero Hunger
    SDG 2 Zero Hunger

Keywords

  • Additive manufacturing
  • Feed forward neural network
  • Fused deposition modeling
  • Particle swarm optimization algorithm
  • Warpage deformation

Fingerprint

Dive into the research topics of 'Warpage Deformation Prediction and Optimization in Fused Deposition Modeling Using a Hybrid FFNN and PSO Approach'. Together they form a unique fingerprint.

Cite this