The student Jose Joaquin Peralta Abadia obtained an EXCELLENT CUM LAUDE with the International Doctorate distinction

Back

The student Jose Joaquin Peralta Abadia obtained an EXCELLENT CUM LAUDE with the International Doctorate distinction

THESIS

The student Jose Joaquin Peralta Abadia obtained an EXCELLENT CUM LAUDE with the International Doctorate distinction

2026·10·02

$titulo.getData()

  • Thesis title: Enhancing smart monitoring in face milling with deep continual learning

Court:

  • Presidency: Giovanna Martínez Arellano (University of Nottingham)
  • Vocal: Alessandra Caggiano (Roma Tre University)
  • Vocal: Maite Beamurgia Bengoa (Fagor AOTEK S. Coop)
  • Vocal: Miren Illarramendi Rezabal (Mondragon Unibertsitatea)
  • Secretary: Gorka Ortiz de Zarate Bengoa (Mondragon Unibertsitatea)

Abstract:

Deep learning (DL)-based tool condition monitoring is effective in milling operations, but its industrial adoption is limited by manual model design, a shortage of labelled machining data, and the use of architectures that are unable to adapt to new input parameters. This thesis addresses these limitations by proposing an automated learning framework to generate DL workflows and a resource-efficient continuous learning (CL) framework. To validate the frameworks in the absence of machining data, the MU-TCM milling dataset was generated. The proposed methodology improves model robustness by mitigating catastrophic forgetting and reducing prediction error by up to 32% under inter-machine and inter-material variations. This facilitates its practical deployment in SMEs that use internal CNC signals, thereby extending tool life and reducing error margins.