Improving Quality and Reliability in Latin American Textile SMEs Through a Lean TPM 4.0 Framework

Mallqui-Alzamora S.; Caparachin-Romero G.; Maradiegue-Tuesta F.; Pinzón-Hoyos F.; Straujuma A. · Advances in Transdisciplinary Engineering (2026) · DOI: 10.3233/atde260684

Resumen

Textile SMEs experience productivity losses due to high defect rates and reactive maintenance in sewing operations. This study proposes a Lean-TPM 4.0 framework integrating a CNN-based Digital Poka-Yoke and an LSTM predictive maintenance model, validated through discrete-event simulation. The Digital Poka-Yoke achieved 59.6% accuracy, suitable for proof-of-concept under limited SME data conditions. Simulation results project defect reduction to 8.1%, MTBF increase beyond 150 hours, and efficiency improvement to approximately 86%. The LSTM model obtained an RMSE of 6.2 hours, supporting preventive maintenance planning. The findings demonstrate the feasibility of combining Lean and accessible Industry 4.0 technologies in textile SMEs.

Instituciones: Universidad Peruana de Ciencias Aplicadas