A component-based theory for sewing thread consumption: modeling overlock stitches via needle and looper kinematics

Abstract

Accurate estimation of sewing thread consumption is critical for cost optimization, material management, and waste reduction in the denim garment manufacturing sector. This study proposes a novel, component-based methodology that decouples thread consumption from specific sewing machine models by focusing on needle and looper kinematics. Four overlock machine configurations (utilizing 3, 4, 5, and 6 threads) were evaluated using three denim variants and a lining fabric, with areal densities. Principal component analysis (PCA) was implemented for dimensionality reduction, explaining 80.11 % of the total empirical variance (V1 = 68.23 %; V2 = 11.88 %) and successfully condensing the 18 initial response parameters into four functional clusters monitored by specific proxy variables (3RNT, 3LLT, 5SLT, and 6HLT). Full factorial regression models revealed high explanatory and predictive power for the right needle thread (3RNT), achieving an R ² of 97.90 % and a predictive R ² of 94.71 %. The 6-thread higher looper didn’t behave as expected because of friction and tiny movements that happen when threads are packed tightly together. But when we tested it in a real-world setting, like sewing trouser seams, the results were very good. The errors were really small – only 0.68 % for the threads that the needle uses and 3.43 % for the threads that the looper uses. This is much better than the 5 % error rate that’s usually allowed in industry. So, the system we developed is a strong tool for planning production automatically. It can help make sure that things are made correctly and efficiently, which is important for making clothes and other textiles.