Statistical optimization of a solvent-free laminating unit for enhanced adhesion strength

The increasing demand for environmentally sustainable and high-performance flexible packaging has accelerated the adoption of solvent-free lamination processes, particularly for multi-layered films such as polyethylene and polyamide bonded with polyurethane adhesives. Achieving optimal adhesion strength (AS) in solvent-free lamination remains challenging due to the complex interplay of processing parameters. This study employs Taguchi’s design of experiments (DOE) methodology to statistically optimize eight key parameters influencing AS, including application temperature, curing temperature, coating weight, machine speed, rewind tension, taper tension, surface energy, and mix ratio. An L ₁₈ orthogonal array was used to reduce experimental runs from 6,561 (full factorial design) to 18 while maintaining balanced parameter representation. Signal-to-noise (S/N) ratio analysis identified surface energy as the most influential factor, followed by machine speed and application temperature. ANOVA confirmed the statistical significance of surface energy (P = 0.047), accounting for 70.25% of the total variance in AS. Linear and quadratic regression models were developed to validate predictive accuracy, yielding R² values of 85.75% and 96.53%, respectively. A confirmation test under the optimized conditions predicted an AS of 646.94 N, closely matching the experimental value of 642 N with an error margin of 0.76%. The results demonstrate the effectiveness of Taguchi-based optimization and regression modeling in improving adhesion performance while minimizing experimental effort in SF lamination systems.