Enhancing adhesion strength of a solvent-free lamination unit for flexible packaging applications using Taguchi based optimization

Abstract

A solvent-free (SF) laminating unit was modelled with the objective to demonstrate the influence of the various process parameters to improve adhesion between polyamide (PA) and polyethylene (PE) films. The effects of the process parameters were investigated using an experimental design based on a fractional factorial method. Taguchi’s L18 orthogonal array (OA) was considered to study eight process parameters. Taguchi’s L18 OA is capable of studying seven factors at three levels, namely: rewind tension (RT), taper tension (TT), surface energy (SE), coating weight (CW), machine speed (MS), application temperature (AV), mix ratio (MR), plus one factor at two levels – curing temperature (CT). In theory, assumptions are made when optimising the manufacturing process; however, in practice, these assumptions can distort the output response, thereby affecting precision and accuracy of the DOE method utilised. To limit the system and part-to-part variation in this study, a Gage repeatability and reproducibility (R&R) was performed to eliminate variability within the system. Furthermore, the Gage R&R study was performed to maximise the output adhesion strength (AS) with a higher degree of accuracy and precision. In this study, the multivariate optimisation technique by Taguchi method (TM) was utilised to reduce the effect of the noise factors (NF) on the response while quantifying the contribution of each design parameter through signal-to-noise (S/N) ratio. A model was derived to predict adhesion bond strength between two flexible films using the solvent-free laminating unit. The two process outputs that were used to predict the model were AS and tensile strength (TS). Based on the result, the highest AS achieved was 640 N; at the same parameter condition, a TS of 21.02 seconds was achieved. The optimum operating conditions for optimising the solvent-free (SF) lamination process and achieving a high output response for AS and TS were: SE = 44 dyne/cm, MS = 150 m/min, AV at 45 oC, MR = 85 %, TT = 25%, RT = 100 N, CW = 1.5 gsm, and CT = 32 oC. The optimum parameters were employed to generate a model and predict the behaviour of the system at different parameter conditions. Using the optimum operating condition, an output of 613.05 N was achieved with an experimental error of 4.2%. The output on the S/N ratio estimated using the model was 56.12, and the predicted value was 55.57 with 0.97% error. The analysis of variance (ANOVA) was also conducted to illustrate the contribution of controllable variables on the output response. Based on the ANOVA results, the contribution of design parameters to the output response was calculated. The result demonstrated that SE was the most statistically significant variable with a contrition of 70.2%. According to the response table, MS is the second most significant parameter ranked at number 2 with a Delta of 46.4. The AV is the third most important variable with a Delta of 60 and ranked at number 3. The estimated constant for the model using the ANOVA was 444.16 with S:57.39, R-Sq of 96.5%, and R-Sq (adj) of 70.5%. The regression equation was also developed, and analysis done with the assumption that there is limited interaction between variables. The two polymeric films used in this study were polyamide (C12H22N2O2) n and polyethylene (C2H4) n materials. A two-component polyurethane (PU) adhesive which consists of resin and hardener were used as SF laminating adhesive heated at 40 oC dosed at 35 oC. The hardener in this case was used as a catalyst to drive the curing process. The relationship and the interaction between the variables were also studied. Based on the result obtained, an increase in surface energy (SE) of the PE causes an increase in the AS and TS. Therefore, it was recommended that the SE of the PE should be always kept at level 3 which is 44 dyne/cm. Under such conditions, the molecules of the film are activated and rearranged in order to prepare them for adhesion. When process parameters were set according to the achieved operating condition, waste was reduced and more material was recovered since adhesive and cohesive forces between the PE and polyamide (PA) were improved. Chapter 6 quantified the value of all resources utilised in this study. The profitability and cost of waste generated were highlighted using the current value stream map (CVSM). A new process is mapped using one of the lean tools called the future value stream map (FVSM). Upon achieving the best operating level, the Taguchi quality loss function (QLF) was used to validate the effectiveness of the generated model. The QLF was examined to reduce process variation and improve consistence. A new experiment was conducted to monitor deviation, and a normal distribution curve was plotted to assess the distribution of the actual output response with respect to the desired new target (640 N).