Multi-objective Optimization Based Viscosity Prediction for Inks in Direct Ink Writing Numerical Simulations

Direct ink writing (DIW) is one of the most popular additive manufacturing (AM) techniques. To fully understand the DIW, numerical simulations are used to model this process. The prediction accuracy of DIW numerical simulations is significantly influenced by the viscosity expression accuracy of inks. However, the previous works failed to realize an accurate viscosity prediction due to the lack of model selection and the low-accuracy parameters determination. Herein, this paper develops a novel multi-objective optimization based viscosity prediction method for inks in DIW numerical simulations. Firstly, rotational shear rate sweep tests are conducted to obtain shear rate-viscosity data. Then, four empirical shear-thinning viscosity models are selected as alternative models. The question of determining model parameters is then transformed into a multi-objective optimization problem (MOOP), and NSGA-II and TOPSIS are deployed to determine the model parameters and the final viscosity prediction result. Afterwards, the proposed method is validated by comparing the simulation and experimental results of profile of freeform extruded filaments (FEF). The proposed viscosity prediction method is validated as the predicted result using a cellulose-based ink has the minimum difference between simulation and experimental profile of FEF with 4.31% relative error. The work overcomes the limitations of the previous works in viscosity prediction of inks and demonstrates an effective and accurate approach for viscosity prediction of inks in DIW numerical simulations.