Flow simulation of microinjection moulding of microfluidic device
Mohammed Faizal Mohd Fauzi
Date of Issue2009
School of Mechanical and Aerospace Engineering
Singapore-MIT Alliance Programme
Simulation has been widely used to predict the performance of an injection moulding process and the part quality derived from it. While simulations for a macro-scale, conventional injection moulding are rather common and accomplished, simulations on the micro-scale level are still extensively being researched. The rheology of polymer flow in the micro-scale level has made simulations of micro-injection moulding ambiguous at times. In this study, using a commercial injection moulding software package called MOLDFLOW MPI 5.0, a comparative study of the accuracy of results based on different mesh types used was performed. A CAD model of a micro-fluidic device with micro-channels of 50 μm by 40 μm was imported into MPI 5.0. 3 models, each with differently assigned edge lengths and each subjected to 2 different mesh types namely Fusion smooth and Fusion match mesh, underwent simulations for filling and packing phases based on modified process parameters obtained from a concurrently running experiment. Fill time, wall shear stress and volumetric shrinkage were selected to provide quantitative and qualitative comparison of the accuracy of the results between the 2 mesh types in each output. Quantitatively, the match mesh model displayed better results accuracy compared to the smooth mesh model. Results predicted by the match mesh model were less sensitive to element quantity variation, hence, more robust. Qualitatively, there was no clear indication of which mesh type yielded a more accurate interpretation of the results. Finally, a simulated Taguchi design of experiment (DoE) for overall part quality was performed. Only the ranking of main effects were used to compare with the significant factors from the experiment. Of the 3 factors, namely melt temperature, hold pressure and mould temperature, highlighted as significant in the experiment, the match mesh model ranked them within top 3. However, the smooth mesh model predicted only melt and mould temperature in its top 3 list of significant factors. This proved the superiority of the match mesh over the smooth mesh in simulating the injection moulding of a micro-fluidic device.