Hello!
We are currently doing research that models deformation of a membrane, and we are using COMSOL Microfluidics 5.2a. We want to know the feasibility of our design we present, as well as feedback on the configuration of our solution. Please find attached a sketch of the diagram we plan to implement (not drawn to scale). We use a large rectangle with dimensions 30 μm (width) by 300 μm (height) to represent a fluid channel, through which air passes. We then created a membrane with dimensions 0.1 μm (width) by 100 μm (height), in the middle of and bordering the left side of the channel. The material used for the membrane is PDMS (Polydimethylsiloxane). Two fillets are added to the top and bottom of the membrane to create smoothness to those sides. We want to measure the deformation of the membrane using a time-dependent study. We have based our simulation on that presented in the FSI handout, and we would like to create similar results (see link below for handout):
www.comsol.com/model/download/...luid_structure_interaction.pdf
We have used the same settings for inlet and outlet as in the handout, with parameters and variables defined accordingly. Because the structure of our channel is vertical as opposed to horizontal, we changed the input velocity so that H (height) was changed to W (width). Other than that, no other modifications were made to the inputs. The output was given pressure with the default of 0 Pa. We applied fixed constraints to the fillets, a free boundary condition to the left side of the membrane, and a fluid-solid interface boundary to the right side of the membrane. In addition, we have included an extremely coarse mesh to decrease runtime, as well as a free triangular condition and size to the mesh as explained in the FSI handout. Finally, in the solver configuration, we have set the range of times from 0 s to 4 s in 0.1 s increments.
Overall, we would like to display signs of deformation and receive results similar to those in the handout, except with deformation in our model happening horizontally. We would also like to see deformation displacement values, as well as measure flow rate and pressure (also, create plots) from the deformation. We have been running the simulation for 2 days, and it will most likely take some more time before computation completely finishes. We would appreciate any input on our design. Is this the correct approach? What modifications would make the model more efficient? How can we fix the design so we can receive results as intended, and what should we change to increase efficiency and decrease runtime?
Please feel free to ask me for any clarifications on our design. Thank you!