[Bmasse]

Hello,


Currently doing my PhD in rocket propulsion, I am trying to find a way to simulate solid propellants burning. I saw this video made in Converge CFD.
I wanted to know if there is any tutorial or help for setting up such User Defined Function for a surface regression model for the solid propellants. I understand that the UDF is replacing the combustion model and only a chemical equilibrium model is used to have the product of the combustion.
I have with my research center, experimental data of a propellant (burning rate in fct of Pressure and density in fct of Pressure). I can see how the equation are used and how it can be theoretically implemented but I am missing the point on how to implement such complex UDF with a moving wall in a transient simulation.
If there is any help or guidelines for that kind of simulation I would gladly take them !

Thanks in advance for any tips !


Baptiste



EDIT_1: I am gonna try to do a follow up of my findings for that problem. If someone can guide me if my reasoning is wrong that would be of big help!
Through more searching with UDFs and more thinking on how to implement it i found some way that may works.
The UDF : gradient_regression with MOTION seems a good lead to move the interface between flow and solid grain. However the experimental data of the regression rate I have is in fct of the pressure. I therefore need cell input on the boundary to calculate and use a surface regression model to find the translation vector through the UDF.

From these calculation I need to add a CONVERGE_PROFILE on each cell to give the normal velocity and the density of the exhaust gases (calculated with continuity equation and Vieille law). That on each cell from the solid propellant boundary.
Then iterate this at each time step.
The chemical equilibrium solver will then be used on the side the calculate the product of the combustion of the solid propellant and interaction with the combustion chamber medium (and temperature ?).
Equations are quite clear (easy part of that job) and I will try to do some tests with those UDFs to see if they may suits my needs.

[ksrivast]

Hello Baptiste,

As you have found out, the simulation you are interested in makes use of User-defined functions (UDFs) to evaluate surface regression rates and move the solid propellant grain surface during the simulation using predictive models. UDFs also also used to source the resulting gaseous fuel as a species source along the grain boundary surface.

We recommend you refer the sample grad_reg_motion.c UDF. The grain surface boundary can be provided with a USER wall motion type and this UDF can be modified to move the grain surface triangles, using a global surface regression rate model based on Vielle’s law which evaluates burn rate as a function of surface pressure. Our many other sample UDFs can be looked into to find out how to access cell-center/boundary pressure values through UDFs.

Calculations of burn rate at the start of every time step can be used to specify a regression velocity on the grain surface boundary and evaluate a volumetric species source term for the gaseous fuel and source it using UDF sources (through source.in) into the cells along the grain surface.

We're happy to further assist you with such setups and provide you with more details if needed as this setup is shareable. Please reach out to support@convergecfd.com. Please use your official email for all correspondence with Convergent Science. Please mention your request, attach your current case setup and add the cfd-online thread, as reference.

Hope this helps.

Sincerely,