Hi everybody

I'm really dissapointed with this problem. I'm trying to simulate a Hydrogen fuelled Internal Combustion Engine, but I'm lost. I have tried many things. I use a Source Point of energy to simulate the ignition-Spark. It works well. I use a Hydrogen Air mixture, and I call the "hydrogen air FLL STP and NO PDF" reaction.I use the Flamelet model. Then i found 3 ways: A) I initialize mixture fraction=0.029. When simulation is finished, combustion process works (i don't know if works correctly but works), but this initialisation means that the 2,9% of the combustion is done at t=0. So I can't initialize each component independently. B) In the Fluid models tab, i put H2 with transport equation instead of automatic (component details). Then H2 initialization; massFraction=0.029. Simulation initializes well but there isn't a combustion process. C) I try calling a more simple reaction, like "hydrogen Air". Combustion happens, initialization is good, but combustion happens in the Sixth part of the total volume.

What can i do?Radation model is necessary?How can I initialize the reaction with NO formation?Fluid models?

A time ago, i talk with Glenn Horrocks about something like this. My problem is: HOW CAN MODEL THIS IN CFX????????

thank you all

Hi,

Have you done a simple validation case for hydrogen combustion? I would make sure you can model a simple case correctly before I model an IC engine combustion.

There should be plenty of validation cases in the open literature.

Glenn Horrocks

hello Glenn,

I have try Hydrogen Air combustion model in a steady state simulation and worked quite well, however, after a time I realize that i wasn't doing a correct initialisation of the mixture (or components). This is the main problem. What kind of validation cases can I probe?? please, help me... I read, and search for something that I may forgot but I don't found anything in the literature.

Hello,

OK, I have found more or less the meaning of the Flamelet libraries after long time. These are built in a way that mixture fraction changes between Mixt.Fract=0 ( Air ) and Mixt.Fract=1 (H2), logic. But middle values are only for reacted fractions of this 2 sides. For example; I want to initialize H2 massFract=0.029 (stechiometric in air) O2mf=0.226 N2mf=0.745 and then simulate combustion. If mixt.Fract=0.029, in the initial mixture we found 15% of H2O, due to a 2,9% of reacted products... But what i want is a pure initialization.

WHAT CAN I DO?????

1- re-do the flamelet models for my initial composition???

2- I have found that in CFX11 there is a SPARK MODEL (Bvm), but i think it uses the same libraries and so the same restrictions, for my problem.CFX11 is better for this problem???

3- Should I create the reaction model and the 18 or 19 reactions that form the H2-Air combustion by my own???How can i do???I will need special coeff. and constants (similar to NASA coeff.)??? Can I simplify this using Hydrogen-Oxygen library and adding others???

!!! PLEASE HELP ME!!!! It's very important for me

thank you all

Hi,

Sounds like you need to talk to a CFX support person. Your question is a little specialised and in-depth for the forum.

Glenn Horrocks

Have a look at here: http://www.websitetoolbox.com/tool/p...ost?id=1732718.

They did a large project on the hydrogen ICE. ANSYS CFX did work on the modeling of such engines. May be it was one of large europe offices (UK or Germany)? You should try to reach them directly.

OK

Thank you very much. Maybe I should leave down this proyect...It's a joke. However seems a high effort research. Maybe they could help me but...Maybe I'm trying to do something very difficult????

Thanks I'll think about it.

Meanwhile, I'll continue with more simple reactions like, Hydrogen Air Library (without NO model). Only 3 elements, and a product (N2,O2,H2,and H2O). Finite rate or Eddy dissipation model??? wich is better?? HOW CAN I MODEL IT????and HOW CAN I CREATE A SPARK SIMULATION IN CFX10???

Maybe too many questions... THANKS ALL

Hi,

Anything involving combustion modelling is complex. They are difficult simulations to run and even more difficult to get accurate.

Glenn Horrocks