Hi everyone,

I'm trying to model fluid flow through the intake and exhaust ports and poppet valves of a simple single cylinder IC engine and need a little advice.

I'm using Solidworks' CosmosFlowWorks which doesn't allow for a dynamic analysis whilst the piston is in motion. One suggestion so far has been to split up the cycle into a finite number of 'snapshots' of the fluid flow at various crank angles.

Is this a common method utilised by professional engineers using CFD? I'm an undergraduate and totally new to CFD too, but that idea seems a little flawed as it will not take into account the volume of gas already inside the cylinder from the previous snapshot - can these values be put in manually?

Any other problems with this method of splitting up a dynamic system into lots of static ones? I get the feeling I'm missing something because it seems so simple!

Thankyou!

http://www.adapco-online.com --> Engine Modelling

and

http://www.adapco-online.com/adapco_...ice/index.html

The task you wish to perform has been routine for a while but requires a CFD code that works with a moving mesh. Several commercial CFD codes can do this but I presume yours doesn't? The obvious solution would be to dump it and get a more appropriate one such as that suggested by William. There are also a few research codes that can solve the problem like KIVA which is usually around somewhere in most engineering universities.

I am not sure what you think the static solutions will be. What is driving the flow? Why is zero everywhere not the solution?

If you just want the flow through the valves and not that within the cylinder then you could impose time varying boundary conditions to drive the flow based on known upstream conditions and an estimate of the average conditions within the cylinder. This would mean attaching a bit of your code to work out the boundary conditions to the commercial solver. Again, many commercial CFD codes can do this but can yours?

We're mainly aiming for flow characteristics determined by the valves, hopefully throughout the entire 4 stroke cycle - so both exhaust and intake strokes will need to be analysed in detail. Certainly a few [average] conditions are being assumed, things like cylinder wall temperate in the exhaust stroke, and known variations for valve lift and piston position according to crank angle.

But the Solidworks CFD add-on does not appear to even mention the idea of inputing your own math models into the system, so I'm guessing changing parameters as I go along can be done, but manually!

I think the issues I'm facing are very Solidworks focused, if anyone has used it to simulate a moving engine and could give me some hints, that would be very useful, but I understand that there are better CFD packages out there so the chance of me finding an answer here is slim. =)