Turbo YouTube 5 - Solution

Hello everybody, this is Meysa Mehr Al-Alanat. I'm from Ozen Engineering and I've been at Ozen for a year and a half and before that I was in an electric car company called Byton and before that I was five years with Ford Motor Company. This is the fifth of our machinery YouTube shorts.

This is mostly going to focus on solution setup, which is done in CFX- 3. A little about us. We are the spineal partner analysis in California.

We have 17 years of successful operation in the Bay Area, a thousand plus happy clients, a good amount of expertise in Multi physics finite analysis or FEA expertise in computational dynamics or CFD, and also on top of that, electromagnetics which covers both high and low frequency.

Our technology gasket is based on synthetic steel Arkt. Our territory includes of course Oregon and California and Nevada, with headquarters in Sunnyvale and offices in Portland, Oregon and Long Beach, California.

We can sell you the following software for structures: mechanical and code elastina and floors, CFX and Fluent for electronics in high and low frequency categories, FHSS and Maxwell, and tools for semiconductor and optics on the material side, grant for embedded software, and different packages for system analysis, such as Twin Builder and Medini.

Whatever industry you are at, we probably have some tools that can make your job easier or more efficient. Feel free to give us a call and we'll be more than happy to guide you to the right tool.

To go over the solution setup, we went through a couple of slides in previous YouTube shorts and went all the way until meshing. We created an initial design or conceptual design and then brought it into Workbench and meshed with Turbo Grid. Now we've got to do the solution part.

Solution usually is done in CFX-Pre, using a rotating frame of reference or RFR. In some tools like Fluent or Icepack, sometimes they call it MRF, which is pretty much the same thing.

To set up the CFX model, we need to define the fluid domain, which means we got to define fluid material properties, domain motion, and also the physical models that we are going to use, such as turbulence model, heat transfer model, etc.

And after that, we define the boundary conditions and solver settings and a couple of monitors to check on convergence. For domain setting, we need to define its rotation. We need to set the angular velocity and also the rotation axis.

And also we need to define material properties and also reference pressures. For modeling turbulence and heat transfer, definitely SST and total energy are the preferred methods for turbo machinery.

For the boundary conditions, we usually have a passage because the geometry is mostly periodic that you're dealing with. There's an inlet and outlet. And also there is hubs and blade, which is also like a no-slip wall. And on the right and left, you have the periodic interfaces.

The good thing about CFX is that it has a turbo mode. So it does a lot of its work. It's a little bit more complex, but it's a little bit more easy to use. Now, let's see the tool. This is the file we worked with in the last YouTube short.

We generated the geometry for the impeller and also stator inside CFTurbo. We brought it into TurboGrid and meshed that. Now the next step is that to solve it. So we go to Transfer Data. Transfer Data to New CFX. And technically, we can drag and drop also this term. So you have both parts.

But just for simplicity, we just model the rotor for this case. You open CFX-PRE. Here you see that you can manually define all the parameters. You can manually define all these boundary conditions, as I mentioned in previous slides. But there is a simpler way to do that. There is a tool for it.

It goes Tool, TurboMult. First, it's asking whether you're dealing with what's the machine type you are trying to model. You say, OK, it's a fan, an actual fan. And it has defined the z-axis. And the solution is steady state. You click Next. This is R1, which is this whole thing.

This is the whole area. It's rotating. You say, what is the RPM? Let's say you define 376 RPM. You can put any value you want. And then just define it. That easy. Then you push Next. So what's the fluid? Of course, for a fan, I'll say you can choose air. That's a good approximation.

And what are the boundary conditions? You can put different values. You can put the pressure, total inlet, outlet. I will say, let's say, pressure, total inlet, and mass flow outlet, just to put some values on it. But you can choose whatever is more correlated to your model.

You have different options for heat transfer and turbulence. You can choose anything that sounds more relevant to you. And then you have the periodic boundary condition. As you see, these are the interfaces. These two walls are the periodic interfaces, which are the ones that are detected.

You don't need to do anything manual. And then these are the other boundaries. Blade, which is a wall. Hub, which is again a wall. Inlet, which is inflow. And outlet, which is outflow. And the shroud, again, which is a wall. So as you see, quite straightforward.

It just happens automatically for you. And then after that, enter general mode. You don't have any options on that. Just finish. You're pretty much all set. As you see, the interface has been placed. The boundary conditions have been placed.

I think you don't need to do anything else just for the solution output. Maybe we need to put a monitor to make sure we are seeing the convergence properly. To do that, I create, let's say, pressure outlet, P out, to see if the convergence has been achieved. And we define expression for that.

We simply can say, for example, the mass flow average pressure at the outlet. To do that, it's quite straightforward. So I look at mass flow average. Pressure at R1 outlet. That's simple. You click OK. And you have this monitor set up.

You can put anything else that you're interested in to just monitor it and see how good it converges. And you're all set. Save. Close. You can definitely run this model. OK, it's done. Now you go to solution. Sometimes it needs a little patience. It's working OK. We are here.

Just select double precision. And parallel two cores. You can start to run. As you see, the process is pretty straightforward. So if you're interested, you can even do it manually. If you have different parts or different meshes that merge into together, you can bring it in separately.

And then you might need to create an interface in between in more complicated cases. And pretty much it's similar. As you see, it's iterating. It's going toward convergence. Everything is going down. There will be some oscillation. And then it's flattened out.

I don't have that much time to wait for it until the convergence is achieved. But I will say it's like a 20 minute run or so. And this is the end of the current YouTube short on how to run a thermal missionary case in CFX. OK. Thanks for attention.

And if you have any questions, I'll be more than happy to answer. Thank you, and have a great rest of the day. Bye.