A well designed ductwork system is essential in delivering the best interior comfort at the lowest cost. Ducts are simply passages that are used to deliver or extract air. Most often you find them in heating, ventilation and air conditioning (HVAC). Computational Fluid Dynamics can be a major asset in designing ducts that are required to achieve certain flow properties. Instead of repetitive prototyping, CFD can significantly reduce design costs and save a lot of time. OnScale Solve provides a top tier CFD Solver that is simple to use and highly parallelizable, allowing you to run fast simulations to any CAD .
So, what is this blog about ?
In this blog, we will discuss how to set up a CFD study on OnScale Solve. Here, we will use a converging duct design to do so. However, more complicated geometries can also be treated the same way. A step-by-step video tutorial is available in the video below.
But before we get started, let’s take a look at the dimensions of the duct we have:
You will see next how these dimensions play a critical role in setting up the study. So let’s go ahead and discuss the six parameters you need to consider when setting up the a CFD study on OnScale Solve.
1. The fluid material you want to use. This material is associated with a default density and kinematic viscosity at standard temperature and pressure. These properties will be fixed throughout your simulation. While the current library only includes water and air, you can simply create your own material by changing the two properties and renaming your material. So, it’s really more of an open library! Here, we want to choose “Air” with default values. This is done under the Modeler tab.
2. Flow Load. This is usually set to be an averaged velocity value (m/s) across the inlet. Here, we want to choose 1 m/s. Onscale Solve requires at least one geometry face to be assigned as a Flow load. If your problem requires a flow rate boundary condition, you need to simply divide that by the area of your inlet and plug in your answer as velocity magnitude. Note that you need to switch from Mechanical Physics to Fluid Physics before you start setting up our boundary conditions.
3. Fluid Pressure constraint. This is usually set to be an average pressure value at the outlet. In OnScale Solve, this is set as a differential pressure and expressed in units of Pa. Here, we want Pressure=0 Pa. At least one geometry face needs to be assigned as a Fluid Pressure Constraint.
The next 3 parameters will require a dimensional insight of our design
4. Contraction Area Ratio. This is defined as the ratio between the inlet area and the smallest cross sectional area. As the smallest cross sectional area is at the outlet, the Contraction Area ratio is calculated to be 1.4.
5. Characteristic length. Before we talk about Characteristic length, let’s talk a little bit about meshing. In OnScale Solve, meshing is handled by an automesher that creates 5 different mesh densities as soon as you click “Generate Mesh”. All mesh cells are constructed of identical hexahedral cubes. For each mesh density, there is a fixed number of cells that fills the characteristic length you select. Below is a table that shows the count of these cells at each mesh density:
| Mesh Size | Number of Cells in a Characteristic Length |
| Very Coarse | 15 |
| Coarse | 21 |
| Medium | 30 |
| Fine | 42 |
| Very Fine | 60 |
Thus, the cell size corresponding to your mesh is a byproduct of the characteristic length and mesh density you choose.
For internal flows, the Characteristic Length is recommended to be the hydraulic diameter of the inlet. In this problem, It can be approximated to be the inlet width of the duct which is 0.7 m.
6. Duration. This is the simulation time that the algorithm is going to solve for. We recommend that your duration satisfies the condition below:
where the Net Distance is the approximate distance that a fluid parcel would traverse. In this problem, it is simply the length of the duct. Here, we chose Duration= 25 seconds.
Now you are ready to mesh and run!
But how do all these parameters tie together?
The number of time steps that the algorithm loops through is directly related to these parameters as follows:
where C=0.577735 and M, the Mach number, is set equal to 0.15 for all internal flows. Thus, for this problem, the solver will iterate through 17,322 timesteps. Note that the minimum number of timesteps that the solver requires to run is 100. So you can use the formula above to guarantee that your simulation satisfies this condition.
Want to run more CFD studies on OnScale Solve? Check out this tutorial of an FDA Nozzle fluid simulation and this webinar on Manifold flow simulation.
