Using Ansys Fluent’s Gradient-Based Optimization

There is a new workflow that has been developed for the Fluent CFD solver.  It is called gradient-based optimization.  It uses the adjoint solver, which computes the linearized derivatives of a single output variable with respect to all the input variables.  It then calculates separate sensitivity fields for the inputs.  Based on the sensitivity fields, it determines which inputs to change to maximize the desired change in the output variable.

The optimization tool is accessed through the Design tab in the Fluent menu.

There are several observable types that can be optimized for:

The first step in the process is to calculate a steady state solution of the problem.  Once a converged solution has been obtained for steady state solution, an adjoint solution is evaluated to either maximize or minimize the desired observable.

Once the evaluation is completed, the adjoint solution is calculated.

The next step is to use the Design Tool menu to define the wall boundaries that will be modified by the optimization process and what portions of those boundaries.

To perform an individual iteration in the optimization process, click on the Calculate Design Change button in the Design Tool window.  If you are looking to achieve a larger change to the observable, series of iterations will need to be run.  This can be done automatically using the Gradient-Based Optimizer tool.

To test out the capability of this new optimization tool, I ran a simple model of a u-bend pipe and optimized it to reduce the pressure drop through the bend by 40%.  The initial solution of the pipe resulted in pressure contours shown below.

When the optimizer was run to reduce the pressure drop through the model by 40%, the optimization history is as follows:

The resulting pressure contours and pipe geometry are shown below.

The change to the shape of the tube is not something that would be easy to determine without this tool.  It is very easy to use and will allow users to quickly optimize the geometry of their designs.

As you can see, this new capability allows one to quickly optimize flowpath shapes to accomplish optimization objectives. Hopefully you have found this useful and we encourage you to explore this and other enhancements to Ansys Fluent.

All Thing ANSYS 054: Talking CFD – Discussion on the Current State of Computational Fluid Dynamics with Robin Knowles

 

Published on: January 13th, 2020
With: Eric Miller & Robin Knowles
Description:  

In this episode we are excited to share an interview done with host and Co-Founder of PADT, Eric Miller and host of the Talking CFD podcast Robin Knowles, regarding the history of PADT’s use of simulation technology as a whole, and the current state of all things CFD.

If you would like to hear more of Robin’s interviews with various other CFD based companies both small and large, you can listen at https://www.cfdengine.com/podcast/.

If you have any questions, comments, or would like to suggest a topic for the next episode, shoot us an email at podcast@padtinc.com we would love to hear from you!

Listen:
Subscribe:

@ANSYS #ANSYS

All Things ANSYS 049: Predicting & Controlling Environmental Pollution with ANSYS Simulation

 

Published on: October 21st, 2019
With: Eric Miller & Clinton Smith
Description:  

In this episode, your host and Co-Founder of PADT, Eric Miller is joined by PADT’s CFD Team Lead Engineer Clinton Smith for a discussion on how ANSYS fluids tools are being used to help predict and control environmental pollution. This information is helping engineers in a variety of ways, such as understanding the formation and dispersion of pollutants such as NOx, SOx, CO and soot.

If you would like to learn more about what this application is capable of, check out our webinar on the topic here: https://www.brighttalk.com/webcast/15747/374571

If you have any questions, comments, or would like to suggest a topic for the next episode, shoot us an email at podcast@padtinc.com we would love to hear from you!

Listen:
Subscribe:

@ANSYS #ANSYS

All Things ANSYS 046: The Founding of CFX

 

Published on: September 9th, 2019
With: Eric Miller, Paul Galpin, & Brad Hutchinson
Description:  

In this episode, your host and Co-Founder of PADT, Eric Miller is joined by Paul Galpin and Brad Hutchinson, two founders of the Computational Fluid Dynamics (CFD) simulation tool now owned by ANSYS, called CFX. They discuss how they initially got into the world of simulation, the current state of CFD, and what is important to be aware of as it continues to grow and develop.

If you would like to learn more about what’s new in the latest version of CFX, check out PADT’s webinar on fluids updates in ANSYS 2019 R3 here: https://www.brighttalk.com/webcast/15747/369903

If you have any questions, comments, or would like to suggest a topic for the next episode, shoot us an email at podcast@padtinc.com we would love to hear from you!

Listen:
Subscribe:

@ANSYS #ANSYS

Gone Skiing: Aerodynamics – Does It Matter Which Way Your Skis Are Pointing On Your Roof Rack?

I was on the gondola up at Keystone for night-skiing a week ago, after a long day at Beaver Creek, because the last thing I am going to do at 3:00 pm is try to make it back to Denver, as everyone knows it’s hardly more than a parking lot at that point. As it gets later, there’s nothing like a solo gondola ride, however, a solo ride would stop this story right about now.

On the gondola, I overheard a conversation where one gentleman was discussing how he was unable to open the hatch of his vehicle when his skis are in his roof rack. That’s fair, I know older WRX wagons with the spoiler would not be able to open with skis on the roof no matter what, so I figured that was the case. It turns out, that was NOT the case. The reason his hatch would not open was that he orients the skis with the tails forward because it is ‘more aerodynamic’ that way… I was skeptical, but held my tongue, knowing that I had the tools at my disposal to investigate!

I decided to make a model that would allow me to simulate various conditions to get to the bottom of this. My initial hypothesis is that the addition of the ski rack and crossbars is what has the largest effect on aerodynamics, and orientation of the skis probably has a negligible effect after that. As a side note, I am solely concerned with aerodynamics in this case, and am not worrying about the amount of the ski’s base material that is exposed for a given orientation. I am of the mindset that tree trunks and hidden rocks on the mountain are more of a danger to your bases than small rocks on the highway anyway. If you are waiting to comment, “Just get a roof box!”, I understand as I own both a box and a rack at this point, and they both have their advantages, and I will not be exploring the aerodynamics of a box…

…yet…

I was able to start by finding some faceted geometry of a Subaru Forester online (I’m from Colorado, can you tell?) and was able to import that into ANSYS Spaceclaim. Once in Spaceclaim, I was able to edit the faceted geometry to get nice exterior panel surfaces, which I then combined to get a single clean faceted exterior for the car.


Faceted Forester Geometry (Equipped with factory side rails)

After that, I used Spaceclaim to generate the remainder of the rack and skis, including crossbars, a ski rack, and a pair of skis (Complete with the most detailed bindings you have ever seen!). I made a combined part of the crossbars, rack, and skis for each one of my orientations, as this allows me to report the forces on each combined part during the simulation.


Added CAD geometry for the crossbars, ski rack, and a pair of skis

For the simulation, I used ANSYS Discovery Live, the newest tool from ANSYS that allows for instant and interactive design exploration. This tool lets me actively add my CAD geometry and shows results in realtime. I was able to start with just the car and then add and swap my ski/rack geometry with simple button clicks. With traditional simulation tools, I would have needed to create a mesh for each one of these cases, analyze them one at a time, and the post-process and compare results after the fact. After launching Discover Live, it’s as easy as selecting the type on analysis I want to run.

The various types of solutions that can be done in ANSYS Discovery Live. For the purpose of this blog, I am using ‘Wind Tunnel’

Once I have selected ‘Wind Tunnel’ for my solution, I can select my geometry, and then am prompted for the direction of flow, as well as selecting the ‘floor’ of my domain. Once that is done, results show up on the screen instantly. I only needed to modify the flow velocity to ~65 mph. I am most interested in the force on the faces of the combined crossbars, rack, and skis in each orientation, so I created Calculations for each one, which is done by simply selecting the part and using the popup toolbar to create the graph.

Popup toolbar allows for the quick creation of solution calculations

I was already off and running. I ran each one of the cases until the force plot had become steady.

Car Only
Skis Tips Forward Orientation

Skis Tails Forward Orientation

Seeing that the force results for the Tips Forward vs. Tails Forward cases were very similar, I decided I should also run a ‘Bases Up’ Orientation, even though I STRONGLY advise against this, as UV wrecks the base material of your skis/snowboard.

Ski Bases Up and Tips Forward Orientation

In addition to the contour plot shown in the images above, you can also use emitters to show streamlines and particle flow, which also give some pretty neat visualizations.

Streamlines shown on the Tips Forward orientation

Particle Emitter shown on the Tips Forward orientation

The graph plots show values for the Total Y Force for Tips Foward, Tails Forward, and Bases Up orientations to be 37.7 N, 39.1 N, and 37.1 N, respectively. Using Discovery Live, I was able to quickly run all 3 of these simulations, showing that there is not a major difference in the forces on the ski rack between the three orientations. So, put the skis on the roof in the direction that makes life easiest for you, and keep those bad boys paired to protect your bases from the sun, because splitting them isn’t going to help with aerodynamics anyway!

Next steps would be taking a specific case and running in 2D, then 3D, in ANSYS Fluent.