Undergraduate and Graduate Use of STAR-CCM+ at Brigham Young University
Steve Gorrell Greg Jensen Tadd Truscott Dan Maynes
STAR Global Conference 2013 18-20 March, 2013
Undergraduate and Graduate Use of STAR-CCM+ at Brigham Young - - PowerPoint PPT Presentation
Undergraduate and Graduate Use of STAR-CCM+ at Brigham Young University Steve Gorrell Greg Jensen Tadd Truscott Dan Maynes STAR Global Conference 2013 18-20 March, 2013 Outline Introduction Coursework Applications of Fluid
STAR Global Conference 2013 18-20 March, 2013
Introduction Coursework
– Applications of Fluid Dynamics – Computer-Aided Engineering Applications
Capstone
– Minimizing drag on locomotive
Graduate Research
– Cavitation in water turbopump inducer
Summary
2
CFD is a useful tool that can help students learn
– Not just fluid mechanics
STAR- CCM+ has features that help students
– CAD to mesh to solution
The risk
– The code gave a solution therefore it is correct – Not understanding why a solution does not converge
Students can do amazing things
– Undergraduate as well as graduate
Warning! Unedited comments included
– Useful to see and appreciate students perspective – Comments do not represent opinion of Faculty or BYU
Class has 3 overarching themes to
– Analytical solutions – Numerical solutions – Experimental solutions
One example sphere in a cross flow
– Use STAR-CCM+ to visualize and predict
drag coefficients and pressure distributions.
4
Student comments: “I appreciated the time we took to learn about CFD, potential flow, and experimental fluid dynamics”
5
Over course of semester use suite of CAx and
Demonstrate the appropriate application of 3D
6
A senior with no CFD or STAR-CCM+ experience
– Preprocessing and Solving based on James Clement’s
Surface Wrapping a Motorcycle video
– Postprocessing based on James Clement’s
Understanding Scenes video
Industries & Investments Workforce Demographics Aging Population US Citizens in Israel Inflation GDP & Wealth Distribution PPP& Exchange Rates
Introduction Strategy Modeling Optimization FEA Mass Motion Sim CFD Visualization
Student Comments about CFD
– Much simpler than we thought it would be, except for
meshing
– Easy way to make awesome images and videos – Makes fluids seem interesting again… barely
Design of Aerodynamic Railroad Trains Objectives
– Design of an aerodynamic freight locomotive – Minimize drag coefficient 9
Student comment “Vector fields allowed for visual confirmation”
Tools Used in STAR-CCM+
– Meshing – Drag analysis – Vector fields 10
Student comments: “Varying mesh sizes saved computational time” “Allowed for amazing pictures during presentations”
Common reasons for failed simulations
– Floating point errors – Meshing issues
Optimate plug-in was unusable due to lack of
Successes
– Able to run iterations on various models – Predict the optimal value through our optimization script 11
Student comments: “Errors during meshing wasted valuable time”
Cavitation
– Local static pressure drops below the
vapor pressure
– Fluid begins to boil and vapor bubbles
form.
– Unsteady two phase flow
Time-accurate simulations an
12
Image taken from July 2004 Newsletter, Lawrence Pumps Inc http://www.lawrencepumps.com/Newsletter/news_v01_i2_july.html
Centrifugal Pump Portion Inducer
Turbopump Inducers
– Designed to work under cavitating conditions
Off-Design Operation
– Inducers are designed for a given flow coefficient – Off-design flow coefficient explored 60% of design 13
Off-Design On-Design 𝜏 ≈ 0.05 𝜏 ≈ 0.02
14
The main reason for the increase in flow
Reversed flow is caused by flow separation on
Off-Design On-Design
15
When the inlet pressure drops to a low enough
16
The cavitation core grows and collapses causing
17
STAR-CCM+ a great tool for university classroom,
Teaching CFD to undergraduates gives them an
– Used while a student – Desirable skill in workforce
Students can do amazing things
– Undergraduate as well as graduate
From CD-adapco
– Lisa Mesaros – Tammy de Boer – James Clement – Bill Kester – Deryl Snyder
BYU Fulton Supercomputing Laboratory
– 12,356 CPU cores and 30.3 TB of memory across 928
compute nodes.