Optimizing Design Performance with Baffle Placement

FLOW-3D (x) case studies

Optimizing Design Performance with Baffle Placement

Optimization Goal

Slosh damping in liquid storage tank with baffles

Optimization Challenge

Generate a workflow that allows a user to run multiple iterations of a sloshing simulation to find the optimal ring baffle position that maximizes damping in a large, cylindrical tank. The case simulated here is based on the physical experiment by Maleki and Ziyaeifar (2008)1.

Liquid storage tank schematic

Optimization Solution

The simulation represents the free decay of sloshing in fluid that is initially oriented 5 degrees from horizontal at a fluid height of 0.6 m in a vertically oriented cylindrical tank. The position of the ring baffle can translate in the z-direction. The objective is to find the location of the baffle that results in the greatest amount of slosh damping. Each simulation runs approximately ten minutes on 12 CPU cores.

A budget, or number of simulation iterations allowed, of 30 iterations is specified. FLOW-3D (x) runs 30 simulations to generate a response surface which represents the behavior of the system. This allows the best solution to be found.  

FLOW-3D (x) Workflow

FLOW-3D (x) uses nodes to construct automated workflows for the optimization. At the start of this workflow an initial baffle location in the z-direction is given. The baffle location is then allowed to translate vertically between prescribed bounds. Each simulation is then fed into a FLOW-3D node that executes iterative simulations. The results from the simulation are then connected into a calculator node that performs the damping calculation. The optimization engine will then choose another z coordinate of baffle based on the ever-improving response surface and continue with another simulation run.  

FLOW-3D (x) optimization workflow baffle performance design

Results

Using FLOW-3D (x)’s built-in data analysis tools, a graphical representation of results immediately reveals that a baffle height of 0.55m provides the maximum damping ratio. The simulations and iterative design features are all automated with the program. Additionally, images and videos of each individual simulation can be set to output.   

Performance design optimization-sloshing

References

1Maleki, A. and Ziyaeifar, M., 2008. Sloshing damping in cylindrical liquid storage tanks with baffles. Journal of Sound and Vibration, 311(1-2), pp.372-385.

Characterizing Contact Tank Performance

FLOW-3D (x) case studies

Characterizing Contact Tank Performance

Performance Goal

Characterize the performance of a chlorine disinfecting contact tank in an activated sludge process with varying bacteria loads. The model is set up based on a paper by Evans and Kothandramm (2000)1.

Chlorine contact tank schematic

Engineering Challenge

Automate a workflow that allows a user to run multiple iterations of a simulation with various bacteria loads and chlorine injection concentrations. This is set up to find the minimum amount of chlorine required to minimize the concentration of bacteria at the system’s outlet.

Characterizing contact tank performance

FLOW-3D (x) Workflow

An initial simulation computes only the hydrodynamics of the system at a given flow rate. A steady state condition is found at 3000s. Using the flow field from this simulation, a restart simulation is created to introduce the reaction kinetics model. Chlorine and bacteria are introduced via the inlet pipe. At the outlet, a flux surface is placed to measure the concentration of the reactants (chlorine and bacteria).

FLOW-3D (x) contact tank performance

FLOW-3D (x) uses nodes to construct automated workflows for the optimization. The inflow bacteria concentration variable is given a range from 1.0e9 to 10.0e9 and the chlorine concentration variable is given a range from 1.0e-6 to 9.0e-5. A simulation is executed using a FLOW-3D node. The post-processing node extracts the concentrations of chlorine and bacteria from the results and builds a Pareto front which represents the minimum chlorine required to minimize the bacteria at the outlet.

The budget, or number of simulations allowed, was set to twenty and then fifty to demonstrate the effect of increased budget on the quality of the Pareto front. The runtime for a single simulation is approximately fifteen minutes.

Performance Results

Using FLOW-3D (x)’s data analysis tools, Pareto fronts of the bacteria concentration with respect to chlorine concentration are compiled for both the twenty-budget case and fifty budget case. The larger budget gives the optimization engine more time and data to report for a more accurate Pareto front.

Budget of 20 Simulations

FLOW-3D (x) 20 simulation budget contact tank

Budget of 50 Simulations

FLOW-3D (x) 50 simulation budget contact tank

With the Pareto front created by FLOW-3D (x), system designers can quickly assess the amount of chlorine needing to be injected into the contact tank to minimize bacteria at the tank outlet under a given flow rate. For other flow rates, the original simulation defining the tank hydrodynamics would be run at the new flow rate and the same FLOW-3D (x) project would be run, possibly with different ranges on the inlet bacteria and chlorine variables.

References

1 Evans, R. and Kothandraman, V., “Design and Performance of Chlorine Contact Tanks, “Circular 119 STATE OF ILLINOIS DEPARTMENT OF REGISTRATION AND EDUCATION 2000.

Calibrating Simulation Parameters

FLOW-3D (x) case studies

Calibrating Simulation Parameters

Goal of Calibration Study

Investigate the impact of various numerical parameters on the simulation of air entrainment on a stepped spillway.

Engineering Challenge

Generate an automated workflow that allows a user to easily study the impact of numerical parameters on the simulation of incipience of air entrainment and help calibrate them to match experimental data1.

The parameters to study are mesh size, turbulence model, turbulent length scale, and dynamic vs fixed turbulent length scale. Additionally, FLOW-3D (x) will create an image of the entrained air concentration at the last time step and an animation showing the evolution of air entrainment in the simulation.

FLOW-3D (x) Workflow

The simulation is set up with sampling volumes on steps number 3, 4, and 5 to report the amount of entrained air. FLOW-3D (x) uses nodes to construct automated workflows. The first node is used to read the simulation parameters from a .csv file. The parameters are then sent to a FLOW-3D node to execute the simulation. The post-processing node extracts the entrained air volume from sampling volumes on each step on the spillway, creates an image of entrained air at last time step, and creates an animation of air entrainment. The last node writes the reported entrained air values from our sampling volumes to a .csv file.

The budget, or number of iterations allowed, was set to 18 as there are 18 sets of parameters specified in the parameter definition input file. The runtime for a single simulation is dependent on the mesh size used in each iteration.

Stepped spillway calibrating simulation parameters
FLOW-3D (x) stepped spillway parameter study

Parameter Study Results

Using FLOW-3D (x)’s data analysis functions and automatic image generation allow for quick visual assessments and verification of results. Additionally, air entrainment values at each step for each simulation run are easily accessed from the reported .csv file. Batch execution was used to save time on the optimization study.

Before Calibration

Stepped spillway entrained air after calibration

Entrained air at the last time step from a simulation with a mesh size at 0.01m, a k-ω turbulence model, and a turbulent length scale equal to 0.005m

After Calibration

Stepped spillway entrained air before calibration

Entrained air at the last time step from a simulation with a mesh size = 0.005m, a k-ω turbulence model and a turbulent length scale equal to 0.005m. Incipience of air entrainment with 2x finer mesh compares better with experimental results than 0.01m mesh

References

1 Felder, Stefan (2013). Air-water flow properties on stepped spillways for embankment dams: Aeration, energy dissipation and turbulence on uniform, non-uniform and pooled stepped chutes. PhD Thesis, School of Civil Engineering, The University of Queensland.

Optimization of a Tilt Pour Casting

FLOW-3D (x) case studies

Optimization of a Tilt Pour Casting

Optimization Goal

Optimize a tilt pour casting of a combustion engine piston to minimize air entrainment.

Engineering Challenge

The objective of this optimization is to minimize the amount of air entrainment and turbulence during a tilt pour casting. This objective will be achieved by modifying the profile of the tilt filling motion. Minimizing air entrainment and turbulence will reduce the possibility of defects being introduced into the casting. Additionally, optimizing the filling parameters can increase quality without an increase in costs.

Tilt pour casting before optimization

Optimization Solution

Generate a workflow that allows a user to run multiple iterations of a tilt pour casting simulation. FLOW-3D (x) uses nodes to construct automated workflows for the optimization. Three process variables (start of rotation, duration of rotation, and volume flow rate) serve as variable inputs and are varied for each iteration of the simulation.

Tilt pour optimization solution

FLOW-3D (x) Workflow

The Excel spreadsheet node is used to define a table of the start and duration of the mold rotation, and the volume flow rate of the filling profile. A calculator node converts the profile description into a movin.inp file that prescribes the ladle motion. Next, a FLOW-3D node is used to execute the simulations. The output of each simulation is the total fill fraction and entrained air volume fraction, which is extracted from the results by the post-processing node. The fill fraction is used as a dynamic termination condition for the simulation, to ensure the mold is completely filled. A budget, or number of simulations allowed for the optimization study, is set to thirty. A single simulation run is around 15 minutes.

FLOW-3D (x) optimization workflow

Optimization Results

Using FLOW-3D (x)’s data analysis tools, a Pareto front graphical representation of the results reveals which simulation corresponds to the optimal filling profile with the least amount of entrained air and the highest fill fraction. The simulations and iterative design features are all generated autonomously by FLOW-3D (x). Additionally, images and videos of each individual simulation can be set to output.

Before and after of tilt pour optimization

Here is a comparison of the original pour rate and pour duration (left side) and the optimized values on the right. Notice that the pour rate is increased slightly and the pour is completed slightly earlier.

Original pour rate
Optimized pour rate

Here is a comparison of the original mold rotation rate and duration (left side) and the optimized values on the right. You can see that the rotation rate increased and the rotation is of a shorter duration than the original.

Original mold rotation rate
Optimized mold rotation rate

To learn more about FLOW-3D (x), contact our technical sales team.

Sign Up for Our Newsletter

Newsletter Registration Form
Privacy *
Linkedin Youtube Twitter

FLOW-3D and TruVOF are registered trademarks of Flow Science, Inc. in the USA and other countries.

COPYRIGHT 2000 – 2023, FLOW SCIENCE, INC. ALL RIGHTS RESERVED.

Request More Information

FLOW-3D AM WELD Request Info

What additive manufacturing processes do you want to simulate? *
What laser welding processes do you want to simulate? *
FLOW-3D News
Privacy *
CSTsiteisloaded