Aerofoil Modelling

Before we can start simulating however, we need to design our aerofoil. This is relatively straightforward, as there is a wealth of aerofoil coordinate data libraries online, and we can import those coordinates into SOLIDWORKS by using the Curve Through XYZ function.

For this tutorial, we will be using a NACA 4415 aerofoil.

You can copy the NACA 4415 aerofoil coordinates from the University of Illinois at Urbana-Champagne aerofoil database website, or you can obtain it from the AirfoilTools website here. Note, the AirfoilTools site has a nice visualization tool that shows you how the shape of the aerofoil geometry changes as you modify the NACA parameters, which is great if you want to know exactly what those NACA numbers mean.

Excel

The raw coordinates need cleaning up a little before we can import them into SOLIDWORKS.

Open up Microsoft Excel and copy/paste them into the first cell. You will notice that both X and Y coordinates have been copied into a single column, so in order to make them usable we need to separate them into individual columns.

Highlight Column A in Excel, click on Data, and select Text to Columns.

On the first page of the Text to Columns Wizard, we want to select Delimited, if it isn’t already selected by default.  Then click Next.

On the second page of the Wizard, check the Tab and the Space delimiter boxes. This should separate the X and Y coordinates into two columns. Then click Finish.

Now that we have two columns with separated X and Y coordinates, we are going to need to create a third column full of zeros so that SOLIDWORKS can import it. These zeros represent the Z coordinate, but as this is a two-dimensional curve, we have no need for a Z coordinate, and, hence, we set them to zero. Type “0” into the top cell in the third column, and click the little box at the bottom of the cell, or drag the box to the bottom of the data set in order to populate the third column with zeros.

Now we have our X, Y and Z coordinates in three columns. We can click File>Save As and select Text (Tab Delimited) from the drop down menu. Select a location to save the file to, pick a name for your file and click Save.

If an Excel warning appears, just click Yes to ignore it.

You can now close Excel and open SOLIDWORKS.

Loading the Aerofoil Coordinates

Open SOLIDWORKS, open a new part file and on the top of the screen select:

Insert> Curve > Curve Through XYZ Points

This will open the Curve File pane. Click Browse, and then locate the text file containing the cleaned up coordinate data that you exported from Excel. It will load the coordinates into the pane, as seen below.

Click OK, and you will see the aerofoil curve appear in the design window, as seen in the image below.

Of course, being a curve, it is still not useful for creating geometry, so select the Front Plane from the design tree and click Sketch from the Sketch tab.

Now click Convert Entities from the Sketch tab, and in the Convert Entities panel, select the aerofoil curve from the design window.

Next, we want to make a centreline from the trailing edge to just beneath the leading edge. This will represent the chord length of the aerofoil, and once we have constrained it we can alter the chord length at will.

After the chord line is sketched, we need to put another line connected to the last line near the leading edge. This new line needs to be tangential to the aerofoil, as shown below.

Then, we can select both the chord line and the tangent line, and constrain them so that they are perpendicular to each other. Why? Because when we rotate the sketch or extend the chord length, we want it to retain shape, and the perpendicular constraint will ensure that the whole thing remains aerofoil-shaped.

Now that the sketch is constrained, we can just double-click on the chord line and enter a value for how long we want the chord to be. In this case, let’s set it to 1.6 meters.

Congrats! You have now converted your aerofoil curve into a sketch entity. Now we can model our solid aerofoil.

2D to 3D

This part is easy. Simply select the aerofoil sketch and extrude it to 4 metres. This will provide us with a basic constant-chord (i.e., non-tapered), rectangular wing. This type of wing, incidentally, is referred to colloquially as a “Hershey Bar”.

And there it is. Our Hershey Bar wing is now ready for some flow simulation!

Flow Simulation Time!

Load up the Flow Simulation add-in by clicking Tools > Add-ins and checking the SOLIDWORKS Flow Simulation box. Once it is loaded, select the Flow Simulation tab and click the Wizard button to start the Flow Simulation Wizard.

On the first page of the wizard (Project Name), name your project and click Next.

On the second page (Unit System), select your preferred unit system. For consistency, we will select SI units here (m-kg-s). Then click Next.

On page three (Analysis Type), we can select Internal or External study. Internal studies are for simulating flows that are constrained by some kind of vessel, such as a pipe, and external studies are for simulating flows around the outside of a body such as a truck or an aerofoil. So, we click External, and then press Next to advance to the next page.

The next page (Default Fluid) allows us to select the fluid in our study. This is an aerodynamic study, so we select Air from the top list and click Add. Once the default fluid has been added, we can click Next.

We can skip over the next page (Wall Conditions) by clicking Next.

The final page that we need to deal with in the wizard is the Initial and Ambient Conditions page. This is where we set the temperature and pressure of the environment and the velocity of the flow in the x-direction. We have set the temperature and pressure to SSL (standard sea level) values and the velocity in x-direction to 55m/s (about 200km/h).

That’s all we need to worry about with the wizard. Click Finish and the wizard will close.

You will notice that the wizard has created a box around the wing. This is our Computational Domain, where all the magic happens. Think of it as the inside of a wind tunnel. Everything inside it is part of the simulation, and everything outside it is irrelevant.

Note that a larger Computational Domain requires more processing.

Click on Computational Domain on the left hand panel (as seen below) and you will notice six handles appear on the box. Drag these handles until the domain box fits just around the wing model. Be sure to leave enough room at the fore and aft of the wing so we can get some sweet visualization of the fluid flow as it passes around the wing.

Next up, we want to set our goals.

The Goals in SOLIDWORKS Flow Simulation serve three purposes:

1. Defines Design Goals and/or other important criteria
2. Used for Convergence Control
3. Finish the calculation

Being an aerodynamic simulation, we want to set goals that are relevant to this domain. So, go into the left-hand project simulation panel again, right click on Goals, and select Surface Goals. This will bring up a list of parameters that we wish to measure and visualise, and we can select the minimum, maximum and average for each goal.

First, we want to select the faces of the wing that we want included in the study. In the Surface Goals panel, click the blue Selection area to activate it and click all of the faces of the wing model.

Next, go down the list and check the minimum, maximum and average for the following parameters:

Static Pressure
Total Pressure
Dynamic Pressure
Density (Fluid)
Mach Number
Velocity
Velocity (X)
Turbulence Time
Turbulence Length
Turbulence Intensity
Turbulence Energy
Turbulence Dissipation

Note that we have selected Velocity (X), because this is the direction that the flow will be travelling in.

Click the green check mark to exit Surface Goals.

Next, we want to go into the study panel on the left, right click Input Data and select Calculation Control Options.

Check the iterations box and ensure it is set to 100 iterations. It may be that your simulation requires less, or even more. But for now, 100 iterations is fine. This should be enough for the goals to reach convergence. More iterations will generally give a better result, but after a point, the trade-off between accuracy and time-taken simply isn’t worth it. You can run the simulation all day long and the gains to accuracy will become very modest. So, 100 is fine in this case. Click OK to exit.

Now that our simulation is set up, we can run it. You can find the Run button in the top ribbon (as seen below). Click it and you will see the solver screen appear, informing you of how many iterations are left.

Grab yourself a coffee and wait. Depending on how fast your computer is, this could take a while. My computer is rubbish. It will absolutely take a while.

Displaying the Results

Now the calculations have finished, we can go into the study panel on the left and expand the Results section to show us a selection of graphs and plots. Right clicking any of these plots will allow you to insert the plot into the main window.

Cut Plots

The first plots we will look at are cut plots. This type of plot will display a 2D slice (a plane) of the model, and you can drag the green arrow to move the slice along any part of the 3D model.

Right click Cut Plot, and select Insert.

In this instance, I select Front Plane, then I select Contours to show a contourplot. In the Contours section, you can see that the default parameter should be Velocity (X). We would like to see the pressure contours here, so we can click the parameters box and select Pressure.

Click the green check mark and you will see your plot appear in the main design window. You can move the slice along the length of the wing by using the green drag handle and you can rotate the plot as you would do your 3D model. The image below shows an isometric view and a side view. The color code shows how the colors relate to differences in pressure.

Because the results are already loaded into your computer, you can easily switch between data types by clicking the parameter name just beneath the colour scale and selecting new results to display.

So, if I want to change from a pressure contour plot to a velocity contour plot, I simply click Pressure beneath the colored scale (as seen above) and switch it to velocity. The main plot will change accordingly.Note, if you want to see the slice scan along the entire length of the wing, you can right click on Cut Plot and select Play for a little animation.

Flow Trajectories

Cut plots are nice, but they don’t show the holistic view of what is going on; they simply show a 2D slice of the 3D whole. The trajectory plot is more useful for showing behavior over the full length of the wing at any given time. This is more like a wind tunnel with smoke injected into the chamber, which you may be more familiar with from university.

Right click on the Flow Trajectory option in the study pane, and select Insert. This will open the Flow Trajectories pane.

In this pane, select the faces that we want to be a part of the study as we did for the Cut Plot.

In the Number of Points box, type 15 and set the Spacing to 0.03m.

In the Appearance section, we select Static option, and then we select the appearance of the trajectory. In this instance we select Pipes, but feel free to play around here and experiment with different appearances.

Again, in this plot we will be looking at Pressure, so select that from the Appearance section, and then click the green check icon. The plot will appear in the main window, as you can see below.

Here we can see the variations of pressure as the air flows over the aerofoil, and also we can get some idea of the turbulence/vortices created by the wing tip.

Summary

OK, so you’ve done your first aerofoil flow visualization in SOLIDWORKS Flow Simulation.

After this tutorial you should now be able to do the following:

• Import aerofoil coordinate geometry
• Create a solid from imported coordinates
• Set up a fluid analysis
• Run the analysis
• Visualize flows in cut plots
• Visualise flows in trajectory plots
• Switch parameters from inside a plot

There are a lot of different visuals that can be created in Flow Simulation (and so little space to write about all the combinations).

The best way to discover them is to experiment with the different parameters and see for yourself!

About the Author

Phillip Keane is currently studying his PhD at the School of Mechanical and Aerospace Engineering at Nanyang Technological University, Singapore. His background is in aerospace engineering, and his current studies are focused on the use of 3D-printed components in spaceflight. He previously worked at Rolls-Royce and Airbus Military and served as an intern for Made In Space and the European Southern Observatory.

The addition of free surface flows to SOLIDWORKS Simulation 2018 means you can simulate how water fills up this tank. (Image courtesy of Dassault Systèmes.)

Another SOLIDWORKS launch has come and gone, and this year has marked some big improvements for the SOLIDWORKS Simulation 2018 portfolio. One of the biggest additions of note is the new topology optimization tool for structural parts.

This tool gives users an optimal design for a part given its design space, loads, constraints and manufacturing methods. The discussion around this advancement could take up a whole article in and of itself, so we have written it here on Engineers Rule (click here).

But there is much more in the realm of simulation for SOLIDWORKS 2018 than topology optimization. Here you will learn about other tools Dassault Systèmes has added to the mix,including cyclic symmetry for computational fluid dynamics (CFD), free surface CFD flows, nonlinear safety factors and displacement controls for nonlinear analysis.

For more improvements to SOLIDWORKS Simulation 2018 not covered in this article, watch this video:

SOLIDWORKS Flow Simulation: Free Surface Flows and Sector Periodicity

According to Stephen Endersby, director of product portfolio management at SOLIDWORKS, one of the biggest simulation additions to this release is the ability to simulate free surface flows.

Any flow where fluids and gas interact is considered a free surface flow. Examples include anything from a half-filled gas tank to a canoe on a river.

The simulation assesses the interface between gases and liquids. The software also considers any solids that might be affecting the flow of either the gas or liquid. To properly assess the interface of the gases, liquids and solids, these locations will need a finer mesh than any bulk areas.

“When you look at energy, power and utilities, there are a lot of flows that are free flows,” said Endersby. “They are challenging to solve computationally so it took us a while to get there.”

The first step to start a free surface flow simulation is to set the water level. Then the workflow changes depending on internal or external flows.

The setup of a free surface flow changes depending on if it is internal or external. (Image courtesy of Dassault Systèmes.)

For internal flows, like the gas tank, the simulation will model how the fuel slushes around. The sneaky way SOLIDWORKS modeled this is by moving the gravity vector around while keeping the tank stationary.

“Changing the gravity is a little smoke and mirrors, but it is physically correct if you are only working with a single physics,” said Endersby. “For now, it is just a single physics so we can get away with moving the gravity. But in the future, we want to incorporate multiphysics so we will need to have a more rigorous setup.”

For external flows, like the canoe on the river, the engineer must first decide if modeling the bottom of the riverbed will be important to the simulation. In a deep river, the riverbed will likely have little effect; however, near the shore, where the canoe is tied up, this is a different story. In this simulation, you can’t really play with gravity to perform the simulation.Instead, the programmers set it up to change the velocity of the liquid. The liquid velocity can also be pulsed to create a wave effect.

Unfortunately, there is no automated way to set up the water level for floating objects. However, Endersby hopes to add a bouncy function in future releases. It would certainly improve the customer experience to not have to break out an equation every time something is floating. For now, users must rely on hand calculations and any macros they can get their hands on.

Additionally, the free flow function is currently incompatible with simulations containing transitions, rotations, porous media or fans. Endersby is pushing for these to be added in future releases.

To reduce the size of the model, the user has simulated only a quarter of the cylinder. The results can then be mirrored across the axis of symmetry. (Image courtesy of SOLIDWORKS.)

Another big addition to SOLIDWORKS Flow Simulation is sector periodicity, or as a structural engineer might call it, cyclical symmetry. This tool is used to cut down on the number of elements in a model.

Instead, the simulation focuses on a portion of the model that is cyclically symmetric. For this to work, the fluid must flow along the path of the axis of symmetry.

Sector periodicity should also be useful to those in the oil and gas or production industries.

Unfortunately,sector periodicity is currently not compatible for phase transitions, cavitation, high Mach and mixing simulations. Endersby hopes to see these functionalities in future releases.

SOLIDWORKS Simulation: Nonlinear Safety Factors and Displacement Control

Nonlinear safetyfactor definitions new to SOLIDWORKS Simulation 2018. (Image courtesy of Javelin.)

The prime SOLIDWORKS Simulation offering has also seen some new additions in the form of displacement control and safety factors for nonlinear simulations.

An important note about the safety factor functions for nonlinear systems is that the user requiresa considerable amount of understanding about their part and its material makeup.

Endersby explains that for traditional materials, the default safety factor is typically all you need. The material is well known and contains homogeneous properties. In this case, you take the yield stress of the material and divide it by the measured stress to get your safety factor. You then optimize the part until said safety factor is at a desired level.

This isn’t true for brittle or nonlinear materials like composites and plastics. For these materials, the user must define a maximum stress value, as the material might yield before it fully breaks or snap before showing signs of failure.

“You want to make sure the part stays in a linear range of behavior,” said Endersby. “If it ratchets and becomes deformed, it will not go back to the original shape. To avoid this, set a maximum stress at a range where the material still acts linear.”

Another addition to SOLIDWORKS Simulation is the use of displacement controls for nonlinear parts. This tool helps to control the iteration when solving a part that experiences large deformations under small forces. The displacement control function ensures that the system won’t jump to a large displacement under a small increase in the force.

“Take a straw,” said Endersby. “You push it on one end and then it collapses. You get a small displacement at first. But as you increase the force, you eventually get one huge displacement. With displacement control, you can set how the system reacts to get a stable result. It controls the force so you don’t move too quickly.”

The system does this by defining a series of displacements and then calculating the force needed to achieve that displacement. This is alternative to increasing the force and hoping it will have a smooth linear displacement.

Other useful additions to SOLIDWORKS Simulation include:

• Improved stress singularity detection in stress hotspot diagnosis
• Single pin connector for multiple coaxial cylinders and hinge definitions
• Ability to copy simulation features of a part or subassembly into a new study of the assembly
  • Importable features include material, element types, contact, connectors, fixtures, loads and mesh controls (which you can import individually or all at once)
• Ability to export deformed geometry for CAE analysis in various CAE tools
  • Formats include Abaqus, STL, NASTRAN andnative SOLIDWORKS
• Ability to exclude area from clamp force (used to simulate plastic part production with slides and undercuts)
• Improved detection of short stops in plastic mold simulations
• Assessment of density results to ensure uniform density in plastic molded parts
• Emailing when analysis is complete

For more on SOLIDWORKS Simulation 2018, check out Dassault Systèmes’ launch page.

About the Author

Shawn Wasserman (@ShawnWasserman) is the Internet of Things (IoT) and Simulation Editor at ENGINEERING.com. He is passionate about ensuring engineers make the right decisions when using computer-aided engineering (CAE) software and IoT development tools. Shawn has a Masters in Bio-Engineering from the University of Guelph and a BASc in Chemical Engineering from the University of Waterloo.

Aerovelo’s Eta zipped past the human-powered land speed record, clocking in at a whopping 144.17 kph (89.58 mph)! Let it into the left lane, folks. (Image courtesy of Aerovelo.)

We’ve all been there, driving down a one-lane street, backed up behind a cyclist that blocks our ambition of hitting the speed limit. Seriously, bikes limp along at 10 mph and hit what—30 mph—max? Well, no. How does 90 mph sound?

That’s right. Thanks to the engineering minds of Aerovelo and its bike, Eta, cyclists could theoretically complain about passing that big bead on your high performance vehicle.

Aerovelo Cofounders Cameron Robertson and Todd Reichert lead their team of University of Toronto (U of T) engineering students and alumni to design, simulate, optimize, build and pilot this escape pod-encased bike into the history books.

After breaking its own human-powered land speed record a few times over, this little tear drop settled on an impressive 144.17 kph (89.59 mph).

“It was a culmination of years of effort,” said Robertson. “There was a lot of excitement and relief that we have taken a good path and all the choices we made showed it could be done. With Eta’s design, we showed the range of improvement. In 2000 to 2015, there wasn’t much change to the [human-powered land] speed record. It incremented 10 mph in 15 years, from 73 to 83 mph. The rate of technological change was small; it was incremental improvements. In the span of two years with Eta, however, we incremented [the record] by 6.5 mph.”

There is no wonder why the team named their bike after the Greek letter us engineers know represents efficiency. And the racing pun asking Eta’s estimated time of arrival at the finish line wasn’t lost either.

How Do You Design a Bike That Can Break a World? Use CFD Simulations!

Consistency is one of the hardest challenges when designing a vehicle to break any speed record. This is because the racing team typically only gets a few kicks at the can on official race days.

“Every day you need to execute as you will only get some days where the environment is what you want,” said Robertson. “We didn’t expect this to be a big point, but we took it based on advice from other teams that have broken records and were always on the ball. We wanted to emulate this.”

So how does one get from the snail on the road to Formula One? And better yet, how do you make the performance of this bullet consistent? The answer is computational fluid dynamics (CFD) simulations, a lot of experience, and trial and error.

“With the bike, it was important to change the design of the outer shell and then slightly modify the simulation results to get a sense of how it performed with respect to that change. Then, we would iterate again,” said Robertson. “Todd [Reichert] did 30 different iterations on the bike’s fairings, and without SOLIDWORKS we would never have been able to do that in an informed way.”

The majority of the Eta simulations were in SOLIDWORKS Flow Simulation. This simulation in-CAD package was a platform that the U of T alumni and students at Aerovelo were well versed in using. They wanted to investigate the airflow around the outer shell. To do this, the team would perform numerous pressure profile simulations.

“We use pressure profiles, which are accurately evaluated in software like SOILDWORKS Flow Simulation. As the air goes over the surface, you can assess the pressure at every point. You can then shape the pressure profile as you go down the bike and have the shape of the profile be maximally conducive to extend laminar flow,” explained Robertson. “From there, we expect, based on the pressure profile, the change to be positive or negative.”

The goal is to maintain laminar flow around the bike as long as possible due to its lower resistance compared to turbulence and transitional flows. One would think that ideally, the goal is to maintain laminar flow around the shell completely. However, if the flow doesn’t transition before the trailing edge then it can fully separate from the surface which will cause tremendous drag.

A red herring of sorts in the optimization and simulation of their bike, according to Robertson, was trying to determine the actual spot on the bike where the air transitions from laminar to turbulent. He said, “When that prediction happens, there is a large margin of error. It’s subject to small variations and it will be [hard] to implement in the real world versus simulation. We see some team point to a spot on their bike and say, ‘we have laminar flow until here and we predict it will run 150kph,’ and then it runs worse than their previous bike.”

This shows the importance of validating your simulation. No engineer should trust their model blindly. This has become a regular practice for Aerovelo. It performs simulations and then tests the bike to compare the turbulence and pressure profile. The team also relies heavily on its experiences and the experiences of the community of engineers working on breaking the human-powered land speed record.

How Does Eta Differ from Traditional Bikes?

So, after all the design changes, optimizations and simulations, what sets Eta apart from that bike in the garage?

“Well, it’s different from a normal bike in almost every way,” said Robertson. “First, Eta is very recumbent. The pilot is almost completely lying down. The bike is fully enclosed for aerodynamics except for controlled intakes for ventilations. Next, it’s steered using cameras on top of the bike connected to screens in front of the pilot’s face, and steering is limited to three degrees to each side. Finally, the tires are not good for turning. So, it’s clearly designed to go straight.”

Robertson also explains that the gearing for the bike is also very different. Eta uses a two-stage drive train to accommodate the bike’s top speed. Due to the reduction created by the two-stage drive train, the wheels of Eta are able to spin many times faster than anyone can spin on a traditional bike.

With the vast differences between Eta and traditional racing bikes, don’t expect to see it on the Tour de France anytime soon (unless there’s a total relaxation of racing rules). However, Robertson is interested in a biking version of Formula1 (F1), similar to Australia’s Pedal Prix, where the engineering behind the equipment is perhaps more important than the driving of said equipment. He said, “F1 is more about the engineering and could be interesting in a bike format.”

But what really interests Robertson is how this technology could affect transportation. He imagines going to work at highway speeds on a vehicle that is human powered and 300 times more fuel efficient than your average car.

“One thing we thought about is how you could use these in the future. Imagine if you were not contending with several thousand-pound cars on the road,” wondered Robertson. “It’s interesting when comparing a small power of the human engine at about two-thirds horsepower. To ride an hour and achieve 60 mph, the bike needs to be very efficient—about 9500 mpg to an average car that is about 30 mpg.”

The Future of Aerovelo’s Record-Breaking Bike Designs

The Aerovelo team is all smiles while huddled around Eta after breaking the record. But how long will they hold the title? How much room for improvement is there? From left to right Tomek Bartczak, Alex Selwa, Victor Ragusila, Todd Reichert, Cameron Robertson and Trefor Evans. (Image courtesy of Aerovelo.)

So, what is next for Aerovelo and the human-powered land speed record?

Well, Robertson believes that there are still many potential areas of improvement for Eta and similar record-chasing bikes.

Unfortunately for Eta, many of these improvements will require a completely new design and even more thinking outside the box from engineers. Two examples involve heat capture and active boundary layers.

The rule books says that the bike must be powered by humans, but does that mean it has to be powered by the legs alone? Remember, that pilot in an enclosed spot will produce a lot of heat. Capturing this human body heat could theoretically help power a bike. This might seem a little like cheating, but remember that it is still human power, and it worked well enough for The Matrix.

“You can’t use energy storage devices on the bike, but you could still drive a bike on an electric motor,” said Robertson. “So, imagine if the driver provides power to the drive train. Humans are 20 to 30 percent efficient on converting energy and the rest goes to heat. Capturing that with some efficiency using heat recapturing tiles or heat pumps could theoretically increase the energy output.”

To test out this theory, engineers at Aerovelo could create a heat transfer simulation of Eta’s replacement and use that data to crunch the electromechanical numbers.

“Simulations would also play a role in active boundary control, which is used to extend the laminar flow around a vehicle,” explained Robertson. “Active boundary control senses and manipulates the boundary layer in order to allow longer runs of laminar flow than would be possible otherwise.”

Active boundary control can be done in two ways: using either wave theory or clever ventilation.

“When air transitions from laminar to turbulent, energy and unstable waves oscillations grow to a point where the stable smooth conditions turn into the chaotic movements,” noted Robertson.“The technology senses the oscillations in the air and then introduces more waves to cancel out those waves that are promoting the turbulence transition.”

Another active boundary layer method would add a system that sucks in the nearly turbulent boundary layer of air. The air is sucked in can then be used for ventilation. However, more importantly, the surrounding flow would fill the gap made by the pre-existing boundary layer. This would effectively give the bike a new laminar boundary layer.

Unfortunately, this method would need precise knowledge of the transitional point. As previously mentioned, this would require simulations with questionable accuracy in this particular application.

Similar to the heat capture option, active boundary layer does use up energy to work the sensors, oscillators and/or pumps and any other equipment. The question is will you get more energy out of the rider than you waste on this added equipment?

“Everything we do is about system design,” noted Robertson. “We look at how we can do more with less or get more from what we have. Reduce the weight and increase strength—that is an important mind-set.”

However, there is another feature of the Flow Simulation toolset that can assist in evaluating fluids interacting with mechanisms in motion, useful for assessing the design and performance of turbomachinery. This involves specifying rotation of the fluid and solid bodies within the computational domain. This article looks at applying these concepts to a pump design.

Pump Model

Consider the pump shown in Figure 1. The motor rotates the impeller at high speed, creating a vacuum to generate suction at the inlet. Pumps and most turbomachinery operate by converting power in the form of an electric motor (P₂ in the diagram below) to fluid power in the form of pressure and flow.

Figure 1 . Typical pump diagram.

The operating parameters we will simulate are as follows:

A SOLIDWORKS solid model representation of the pump internals is shown in Figure 2 below.

Figure 2. Pump solid model.

A key parameter in simulating turbomachinery is identifying the rotating components in the model and using boundary conditions to specify stationary components. As with many Flow Simulation problems, the Wizard option gives a good starting point for setting all the major options. The Wizard starts with the dialog box in Figure 3. Here we name the project and add any comments.

Figure 3. Wizard start dialog box.

The next dialog box is for specifying units. For this example, we will perform the analysis using SI units.

Figure 4 is the first dialog box that allows us to indicate that some of the components in the solid model will be rotating during operation. We activate the Rotation feature and select “Global rotating” from the dropdown.

Figure 4. Specifying rotation.

 We set the angular velocity to 2,000 rpm (209.5 radians/second) about the Z-axis. Since we have set a global rotation, all components are considered rotating with the specified angular velocity. We will specify the stationary components later in the analysis setup.

The fluid that we are pumping is air, and we add it to the Project Fluids definition as shown in Figure 5.

Figure 5. Adding air to the project.

All other dialog options in the Wizard can be left at the default.

To ensure the geometry is modeled correctly for the Flow Simulation to continue, a model check is performed (Tools → Flow Simulation → Tools → Check Geometry). With the Show Fluid option of the geometry check activated, we get the display as shown in Figure 6. The air can be seen contouring the blades of the impeller.

Figure 6. Air volume.

Boundary Conditions

The pump will be required to produce a flow of 0.3 m³/s. We will define this boundary condition at the inlet lid face of the model as shown in Figure 7.

Figure 7. Specifying inlet flow.

 The outlet is defined on the outlet lid surface as shown in Figure 8 by setting the Environment Pressure to ambient (10,325 kPa) on this surface.

Figure 8. Specifying the outlet conditions.

 The final boundary condition is the most critical. Recall that we specified a global rotation of 2,000 rpm in the model setup. This will, by default, impart a rotation on all of the fluid and solid components. We need to make the pump housing stationary to mimic the actual operation of the equipment. From the boundary condition dialog box (Tools → Flow Simulation → Insert → Boundary Condition), we select the pump casing as the stationary item (see Figure 9).

Figure 9. Fixing the nonmoving components.

 We are now ready to solve the model by selecting Tools → Flow Simulation → Solve → Run from the menu.

Post-Processing

The function of a pump is to drive a given flow from a lower pressure to a higher pressure using energy from an attached motor. The pump in our example uses a motor to drive the impeller (Figure 1), which pulls suction on the inlet (P_s). We will evaluate the pressure drop by querying surface results at the inlet and outlet surfaces.

Figure 10. Flow Simulation menu tree.

Right-clicking the Surface Parameters in Figure 10 brings up the dialog box in Figure 11. We are interested in the inlet pressure since that will give the amount of vacuum the pump must draw to move the specified 0.3 m³/s of air.

Figure 11. Getting inlet pressure results.

To calculate the inlet pressure, the inner surface of the inlet lid is selected as the reference geometry, and we select “Pressure” as the required quantity. The average pressure over that face is 100.4 kPa, as shown in Figure 12. This is a vacuum relative to the specified 101.3-kPa ambient pressure.

Figure 12. Inlet pressure results.

 The power input from the motor required to pump that amount of air is determined by the following calculation:
Power (Watt) = Torque(N-m) × Angular Velocity (radians/second)

The torque on the impeller is determined by again bringing up a Surface Parameters dialog box similar to Figure 11. However, in this case, we select all the faces of the impeller for the Selection box and select “Torque” for the parameter to display (Figure 13).

Figure 13. Getting impeller torque.

The results are shown in Figure 14.

Figure 14. Impeller torque.

 Plugging the results into the previous formula relating power to torque and rpm, we have:

Power = 1.507 (N-m) * 209.5 radians/second = 315.7 Watts

Another useful insight to the design of a pump impeller is the efficiency. As shown in Figure 14, Flow Simulation can extract the contribution to the torque requirements between the normal and friction force on the blades. Modifying the blade profile will affect the ratio of the components to the overall torque. This will impact the pump efficiency, which is defined as the power output to the air stream relative to the power input by the pump motor to the impeller. The power in the pressurized air stream is the pressure rise multiplied by flow:

Air Power = Pressure Rise × Flow = (101,325–100,424) Pa × 0.3 m³/s = 270.3 Watts

The pump efficiency is defined as the air power divided by the power delivered to the impeller. The overall efficiency of the pump in our simulation is then 270.3/315.7 = 85.6 percent.

Visualizing Results

The airflow pattern within the impeller can be visualized by selecting “Flow Trajectories” from the Results menu. This brings up the dialog box in Figure 15.

Figure 15. Defining flow trajectories.

Selecting the outer lid as the reference surface and specifying velocity for the parameter to plot, we get the flow trajectories from the pump inlet through the impeller as shown in Figure 16.

Figure 16. Flow trajectory plot.

Parametric Analysis

The previous Flow Simulation was modeled by specifying a global rotation in the settings and then selecting the stationary surfaces (stators) by applying the appropriate boundary conditions. This makes determining the effect of increasing the rotational speed on the pump performance very easy. We will increase this global rpm by 50 percent from 2,000 to 3,000 (314 radians/second) and rerun the analysis.

The simulation with these parameters gives the following results:

Impeller Torque = 2.88 N-m

Power = 2.88 (N-m) * 314 radians/second = 904.3 Watts

Pressure Rise = 101,325–99,124 Pa = 2201 Pa

Air Power = Pressure Rise × Flow = 2201 Pa × 0.3 m³/s = 660.3 Watts

The efficiency in this case is 660/904 (73 percent) compared to the previously calculated 85.6 percent for the 2,000-rpm base case.

Conclusion

SOLIDWORKS Flow Simulation can assist in the design and analysis of various types of turbomachinery (fans, pumps and turbines). In this article, we analyzed an air pump and set up boundary conditions to simulate a specified airflow rate and determine the power required to achieve that performance. Overall pump efficiency was determined by calculating the delivered power to the impeller versus the power in the outlet air stream.

About the Author

Attilio Colangelo has more than 25 years of experience in engineering and project management in the chemical, process, ceramic and advanced-materials industries. His specialties include CAE, with an emphasis on FEA, high-temperature and heavy industrial design. His software skills include SOLIDWORKS Simulation, NASTRAN, Caesar II, ANSYS and iOS programming.

Setting Up the Model

We will start with a simple model of water flow in a pipe. This is a model I used when I was evaluating various CFD packages. It is a 100-ft horizontal length of 4-in-diameter pipe. It’s nothing exciting but a result I knew I could verify from having done these calculations manually and from other technical references (first rule of simulation, trust but verify!).

Figure 1. Solid model of pipe geometry.

 Figure 1 shows a simple solid model of a hollow cylinder to represent our pipe. The next step is to bring this model to the flow environment. The easiest way of doing this is to use the Flow Simulation Wizard. With the SOLIDWORKS Flow Simulation add-in activated, the display should appear similar to Figure 2.

Figure 2. Flow Simulation menu ribbon.

Selecting the wizard option from the menu brings up the dialog box shown in Figure 3.

Figure 3. Flow Simulation Wizard dialog box.

Clicking “Next,” the dialog box in Figure 4 appears for unit system selection. Conveniently, we can mix and match units. This is particularly useful when performing a flow/thermal simulation using U.S. units. We will continue with the default IPS units.

Figure 4. Unit System Wizard dialog box.

The next dialog is the analysis type, as shown in Figure 5. The type of analysis can usually be determined intuitively. Internal flow is bound by a solid at the flow outer boundary. Our current model is internal and the fluid is bound by the pipe walls. An external flow example would be airflow over an airplane wing.

Figure 5. Selecting the analysis type.

We then add the fluid we are simulating to the project. Selecting water in Figure 6 adds it to the project fluids section as the default fluid.

Figure 6. Defining the project fluids.

After completing the wizard, we are greeted with our first error, as seen in Figure 7.

Figure 7. Error message due to nonwatertight model.

Yes, the pipe must actually have its ends “sealed” and be watertight to be able to simulate flow through it. All flow simulation must happen over some contained volume—the “fluid volume.” The software does not know where to end the problem if we don't cap the ends, so we select “Yes” in the dialog box. We are then asked to select the open ends that we need to close, and lids are created as shown in Figure 8. The purpose of the lids and closing the model will be clearer when we discuss boundary conditions.

Figure 8. Adding lids to the pipe ends.

Boundary Conditions

Prior to placing boundary conditions, we basically have a water bottle. The model requires boundary conditions to define the inlets and outlets. However, prior to defining them, we will perform a model check using the Tools → Flow Simulation →Tools → Check Geometry command. This checks that the model has valid geometry to proceed with the analysis. We can also enable a Show Fluid option, which gives us the graphic shown in Figure 9.

Figure 9. The fluid volume.

Referring to Figure 9, we make the following definitions:

• This is the boundary referred to as the computational domain. It represents the mathematical boundary of the flow problem. For internal flow, it closely, if not identically, corresponds to the fluid volume. At a minimum, it must envelop the fluid volume.
• This is the fluid region/volume that the software recognized. It is important to note that this is the only volume in the model that the analysis is concerned with. The solid bodies (pipe wall and lids) are there as a convenient way of defining boundary conditions. They do not participate in the flow simulation.
• This is the lid that we added to close the ends of the pipe. It does not participate in the analysis but serves as a reference to define boundary conditions, as we will see next.

Applying Boundary Conditions

Boundary conditions are where we define inlets and outlets for the flow. For our problem, we know the flow rate and are interested in flow distribution and overall pressure drop. Figure 10 shows the explorer pane for the Flow Simulation environment.

Figure 10. Flow Simulation explorer pane.

Right-clicking on the Boundary Conditions item brings up the dialog box in Figure 11.

Figure 11. Boundary Condition dialog box.

The face that we select for the boundary is the interior of the lid we added to seal the model. We are not able to select the fluid directly. The software assigns those lid surface conditions to the fluid in contact with that face (see Figure 9). If the lid face is not coincident with the fluid, there will be an error applying the boundary condition.

We now have an inlet flow (aka volumetric flow rate) defined as 40 ft3/min. We need to define an outlet by putting a pressure boundary condition at the other end of the pipe, again at the internal lid face. As shown in Figure 12, the outlet pressure is defined as ambient (101 kPa or 14.7 psi).

Figure 12. Defining the outlet conditions.

With these boundary conditions, the problem statement can be written as follows: “Calculate the pressure required to move 40 ft³/min of water through a 4-in-diameter pipe, 100 ft long, with an open discharge.”

Finally, we add goals to the model. Goals give the solver guidance on our solution objective. In this case, we add the inlet pressure as a goal. This is done by right-clicking “Goals” in Figure 10 and inserting a surface goal on the inlet lid face. This is the same face we used to define the inlet flow.

Meshing

The fluid volume is meshed into a grid for the simulation to proceed. This is analogous to the mesh in finite element analysis. The default automatic mesh settings work very well for most flow simulation problems. However, even in the default automatic mode, there are refinement options available to the user. Figure 13 shows the dialog for setting mesh refinement from a scale of 1 to 7. The mesh preview is immediately updated in the model.

Figure 13. Mesh/grid settings.

With the problem properly defined and bound, we are ready to run the simulation. The Tools → Flow Simulation → Solve → Run command brings up the dialog in Figure 14.

Figure 14. Run dialog box.

We will choose “New calculation” and select “Run.” If there was a previous partial analysis performed, there is an option to continue the calculation so as to not have to rerun prior iterations.

The analysis progress and convergence can be monitored by selecting a goals plot. For our analysis, we are solving for the inlet pressure and we graph the convergence over time/iterations. This is useful to troubleshoot convergence issues. Figure 15 shows the value of the inlet pressure over the final iterations of the simulation.

Figure 15. Convergence plot.

Postprocessing

There are many options for querying and viewing the results. Figure 16 shows the various graphics and result quantities that can be evaluated.

Figure 16. Results options.

Two of the most common are Flow Trajectories and Surface Parameters. For our problem, we are interested in a surface parameter, namely the pressure required at the inlet face/surface of the pipe to flow 40 ft³/min of water 100 feet. Right-clicking on the Surface Parameters option brings up the dialog box in Figure 17. We select the face of the lid to represent the inlet surface of the fluid and select “Show” to bring up the results in Figure 18.

Figure 17. Getting inlet pressure result.

Figure 18. Listing of calculated inlet pressure.

The results show that the average inlet pressure is 136,102 Pa. The outlet was set at ambient (101,325 Pa), giving a pressure drop of 35 kPa (5 psi).

The flow trajectories for our simple pipe model are essentially straight lines. Figure 19 better illustrates a flow trajectory plot. It’s taken from a sample model that combines flow and thermal effects.

Figure 19. Sample flow trajectory plot.

Thermal Analysis

The flow model can be easily converted to a thermal model. We will modify the previous model to simulate a water heater by setting the pipe temperature to 500 °F and determining the water outlet temperature. Going to the General Settings dialog box and selecting “Wall conditions” brings up the dialog shown in Figure 20. We set the pipe wall to 500 °F (533 K).

Figure 20. Setting pipe wall temperature.

The steps to obtain a solution are the same as in the previous flow simulation.
The result we are interested in is the water discharge temperature to determine the amount of heat the 500 °F pipe transmits to the water. We right-click on the surface parameters item in Figure 16, which brings up the dialog box in Figure 21.

Figure 21. Water outlet temperature.

The outlet lid face is selected as the reference surface over which the water temperature will be evaluated.

Figure 22. Heat Transfer Coefficient parameter.

Figure 21 shows the results, indicating an average water discharge temperature of 316 K (109 °F). Other heat transfer–related parameters, such as the heat transfer coefficient (HTC), can also be determined. Figure 22 shows the results from selecting “Heat Transfer Coefficient” and the interior face of the pipe wall.
This shows that the average HTC acting at the pipe/water interface is 19,252 W/m²-K (3,390 Btu/hr/ft²-°F).

Conclusion

SOLIDWORKS Flow Simulation 2016 has capabilities for solving various flow and thermal problems. It uses the SOLIDWORKS modeling engine to define the physical geometry, and then the Flow Simulation environment defines boundary conditions and examines simulation results. In this article, we set up and solved a flow problem from a solid model through analysis and post-processing. The model was then converted to include thermal effects through a simple boundary condition change on the pipe wall. We also evaluated results pertinent to both flow and thermal simulations.

About the Author

Attilio Colangelo has more than 25 years of experience in engineering and project management in the chemical, process, ceramic and advanced-materials industries. His specialties include CAE, with an emphasis on FEA, high-temperature and heavy industrial design. His software skills include SOLIDWORKS Simulation, NASTRAN, Caesar II, ANSYS and iOS programming.