Experimental Aerodynamics

Reflections on wind tunnel testing at Cal Poly

[troywollman]

Reflections on 307

What did you think was the most challenging aspect from all 10 weeks? (other than time... yes.... time... always need more time....)

Testing is problem solving. Something will always not go according to the original plan. The most challenging part of this class has definitely been the process of making due with the struggles of real life testing. It is easy to say that, “given more time… better results may have been obtained.” I think the reality of the situation is that more time doesn’t always mean enough time to work out all the kinks. There will always be challenges faced in real world testing that must be worked around. Do you feel better prepared for the no-single-answer, sometimes-don't-even-know-what-the-question-is stage of senior year and the workplace? And/or has the course helped you better get to know yourself and how to manage your approach to such situations even if that's not the environment in which you feel you thrive?

I would say that this class has served as a fantastic introduction to the open-ended stage of my education/career. I would argue that this is the first time in my education that thoughtful, careful planning early on in the development of a project, as well as communication with group members has been essential to successful work. I have really been forced to but lots of “behind the scenes” time into every project. It seems that for every hour of testing, there is somewhere between 5 and 10 hours of planning, building, fixing, revising, etc. to be done. Knowing that moving forward will be a critically useful skill.

As a side note, one worry that I have moving forward, especially in the workplace, is that when significant time is spent not running the tunnel, it seems that it is hard to demonstrate that actual work, though non-tangible, has happened. I think that these e-logs serve as a great way to document the problem solving and planning that goes into our projects, but I am curious as to how such time will be documented in a formal workplace environment.

You've now worked in 3 or even 4 different groups in the lab setting - what are your main takeaways for how to manage group situations and teamwork and team dynamics going into senior design or other group projects next year?

Communicate with your group members. Early and consistently. It may be uncomfortable sometimes, but it will get worse if you ignore it. The large scale the project, the more critical this communication will be. Regular meetings are a great way to encourage this process. Do you feel like the course made you more interested in aerodynamics and testing, or did you "get your fill" from this course and will strike out in another direction?

A lot of the time it seems like the end-goal of an aerospace engineering degree is the ability to generate valid CFD results. In general, that seems like a very unsatisfying career without the real life component we experience with wind tunnels. While I understand that there is no longer such a thing as a CFD person, or a wind tunnel person, I can definitely see my self maintaining interest in wind tunnel testing and/or flight testing over CFD testing. I may need to get my feet wet in the CFD arena to make a better judgement.

The course (and physical laboratory) has changed a monumental amount in the last 2 years, and will continue to do so into next year (when hopefully almost everything will work properly! We're a bit too brand new right now...). What would be one experiment that you'd really liked to have done in the lab, if the cylinder gets bumped to an earlier course (e.g. full aircraft config for stability testing, something involving another technique.... )?

One of my primary selection criteria when looking for colleges to transfer to was “Do they have a wind tunnel”. This definitely narrowed my choices. Admittedly, I did not really investigate what specific accommodations each school had. I guess I thought a wind tunnel is a wind tunnel is a wind tunnel. I know that to not be the case now. That being said, I am definitely happy with the current (and planned future state) of the wind tunnel at Cal Poly. It comes across as a living breathing facility that we all interact with. We break it, we fix it, we rearrange it, mount things to it, drill into the floor, problem solve. I love that.

The idea of testing a full aircraft configuration sounds really appealing to me. We need to get that pesky load cell worked out first.

[troywollman]

Here’s a quick photo of the live readout from the load cell. You green curve represents the force in the z direction (up and down relative to the model). As you can see, the value is fluctuating (unsteady flow) locally, but also decreasing globally (load cell drift). During the span of several seconds represented in this image, the assembly was static. We believe that the drift was the result of convective cooling and therefore contraction of the load cell. We attempted to mitigate this issue by allowing the load cell to cool for ten minutes before collecting data, as well as shielding the load cell from the wind using aluminum tape. Neither technique proved to make much difference. 

[troywollman]

We ran our wind tunnel test of the Puma model yesterday. Our goal was to characterize the Cl/Cd characteristics of the aircraft with respect to angle of attack. Unfortunately, we encountered some significant drift issues with the load cell, so we may not have collected viable information. With the amount of time we had available, we ran force balance tests for angles of attack of -5, 0, 5, and 10 degrees. Because the load cell was so erratic, we moved on to our second plan of action and began doing some flow visualization of the model. Using smoke, and a laser sheet, we aimed to determine the stall angle of the model. Due to the lighting conditions of the wind tunnel, we were not able to obtain clear images of the flow over the model before the smoke machine overheated and was deemed inoperative. 

As always, things do not go as planned in the wind tunnel. The real world is a challenging environment to test in. As things deviate from the plan, a good engineer must be able to adapt, and find new ways to collect the information required. Given more time in the lab, my team would be able to continue working on this project, but I anticipate we will analyze the data we were able to obtain, and try to quantify the (massive) amount of error that we will show in our force readings. 

[troywollman]

Puma Workday

Yesterday was a Puma work day. We received our 3-D printed symmetric airfoil sting mount adapter, discovered that it printed out of tolerance and needed to be touched up. After a bit of drilling, reaming, and filing, the part now fits well with our Puma model. Now we need to epoxy the adapter to the model, and confirm that everything will install into the wind tunnel easily. The plan is to setup everything for testing Monday night, and to test first thing Tuesday morning. 

As a side note, my team discussed techniques to characterize the effects of our sting mount on the model in order to remove those effects from our results. It seems like we won’t have the time and resources to accomplish this properly, but we may need to do this to make sense of our data.

[troywollman]

In the last weeks of the course, I will be working with a small team to collect performance data for the RQ-20 Puma, a light-weight hand launched UAV designed by AeroVironment. The aircraft was designed as a low cost, easy to use platform, without the need for performance studies. As a result, AeroVironment does not have a well defined understanding of the performance characteristics of the aircraft. My team will be working on generating Cl vs alpha, Cd vs alpha, Cm vs alpha, and Cn vs beta curves for the aircraft using a 5.5:1 scale model in the wind tunnel. 

In order to accomplish this, we will be designing and building an adapter to mount the model in the wind tunnel via the existing force balance assembly in order to interface with the (problematic) load cell. We will also be doing some laser assisted, smoke flow visualization of the aircraft.

[troywollman]

Reflections on Flow Visualization

My apologies for the late post. There’s a lot going on lately.

The flow visualization project was a great experience. It was great to be involved in fully planning out an experiment, building our apparatus, and executing our project. Although we encounter some difficulties (as expected with anything really) I feel like my team gathered some appealing results. 

I think that my favorite part of this challenge was when we looked at the first long exposure image from the setup in the wind tunnel. That was when we realized that we were going to be able to collect some worthwhile images and be able to quantify something about the flow. There was definitely concern that our setup was not going to work as planned.

We were very disappointed to find out that the wind tunnel force balance assembly was acting up, and we would not be able to change the angle of attack of the setup (which was a primary part of our original plan). Nonetheless, we were able to image the tip vortex at varying tunnel speeds, although the change in size was likely smaller than what would have been observed by changing the angle of attack.

I must say that 307 has been one of the most useful classes in my professional development as an aspiring engineer. It has definitely shed light on what it means to design, build, implement, troubleshoot, and analyze for a project in a meaningful way. I honestly wish I had experiences like this earlier on in my education, as the skills are invaluable. It seems like this course has gone by in a flash and I definitely want to have more time to put into “real stuff”. I’m looking forward to our last project in the tunnel for the next few weeks.

[troywollman]

“The Jellyfish” is here. Our flow visualization was successful. Josue was nice enough to stay at the wind tunnel into the wee hours of the night. This is a 2 second exposure of our florescent dyed tufts under UV light in the wake of a NACA 4412. You can clearly see a tip vortex on the lower left of the grid. The high concentration of tufts is a 2 cm grid, while the more spaced out tufts are a 4cm grid. In the top right corner you can see the wake behind the force balance assembly. I’m very pleased with the images we captured, and will be analyzing them more thoroughly with my team over the weekend. Its 2am and I’m tired.

[troywollman]

I did some preliminary testing for my flow viz this week. We will be placing a grid of fluorescent tufts downstream of the finite NACA 4412 in hopes of capturing some nice visuals of the tip vortices. If all goes well, we will be able to locate and determine the size of the tip vortices. It seems that a 2 second exposure on a GoPro will work well. I learned today that the UV light will need to be diffused, not direct from the small flashlights we will be using. 

[troywollman]

Reflections on 4412 Testing

At the close of the 4412 testing series, I have a few thoughts worth sharing:

1) This was my first experience working with heaps of data to analyze. Although the data was pretty simple and straightforward to process on a case by case basis, it was definitely a challenge to organize the data collected across a month of testing. In the end, my strategy aligned more with processing the data in a separate script for each testing week, and outputting new data files that contained what I really wanted. For example, a full lab section of testing would supply enough raw data to generate Cl vs alpha plots for a couple Reynolds numbers. I would run my script that would output just the vectors of Cl and Alpha from the test day, and store those all in one repository on my computer. I found this to be a way to avoid writing a master code that would process all of the data at the same time. If I wanted to compare two separate days of data, I could just plot the output results for each day.

2) I thought it was really awesome that we were able to play such a hands on role in the operation of the wind tunnel. I would say that this has been my first real “learn by doing” experience at Cal Poly, and I really appreciate that.

3) I am fairly disappointing in the level of communication that existed within my lab group. Communication was a constant source of frustration for me. I think I should have addressed my concerns much earlier on and not let them develop over the course of the testings. I am hoping that I can walk away from this experience as a better communicator and team member in the future.

4) I was not happy with a lot of the results my team collected, especially with the force balance. It seems there is a lot to be learned about how to properly implement that tool. I will be working in the wind tunnel over the next couple of weeks with Prove, and I think that I will be able to collect much better force balance data for that project. I will make a point to tare the setup at before and after running the tunnel for EVERY configuration that we test. Also, the load cell in the Prove model does not rotate with the model; it remains fixed regardless of the orientation of the vehicle. This should definitely make the processing easier.

5) All in all, I am very happy that I had the experience of working in the tunnel for the past month on this project. Results obtained from a wind tunnel feel so tangible and REAL! It really brings all the theory to life. I look forward to working on other projects in the tunnel, and I will be looking for internship and/or career opportunities to work in a wind tunnel. Now I just need to learn to appreciate CFD. 

[troywollman]

Another thought about gliders:

Gliders get towed to altitude on the end of a 200′ rope behind a powered plane. I was thinking that with a strain gauge installed along the rope, you might be able to get a useful calculation of drag on a full scale glider without placing a model in the wind tunnel. Assuming steady level flight, the tension in rope will be exactly equal to the drag force acting on the glider. Also, we know very precisely how much the glider and cargo weigh, which means that I would’t be too difficult to calculate L/D for the aircraft. This would be a fun experiment. I am interested in comparing flight test data to the published performance of the aircraft. 

Pictured above is me piloting an SGS 2-33 from the rear seat.

[troywollman]

Here’s another photo of the force balance without the wing installed. The airfoils mount upside down to the rod on the far left. We can mount them upside down as long as we keep track of the axes on the load cell. The air doesn’t care which way is “up” after all.

We found out in lab today that the load cell was spewing out garbage data on Monday, so my suspicions are correct. From what I understand, there were issues with its calibration, as well as EM interference. If we get a chance this week, we will retest. Otherwise, we will get data from another group, and make a note of the Reynolds number being off.

[troywollman]

This plot shows our force balance data plotted against the results from Theory of Wind Sections. It is obvious that there is a flow in our data, or its handling. After a fair amount of inspection, and consulting with classmates, I am confident that my code is working properly. I am hoping to hear that there was an issue with the force balance or Labview. 

Our trend for Cl is roughly correct. It increases with alpha and appears to drop off around 13 degrees angle of attack, indicating a stall. The magnitudes are very far from target however. 

[troywollman]

This week, we are testing two finite span 4412s. The 4412 have wingtips in the flow field, which allows tip vortices to be generated. This will induce downwash on the flow, rotating the lift vector backwards compared to a 2-D airfoil. This better models the reality of aerodynamics of airfoils. We will now be able to capture the resultant “induced drag” or drag due to the downwash. 

For this experiment, the two finite span NACA 4412s are mounted to a force balance. In our lab, we have a digital load cell. Its basically a collection of strain gauges used to measure force in x, y, and z, as well as moments about x, y, ans z. This very sensitive (and expensive) equipment is incredibly useful as it allows us to determine forces on the airfoil directly, rather than interpreting it from pressure data.The airfoils are mounted to the the load cell via a sting. The load cell enters the test section downstream of the airfoil through the roof ceiling. 

When we arrived at the lab today, worked with our lab technician to calibrate the load cell by hanging known weights from the sting, at a predetermined location, and confirming that the software was registering the appropriate forces and moments. This took considerably longer than expected, and our testing was delayed by about 45 minutes.

With the short period of testing time remaining, we collect 9 data points of varying angle of attacks. We ran the tunnel sufficiently fast so that we matched Reynolds number with the fullspan testing from prior weeks. While checking the data in the lab, I observed the appropriate trends, but incorrect values for our forces and moments. I will investigate my code tonight to make sure I did not handle data poorly, or misplace a decimal, etc. We expect to test again later this week, or if we can’t secure more tunnel time, to get data from another group.

[troywollman]

Outside of school, I am a glider pilot. Lately, I have been contemplating putting a scale model the glider I fly (SGS 2-33) in the wind tunnel. Although the performance characteristics of the aircraft are published in the manuals for it, I have been very interested in verifying that data in the wind tunnel. I think the best way to approach this would be to 3D print a small model. The full scale has a 50ft wing span and about a 4ft. chord length. The airfoil is a NACA 633-618. Flying at 66 ft./s at 6000′ gives me a Reynolds number of approximately 1.45*10^9. If I wanted to match this Reynolds number in the wind tunnel running at 120mph, I would need a scale model with a 2.6ft. wing span. That’s about a 1:20 model. At that size, it would probably be better to make it out of wood or carbon rather than trying to 3D print it. Hmmmm....

I think that it is very important to combine theoretical learning (textbooks, lectures) with hands on learning (aviation, wind tunnel, PROVE) to become a well rounded engineer. By immersing yourself in both realms, you an get a fuller understanding of any field.

I think it would be great if more of the aero students could get experience flying. Its one of my goals to get more people out flying with me on the weekends!

Note: My apologies for the (stupid) English units.

[troywollman]

Reflection on Week 4

This week, the most challenging part about working in the lab was using our time as efficiently as possible. I feel that compared to the first week, I have a much better understanding of how to handle data in  Matlab that comes from the Scanivalve system. My group will need to communicate roles and expectations to each other better in the future. We are planning on meeting before our next time in the lab.

It seems that every group member has a appropriate understanding as to what needs to be accomplished at every "station" in the wind tunnel. Once we are in the flow of running tests, everything seems to run smoothly and efficiently. Although it is really simple, the biggest improvement that I have personally made is taking the time to write code before the lab that can be used to make sense of what is happening in real time. Its very helpful in identifying buggy pressure ports, and goofy data. It is a lot easy to recollect immediately after noticing someone stepping on a pressure line (as evident by wonky data) rather than grappling with the strange results the following week during our analysis.

It seems that my group is still struggling to get our Cl versus alpha plots to make sense. We are always getting the correct trends, but the magnitudes are way off. There must be a fundamental issue in the way we are trying to convert our collection of gauge pressures into a lift coefficient. I will be investigating this further, reading the Barnard book, and talking with the TAs.

In terms of the upcoming flow viz experiments, my group has discussed investigating the vortex shedding frequency of the finite wing airfoils in the lab, using smoke and an image capturing device. We would like to compare the results we get to the theoretical values that we get out of a Matlab panel code written for our Aerodynamics class.

I am definitely looking forward to continuing our studies in the wind tunnel lab. It seems that we will be moving the pressure rake around next week for a better understanding of the 3-D effects of the airfoil wake. 

[troywollman]

It appears that our code is outputting viable plots. The above plot shows the pressure coefficient for every port on the pressure rake. Notice the boundary layer is visible in the pressure change for the lowest and highest 3 ports. We expect the zero slip condition to slow the flow near the wall below free stream. 

The rake we are using is not a complete set of ports. No physical system could collect data at every location in the flow field. There is a high concentration of ports in the center of the test section (measured vertically) and a three on each extreme end (top and bottom) to capture the effects of the boundary layer. We end up with a large section of unmeasured flow velocity. On the plot, this is represented by the straight line sections where two far-spread data points are connected with a line. To better capture the reality of the data, I might try using a spline fit between points, or I could make a note of the expected error that this will produce.

The main thing that I learned from the initial analysis of this data is that we can obtain useful information about an object in a flow field by measuring how it can change the flow around it, rather than directly measuring pressures and forces on the body. This is a very interesting and exciting concept. I will be using the wake information obtained off of the body to determine the drag force exerted by the flow onto the NACA 4412. 

[troywollman]

This week we used a cool device called a pressure rake to record information about the wake behind out beloved NACA 4412. The airflow behind the airfoil is expected to be non-uniform. Because we know that we can measure flow velocity indirectly by measuring its pressure and then converting, we can use a collection of pitot tubes to measure flow velocity. In this case, the rake is a vertical collection of pitot tubes that we can use to determine the downstream velocity profile for a single cross-section of the airfoil. To assume that this profile would represent the entire flow field behind the airfoil for any cross section would be to assume 2-D flow, which is indeed false. Next week we will be moving the rake to another location to capture the changes in velocity for the 3-D flow case.

After running the testing with the pressure rake, I have a few thoughts:

1) My testing group is lacking in effective communication. We came to lab with different ideas of how to collect data, which should have been discussed and formalized beforehand. Additionally, there was a miscommunication regarding who was responsible for writing the code to check our data in real time. In total, these led to a delay of nearly 30 minutes to our testing for the day. Ouch! This is not a trend I wish to continue. 

2) There seemed to be a lacking of communication between the lab tech and the students with respect to the current functionality of the lab equipment, specifically our ability to determine tunnel velocity. While I understand that the wind tunnel is a living facility under constant renovation (which is great for all of us), it has been a point of confusion each week as different work-arounds have been presented for each flaw encountered. It should be noted that knowing how to make use of faulty equipment is an incredibly useful skill and I am pleased that we are exposed to the reality of testing.

3) The formulation of well thought out procedures and roles for people working in a lab are critical to a functional and efficient working environment. There still seems to be a bit of confusion as to each persons role in the tunnel, which I hope to see resolved in the future. 

Results from the initial testing to follow.