Flying by the Seat of Your Pants

This is a transcript of a non-technical presentation I gave to the Nehemiah Fellowship at the North Toronto Chinese Alliance Church.  Most of the attendees were young professionals, but have no technical background in aerospace engineering or aircraft design, or in engineering in general. The purpose of my talk was to discuss the work that I did for my Ph.D. research, and some overview about what aviation is all about. The talk was given in Cantonese, and the transcript shown here was subsequently translated back into English by myself. The Chinese language transcript may be available in the future.

Shall we start then?

I want to thank the Nehemiah Fellowship committee for inviting me to speak about my work tonight. Admittedly I am in fact very nervous: most of the talks I have given so far have been for a technical audience, and it's only the second time that I had to give a talk in Chinese. I have titled tonight's talk “Flying by the Seat of Your Pants”, which I feel is an appropriate description of how things are when it comes to aviation, and when it comes to engineering research in general.

I'm sure everyone here tonight have been on an airplane before, and most of you have probably been on an airplane within the past year. Indeed, aviation is so common that we no longer think twice before we fly. We complain about delays, services, how much we have paid for the tickets, but most of us don't think about how safe flying actually is.  Which brings me back to the title of the talk, “flying by the seat of the pants” was a term used by early aviators—back when flying was dangerous—to describe them “using intuition as you go along...” In a lot of ways, doing research in aerospace engineering is a lot like those days, only without the dying part.

Before we go further, you must decide whether to believe or agree with what I will say. My perspective is, of course, incomplete and biased. And you must remember that there are many things that I don't know, like the fact that I don't know how to fly an airplane, how to do maintenance on one, how to build an airplane, or how airlines work. In fact, it has been pointed out numerous times that I don't even really know how to design an airplane. And, if you known me for awhile—as many of you do—you may recall that when I graduated from university, I was ranked 21 out of 22 students. But while I don't know how to design an airplane, my tools are used by those who do, for example, the University of Toronto Institute for Aerospace Studies, Ryerson University, Bombardier Aerospace, and occasionally, the folks at a little known organization called the National Aeronautics and Space Administration, “NASA” as we call it, down in the United States of America. The first “A” is the little known part. All the publicity and money go to the space program, there are actually lots of research down on the aeronautics, the part related to aviation.

While flying isn't exactly a new concept, we didn't actually achieve powered flight until the Wright Brothers' "Flyer” in 1903.  Their very first flight lasted 12 seconds, and the airplane flew for a total distance of 120 feet. I have highlighted the number; you will know why. Actually by the end of that first day, the Wright Brothers have made a number of flights that lasted longer and few further. What is really remarkable is that when they flew the Flyer, we were still more than a decade away from discovering the modern theories of aerodynamics and flight (Tim's note: e.g. Ludwig Prandtl's lifting-line theory was discovered in 1919). The reason they succeeded was because they did good fundamental research, albeit empirically. In the short time since the Wright Brothers, the sophistication of aircraft has greatly accelerated. I am showing just a few of the aircraft here. The Flyer of Wright Brothers, the tri-plane of the Red Baron, a WWI ace, Charles Lindbergh;'s "Spirit of 'St. Louis"—first non-stop flight, from the US to Europe, Constellation—with its pressurized cabin, Boeing 707—the first commercially successful jetliner, Boeing 747—the largest commercial airliner until recently, Concorde—the first supersonic airliner, Voyager—the first aircraft to fly around the world non-stop and without refueling, and the Airbus A380—currently the largest commercial jetliner.

I want to point out that the Boeing 777 is one of my favourite airplanes. This is the airplane that you would be on if you fly direct from Toronto to Hong Kong. But I really want you to pay attention to the length of the aircraft: the airplane itself is longer the entire first flight of The Flyer. So, if you think about the way we fly now, people from 100 years ago would think that it is absurd. We stuff a few hundred people in a pressurized aluminum tube, at more than 10 kilometers above the surface of the earth, travelling near the speed of sound, yet we breathe normally inside this airplane, eating, drinking, watching TV or movies, or sleeping, or if you have enough money to pay for a first-class ticket to Dubai, you can even take a shower. And, with the 777 that I have just showed you, you can practically get on an airplane anywhere in the world, and land anywhere in the world.

Indeed, aviation has allowed us to go many good things. It has made travel and tourism easier—most of you would have travelled somewhere recently on an airplane; it has fostered the exchange of knowledge and culture; and it has increased commerce and trade; and most importantly, allowed for speedy delivery of humanitarian aids when disaster strikes. The picture shows a Antropov cargo plane. It is often used to send relief supplies to those disaster areas.

But aviation also has its bad side. For as long as we had airplanes, we have quickly learned how to kill people with it. In World War I—while aviation was still in its infancy—people already used airplane to drop bombs on their enemies. In WWII and the Vietnam War, we began seeing large-scale carpet bombing using very large bombers. When you can't find your enemy, you can just drop enough bombs to obliterate the whole area. Problem solved, nevermind the collateral damage. The B-52 bomber has been in this role for many decades, and will remain in the United State's arsenal of weapons for many more decades to come. The new generation of warplanes will not even be manned; they can drop bombs on people while the pilot is thousands of miles away. That's not even it's ugliest side. To end the Second World War, the Enola Gay delivered the first atomic bomb on Hiroshima, Japan and it killed thousands and thousands of people. This was followed by another atom bomb two days later. During the cold war, each B-52 bombers that you saw on the last slide carried many nuclear weapons that are hundred times more powerful than the Hiroshima bomb. Well, the Soviet Union had similar weapons too. When we're engaged in engineering research work, we must have a clear idea that what we find can be used for both good and evil.

Let us focus for a minute on how airplanes work. If you want to fly, you have to generate as much lift from the wing as the weight of the aircraft—save weight on the plane and your wing won't have to work as hard. You must also generate as much thrust as the drag of the aircraft—decrease drag and you can fly further with the same amount of fuel. The website How Stuff Works has some information that you may find interesting.


Before I go further, let me tell you how airplanes work...and how they don't. A lot of first-year textbooks will tell you that because of the shape of the wing, all the air particles flowing on the top surface have a longer path to travel, and therefore will have to speed up to meet up with the air particles on the bottom, and according to the Bernoulli equation, the pressure on the top surface is lower, and lift is generated. This is wrong because the particles on the top surface won't “know” that the particles on the bottom surface has a shorter path (the particles don't meet up at the same time anyway). But more importantly, I can have a wing where the top and bottom surfaces have the same “distance” and still generate lift. What really happens is that the presence of the wing forces the air particles to be deflected, and from the conservation of momentum, any change in momentum in the fluid must have an equal and opposite force acting on the object that causes the change, hence we have both lift and drag.

Now just for fun, let's build an airplane using 1-2 pieces of paper, which I have here, 1 paperclip, and 2 strips of tape, to build a paper airplane that can do a U-turn.

We finally get to the part about what I actually do. As a Christian, I have said many times, “In God We Trust”, and we must trust God in our daily lives. But as a researcher, we trust nothing, so everyone else, please show some background information about your research, justifications, mathematical proofs, if you have any, the description of the methods you use, models, data, plots, figures, tables, observations, conclusions, references and citations. Then, may be, I'll take notice of your work. Missing any of these, and the strength of your research is greatly diminished.

As for how exactly I do research, well, we start with The Research Question. Anyone with different background will view this a little bit differently. When the manager of a bank asks his/her staff to “research on this topic”, he means something very different from when a scientist (e.g. chemist or physicist) asks this question, which can be different from when an engineer asks the same question. The Research Question may be an old question, or it may be a completely new question. As engineers, we must also asks, what innovation do we bring to solve The Research Question. Something novel, something never been tried before. What do we discover along the way? And what other new Research Questions do we bring up?

Take for example, my work. We want to ultimately ask the question, “what is the best configuration for a commercial aircraft?” Of course, what is best depends on what you call best. Do you want to reduce operating cost? Or do you want to reduce the amount of fuel burn? Or do you want to extend the range of an aircraft? To answer each of those questions, we must now find that particular configuration. Can our tools be used to provide all the answers to those questions? How can we make these tools work faster and cheaper? What are the fundamental mathematics that governs how well our tools work? Can we now exploit the math to make our tools work better? What other design questions can we answer now that we have these tools?

My work is motivated in part by some perceived environmental concerns. The image is from the poster for the movie An Inconvenient Truth. I actually had the pleasure of seeing Mr. Gore present this talk at University of Toronto a few years ago, and while I didn't agree with everything he said, he did have some valid points to ponder.

Brothers and sisters, do we know what the global contribution of greenhouse gas emission from civil aviation? Is it 2%? 6%? 10%? or 30%? If you answered 2%, you are right.  But because in aviation, we are burning fossil fuel at high altitudes, the effects are about 3 times stronger, so if you picked 6% you're kind of right too. This problem is small compared to GHG from factories and automobiles, but it is not insignificant, because unlike other forms of transportation like cars, there are no alternatives for the way we fly...at least not in the next 20 years.

We can compare the Boeing 707—the first commercially viable jetliner—with the Boeing 767-400ER introduced 40 years later; both carry about the same number of passengers.  The 767 flies slightly faster; it's quieter, and uses less fuel and generate less pollution, and was cheaper to develop. The reason is the improvement in engine design, airframe structure, and the most important because of my work, improvements in aerodynamics. Part of engineers being able to improve on the aerodynamics is because we have something called "CFD" that has replaced wind-tunnel testing. CFD means Computational Fluid Dynamics, which I assume means absolutely nothing to you.

At the risk of sounding too technical, this is the same talk that I give to Grade 12 students when they visit the lab, and most of them get the big picture. What I am going to do is to summarize 40 years of research into one slide.  We start with a geometry. In this example, it's the cross section of a wing. We call this an airfoil. We put a computational grid around this geometry and we solve the equations at the points where these grid lines cross. Think of this problem as the same as the one you had in Grade 9, where you have to solve something x plus something y equals something, and another equation of something-else x plus something-else y equals to something else. Only in our case, the problem we try to solve is not 2 equations, but 100,000 equations all at once, and that's in two-dimensions. When we get to 3D—actually solving the problem around a real aircraft, the size of the problem goes up to 100's of millions. And we're quite good at solving these problems. We have one of the fastest, if not the fastest algorithm in the world, and we can prove that our solution is at least as accurate as wind tunnel testings. For comparison, this is the type of wind tunnel that we want to replace. This one is found at NASA's Ames Research Center, near Standard California. It is one of the biggest wind tunnels in the world: you can fit an entire airplane inside. And it comes with the huge cost too; doing an experiment with the wind tunnel costs 10 times more than using the computer, and doing experiments well is incredibly tricky.

But even that is only part of answering the research question, because knowing how the air flows around an aircraft only tells you so much, it doesn't tell you exactly how to improve on it to mathematically find the best design for a given objective.

And this is where my research work came in. What we're trying to do is to take advantage of the information we gathered during the process of solving the flow problem, and applying it in something called "optimization". Back in the days when designer had to rely on wind tunnels, the design process is mostly a cut-and-try approach. Literally. Once they gathered performance information from a wind-tunnel test, they cut a new shape that they think will be better, based on their own expertise and experience, and try the new shape in the wind tunnel again. Needless to say, that takes time and it costs money. Lots of money. And you can really never proof that you've got the optimal shape; all you can say is that there are improvements. But what we are able to do now is that quite different.

If you look at airplanes since the 1930's, you see an emerging shape that we now call the "dominant configuration". This shape emerged when engineers started pressurizing the cabin. It consists of a tube, call the fuselage, two large wings, two smaller wings near the tail, and a vertical wing, called the stabilizer. We are incredibly good at building airplanes like this, but are there any other alternatives? May be the answer lies with a flying wing, or blended wing body configuration. The pictures I have here was actually made by a high-school summer intern...we didn't even pay her, and she did great work plotting these fantastic graphics for us.

I want to show you of a real-life example that worked. My research group was contacted by a small aircraft company in British Columbia. They wanted to redesign one of their bush planes for better take-off performance. You see, these airplanes take off from lakes in BC, and if you can't climb fast enough, well, you end up in the mountains, and that's very bad for you. Anyway, we used our optimization tool to redesign the leading-edge of the wing, and also relocated the flaps. We had expected the company to do some wind-tunnel testing, instead, they went to the factory floor and built the wing that we design. Surprisingly, or may be it shouldn't come as a surprise to us, the result was beyond the company's expectation, and the work was presented at various conferences for a few years.

Here's an example of my work. I have this wing here, called the ONERA M6, and I am trying to decrease the drag on it as much as possible, while maintaining the lift coefficient. It took almost 2 days, but the result is that I optimized a wing to operate at a condition that no one thought was possible, and the final design was without shocks, which is the main contribution of drag at this Mach number.

To wrap up my talk tonight, I want to show you some of the ideas that we are looking at in terms of "wingtip devices", such as this split-tip configuration. Some of the ideas that we're exploring may appear on airplanes in the next couple of decades. So the next time you ask me for a presentation, perhaps I will be much more exciting things that I can show you!

Thank you for your time.  It's now okay to ask me questions.