Base: A New Definition
Mitochondria: Aerobic Engines
Stripped Down
More
Zipp 808 vs TriSpoke
Wheels
Forks
Aero Bars
Frontal Area
Skinsuits
More
Roller Crr Data
Tire Wear
Project: Rolling Resistance
More
Stems
Power Meters
Wheels
More

What do Skinsuits have to do with Boundary Layers?


7/15/2002


words and images by Kraig Willett

 

1  Prologue – The Tour

 

A week ago Saturday, I sat in front of the TV for a solid two plus hours.  I had been looking forward to this self-indulgent, yet taken for granted, treat for nearly a year.  It wasn’t so long ago that all of us had to wait a week to get any coverage of the Tour de France – even then, it was watered down and tainted with John Tesh-ian keyboard rhythms and Phil Ligget trying really hard to be a wordsmith, rather than a spirited bike race announcer.  Who could forget the ‘87 Tour coverage of the Ventoux TT stage – “How far… How high… How fast…”  The images of Jean-Francois Bernard suffering up that climb with Phil doing the voiceover haunt me to this day.

Nowadays, we can turn on OLN at any of four times during the day to catch all the tour action.  And so it was last week, during the opening prologue that I noticed something peculiar when Ekimov’s ride was briefly covered by the live feed.  Something seemed a little off about the skinsuit he was wearing.  The shade of blue was a little darker.  The same Postal, TREK, and Nike graphics weren’t all there.  Then I noticed that the seams along the side of Eki’s torso trailed off into the center of his back.  This all seemed a bit odd and I figured Nike had been up to something in the wind tunnel during the off-season.   Sure enough, reports from several online media sources, OLN’s Frankie Andreu, and an official Nike Press Release confirmed that some serious wind-tunnel time had indeed been invested in the Emperor’s new time trial clothes.


The new skinsuit that the US Postal Service squad put into use in the opening prologue and yesterday’s ITT isn’t even really a “skinsuit”.  According to Nike, the clothing is called a “Swift-Spin Body Suit”.  The technology is a carryover from Project Swift, which began in 1998 as an attempt by Nike to improve the performance of runners during the 2000 Summer Olympic Games.  Swift body suits were developed at that time to improve the aerodynamics and muscle performance during the track and field sprinting events.  Similar suits were again used during the 2002 Winter Olympic Games in the speed skating events.  It has been claimed that during the skating events, the suits improved performance by approximately 0.9%.


Now, 0.9% may not seem like very much, but during an hour-long effort like the TT yesterday, this could mean timesavings of nearly 35 seconds.  If this magnitude of effect were real, and if Lance would have used this new skinsuit from Nike, he would be in the yellow jersey today.  Maybe we will see Lance donning this peculiar new suit in the next ITT.


These what-ifs and hand wavings are nice and all, but what kind of aerodynamic fundamentals is the whole concept of the suit based on.  Nike says that the suits benefit the athlete based on the same reasoning that golf balls benefit from dimples.  This can mean only one thing – Nike is claiming that cyclists can reduce aerodynamic drag by controlling the boundary layer.


Furthermore, this new clothing technology is a departure from the historical testing and the historical products offered to athletes at the highest level.  High tech skinsuits in the past were traditionally smooth surfaced and rubbery in appearance.  The track cycling program during the 1984 Olympics had such apparel.  Pearl Izumi has also offered smooth surfaced skinsuits in the past, and one can only assume that this design was a result of their participation/sponsorship of the infamous Project 96 program that was a bitter disappointment when the final medal count was tallied.


Which line of thinking is the correct one – smoother, or rougher?  After reading the Nike literature, it appears that Nike thinks that “rougher is better”.  The focus of the following article will be in discussing some basic fundamentals of aerodynamics and, in particular, boundary layer control and its manifestations in the pro peloton.


2    What is Aerodynamic Drag?

                   
Fundamentally, aerodynamic drag is a function of the geometry of the object and the physical properties of the fluid (air is a fluid) it is traveling through.  For example, an object on top of Mount Everest experiences less drag than an identical object at sea level because the air is less dense.  Drag is often times defined by the following equation:

 


Equation 1


Where r is the fluid density, u is the fluid velocity, Cd is the coefficient of drag, and A is the frontal area.
Historically, designers have decreased drag by two methods: decreasing the frontal area or decreasing the drag coefficient by streamlining.  For example, the Cd of a round bicycle spoke is approximately 1.2.  Efficient streamlining can reduce this value to around 0.3 - 0.5 (for example, the Hed3/Specialized Tri-spoke significantly reduces spoke drag coefficient by tapering its spoke geometry in much the same way as an airplane wing).  Optimal boundary layer control might be able to reduce the Cd of a bicycle spoke to nearly 0.6 to 0.7.


3   2-D Cylindrical Flow, 3-D Spherical Flow


The 2-d flow around a cylinder is one of the fundamental flows in fluid dynamics.  Much research has been done to investigate this flow since it is simple, yet filled with many unique features.  For purposes of this discussion, it can be assumed that the flow around a 2-d cylinder adequately represents the basic features of the flow around a cyclist.


One of the most interesting features of the 2-d cylinder flow is the dramatic drop in drag it experiences when the boundary layer transitions from laminar to turbulent flow (see Figure 1 – the same is true for the flow around a sphere).  The point at which this phenomenon occurs is known as the critical Reynolds number, and it stands to reason that it would be desirable to design a skinsuit (or a golf ball) so that the flow resides in the corresponding low drag region above this critical value.  When Nike says that their “Swift-Spin” suit is based on the same reasoning as why golf balls have dimples, they are saying that they have induced the low drag flow above the critical Reynolds number.  But this is a lot of information to digest without a little background information.

 
Figure 1.  Drag on 2-d cylinder, with the critical Reynolds number located near 2E+05 (based on data from Lindsey, W.F – NACA Rept., 619 – 1938).


4    Boundary Layer Fundamentals


All solid bodies exposed to a viscous fluid flow have a thin layer near its surface called the boundary layer (see Figure 2 below).  In this region, the fluid velocity changes from the free-stream velocity (the fluid velocity away from the surface) to a velocity of zero at the body surface.  The boundary layer is the cause of the aerodynamic drag that all bodies are subject to in a viscous fluid.


Figure 2:  Boundary layer on a flat plate.


Aerodynamic drag forces can be divided into two components: skin friction drag and pressure, or form drag.  Skin friction drag dominates in streamlined flows like airplane wings.  Bluff bodies, like cyclists in the TT position, or objects that create large wakes are susceptible to large pressure drag.

5     Skin Friction Drag and Streamlined Flows


Fluids have a property commonly known as viscosity, which characterizes the relative ease with which the fluid can be deformed (i.e, viscosity is a measure of how well the fluid molecules “stick” to one another).  A high viscosity fluid requires a larger shear force to deform than a low viscosity fluid.  There is a relationship that describes the effects of viscosity on that shear force:

 


Equation 2

 

Where tau is the shear force, m is the fluid viscosity, and the ratio du-dy is the velocity gradient (change in velocity relative to the change in distance) perpendicular to the surface of the solid body.  Looking at Equation 2, we can see that as the viscosity of the fluid increases, the shear force, or skin friction drag, increases.  Similarly, when there is a higher velocity gradient there is higher skin friction. 

 

5.1.1       Reynolds Number Dependence


There is a common term used in aerodynamics that describes the relationship between the viscous forces and the inertial forces for a given fluid and solid object.  Viscous forces have been briefly described already.  Inertial forces can be thought of as those forces that are related to the fluid’s velocity field.  In equation form, the Reynolds number is defined as:


Equation 3

 

Where r is the fluid density and L is a length scale (diameter of cylinder, cyclist chest width, chord length of a 2-d airfoil, length of a flat plate in the direction of the flow, etc...). 


Low Reynolds numbers (not necessarily less than 1, anything on the order of 1,000 is a low Reynolds number) mean that viscous forces will dominate the flow. For example, the flow over a flat plate with a Reynolds number of 1 (r, m, L, and u equal to 1) means that the viscous forces of the fluid are large compared to the inertial forces.  The low Reynolds number means that boundary layer has a smooth and gradual transition from the surface to the free-stream.  The viscous fluid forces “suck” the molecules of the fluid together resisting the inertial shearing forces, and overall drag remains low.  There is negligible pressure drag due to the absence of a wake, and therefore, the shear force developed by the boundary layer is the primary contributor to the overall drag on the object.  Low Reynolds number flows have very small overall drag due to the small skin friction drag and negligible pressure drag.

 

Inertial forces, however, dominate high Reynolds number flows where large velocity gradients are present.  In high Reynolds number flows, the viscous forces of the fluid molecules are too small to successfully resist the shearing action of the high velocity (and high inertia) of the free-stream flow.  Large changes in velocity over very small distances are present in these flows and, as a result, high skin friction drag is created.


Using the same flat plate example as before but with a Reynolds number of 300,000 (r, m, L = 1, u = 300,000) it is easy to see that a much larger change in velocity from the wall to the free-stream takes place (zero to 300,000 as opposed to 0 to 1).  Recalling Equation 2, it can be shown that the large velocity gradient will result in large shearing forces.  The skin friction drag generated in high Reynolds number flows is significantly greater than in low Reynolds number flows.  These types of high Reynolds numbers are never generated during normal bike riding conditions.  Skin friction concerns should, therefore, not drive design feature choices in the bluff body geometries encountered on bikes/riders.


5.1.2       Transition to Turbulence

 

When the boundary layer changes from a laminar flow to a turbulent flow it is referred to as transition (see Figure 3).  Everyone who has seen a candle, match, or cigarette burn in still air has witnessed a transition to turbulence.  Near the flame, the smoke streams upward in a steady smooth flow.  The stream of hot air remains organized in laminar (tightly layered) sheets of smoothly varying velocities.  As the smoke travels away from the flame, the Reynolds number of the flow increases (because the “L” in Equation 3 is increasing).  At a certain distance away from the flame, the smoke begins to swirl and mix violently with the rest of air in the room.  This violent mixing of the smoke and the still air surrounding it signals that a transition from the smooth laminar flow near the flame to the turbulent flow away from the flame has occurred.



Figure 3.  Transition to turbulence in a plume of smoke.

 

For a large number of flows, transition takes place at a Reynolds number of approximately 500,000.  It should be realized though that transition does not naturally occur at one single Reynolds number, rather, it occurs over a range of Reynolds numbers (Figure 4 below).  500,000 is a widely accepted value that ensures transition has occurred for many flows (flat plates, 2-d cylinder, sphere, etc..).


5.1.3       Laminar Flow

 

Laminar flow generally refers to the properties of the boundary layer surrounding an object in a moving fluid.  In a laminar flow, there is no mixing of the boundary layer flow and the free-stream flow.  The velocity gradients within the layer are small.  As a result, skin friction drag in a laminar flow is minimal.

 

For streamlined bodies not susceptible to large pressure drag, laminar flow is desirable (highly tapered head tubes, downtubes, leading surfaces of helmets).   The overall drag of the object will be less since the skin friction for laminar flow is less than the skin friction for a turbulent flow.  There are several things that can be done to help maintain a laminar flow over a solid body for as long as possible.


5.1.3.1       Promoting Laminar Flow

 

The most important variable in promoting a laminar flow over a streamlined body is the surface roughness.  The smoother the surface of the object is, the longer the boundary layer will remain laminar.  Surface roughness churns up the fluid within the boundary layer and hastens the transition to turbulence.  For example, if a uniform flow were blown over a smooth flat plate and also a flat plate that had tiny grains of sand glued to the surface, the smooth plate would maintain a laminar boundary layer, and correspondingly less drag, for a longer distance.  This theory is what has been driving helmet design and skinsuit design up to this point (with a few notable exceptions in the helmet arena).

 

Another tool at the disposal of engineers is altering the geometry of the object.  Shaping the object in such a way that a favorable pressure gradient (as the flow travels along the surface the pressure is decreasing) is set up on the surface for as long as possible is desirable.   A favorable pressure gradient is usually formed up until the point of maximum thickness on the object.  After the point of maximum thickness an unfavorable, or adverse, pressure gradient is formed which tends to induce turbulence.  Laminar flow wings have a peak thickness that is further downstream than a typical commercial aircraft wing due to its longer favorable pressure gradient and laminar flow inducing characteristics.

 

The length of the object in the direction of the flow (chord length for wings) also affects laminar flow properties.  A shorter length means the flow cannot reach high Reynolds numbers and transition cannot take place.  There is a practical limit that this technique can be employed, but in general, the shorter the streamwise length, the more laminar the flow.

 

Engineers of the long distance airplane “Voyager” developed laminar flow wings, which used smooth, short chord, favorable pressure gradient wings to fly around the world on one tank of gas.  Gliders, since they have no on board power, also use these techniques on their wings. 


5.1.4       Turbulent Flow

 

Turbulent flow is characterized by the unsteady mixing of the free-stream flow and the boundary layer.  The injection of high momentum fluid into the boundary layer increases the velocity gradient near the surface, which increases the skin friction drag.  With streamlined objects, this turbulent flow is undesirable since the skin friction component of drag is increased.  However, with bluff bodies, (where wakes and large separated regions are present) turbulent flows can be advantageous.  It is this theory that Nike is basing its “Swift-Spin Body Suit” on.

 

6     Pressure Drag and Bluff Body Flows

 

All prior discussions have focused primarily on flows that are dominated by skin friction drag.  Flows that are dominated by the pressure, or form, component of drag are sometimes referred to as bluff bodies.  The flow around a cyclist’s torso, upper arms, and legs can be described as bluff body type of flows.  In these types of flows, the skin friction drag is negligible since there are large regions of the flow on the leeward side (side of the body that is downstream of the maximum point of thickness) that are separated.  These separated regions on bluff bodies are also referred to as wakes.


 
Figure 5.  Separated flow about a cylinder – leeward side is to the right, windward to the left.


The leeward side of the body has a lower pressure than the windward side of the body.  This pressure differential results in a force that is directed against the direction of travel, thus, the term “pressure drag”.


6.1        Velocity Profile in the Boundary Layer


The fundamental process that causes a wake to form is an adverse pressure gradient.  The boundary layer on a flat plate will never leave the surface and form a wake.  A flat plate has no pressure gradient to cause this to happen.  However, a bluff body, such as a cyclist’s torso, has definite, and severe, pressure gradients.

 

On the windward side of a bluff body (upstream of the point of maximum thickness) a favorable pressure gradient is in place.  As a particle in the flow passes along the windward surface, the pressure is dropping, causing the particle to accelerate.  This favorable gradient keeps the boundary layer energized and the flow is “sucked” on to the body.  Once the point of maximum thickness is reached the flow continues onto the leeward side of the body.  It is on the leeward side of the object that an adverse pressure gradient is present and separation of the flow becomes possible


6.1.1       Separation and Wakes

 

Separation occurs when an adverse pressure gradient slows the flow in the boundary layer such that the stream-wise component of velocity (velocity parallel to the surface) goes to zero, as shown in Figure 6.  The perpendicular component of velocity “lifts” or “separates” the flow off the surface and a wake is created.  Downstream of this separation point is a region of flow reversal, in which there is a re-circulation of the fluid in a direction opposite of the free-stream and the low pressure on the surface of the body does not recover.  With a bluff body, flow separation is difficult to avoid and minimization of this phenomenon through streamlining and/or boundary layer control is highly advantageous.



Figure 6.  Velocity profile in the boundary layer and separation.


6.2    Reynolds Number


Since the Reynolds number is the ratio of inertial forces to viscous forces, a flow in which the viscous forces dominate (low to moderate Reynolds number) will lead to lower overall drag.  Low Reynolds numbers flows, however, are not very practical and for the most part are not widely seen in commercial applications.  For the majority of flows (with a few important exceptions), bluff body and streamlined included, as the Reynolds number increases the overall drag increases (see Equation 1).


6.3     Laminar Flow


With laminar flow and streamlined bodies, the Reynolds numbers are low and the viscous forces are dominant.  When the geometry of the object changes to a bluff body, in which large adverse pressure gradients are present, a laminar flow is undesirable.  The low energy contained in the laminar boundary layer is not able to resist these large pressure gradients effectively, and as a result, separation with a laminar boundary layer occurs much sooner than with a turbulent boundary layer.  Laminar boundary layers should be avoided when the geometry of the object is considered to be a bluff body due to the large wakes and correspondingly high-pressure drag values that result.


6.4        Turbulent Flow


Turbulent flow is desired on bluff-bodied objects.  Even though there is a corresponding increase in the skin friction component of drag with a turbulent boundary layer, the pressure drag can be significantly decreased.  The turbulent boundary layer and its violent mixing with the high momentum free-stream fluid is better able to resist adverse pressure gradients.  As a result, separation is delayed and the size of the wake on the leeward side of the body is reduced.


With a reduced wake, the high-pressure zone on the leeward side is smaller and the overall drag force is decreased.  Several important geometries in fluid dynamics clearly show the effects of a turbulent boundary layer and its effect on the reduction of overall drag.  The sphere and the 2-d cylinder are two classic examples of how a turbulent boundary layer leads to a smaller wake and less overall drag.


A natural transition to turbulence occurs once a Reynolds number of ~500,000 has been achieved.  On streamlined surfaces or surfaces where laminar flow is desired, the boundary layer is allowed to develop naturally.  With bluff bodies, however, it is sometimes desirable to induce a turbulent flow to hasten the benefits of reduced pressure drag.   This portion of the field of fluid dynamics is sometimes referred to as boundary layer control.


7    Boundary Layer Control – Inducing Turbulence


Often times it is desirable to prematurely induce turbulence in the boundary layer.  This procedure, called tripping the boundary layer, involves placing artificial flow turbulence creators on the surface of the body.  This technique is used in the small-scale wind tunnel setting to ensure full-scale flow behavior on lower Reynolds number models.  There are several methods of controlling the boundary layer that are used, which include vortex generators, trip strips, and surface roughness.


7.1    Vortex Generators


Vortex generators work by creating a streamwise vortex, which mixes the high momentum free-stream flow into the low momentum boundary layer.  A large-scale streamwise vortex can be seen on the wing tips of aircraft as seen in Figure 7).  The high-pressure undersurface air wraps around the tip of the wing to meet the low-pressure top surface air.  When this curling action is added to the streamwise velocity of the aircraft (forward velocity), a 3-d helix of spinning air is created.  This swirling air is called a streamwise vortex.
 Vortex generators operate on a much smaller scale, but the streamwise vortexes they create, inject high momentum free-stream air into the boundary layer through their twirling action (see Figure 8).  These features are placed on aircraft wings to improve lift, by delaying separation, at the high angles of attack seen during take-off and landing.  The vortex generators on wings are usually placed in a row just forward of the maximum thickness and take the shape of a small sheet metal tab placed at an angle relative to the oncoming flow.
The drawback to this method is that it is dependent on orientation.  The layout of the vortex generators must be carefully designed and flow misalignments will decrease performance.


7.2       Trip Strips


A less elegant and much lower tech method for inducing boundary layer turbulence is a trip strip.  A trip strip is a surface feature that extends past the boundary layer into the free stream and is usually place perpendicular to the flow.  As the free stream flow passes over the strip (usually a wire, wall, or series of 3-d columns of material) a localized separated flow is created that mixes the high momentum free-stream fluid into the boundary layer.  Descente had a skinsuit available in the early to mid 90’s that incorporated some of these types of features in the shoulder region.  This product was never really accepted by the marketplace.
Trip strips are usually used in the wind tunnel where orientation and placement are carefully controlled.  However, there are underwater applications where this method was used with success in the real world.

 

Recent experiments with a series of longitudinal trip strips on an underwater cable proved successful in reducing the size of the wake (as measured by accelerated submersion velocities).  US Patents have been issued that describe this method (see patent number 3884173) as it pertains to underwater cables.
Depending on the design, this method can also be dependent on orientation of the trip strip.  If incorrectly designed, the boundary layer may not be tripped or the overall drag may actually be increased.


7.3     Surface Roughness


The final boundary layer method to be discussed is the one that is the least dependent on orientation and it is also the one that is implemented by Nike engineers in their new “Swift-Spin” skinsuit.  It is also the theory behind the large dimples in the current Lazer helmets and the Troxel radius Ti helmets of the past.  A rough surface finish will cause the transition to turbulence to occur at lower Reynolds numbers.  This method should be completely independent of orientation assuming the surface roughness is completely random.


Careful attention must be taken when determining the size of the surface roughness features, however.  Too large of a roughness and the drag might actually increase, but too small and no gains will be made.  Nike claims that over 50 types of fabric were investigated during the development of their suits.  It is apparent that they were trying to dial in the surface roughness for each particular body feature (torso, upper arms, legs, etc..).


7.4        2-d Cylinder Flow Redux


Remember the plot that started this whole thing?  The one that showed a large drop off in drag just past the critical Reynolds number for a 2-d cylinder?  The one that Nike is banking its R&D dollars on?  Now that there has been some detailed discussion about some fundamental aerodynamics we can re-investigate this flow and tie it all together.  The flow region with Reynolds numbers below the critical value is termed sub-critical.


7.4.1       Sub-Critical Flow


Since the boundary layer in a sub-critical flow is still laminar it cannot resist the adverse pressure gradients introduced past the point of maximum thickness.  As a result, the laminar boundary layer becomes separated from the surface of the cylinder just aft of the point of maximum thickness (approximately 90° clockwise and counter-clockwise from the stagnation point) and a large wake and a large low-pressure zone on the leeward side of the body results.  This point on a cyclist is analagous to where the “love handles” are located on the lower torso, and just below the armpits.  Even though the boundary layer is laminar and skin friction is low, the drag on the cylinder is high due to the large wakes and the subsequent dominant pressure drag.


7.4.2       Super-Critical Flow


The boundary layer in the super-critical 2-d flow around a cylinder is turbulent.  The highly energized turbulent boundary layer is better able to resist the adverse pressure gradient on the leeward side of the cylinder.  Separation is delayed past the point of maximum thickness, up to approximately 120° clockwise and counter-clockwise from the stagnation point.  The resulting smaller wake decreases the size of the low-pressure zone on the leeward side of the cylinder and the overall drag is decreased.  Even though the boundary layer has transitioned into turbulence and the skin friction has increased, the drag is lower due to the significantly reduced pressure drag.  Nike engineers have apparently induced a turbulent boundary on the appropriate body parts of the USPS riders through clever utilization of finely tuned and selected surface roughness lycra.  They have therefore claimed to reduce the size of their wakes and, therefore, their total drag.


8       Summary


Nike implemented boundary layer control features on selective panels of their new “Swift-Spin Body Suit” that induce a super-critical flow around a cyclist’s body.   Overall drag has been claimed to be decreased by simply putting on their new product.


With the release of the Nike product, Nike is saying that for the bluff body flows around a cyclist, where large wakes are present, it is advantageous to induce turbulent boundary layers (thus increasing skin friction drag) in order to take advantage of the super-critical flow reduction in pressure drag.  It is for this fundamental aerodynamic reason that golf balls have dimples, and Lance and his US Postal mates are being offered rougher than normal surface textured skinsuits (“Swift-Spin Body Suits”).  Cycling history says that Nike engineers have gotten it all wrong, but Lance and his team may just prove them right when they roll onto the streets of Paris - but first, Nike has to convince Lance to slip one of these bad boys on!!!

 

What did the inventors of the Nike Suits had to say about this article?  Read it here!