Wind tunnel

English military engineer and mathematician Benjamin Robins (1707–1751) invented a whirling arm apparatus to determine drag[4] and did some of the first experiments in aviation theory.

Sir George Cayley (1773–1857) also used a whirling arm to measure the drag and lift of various airfoils.[5] His whirling arm was 5 feet (1.5 m) long and attained top speeds between 10 and 20 feet per second (3 to 6 m/s).

Otto Lilienthal used a rotating arm to accurately measure wing airfoils with varying angles of attack, establishing their lift-to-drag ratio polar diagrams, but was lacking the notions of induced drag and Reynolds numbers.[6]

Replica of the Wright brothers' wind tunnel

Eiffel's wind tunnels in the Auteuil laboratory

However, the whirling arm does not produce a reliable flow of air impacting the test shape at a normal incidence. Centrifugal forces and the fact that the object is moving in its own wake mean that detailed examination of the airflow is difficult. Francis Herbert Wenham (1824–1908), a Council Member of the Aeronautical Society of Great Britain, addressed these issues by inventing, designing and operating the first enclosed wind tunnel in 1871.[7] Once this breakthrough had been achieved, detailed technical data was rapidly extracted by the use of this tool. Wenham and his colleague John Browning are credited with many fundamental discoveries, including the measurement of l/d ratios, and the revelation of the beneficial effects of a high aspect ratio.

Konstantin Tsiolkovsky built an open-section wind tunnel with a centrifugal blower in 1897, and determined the drag coefficients of flat plates, cylinders and spheres.

Danish inventor Poul la Cour applied wind tunnels in his process of developing and refining the technology of wind turbines in the early 1890s. Carl Rickard Nyberg used a wind tunnel when designing his Flugan from 1897 and onwards.

In a classic set of experiments, the Englishman Osborne Reynolds (1842–1912) of the University of Manchester demonstrated that the airflow pattern over a scale model would be the same for the full-scale vehicle if a certain flow parameter were the same in both cases. This factor, now known as the Reynolds number, is a basic parameter in the description of all fluid-flow situations, including the shapes of flow patterns, the ease of heat transfer, and the onset of turbulence. This comprises the central scientific justification for the use of models in wind tunnels to simulate real-life phenomena. However, there are limitations on conditions in which dynamic similarity is based upon the Reynolds number alone.

The Wright brothers' use of a simple wind tunnel in 1901 to study the effects of airflow over various shapes while developing their Wright Flyer was in some ways revolutionary.[8] It can be seen from the above, however, that they were simply using the accepted technology of the day, though this was not yet a common technology in America.

In France, Gustave Eiffel (1832–1923) built his first open-return wind tunnel in 1909, powered by a 50 kW electric motor, at Champs-de-Mars, near the foot of the tower that bears his name.

Between 1909 and 1912 Eiffel ran about 4,000 tests in his wind tunnel, and his systematic experimentation set new standards for aeronautical research. In 1912 Eiffel's laboratory was moved to Auteuil, a suburb of Paris, where his wind tunnel with a two-metre test section is still operational today.[9] Eiffel significantly improved the efficiency of the open-return wind tunnel by enclosing the test section in a chamber, designing a flared inlet with a honeycomb flow straightener and adding a diffuser between the test section and the fan located at the downstream end of the diffuser; this was an arrangement followed by a number of wind tunnels later built; in fact the open-return low-speed wind tunnel is often called the Eiffel-type wind tunnel.

German aviation laboratory, 1935

Subsequent use of wind tunnels proliferated as the science of aerodynamics and discipline of aeronautical engineering were established and air travel and power were developed.

The US Navy in 1916 built one of the largest wind tunnels in the world at that time at the Washington Navy Yard. The inlet was almost 11 feet (3.4 m) in diameter and the discharge part was 7 feet (2.1 m) in diameter. A 500 hp electric motor drove the paddle type fan blades.[10]

In 1931 the NACA built a 30-foot by 60-foot full-scale wind tunnel at Langley Research Center in Langley, Virginia. The tunnel was powered by a pair of fans driven by 4,000 hp electric motors. The layout was a double-return, closed-loop format and could accommodate many full-size real aircraft as well as scale models. The tunnel was eventually closed and, even though it was declared a National Historic Landmark in 1995, demolition began in 2010.

Until World War II, the world's largest wind tunnel, built in 1932–1934, was located in a suburb of Paris, Chalais-Meudon, France. It was designed to test full-size aircraft and had six large fans driven by high powered electric motors.[11] The Chalais-Meudon wind tunnel was used by ONERA under the name S1Ch until 1976 in the development of, e.g., the Caravelle and Concorde airplanes. Today, this wind tunnel is preserved as a national monument.

Ludwig Prandtl was Theodore von Kármán’s teacher at Göttingen University and suggested the construction of a wind tunnel for tests of airships they were designing.[12]^:44 The vortex street of turbulence downstream of a cylinder was tested in the tunnel.[12]^:63 When he later moved to Aachen University he recalled use of this facility:

I remembered the wind tunnel in Göttingen was started as a tool for studies of Zeppelin behavior, but that it had proven to be valuable for everything else from determining the direction of smoke from a ship’s stack, to whether a given airplane would fly. Progress at Aachen, I felt, would be virtually impossible without a good wind tunnel.[12]^:76

When von Kármán began to consult with Caltech he worked with Clark Millikan and Arthur L. Klein.[12]^:124 He objected to their design and insisted on a return flow making the device "independent of the fluctuations of the outside atmosphere". It was completed in 1930 and used for Northrop Alpha testing.[12]^:169

In 1939 General Arnold asked what was required to advance the USAF, and von Kármán answered, "The first step is to build the right wind tunnel."[12]^:226 On the other hand, after the successes of the Bell X-2 and prospect of more advanced research, he wrote, "I was in favor of constructing such a plane because I have never believed that you can get all the answers out of a wind tunnel."[12]^:302–03

In 1941 the US constructed one of the largest wind tunnels at that time at Wright Field in Dayton, Ohio. This wind tunnel starts at 45 feet (14 m) and narrows to 20 feet (6.1 m) in diameter. Two 40-foot (12 m) fans were driven by a 40,000 hp electric motor. Large scale aircraft models could be tested at air speeds of 400 mph (640 km/h).[13]

The wind tunnel used by German scientists at Peenemünde prior to and during WWII is an interesting example of the difficulties associated with extending the useful range of large wind tunnels. It used some large natural caves which were increased in size by excavation and then sealed to store large volumes of air which could then be routed through the wind tunnels. This innovative approach allowed lab research in high-speed regimes and greatly accelerated the rate of advance of Germany's aeronautical engineering efforts. By the end of the war, Germany had at least three different supersonic wind tunnels, with one capable of Mach 4.4 (heated) airflows.[14]

A large wind tunnel under construction near Oetztal, Austria would have had two fans directly driven by two 50,000 horsepower hydraulic turbines. The installation was not completed by the end of the war and the dismantled equipment was shipped to Modane, France in 1946 where it was re-erected and is still operated there by the ONERA. With its 8m test section and airspeed up to Mach 1 it is the largest transonic wind tunnel facility in the world.[15]

On 22 June 1942 Curtiss-Wright financed construction of one of the nation's largest subsonic wind tunnels in Buffalo, N.Y. The first concrete for building was poured on 22 June 1942 on a site that eventually would become Calspan, where the largest independently owned wind tunnel in the United States still operates.[16]

By the end of World War II, the US had built eight new wind tunnels, including the largest one in the world at Moffett Field near Sunnyvale, California, which was designed to test full size aircraft at speeds of less than 250 mph[17] and a vertical wind tunnel at Wright Field, Ohio, where the wind stream is upwards for the testing of models in spin situations and the concepts and engineering designs for the first primitive helicopters flown in the US.[18]

Later research into airflows near or above the speed of sound used a related approach. Metal pressure chambers were used to store high-pressure air which was then accelerated through a nozzle designed to provide supersonic flow. The observation or instrumentation chamber ("test section") was then placed at the proper location in the throat or nozzle for the desired airspeed.

In the United States, concern over the lagging of American research facilities compared to those built by the Germans led to the Unitary Wind Tunnel Plan Act of 1949, which authorized expenditure to construct new wind tunnels at universities and at military sites. Some German war-time wind tunnels were dismantled for shipment to the United States as part of the plan to exploit German technology developments.[19]

For limited applications, Computational fluid dynamics (CFD) can supplement or possibly replace the use of wind tunnels. For example, the experimental rocket plane SpaceShipOne was designed without any use of wind tunnels. However, on one test, flight threads were attached to the surface of the wings, performing a wind tunnel type of test during an actual flight in order to refine the computational model. Where external turbulent flow is present, CFD is not practical due to limitations in present-day computing resources. For example, an area that is still much too complex for the use of CFD is determining the effects of flow on and around structures, bridges, terrain, etc.

Preparing a model in the Kirsten Wind Tunnel, a subsonic wind tunnel at the University of Washington

The most effective way to simulative external turbulent flow is through the use of a boundary layer wind tunnel.

There are many applications for boundary layer wind tunnel modeling. For example, understanding the impact of wind on high-rise buildings, factories, bridges, etc. can help building designers construct a structure that stands up to wind effects in the most efficient manner possible. Another significant application for boundary layer wind tunnel modeling is for understanding exhaust gas dispersion patterns for hospitals, laboratories, and other emitting sources. Other examples of boundary layer wind tunnel applications are assessments of pedestrian comfort and snow drifting. Wind tunnel modeling is accepted as a method for aiding in Green building design. For instance, the use of boundary layer wind tunnel modeling can be used as a credit for Leadership in Energy and Environmental Design (LEED) certification through the U.S. Green Building Council.

Fan blades of Langley Research Center's 16 foot transonic wind tunnel in 1990, before it was retired in 2004

Wind tunnel tests in a boundary layer wind tunnel allow for the natural drag of the Earth's surface to be simulated. For accuracy, it is important to simulate the mean wind speed profile and turbulence effects within the atmospheric boundary layer. Most codes and standards recognize that wind tunnel testing can produce reliable information for designers, especially when their projects are in complex terrain or on exposed sites.

In the United States, many wind tunnels have been decommissioned in the last 20 years, including some historic facilities. Pressure is brought to bear on remaining wind tunnels due to declining or erratic usage, high electricity costs, and in some cases the high value of the real estate upon which the facility sits. On the other hand, CFD validation still requires wind-tunnel data, and this is likely to be the case for the foreseeable future. Studies have been done and others are underway to assess future military and commercial wind tunnel needs, but the outcome remains uncertain.[20] More recently an increasing use of jet-powered, instrumented unmanned vehicles ["research drones"] have replaced some of the traditional uses of wind tunnels.[21] The world's fastest wind tunnel as of 2019 is the LENS-X wind tunnel, located in Buffalo, New York.[22]

Pressure across the surfaces of the model can be measured if the model includes pressure taps. This can be useful for pressure-dominated phenomena, but this only accounts for normal forces on the body.

A typical lift coefficient versus angle of attack curve

With the model mounted on a force balance, one can measure lift, drag, lateral forces, yaw, roll, and pitching moments over a range of angle of attack. This allows one to produce common curves such as lift coefficient versus angle of attack (shown).

Note that the force balance itself creates drag and potential turbulence that will affect the model and introduce errors into the measurements. The supporting structures are therefore typically smoothly shaped to minimize turbulence.

Compilation of images taken during an alpha run starting at 0 degrees alpha ranging to 26 degrees alpha. Images taken at the Kirsten Wind Tunnel using fluorescent mini-tufts. Notice how separation starts at the outboard wing and progresses inward. Notice also how there is delayed separation aft of the nacelle.

Fluorescent mini-tufts attached to a wing in the Kirsten Wind Tunnel showing air flow direction and separation. Angle of attack ~ 12 degrees, speed ~120 Mph.

China clay on a wing in the Kirsten Wind Tunnel showing reverse and span-wise flow

Oil flow visible on a straight wing in the Kirsten Wind Tunnel. Trip dots can be seen near the leading edge.

Fog (water particle) wind tunnel visualization of a NACA 4412 airfoil at a low-speed flow (Re=20.000)

• Smoke
• Carbon Dioxide Injection
• Tufts, mini-tufts, or flow cones can be applied to a model and remain attached during testing. Tufts can be used to gauge air flow patterns and flow separation. Tufts are sometimes made of fluorescent material and are illuminated under black light to aid in visualization.
• Evaporating suspensions are simply a mixture of some sort or fine powder, talc, or clay mixed into a liquid with a low latent heat of evaporation. When the wind is turned on the liquid quickly evaporates, leaving behind the clay in a pattern characteristic of the air flow.
• Oil: When oil is applied to the model surface it can clearly show the transition from laminar to turbulent flow as well as flow separation.
• Tempera Paint: Similar to oil, tempera paint can be applied to the surface of the model by initially applying the paint in spaced out dots. After running the wind tunnel, the flow direction and separation can be identified. An additional strategy in the use of tempera paint is to use blacklights to create a luminous flow pattern with the tempera paint.
• Fog (usually from water particles) is created with an ultrasonic piezoelectric nebulizer. The fog is transported inside the wind tunnel (preferably of the closed circuit and closed test section type). An electrically heated grid is inserted before the test section, which evaporates the water particles at its vicinity, thus forming fog sheets. The fog sheets function as streamlines over the test model when illuminated by a light sheet.
• Sublimation: If the air movement in the tunnel is sufficiently non-turbulent, a particle stream released into the airflow will not break up as the air moves along, but stay together as a sharp thin line. Multiple particle streams released from a grid of many nozzles can provide a dynamic three-dimensional shape of the airflow around a body. As with the force balance, these injection pipes and nozzles need to be shaped in a manner that minimizes the introduction of turbulent airflow into the airstream.
• Sublimation (alternate definition): A flow visualization technique is to coat the model in a sublimatable material where once the wind is turned on in regions where the airflow is laminar, the material will remain attached to the model, while conversely in turbulent areas the material will evaporate off of the model. This technique is primarily employed to verify that trip dots placed at the leading edge in order to force a transition are successfully achieving the intended goal.

High-speed turbulence and vortices can be difficult to see directly, but strobe lights and film cameras or high-speed digital cameras can help to capture events that are a blur to the naked eye.

High-speed cameras are also required when the subject of the test is itself moving at high speed, such as an airplane propeller. The camera can capture stop-motion images of how the blade cuts through the particulate streams and how vortices are generated along the trailing edges of the moving blade.

• Pressure Sensitive Paint (PSP): PSP is a technique whereby a model is spray coated with a paint that reacts to variations in pressure by changing color. In conjunction with this technique, cameras are usually positioned at strategic viewing angles through the walls, ceiling, and floor of the wind tunnel to photograph the model while the wind is on. The photographic results can be digitized to create a full distribution of the external pressures acting on the model, and subsequently mapped onto a computational geometric mesh for direct comparison with CFD results. PSP measurements can be effective at capturing pressure variations across the model however often require supplemental pressure taps on the surface of the model to verify the absolute magnitude of the pressure coefficients. An important property of well behaved PSP paints is they also should be insensitive to temperature effects since the temperature inside the wind tunnel could vary considerably after continuously running. Common difficulties encountered when using PSP include the inability to accurately measure the leading and trailing edge effects in areas where there is high curvature due to limitations in the cameras ability to gain an advantageous viewing angle. Additionally application of PSP on the leading edge is sometimes avoided because it introduces a finite thickness that could cause early flow separation thus corrupting results. Since the pressure variations at the leading edge is typically of primary interest, the lack of accurate results in that region is very problematic. Once a model is painted with pressure sensitive paint, certain paints have been known to adhere and continue to perform for a matter of months after initially applied. Finally PSP paints have been known to have certain frequency characteristics where some require a few moments to stabilize before achieving accurate results while others converge rapidly. In the latter instance paints that have ability to reflect rapid changes in pressure can be used for Dynamic PSP applications where the intent is to measure unsteady flow characteristics.
• Particle Image Velocimetry (PIV): PIV is a technique in which a laser sheet is emitted through a slit in the wall of the tunnel where an imaging device is able to track the local velocity direction of particles in the plane of the laser sheet. Sometimes this technique involves seeding the airflow with observable material. This technique allows for the quantitative measurement of the velocity and direction of the flow across the areas captured in the plane of the laser.
• Model Deformation Measurement (MDM): MDM works by placing markers at known geometric locations on the wind tunnel model and taking photographs of the change in the marker's location as the wind in the tunnel is applied. By analyzing the change in marker positions from different camera viewing angles, the translational change in location of the marker can be calculated. By collecting results from a few markers, the degree to which the model is flexibly yielding due to the air load can be calculated.

The main subcategories in the aeronautical wind tunnels are:

Reynolds number is one of the governing similarity parameters for the simulation of flow in a wind tunnel. For mach number less than 0.3, it is the primary parameter that governs the flow characteristics. There are three main ways to simulate high Reynolds number, since it is not practical to obtain full scale Reynolds number by use of a full scale vehicle.

• Pressurised tunnels: Here test gases are pressurised to increase the Reynolds number.
• Heavy gas tunnels: Heavier gases like freon and R-134a are used as test gases. The transonic dynamics tunnel at NASA Langley is an example of such a tunnel.
• Cryogenic tunnels: Here test gas is cooled down to increase the Reynolds number. The European transonic wind tunnel uses this technique.
• High-altitude tunnels: These are designed to test the effects of shock waves against various aircraft shapes in near vacuum. In 1952 the University of California constructed the first two high-altitude wind tunnels: one for testing objects at 50 to 70 miles above the earth and the second for tests at 80 to 200 miles above the earth.[23]

V/STOL tunnels require large cross section area, but only small velocities. Since power varies with the cube of velocity, the power required for the operation is also less. An example of a V/STOL tunnel is the NASA Langley 14' x 22' tunnel.[24]

Aircraft have a tendency to go to spin when they stall. These tunnels are used to study that phenomenon.

Automotive wind tunnels fall into two categories:

• External flow tunnels are used to study the external flow through the chassis
• Climatic tunnels are used to evaluate the performance of door systems, braking systems, etc. under various climatic conditions. Most of the leading automobile manufacturers have their own climatic wind tunnels

Wunibald Kamm built the first full-scale wind tunnel for motor vehicles.[25]

For external flow tunnels various systems are used to compensate for the effect of the boundary layer on the road surface, including systems of moving belts under each wheel and the body of the car (5 or 7 belt systems) or one large belt under the entire car, or other methods of boundary layer control such as scoops or perforations to suck it away.[26]

These tunnels are used in the studies of noise generated by flow and its suppression.

Vertical wind tunnel T-105 at Central Aerohydrodynamic Institute, Moscow, built in 1941 for aircraft testing

A high enthalpy wind tunnel is intended to study flow of air around objects moving at speeds much faster than the local speed of sound (hypersonic speeds). "Enthalpy" is the total energy of a gas stream, composed of internal energy due to temperature, the product of pressure and volume, and the velocity of flow. Duplication of the conditions of hypersonic flight requires large volumes of high-pressure, heated air; large pressurized hot reservoirs, and electric arcs, are two techniques used.[27]

The aerodynamic principles of the wind tunnel work equally on watercraft, except the water is more viscous and so sets greater forces on the object being tested. A looping flume is typically used for underwater aquadynamic testing. The interaction between two different types of fluids means that pure wind tunnel testing is only partly relevant. However, a similar sort of research is done in a towing tank.

Air is not always the best test medium for studying small-scale aerodynamic principles, due to the speed of the air flow and airfoil movement. A study of fruit fly wings designed to understand how the wings produce lift was performed using a large tank of mineral oil and wings 100 times larger than actual size, in order to slow down the wing beats and make the vortices generated by the insect wings easier to see and understand.[28]

Wind tunnel tests are also performed to precisely measure the air movement of fans at a specific pressure. By determining the environmental circumstances during measurement, and by revising the air-tightness afterwards, the standardization of the data is ensured.

There are two possible ways of measurement: a complete fan, or an impeller on a hydraulic installation. Two measuring tubes enable measurements of lower air currents (< 30,000 m³/h) as well as higher air currents (< 60,000 m³/h). The determination of the Q/h curve of the fan is one of the main objectives. To determine this curve (and to define other parameters) air technical, mechanical as well as electro technical data are measured:

Air technical:

• Static pressure difference (Pa)
• Amount of moved air (m³/h)
• Average air speed (m/s)
• Specific efficiency (W/1000 m³/h)
• Efficiency

Electro technical:

• Tension (V)
• Current (A)
• Cos φ
• Admitted power (W) fan / impeller
• Rotations per minute (RPM)

The measurement can take place on the fan or in the application in which the fan is used.

In wind engineering, wind tunnel tests are used to measure the velocity around, and forces or pressures upon structures.[29] Very tall buildings, buildings with unusual or complicated shapes (such as a tall building with a parabolic or a hyperbolic shape), cable suspension bridges or cable stayed bridges are analyzed in specialized atmospheric boundary layer wind tunnels. These feature a long upwind section to accurately represent the wind speed and turbulence profile acting on the structure. Wind tunnel tests provide the necessary design pressure measurements in use of the dynamic analysis and control of tall buildings.[30][31]