The flow of fluids can be analyzed by theory, numerical computation, and experiment. Visualization is one of many experimental tools for surveying or measuring the flow of a fluid that is normally invisible due to its transparency. By applying the methods of flow visualization, a flow pattern is made visible and can be observed directly or recorded with a camera. The information on the flow is available for the whole field of view at a specific instant of time. This information can be either qualitative, thus allowing for an interpretation of the mechanical and physical processes involved in the development of the flow, or quantitative, so that certain properties of the flow (e.g., velocity, density) can be measured. The methods of flow visualization can be classified according to three basic principles: light scattering from tracer particles; optical methods relying on refractive index changes in the fluid; interaction processes of the fluid flow with a solid surface. While the two former methods serve to visualize the pattern in the interior of a flowing fluid, the third method provides information on the transfer of momentum, heat, or mass between the fluid and a solid body.
The scattering and optical methods are based on the interaction of the fluid with light. A light wave incident into the flow field (Figure 1) may interact with the fluid in two different ways:
light can be scattered from the fluid molecules or from tracer particles with which the fluid is seeded; and
the properties of the light wave can be changed due to a certain optical behavior of the fluid and, as a consequence, the light transmitted through the flow is different from the incident light.
The visualization methods based on these two interaction processes are totally different in nature and are applicable to different types of flow.
The principles of flow visualization and a great number of respective methods are discussed by Merzkirch (1987). Examples of visualized flow are presented in collections compiled by Van Dyke (1982) and the Japan Society of Mechanical Engineers (1988). The progress in this field is well described in the Proceeding of the International Symposia on Flow Visualization edited by Asanuma (1979), Merzkirch (1982), Yang (1985), Veret (1987), Reznicek (1990), and Tanida and Miyashiro (1992).
Since the light scattered from the fluid molecules (Rayleigh scattering) is extremely weak, the flow is seeded with small tracer particles (e.g., dust, smoke, dye), and the more intense radiation scattered from these tracers is observed instead. It is thereby assumed that the motion of the tracer is identical with the motion of the fluid, an assumption that does not always hold, particularly in unsteady flows. The scattered light carries information on the state of the flow at the position of the tracer particle, that is, the recorded information is local. For example, if the light in Figure 1 is incident as a thin light sheet normal to the plane of the figure, an observer could receive and record information on the state of the flow (e.g., the velocity distribution) in the respective illuminated plane.
The signal-to-noise ratio in this type of flow visualization can be improved if the tracer does not just rescatter the incident light but emit its own, characteristic radiation (inelastic scattering). This principle is realized by fluorescent (e.g., iodine) or phosphorescent (e.g., biacetyl) tracers which may emit bright fluorescing (or phosphorescent) light once the fluorescence is induced by an incident radiation with the appropriate wavelength (laser-induced Fluorescence or Phosphorescence, respectively).
That a flow becomes visible from foreign particles that are floating on a free water surface or suspended in the fluid is a fact of daily experience. This crude approach has been refined for laboratory experiments. The methods of flow visualization by adding a tracer material to the flow are an art that concerns the selection of the appropriate tracers, their concentration in the fluid, and the systems for illumination and recording. (See also Tracer Techniques.)
The trace material, after being released, is swept along with the flow. If one does not resolve the motion of single particles, qualitative information on the flow structure (streamlines, vortices, separated flow regimes) becomes available from the observed pattern. The identification of the motion of individual tracers provides quantitative information on the flow velocity, provided that there is no velocity deficit between the tracer and the fluid. Only in the case of fluorescent (or phosphorescent) tracers it is possible to deduce data on quantities other than velocity (density, temperature). Using special optical filters, the radiation of this inelastic scattering can be separated from any other radiation in the test field, which makes the fluorescence methods attractive for studying combustion processes.
Besides some general properties that any seed material for flow visualization should have (e.g., nontoxic, noncorrosive), there are mainly three conditions the tracers should meet: neutral buoyancy, high stability against mixing, and good visibility. The first requirement is almost impossible to meet for air flows. Smoke or oil mist are the most common trace materials in air, with the particle size of these tracers being so small (<1 μm) that their settling velocity is minimized, A number of neutrally buoyant dyes are known for the visualization of water flows, the colors introducing an additional component of information [Merzkirch (1987)].
Specific liquids allow a visualization by photochromic reaction. For a short period of time the fluid molecules are converted into an opaque state, and their motion becomes visible. The reaction is reversible, and the fluid is again transparent at the end of this period. The visible fluid particles are neutrally buoyant, but the experiment is restricted to the special properties of these liquids. (See also Photochromic Dye Tracing.)
Special arrangements of illumination and recording as well as timing are necessary if the goal is to measure the velocity of individual tracer particles. A time exposure (Figure 2) is a convenient way for visualizing the instantaneous velocity distribution in a whole field (plane) of the flow. This plane section is realized by expanding a thin (laser) light beam in one plane by means of a cylindrical lens, so that all tracer particles in this plane light sheet are illuminated. The velocity component normal to the plane is not recovered. Recordings of the motion of tracer particles, taken either with a photographic or electronic camera, contain a great number of quantitative data on the planar velocity distribution. Methods employing the technique of digital image analysis have been developed for extracting and presenting the information on the velocity distribution that is often shown in the form of a vector plot (Figure 3). These methods are known as Particle Image Velocimetry and Particle Tracking Velocimetry [Adrian (1991)].
Figure 2. Stroboscopic exposure of tracer particles in a water flow. (S. Hilgers, Universitat Essen).
The properties of the light wave transmitted through the fluid flow as indicated in Figure 1 may be changed due to changes of the refractive index of the fluid. The refractive index (or index of refraction) is a function of the fluid density. The relationship is exactly described by the Clausius-Mosotti equation; for gases, this equation reduces to a simple, linear relationship between the refractive index, n, and the gas density, p, known as the Gladstone-Dale formula [see, e.g., Merzkirch (1981)]:
where the Gladstone-Dale constant K is different for each gas and weakly dependent on the light wave length. A light wave transmitted through the flow with refractive index changes is affected in two different ways; it is deflected from its original direction of propagation, and its optical phase is altered in comparison to the phase of the undisturbed wave. These alterations of the wave properties can be made visible in a recording plane at a certain distance behind the flow field under study. A particular method requires the use of an optical apparatus transforming the measurable quantity (light deflection, optical phase changes) into a visual pattern. The pattern is either qualitative (shadowgraph, schlieren) or quantitative (moire, speckle photography, interferometry), thus allowing for a deduction of data values of the refractive index or density distribution in the flow. (See also Photographic Techniques, Shadowgraph Technique, and Interferometry.) Refractive index variations occur in a fluid flow in which the density changes, e.g., because of compressibility (high-speed aerodynamics or gas dynamics), heat release (convective heat transfer, combustion), or differences in concentration (mixing of fluids with different indices of refraction).
A standard case of applying optical flow visualization is convective heat transfer. Variations of the fluid density are caused by the respective temperature distribution. Figure 4 is a shearing interferogram of the vertically rising plume (natural convection) above a candle flame that serves as a heat source. The distortion of the oblique, parallel, equidistant fringe system is a measure of the temperature gradient in the ascending gas flow. Quantitative evaluation of the temperature field is in most cases restricted to laminar flow.
Figure 4. Shearing interferogram of the plume above a candle flame. (H. Vanheiden, Universitat Essen).
Figure 4 is a two-dimensional (plane) projection of a three-dimensional flow field. The information on the density distribution is integrated along the path of the transmitted light("line-of- sight methods"), i.e., the obtained information is not local as in the case of the techniques using light scattering (see Figure 1). For the purpose of obtaining quantitative results on the three-dimensional density distribution, it is necessary to record with the optical setup several projections in different directions through the flow and to process the optical data by methods known as computer tomography [Hesselink (1989)].
The interaction of a fluid flow with a solid body is the subject of many experimental investigations. Such studies are aimed at determining, e.g., the shear forces, pressure forces, or heating loads applied by the flow to the body. A possible means of estimating the rates of momentum, mass, and heat transfer is to visualize the flow pattern very close to the body surface. For this purpose, the body surface can be coated with a thin layer of a substance that, upon an interaction with the fluid flow, develops a certain visible pattern. This pattern can be interpreted qualitatively, and in some cases it is possible to measure certain properties of the flow dose to the surface. Three different interaction processes can be used for generating different kinds of information:
In the most common technique, which applies to air flows around solid bodies, the surface is coated with a thin layer of oil mixed with a finely powdered pigment. Because of frictional forces, the air stream carries the oil with it, and the remaining streaky deposit of the pigment gives information on the direction of the flow close to the surface. The observed pattern may also indicate positions where the flow changes from laminar to turbulent and positions of flow separation and attachment (Figure 4).
The solid surface is coated with a substance that changes color upon the chemical reaction with a material with which the flowing fluid is seeded. The reaction, and therefore the color change, is the more intense, the higher the mass transfer from the fluid to the surface. Separated flow regimes with little mass transfer rates can be discriminated from regions of attached flow.
Coating materials that change color as a function of the surface temperature (sensitive paints, liquid crystals) are known. Observation of the respective color changes allows for determining the instantaneous positions of specific isothermals and deriving the heat transfer rates to surfaces, which are heated up or cooled down in a fluid flow. Equivalent visible information is available, without the need of surface coating, by applying an infrared camera.
Adrian, R. J. (1991) Particle imaging techniques for experimental fluid mechanics, Ann. Rev. Fluid Mech., 23, 261-304.
Asanuma, T., Ed. (1979) Flow visualization, Proc. 1st Int. Symp. Flow Visualization, Hemisphere, Washington, D.C.
Hesselink, L. (1989) Optical tomography, Handbook of Flow Visualization, W.-J. Yang, Ed., Hemisphere, Washington, D.C.
Japan Society of Mechanical Engineers, Ed. (1988) Visualized Flow, Pergamon, Oxford.
Merzkirch, W. (1981) Density sensitive flow visualization. Fluid Dynamics, Vol. 18 of Methods of Experimental Physics, Ed. R. J. Emrich, Academic Press, New York.
Merzkirch, W., Ed. (1982) Flow visualization II, Proc. 2nd Int. Symp. Flow Visualization, Hemisphere. Washington, D.C.
Merzkirch, W. (1987) Flow Visualization, 2nd edn. Academic Press, San Diego. Reznicek, R., ed. (1990) Flow visualization V, Proc. 5th Int. Symp. Flow Visualization, Hemisphere, Washington D.C.
Tanida, Y, Miyashiro, H., Eds. (1992) Flow visualization VI, Proc. 6th Int. Symp. Flow Visualization, Springer, Berlin.
Van Dyke, M., Ed. (1982) An Album of Fluid Motion, Parabolic Press, Stanford, California.
Veret, C, Ed. (1987) Flow visualization IV, Proc. 4th Int. Symp. Flow Visualization, Hemisphere, Washington, D.C.
Yang, W.-J., Ed. (1985) Flow visualization III, Proc. 3rd Int. Symp. Flow Visualization, Hemisphere, Washington, D.C.