An overview of room air motion measurement: technology and application.

INTRODUCTION

It is widely believed that the Navier-Stokes equations govern detailed behaviors of fluid flow. Unfortunately, there are no analytical solutions using these equations even with some of the simplest turbulent flows (Moin and Mahesh 1998). To solve the Navier-Stokes equations numerically, the solutions of which are known as direct numerical simulations (DNSs), is apparently the only way to obtain a complete description of a turbulent flow. However, the current computer technology, even with the fastest supercomputers, is still insufficient to simulate turbulent flow in a large space such as a room using an efficient DNS code (Graham 2003; Gharib 1996). The computing power limitation of direct simulations is usually addressed by the use of simplified mathematical flow models combined with experimental validation. To obtain reliable and precise information on actual airflow patterns and velocity distributions, experimental measurements are still a must at present.

The great need for experimental data has led to the development of many flow measurement systems for flow velocity and the velocity distributions. A flow velocimetry technique can be regarded as point-wise or global-wise. Point-wise measurement techniques, such as Pitot tubes, thermal anemometers, and laser Doppler velocimetry, can be used to obtain the velocity information at the point of the probe. In order to measure the global velocity distribution in a whole flow field, arrays of these point-wise measurement probes have to be used, which may either disturb the flow field, such as with intrusive methods (e.g., thermal anemometers) or excessively increase the complexity of equipment settings and the measurement cost for the nonintrusive or partially intrusive methods, such as laser Doppler velocimetry. Particle image velocimetry (PIV), particle tracking velocimetry (PTV), and other global-wise techniques, which have been intensively studied in recent years, can measure fluid velocities over the flow field or, in other words, a global domain. The global domain can be either a plane (e.g., PIV technique) or a volumetric space (e.g., holographic PIV and volumetric PTV).

The room airflow is normally characterized by relatively low velocities, by high turbulence levels, and, in most cases, by unsteady behavior in a relatively large enclosure. The commonly used methods at present and the potential techniques for the near future for indoor airflow measurements include rotating vane anemometers, thermal anemometers, ultrasonic anemometers, Laser-Doppler velocimetry, visualization techniques, particle image velocimetry, particle tracking velocimetry, and molecular tagging velocimetry. These techniques are reviewed in the remainder of this paper.

ROTATING VANE ANEMOMETERS

Rotating vane anemometers (Figure 1) sense the air speed from the pressure differentials. Airflow causes the anemometer rotor to rotate with an angular speed that is directly proportional to the airflow speed. The typical measuring range of a modern vane anemometer for ventilation application is from 0.25 to 30 m/s (50 to 6000 ft/min) with a resolution of 0.01 m/s (2 ft/min), and the accuracy of [+ or -]0.05~0.1 m/s ([+ or -]10~20 ft/min) is in the low measurable velocity range. For low-speed airflows under 0.25 m/s (50 ft/min), the measurement accuracy of vane anemometers is very low and cannot be used. The anemometers are simple, but their responsiveness, typically longer than 0.5 seconds, is not good enough to measure a turbulent flow in a common fluid field.

[FIGURE 1 OMITTED]

Rotating vane anemometers are point-wise and intrusive. The sensing head has a large size. The probe intrusiveness will unavoidably disturb the flow field. When used in an array, the distance between two probes should be larger than two diameters of the vane anemometer. Because of the requirement for alignment with the airflow direction and their poor performance for low-speed airflows, rotating vane anemometers typically have been applied in in-situ measurements of face velocities at the room air inlets or outlets where the airflow has a relatively high speed and a steady direction. The price of a rotating vane anemometer ranges from less than $100 to $1000 at present.

THERMAL ANEMOMETERS

Thermal anemometers measure the local flow velocity through its relationship to the convective cooling of electrically heated sensing elements (Webster 1999). Thermal anemometers can be further classified as constant-current anemometers (CCAs), constant-temperature anemometers (CTAs), and constant-temperature-difference anemometers (CTDAs). The operating principle of a typical commercial thermal anemometer is shown in Figure 2. The heat transfer from the sensing elements is a function of the air velocity and the temperatures of the surface and the air. The dimensions of thermal anemometer sensing parts are very small (e.g., the hot-wire diameter is only several micrometers), and current manufacturing technology is incapable of maintaining sufficiently small tolerances to ensure sensor reproducibility. Therefore, each thermal anemometer must be calibrated individually. One of the most significant error sources of thermal anemometers is the difference between measurement and calibration conditions. In low-velocity flow fields, the free convection, or thermal buoyancy, generated by the heat from the sensing elements becomes significant around the sensor itself, and the resulting output is consequently prone to considerable inaccuracy. The shape of the sensors could be wires (hot-wire anemometry), films (hot-film anemometry), or spheres (thermisters).

[FIGURE 2 OMITTED]

Hot-Wire Anemometers

The sensing element of a hot-wire anemometer is a tungsten or platinum-alloy wire 0.8~1.5 mm (0.03~0.06 in.) long and 2.5-7.5 [micro]m in diameter. The wire is mounted at its two ends on thin metallic prongs, usually tapered and having diameters of less than 1 mm. The wire is heated by electric current and cooled by the local airflow. This thermal effect is recorded and converted together with the current information to find the local flow velocity. The small sensing elements

allow hot-wire anemometers to have fast time responses; thus, they are capable of measuring velocity fluctuation frequencies up to 10,000 Hz and have a millimeter range spatial resolution. Although it is possible to measure airflow as low as 0.01 m/s (2 ft/min) with a single-component hot-wire probe, the accuracy is subject to the meter's design capability. The inaccuracy is on the order of 10%-25% at 0.05-0.5 m/s (10-100 ft/min) for a single-component hot-wire anemometer. The hot-wire sensors are fragile and thus suitable only for clean airflows.

More than one component of velocity can be measured if multiple wires are mounted. For two-or three-dimensional airflow studies, probes with, respectively, two or three perpendicular wires in an "X" pattern are used. Usually, a thermal anemometer also includes a sensor for the fluid temperature and a bridge circuit to compensate for temperature variations. Air velocity components cannot be measured simultaneously with multicomponent hot-wire probes if the air velocity is below 0.2 m/s (40 ft/min) (Zhang 1991) due to the thermal buoyancy generated by the heat from the sensing elements.

The price range for a single-component hot-wire anemometer is $1000 to $15,000. For a three-component hot-wire anemometer system, the price ranges from $15,000 to $250,000.

Hot-Film Anemometers

The sensors of hot-film anemometers are fabricated by depositing a thin film of conducting metal (usually platinum or alumina) over the surface of a firm nonconducting substrate (usually quartz). Cylindrical film sensors are typically 50 [micro]m (0.002 in.) in diameter. Compared to hot-wire sensors, these larger diameter film sensors generally have a smaller signal-to-noise ratio and slower frequency response. The response of hot-film sensors is about 10 to 20 Hz, which is still sufficient for turbulence measurements in many indoor airflow applications. The primary advantages of hot-film sensors over hot-wire sensors are that film sensors are sturdier and more stable in retaining their calibrations than hot-wire sensors. Therefore, they may be used in dirty airflows with particulates. Film sensors generally are more expensive than wire sensors.

Hot-Sphere Anemometers

Hot-sphere anemometers were specially designed for indoor airflow applications. The sensors are of a spherical shape with a diameter of 1 to 3 mm (0.04 to 0.125 in.) and operate at a lower temperature. Unlike hot-wire or hot-film sensors, hot-sphere sensors are omnidirectional, i.e., they are insensitive to the ambient airflow direction and therefore essentially measure the total speed. The typical measuring range is from 0.1 to 5 m/s (20 to 1000 ft/min) with a resolution of about 0.01 m/s (2 ft/min). Due to the relatively large size of the sensing elements, this type of sensor has a time constant of typically about two seconds. Therefore, they are mainly for measuring mean velocity and not suitable for measuring velocity fluctuations. However, in one study (Li 1995), a spherical thermister used as a sensing unit was able to detect room airflow fluctuations up to 20 Hz. The inaccuracy of hot-sphere anemometers is on the order of 10%-20% at 0.1~0.5 m/s (20~100 ft/min), and there is a need for regular calibration. The sensors may not be completely omnidirectional. Directional sensitivity exists due to the probe support, the sensor guard, and the orientation.

Hot-sphere anemometers cost less than hot-wire anemometers. The price of a hot-sphere anemometer system ranges from $500 to $4000.

Thermal anemometers are point-wise and intrusive. A typical measuring time for one point is about two to ten minutes. To define the flow patterns in a large fluid space, a sufficient number of points have to be measured simultaneously. The probe intrusiveness will unavoidably disturb the flow field. This effect is magnified when multiple points are measured, especially when the probes are close to each other. Compared to rotating vane anemometers, thermal anemometers can measure lower velocities, have much better spatial and time resolution and a wider range of measurement, and can be made smaller. However, they generally have a much higher cost.

ULTRASONIC ANEMOMETERS

Popular measurement systems using the ultrasonic technique include transit-time-type and Doppler-type systems (Webster 1999). Ultrasonic anemometers of the transit-time type (Figure 3) are typically used for collecting meteorological data, especially wind speed and direction. The principle for ultrasonic anemometers is that the speed of a sound wave varies with the local air velocity. The measured velocity is a mean value for the measurement volume of a typical dimension of about 50 mm (2 in.) and ranges from 0.01 to 20 m/s (2 to 4000 ft/min). Three-axis anemometers can resolve the three orthogonal velocity components ([U.sub.x], [U.sub.y], and [U.sub.z]). The measurement of three-dimensional air velocity vectors relies on an ultrasonic pulse traveling back and forth between three sensor probes. The spatial resolution is given by the path length between sensors, which is typically about 35 to 100 mm (1.4 to 4 in). The temporal resolution is 20 Hz or better.

[FIGURE 3 OMITTED]

An ultrasonic anemometer of the transit-time type is relatively simple, easy to operate, and, once properly calibrated at the factory, needs no further calibration. It can measure all three velocity components simultaneously and evaluate mean velocities and turbulence intensities. There is no influence of temperature. The inaccuracy could be as low as 0.01 m/s (2 ft/min).

Ultrasonic anemometers are also point-wise and intrusive. Their probe head size is relatively large and mechanically disturbs the airflow field more strongly than thermal anemometers. Although data collection is simple, the preparation required to ensure proper orientation of ultrasonic probes is labor intensive (Taylor et al. 2004). The price for an ultrasonic anemometer ranges from $2000 to $20,000.

Sonic Doppler measurement systems make use of the Doppler frequency shift caused by sound reflected or scattered from suspensions in the flow path to detect the speed of the suspensions, which is assumed to be the same as the flow. The operating principle is similar to radar systems except that the sonic Doppler technique uses sound waves instead of radio waves, which are used in the radar technique. The sonic Doppler technique is widely used in liquid flowmeters and sonic detection and ranging (SODAR) systems for the measurements of wind profile in the lower layer of the atmosphere. The sonic Doppler systems are not considered as accurate as the transit time systems. In liquid flowmeter applications, the technique requires that the scattering particles must be larger than 100 [micro]m (0.004 in.) and have a concentration greater than 100 ppmv. It is difficult in a typical size room to seed the airflow with tracer particles (e.g., soap bubbles) to reach the required concentration. Additionally, similar to the ground clutter problem with SODAR systems for wind measurements, the reflection of the sound waves from the room interior surfaces and objects may overwhelm the return signal from the scattering particles and cause the measurement to be unreliable. These two principal hurdles must be overcome before the technique can be applied in indoor air movement measurements.

LASER DOPPLER VELOCIMETRY

Laser Doppler velocimetry (LDV) is a method of measuring fluid velocities by detecting the Doppler shift of a laser light that has been scattered by small particles (typically 1~10 [micro]m [0.00004~0.0004 in.]) moving with the fluid (Adrian 1996a). It is assumed that the particles faithfully follow the flow and that their velocity represents the velocity of the flow. The fundamental principle of LDV (see Figure 4) is as follows. A monochromatic, coherent, linearly polarized beam of a laser is split into two parallel beams by a beam splitter. The two parallel beams are focused onto a small control volume by the transmitting lens. When a small particle moves from that volume, it will scatter the light of both beams. Known as the Doppler effect, the scattered light has a wavelength difference, which is determined by the velocity of the particle, from the wavelength of the incident light (the laser). A photodetector is used to collect the scattered light, and the wavelength is analyzed to obtain the particle velocity (Durst et al. 1976; Tavoularis 2005).

[FIGURE 4 OMITTED]

Using LDV, up to three components of flow velocity ranging from micrometers per second to hypersonic speeds can be measured at any time over a small volume, which is in any case small enough to be considered as a point for applications in room air. It is well developed and becomes a well-established optical technique. The advantages of the LDV technique include (Menon 1999): (1) it has high measurement accuracy (Goldstein and Kreid [1967] reported that 0.1% absolute accuracy was achieved for the measurements of flow development in a square duct, typically about 1% error in mean velocity for indoor air motion measurements); (2) it has the ability to measure any desired velocity component; (3) it is a partly intrusive method; (4) it has accurate measurement of high-turbulence intensities, including flow reversals; (5) it is independent of temperature and fluid properties; (6) it has a large dynamic range; (7) there is no required velocity calibration; and (8) it has good frequency response, which is basically limited by the time constants and response frequencies of scattering particles used in room air measurements. The natural particulate matter in the dusty room air can be conveniently used as the light scattering particles. The air in a clean room can be seeded with artificial particles, including silicon carbide particles, alumina particles, polystyrene spheres, microballoons, peanut oil droplets (Adrian 1996a), or aerosols of 10% glycol in water (Johnson et al. 1996).

Because conventional LDV can only furnish the instantaneous velocity at a single point in a flow, to obtain the complete flow pattern requires a time-intensive scan of the region of interest. When the flows are very voluminous, then the effort or the cost involved in scanning the flow with a single-point probe and a high-powered laser are enormous and may even be prohibitive. That is why the successful applications of LDV are mainly found in reduced models in room airflow measurements (Murakami and Kato 1989) and it is less suited for extensive measurements for full-scale room indoor airflows. LDV systems are expensive. Depending on the complexity of the system, from simple one-component to three-component velocity measurements and depending on the laser power and data acquisition, the price ranges from $50,000 to $150,000.

Doppler global velocimetry (DGV), also known as planar Doppler velocimetry, overcomes the drawback of LDV's single-point measurements (Meyers 1995). Currently, however, DGV is considered to be more suitable for the measurement of high-speed flows (higher than 10 m/s [2000 ft/min]) because of the limit on velocity resolution, which is approximately equal to 0.5~1 m/s (100~200 ft/min), the absolute error of current state-of-the-art DGV systems (Meyers et al. 2001; Willert et al. 2002).

VISUALIZATION TECHNIQUES

Visualization techniques can obtain extensive information on flow patterns and have been used to verify existing physical principles (Mueller 1996) or to support measurement results obtained from other techniques (Loomans 1998). The underlying strategy of these techniques is to make the fluid movement visible by seeding special tracers into the flow field. The required equipment and materials include a lighting system, tracer media, and an image recording system (e.g., a video camera). There are some criteria for choosing an appropriate tracer media. The density of the tracer should be similar to that of the fluid; otherwise it cannot follow and represent the fluid flow accurately. The tracer should also be easy to trace (i.e., to detect). Common tracers are smoke, bubbles, microballoons, and fluorescent particles. For example, fluorescent dyes and a laser detector were used to study the flow mixing patterns of an automobile air-conditioning unit (Nobuyuki et al. 1997). In a room ventilation simulator (Zhang 1991), smoke was used as the tracer to observe airflow patterns in a typical animal building section. After being employed for more than a century, flow visualization techniques are considered practical and effective methods of obtaining qualitative fluid information. The primary limitation of conventional visualization methods is that they only give an approximation of flow pattern and cannot provide quantitative velocity data. If the visualization techniques are combined with digital image processing techniques, quantitative information of flow velocities can be extracted from the recorded images.

PARTICLE IMAGE VELOCIMETRY

Particle image velocimetry (PIV) is a technique that makes an instantaneous velocity measurement in a plane across the whole flow field (see Figure 5). It uses a multiple-pulsed laser at a high intensity for a short duration (normally around 0.1-100 ns) to form a narrow two-dimensional light sheet placed in the plane of interest of the seeded flow. The displacements of seeded particles in a known time interval between two laser pulses can be recorded by a photographic or digital video camera synchronized with the light sheet. The cross-correlation between the two images, each from one exposure, is calculated to find the similar pattern of particle distribution and local displacement. Hence, velocity values across in the plane can be determined by dividing the displacement by the time interval between two exposures (Adrian 1991). PIV combines whole-field visualization with the instantaneous capture of the data, and is a robust technique for industrial applications. It has become a routine experimental technique for a broad variety of flows. A number of commercial PIV packages, furnished with the latest digital technologies and being almost plug-and-play, are available on the market. The two main components of a PIV system are the imaging acquisition system, which produces a double-exposed photographic image of the particles in the flow field, and the interrogation system, which extracts and presents the velocity field information contained in the photographs.

[FIGURE 5 OMITTED]

Coupled charged device (CCD) cameras are usually used in PIV systems for their convenience of data transmission and image processing, although their resolutions are much lower than those of professional film cameras. Ideally, the resolution of a PIV camera should be more than 2-4 pixels per tracer particle diameter. With a high-speed digital video camera of the current state of the art, the airflow in the measurement area can be recorded at a high frame rate. Then the detailed cinematic sequence of the flow provides the ability to study the characteristics of unsteady airflow fields, to track Lagrangian paths, and to evaluate mean velocities and turbulent quantities. This technique is referred to as cinematic PIV (CPIV) by some researchers (Lin and Rockwell 1994). It requires powerful capabilities of cameras and computers and cannot yet obtain measurement results in real time.

The illumination system is also a key component in a PIV system. It allows the camera to record clear images of the particles within the measurement volume and prevent the particles outside the measurement volume from forming image noise. The most widely used light source in PIV systems is the laser. The laser light sheet, typically about 1-3 mm (0.04-0.125 in.) thick, is the illumination source employed in most PIV systems. The measuring space has the same thickness as that of the illuminating sheet. Nevertheless, using a thin laser sheet as the illumination source causes a significant challenge for measuring the instantaneous flow field in spaces with sizeable depth. In a strong three-dimensional flow field, many particles, whose images are captured by the camera at the first laser pulse, may move out of the thin-depth measurement volume and cannot be seen by the camera at the next laser pulse. This limits the measurement accuracy and restricts the PIV measurements to regions of nearly planer flow (Hart 2000). The light sheet should be oriented parallel to the direction of the planer flow. The maximum allowable measuring area depends on the light intensity, the camera resolution, and the size of the tracer particles. The full-scale room airflow field has such a large dimension that most currently available laser light generators are not capable of covering the whole measuring area uniformly. Currently, the measurement area of PIV systems is restricted to a small area, typically less than 1 [m.sup.2] (11 [ft.sup.2]).

Compared with lasers, conventional light sources appear to be more practical for full-scale indoor airflow measurement. Different types of conventional light sources have also been employed in PIV experiments. When light pulses are desired, a common photographic flash lamp is considered one of the simplest and least expensive light sources (Adrian 1986). For steady-state illuminations, certain white light sources, such as halogen lamps (Zhao 2000) and arc lamps (Scholzen and Moser 1996), have proven sufficiently efficient and can be manipulated to form a light sheet with a thickness between 5 and 30 cm (2 and 12 in.) in a full-scale room. When using large particles, which are commonly employed in full-scale room test cases, the disadvantage of white light sources being relatively low in intensity is somewhat compensated for because, when reflectivity is fixed, particles with larger cross-sectional areas lead to a higher light reflection. Most traditional light sources are capable of offering an array illumination with one simple set of cylindrical lenses. To decrease the heat gain effect from light beam radiation on the airflow, the illumination lights should be switched on only when the photos are being taken.

PIV is a partially intrusive measurement method. The fluid must be seeded with tracer particles with an appropriate particle concentration and illuminated by a properly shaped light beam. The tracer particles play a critical role in PIV measurements since the PIV technique actually measures the velocity of the particles rather than that of the fluid. The physical density of the tracer particles should be comparable to that of the ambient fluid, and their size should be as small as possible to guarantee good following behavior but large enough to be visible with suitable illumination. Their shape should be spherical and the surface should be highly reflective. If a particle is steadily moving in a steady airflow in a room, its slip velocity can be estimated as

[V.sub.slip] = [[g[d.sub.p.sup.2][C.sub.c]([[rho].sub.p] - [[rho].sub.f])]/[18[[eta].sub.f]]] (1)

where g is the acceleration of gravity, [d.sub.p] is the particle diameter, [C.sub.C] is the slip correction factor (Zhang 2005), [[rho].sub.p] is the particle density, and [[rho].sub.f] and [[eta].sub.f] are the density and viscosity of the air, respectively. From Equation 1 it can be seen that there is a difference between the particle velocity and the actual air velocity when the particle is not neutrally buoyant in the air. This slip velocity principally determines the uncertainty of PIV measurements.

When a particle is injected into the airflow with an initial velocity, [V.sub.p0], the velocity difference between the particle and the air will change with time t as

[V.sub.f] - [V.sub.p] = [V.sub.slip] + ([V.sub.f]-[V.sub.p0] - [V.sub.slip])exp(- t/[tau]), (2)

where [tau] is the relaxation time of the particle:

[tau] = [[(2[[rho].sub.p] + [[rho].sub.f])[d.sub.p.sup.2][C.sub.c]]/[36[[eta].sub.f]]] (3)

Only when t[much greater than][tau] is the effect of initial particle velocity negligible. In an accelerating flow, the less the relaxation time of the particle, the less the velocity difference between the particle and the flow. In a turbulent flow, the neutrally buoyant particles can track the fluctuating behavior of the flow well. If the particle density is much larger than the fluid density in a turbulent flow, the ideal Stokes number of the particles should be less than 0.1. The Stokes number, St, represents the ratio of the particle relaxation time to the characteristic time of the fluctuating flow:

St = [omega][tau] (4)

where [omega] is the frequency of the velocity fluctuation of the turbulent flow.

In the indoor environments, it is necessary that the tracer particles have a low environmental impact (no corrosion, no health hazard, etc.). Furthermore, it is an advantage if they can be easily stored, handled, and injected into the flow. The widely used particles for room air motion measurements with PIV techniques are fog oil droplets.

The ideal concentration of the tracer particles is 10 to 20 particles in an interrogation window (Adrian 1996a). For measurements of air motion in a normal room, a great effort is usually needed to reach the concentration of the seeding particles.

In general, basic PIV systems do not require external spatial calibration. Pixel coordinates of particle images viewed by the camera are related to particle positions in units of millimeters or meters in the actual plane by the camera magnification, which can be calculated from the system geometry and camera parameters. However, a more accurate relationship between the pixel coordinates of particle images and the particle positions in the actual space may be created through calibration. An accuracy of 5% to 10% is estimated for the measurements using the two-component technique.

The PIV technique has lower limits of speed detection and can present more detailed flow structures than conventional anemometers such as hot-wire anemometers. The time resolution of a PIV system is dependent on the setting of laser pulse intervals. The measurable velocity range is mainly dependent on the pulse interval, camera resolution, tracer particle size, and system configuration. For a fixed interrogation size, the velocity gradients in the measurement area must be low. The spatial resolution in the light sheet plane is usually small enough for the room airflow application. The spatial resolution in the out-of-plane direction is the thickness of the light sheet. When conventional light sources instead of laser are used, the light sheet could be thick, and the resolution in this direction would be poor. PIV systems specially for indoor airflow measurements are not available commercially. The price of a common PIV system ranges from $50,000 to $150,000, comparable to that of the LDV systems.

Basic PIV systems can only make a planar measurement of two-velocity components, and the flow to be measured should be mainly two dimensional. To measure three velocity components in a plane or volume, some extension techniques to basic PIV have been developed.

Extensions to Basic Particle Image Velocimetry

In order to obtain three components of flow velocities in the fluid domain, several techniques have been proposed as extensions of the basic PIV system. Among them, scanning PIV, color PIV, defocus PIV, and stereoscopic PIV are relatively more commonly used in fluid measurements.

Scanning PIV. Scanning PIV is a multiple-plane technique as a straightforward extension of basic single-light-sheet PIV systems to measure volumetric flows by moving a light sheet rapidly across the volume of interest in the flow field (Brucker 1997). The sampling in depth adds information about the volume and/or the out-of-plane component of the velocity vectors, which can be extracted by using correlation processing techniques (Raffel et al. 1995; Brucker 1996). This method requires proper beam-shaping optics and a high-speed camera to record the particle images in the corresponding light sheet along the depth of the measurement field.

Color PIV. Color PIV uses multiple color recording and/or multiple color light sources to obtain additional information of the particle images and, hence, to measure more flow properties or to enhance data-processing methods (Lawson 2004). The technique has been used to resolve directional ambiguity (Rivir et al. 1995; Gogineni et al. 1998) and to separate different phases in a two-phase flow (Towers et al. 1999). Color PIV can be combined with scanning PIV to improve the performance of measurements (Cenedese and Paglialunga 1989; Brucker 1996).

Defocus PIV. Defocus PIV can resolve three velocity components in a substantial depth of the measurement volume by placing an aperture mask with a minimum of two holes in front of the lens of the imaging camera (Willert and Gharib 1992; Pereira et al. 2000; Pereira and Gharib 2002). However, the measurement accuracy of out-of-plane components of velocities is much lower than that of in-plane components, and a high level of light power is required.

Stereoscopic PIV. Stereoscopic PIV has been well developed in recent years and is the most commonly used technique to measure instantaneous three-dimensional velocity fields in a plane. Many researchers have studied and applied it in various flow environments. It employs two cameras to simultaneously observe a common area of interest in the illuminated plane from two different views and, thus, to extract three velocity components in a two-dimensional domain (Prasad 2000). The two cameras can be set in translational (Figure 6a) or angular displacement (Figure 6b) configuration. The translational systems (Arroyo and Greated 1991; Prasad and Adrian 1993) require that the object plane, lens plane, and image plane are all parallel to each other and the lens of all the cameras must be in the same plane. The common area viewed by both cameras is usually small unless some special techniques are adopted, which increases the difficulties in the system setups and produces additional errors of the image recording. Angular displacement systems (Prasad and Jensen 1995; Lawson and Wu 1999) have much fewer requirements for the alignment of system setups, have more accurate measurements across the entire field of view, and are more widely used than translational systems. Some researchers extend the basic angular system with a dual plane setup to obtain three-dimensional vorticity data (Kahler and Kompenhans 2000; Hu et al. 2001). The disadvantages of angular displacement include nonuniform focus over the viewed area and viewing image area through media with varying indices of refraction. The measurement volume of both configurations is a mere sheet or a very limited depth. To resolve directional ambiguity, some special techniques, e.g., image shifting or an additional camera, are employed, which multiplies the work for system setups and/or image processing and adds a source of errors. The measurement error of the out-of-plane component is usually several times higher than that of the in-plane components. To obtain the three-dimensional data from the two-dimensional images, system calibration is generally needed.

[FIGURE 6 OMITTED]

All of the extension methods mentioned above require some specific lighting and/or imaging configurations, careful camera calibration, and specific evaluation strategies (Hinsch 1995). Although the measurement volume may be extended beyond a mere sheet to a significant depth, the depth is still not large enough to cover meaningful three-dimensional flow fields. And they cost much more than the basic PIV systems.

PARTICLE TRACKING VELOCIMETRY

As an extension of flow visualization, particle tracking velocimetry (PTV) determines particles' trajectories and velocities by capturing and analyzing each particle image to locate its center and connecting image tracks (possibly multiple exposures) (Adrian 1996a), therefore offering an approach to investigate Lagrangian characteristics of fluid motion (Agui and Jimenez 1987; Virant and Dracos 1997; Cenedese and Querzoli 1997).

A PTV system setup is similar to PIV system setups and also includes image acquisition hardware and image and data processing software. During a measurement, the flow is seeded with tracer particles, which should follow the flow faithfully. Helium-filled bubbles are widely used as tracer particles for room air motion measurements with a PTV system. A lighting system, such as a laser or an incandescent light source, is required to illuminate the particles in the observation region. One or multiple cameras are used to record the images/videos of the particle motion (Adrian 1991). Then the image and data processing software are used to identify the image of an individual particle, track the particle image in the temporal dimension, match the images of the particle from all of the cameras (if multiple cameras are used), and calculate the velocity of the particle. The particle identification includes detection of the particle images from the background and their position recording. The temporal tracking links the particle images in the consecutive time frames to find the path that the particle was moving along. The image matching fuses the two-dimensional images from all of the cameras together to form the stereoscopic vision.

PTV is usually applied to low seeding density fields in which the distance a particle travels between exposures is small compared to the average distance to its nearest neighbor particle (Uemura et al. 1989; Adrian 1996a). Hence, it has a significant limitation for spatial resolution and cannot be used to investigate complicated fine flow structures. Although PTV's imaging setup is similar to PIV's, it is more strongly affected by out-of-plane noise than correlation-based methods. To obtain three-dimensional vectors of velocities, PTV can also use techniques similar to those of PIV. One similar PTV method to color PIV is to tailor the light sheet for a well-defined intensity profile in depth so that the brightness of a particle image reveals its depth coordinate (Dinkelacker et al. 1992). In another, similar way to defocus PIV, PTV can use the size of an ordinary out-of-focus image to indicate a particle's depth location (Stolz and Kohler 1994). When multiple cameras are used to capture particle images from different directions or locations simultaneously, three-dimensional vectors can be extracted based on some data processing algorithms (Hardalupas et al. 2000; Doh et al. 2000; Doh et al. 2004). High-speed video cameras with a high resolution and high frame rates, like in CPIV systems, are desired to track the airflow paths during a long period of time for room air measurements. However, the ability of the available PTV systems to track flow in three dimensions is still problematic. For example, meeting the Scheimpflug condition (Altenhofen 1952) for all the cameras, enhancing signal-to-noise ratio in the measurements of flows inside complex geometries, and matching particle images among all the cameras pose challenges to both the hardware setup and the image and data processing algorithms of a PTV system for measurements in a three-dimensional flow volume.

As the simplest single-frame mode of PTV, particle streak velocimetry (PSV) captures particle streak images by exposing particles in a light sheet over a period of exposure time. The major limitations of PSV include that the streak lengths cannot be too short or too long and that the image processing of the particle streak lines is not easy because of overlapping and streak deformation. Zhao et al. (2001) used PSV to measure two-dimensional airflow in a full-scale room. The directional ambiguity of streaks was eliminated by employing the image shifting technique. An ordinary light source was used to form a 0.7 m (28 in.) thick light sheet. Airflow patterns, velocity vector maps, and iso-velocity contours under different ventilation and thermal conditions were successfully obtained in this study. PSV performs excellently on velocity measurements for two-dimensional flow fields, but poorly in three-dimensional fields if a narrow laser sheet is used and the streaks may be truncated (Adrian 1991). To avoid this problem, researchers have tried thick light sheets with the three-dimensional measurements. Muller et al. (2001) applied three light sheets of two different colors together to measure the third velocity component. Scholzen and Moser (1996) developed a three-dimensional PSV with a 120 mm (0.4 ft) thick white light sheet to acquire the particle streak images with three cameras in a 2.4 x 1.7 x 1.2 m (78.7 x 5.6 x 39.4 ft) ventilated space.

A unique measurement method, volumetric particle tracking velocimetry (VPTV) (Sun and Zhang 2003; Sun et al. 2004), has been successfully developed to measure all three spatial components of flow velocities in a whole three-dimensional flow volume of a full-scale room using ordinary light illumination. The VPTV system uses two cameras to view the illuminated flow field and capture particle displacement images with the parallax effect. The technique can acquire three velocity components and flow directions in a deep three-dimensional volume rather than on a thin light sheet. Helium-filled bubbles of neutral buoyancy were used as the airflow tracer particles. Two digital cameras in an angular configuration with a synchronizer record quantitative and instantaneous global views for the flow field volume. The calibration of the system showed that the average absolute error was 0.05 m/s (10 ft/min), and the largest relative error was below 30% for the measurements of room airflows. The system has been used in the experimental characterization of airflows in a aircraft cabin mock-up (Sun et al. 2005; Zhang et al. 2005).

Even though one can find every hardware component needed for a PTV system on the market, there are, unfortunately, not off-the-shelf PTV systems available commercially at present because of the complex computational algorithms involved. The total system is still under research and development.

MOLECULAR TAGGING VELOCIMETRY

Molecular tagging velocimetry (MTV) tracks the displacement of the tagged flow regions of interest within a given time delay through imaging flow-tracing molecules excited by a pattern of a pulsed laser to determine the flow velocity (Miles and Lempert 1997; Koochesfahani 1999; Zheng and Klewicki 2000; Lempert et al. 2002). If MTV is applied with the fluorescent emission only from the molecules intrinsically present in the flow itself, e.g., oxygen in airflow, it is totally nonintrusive and very useful when external particle seeding is undesirable or may lead to complications connected to tracking the flow, density mismatch, particle seeding density, or strong out-of-plane motions, such as in combustion and microfluidics applications (Gendrich et al. 1997; Miles et al. 2000; Lempert et al. 2003). The applications of excited-state oxygen fluorescence under the acronym RELIEF (Raman excitation plus laser-induced electronic fluorescence of oxygen) have been mainly in high-speed flows and for small regions of a fluid field (Koochesfahani 1999; Miles et al. 2000). MTV has been commonly used to make a measurement of one or two components of the velocity over a plane in either a liquid or a gas flow (Stier and Koochesfahani 1999; Koochesfahani et al. 2000). By using a stereo-imaging technique similar to stereoscopic PIV, all three components of the velocity vectors over a plane can also be measured using the MTV method (Bohl et al. 2001). The technique has not been applied in room airflow measurements, and there is no commercial product available at present.

SUMMARY

Table 1 summarizes the features of the commonly used techniques discussed in this paper for the measurement of room air motion. Defocus PIV and MTV are not included in the table because they have not been applied in room air measurements and are currently not available commercially.

Table 1. Features of Techniques for the Velocity Measurement
of Room Air Motion

Technique         Velocity Range       Accuracy      Time
                                                  Resolution

Rotating vane   0.25-30 m/s        [+ or -]0.02   0.5 s or
anemometer      (50-6000 ft/min)   m/s ([+ or -]   longer
                                   4 ft/min)
                                   in lower
                                   measurable
                                   velocity
                                   range

Hot-wire        0.05-30 m/s        [+ or -]0.02   0.1 ms or
anemometer      (10-6000 ft/min)   m/s ([+ or -]  longer
                with               4 ft/min)
                single-component   or 3% of
                probes; 0.2-30     reading,
                m/s (40-6000       whichever is
                ft/min) with       greater
                multi-component
                probes

Hot-film        Similar to         Similar to     0.1 s or
anemometer      hot-wire           hot-wire       shorter
                anemometer         anemometer

Hot-sphere      0.05-5 m/s         Similar to     1-4 s
anemometer      (10-1000 ft/min)   hot-wire
                                   anemometer

Ultrasonic      0.01-20 m/s        [+ or -]0.01   50 ms or
anemometer      (2-4000 ft/min)    m/s ([+ or -]  shorter
                                   2 ft/min)
                                   or 1% of
                                   reading,
                                   whichever is
                                   greater

Laser Doppler   0 to >100 m/s (0   [+ or -]1%     From
velocimetry     to >20,000                        <1[micro]s
                ft/min)                           to 10 ms

Visualization   --                 --             --

Particle image  Highly dependent   [+ or -]       Dependent
velocimetry     on each            5%-10%         on the
                individual system                 pulse
                and the operating                 interval;
                parameter                         possibly
                settings                          from less
                                                  than 1 ms
                                                  to 500 ms

Scanning PIV    Same as PIV        [+ or -]       Same as
and color PIV                      5%-10% for     PIV
                                   the in-plane
                                   components;
                                   the
                                   out-of-plane
                                   componentis
                                   higher

Stereoscopic    Same as PIV        [+ or -]       Same as
PIV                                5%-10% for     PIV
                                   the in-plane
                                   components;
                                   the
                                   out-of-plane
                                   Componentis
                                   Several times
                                   higher than
                                   the in-plane
                                   components

Particle        Similar to PIV;    Same as PIV    Same as PIV
tracking        in PSV mode, the   or SPIV
velocimetry     factor between
                the maximum and
                Minimum
                velocities is
                less than 15

Technique          Spatial             Pros                Cons
                  Resolution

Rotating vane   35-100 mm      * Simple and easy to  * Point-wise
anemometer      (1.4-4 in.)    operate

                               * Low cost            * One
                                                     dimensional

                                                     * Significant
                                                     intrusiveness

                                                     * Must be
                                                     calibrated

                                                     * Not accurate
                                                     for low
                                                     velocity

                                                     * Low spatial
                                                     and time
                                                     resolution

Hot-wire        1 mm (0.04     * Up to               * Point-wise
anemometer      in.)           three-dimensional

                               * High spatial        * Intrusive
                               resolution

                               * High time           * Must be
                               resolution            periodically
                                                     calibrated

                               * Wide rangeability   * Not accurate
                                                     for low room air
                                                     velocities

                                                     * Expensive

Hot-film        Similar to     * Up to               * Point-wise
anemometer      hot-wire       three-dimensional
                anemometer

                               * High spatial        * Intrusive
                               resolution

                               * Good time           * Must be
                               resolution            periodically
                                                     calibrated

                               * Able to measure     * Not accurate
                               airflows with         for low room air
                               particulates          velocities

                               * Wide rangeability   * Expensive

Hot-sphere      A few mm       * Omni-directional    * Point-wise
anemometer      (tenths of an
                inch)          * Less cost than      * Intrusive
                                hot-wire anemometer
                                                     * Must be
                                                     periodically
                                                     calibrated

                                                     * Unable to
                                                     measure velocity
                                                     fluctuation

                                                     * Not accurate
                                                     for low room air
                                                     speeds

Ultrasonic      35-100 mm      * Up to               * Point-wise
anemometer      (1.4-4 in.)    three-dimensional

                               * No calibration      * Intrusive
                               needed during use

                               * Good time           * Low spatial
                               resolution            resolution

                               * Wide rangeability   * Not accurate
                                                     for low room air
                                                     velocities

Laser Doppler   1 mm (0.04     * High measurement    * Point-wise
velocimetry     in.) or        accuracy
                smaller
                               * Partly-intrusive    * Expensive

                               * No calibration
                               required

                               * Up to
                               three-dimensional

                               * High tempal and
                               spatial resolution

Visualization   --             * Simple              * Nonquantitative

                               * Inexpensive

                               * Viewable

Particle image  Mainly         * Planar              * Two-dimensional
velocimetry     dependent on     measurement
                the tracer
                particle size
                in the light
                sheet plane
                and the light
                sheet
                thickness in
                the
                out-of-plane
                direction

                               * Partially           * Limited-size
                                 -intrusive          window
                                                     measurement

                               * High time and       * Difficult to
                               spatial resolutions   seed the room
                                                     flow with tracer
                                                     particles at the
                                                     appropriate
                                                     concentration

                               * Low limit of        * Expensive
                               velocity detection

Scanning PIV    Same as PIV    * Volumetric          * Limited zone
and color PIV                  measurement           measurement

                               * Three-dimensional   * Difficult to
                                                     seed the room
                                                     flow with tracer
                                                     particles at the
                                                     appropriate
                                                     concentration

                               * Partially           * Not easy to
                                 intrusive           use

                               * High time and       * Expensive
                                spatial resolutions

                               * Low limit of
                               velocity detection

Stereoscopic    Same as PIV    * Planar              * Limited zone
PIV                            measurement           measurement

                               * Partially-          * Difficult to
                                 intrusive           seed the room
                                                     flow with tracer
                                                     particles at the
                                                     appropriate
                                                     concentration

                               * High time and       * Not easy to
                               spatial resolutions   use

                               * Low limit of        * Expensive
                               velocity detection

                               * Three-dimensional

Particle        A few mm to    * Planar or           * Not easy to
tracking        several        volumetric            use
velocimetry     hundred mm     measurement

                               *                     * Difficult
                               Partially-intrusive   image
                                                     processing

                               * High time           * Expensive
                               resolution

                               * Low limit of
                               velocity detection

Technique         Applicability        Cost               Commercially
                                                        Available?

Rotating vane   * In-situ            $50-$1000          Yes
anemometer      measurement of face
                velocities at the
                room air inlets or
                outlets

Hot-wire        * At single or       $1000-$15,000 and  Yes
anemometer      several points for   $15,000-$250,000
                mean velocities and  for
                turbulence           single-$250,000
                quantities           component and a
                                     three-component
                                     hot-wire
                                     anemometer
                                     systems,
                                     respectively

                * Array of
                measurements
                throughout the
                room

Hot-film        * Similar to         More expensive     Yes
anemometer      hot-wire             than hot-wire
                anemometer           anemometer
                                     systems

Hot-sphere      * At single or       $500-$4,000        Yes
anemometer      several points for
                mean velocities

                * Array of
                measurements
                throughoutthe room

Ultrasonic      * At single or       $2000-$20,000      Yes
anemometer      several points for
                mean velocities and
                turbulence
                quantities

Laser Doppler   * Only used in lab   $50,000-$150,000   Yes
velocimetry     research

                * At single point
                for mean velocities
                and turbulence
                quantities

                * Scanning a large
                space in the room

Visualization   * Usually used in    --                 All the
                lab for verifying                       needed
                simulated results                       equipment and
                or measured results                     materials are
                with  other                             available,
                techniques                              but not as a
                                                        whole system

Particle image  * To obtain airflow  $50,000-$150,000   Yes, but some
velocimetry     pattern or flow                         modification
                details in an area                      may be needed
                                                        to fit the
                                                        room air
                                                        measurement
                                                        cases

                * The measured
                airflow should
                mainly be
                two-dimensional

Scanning PIV    * To obtain airflow  May be much        No
and color PIV   pattern or flow      higher than basic
                details in an area   PIV systems

                * The measured
                airflow can be
                three-dimensional,
                but the
                out-of-plane
                component should be
                several times
                smaller than the
                in-plane
                components

Stereoscopic    * Same as scanning   $70,000-$350,000   Yes
PIV             PIV and color PIV

Particle        * To obtain airflow  May be higher      No
tracking        pattern in an area   than SPIV
velocimetry     or volume

In a specific measurement case, several questions need to be answered before the appropriate technique and instruments can be chosen:

* Where in the room should the measurement be made: several special points or a large area or space of the room?

* What information about the air motion is needed: flow patterns, mean velocities, or both velocities and turbulence quantities?

* What accuracy is required about the information?

When the type of technique is determined according to the application need, the user needs to pay special attention to the specifics of the measurement instruments. The price range is wide for measurement systems or units based on the same technique, and their performances also vary in a wide range. Their specifications and use need to be well understood. Some modification of the system may be needed to meet the specific requirements of a specific room-air application. Constant improvement of the measurement methods, especially those relatively new techniques such as PIV, PTV, and MTV, is expected through further research and development. These new techniques will gain more and more applications in room-air motion measurements.

REFERENCES

Adrian, R.J. 1986. Multi-point optical measurements of simultaneous vectors in unsteady flow--A review. International Journal of the Heat and fluid Flow 7:127-45.

Adrian, R.J. 1991. Particle-imaging techniques for experimental fluid mechanics. Annual Review of Fluid Mechanics 23:261-304.

Adrian, R.J. 1996a. Laser velocimetry. In Fluid Mechanics Measurements. R. Goldstein, ed. Washington, DC: Taylor & Francis.

Adrian, R.J. 1996b. Strategies for imaging flow fields with PIV. AIAA-96-1988. 27th AIAA Fluid Dynamics Conference, New Orleans, LA, June 17-20.

Agui, J.C., and J. Jimenez. 1987. On the performance of particle tracking. Journal of Fluid Mechanics 185:447-68.

Altenhofen, R.E. 1952. Rectification. In Manual of Photogrammetry, American Society of Photogrammetry. Menasha, WI: George Banta Publishing Co., pp. 457.

Arroyo, M.P., and C.A. Greated. 1991. Stereoscopic particle image velocimetry. Measurement Science and Technology 2:1181-86.

Bohl, D.G., M.M. Koochesfahani, and B.J. Olson. 2001. Development of stereoscopic molecular tagging velocimetry. Experiments in Fluids 30:302-8.

Brucker, C. 1996. 3-D PIV via spatial correlation in a color-coded light-sheet. Experiments in Fluids 21:312-14.

Brucker, C. 1997. 3D Scanning PIV applied to an airflow in a motored engine using digital high speed video. Measurement Science and Technology 8:1480-92.

Cenedese, A., and A. Paglialunga. 1989. A new technique for the determination of the third velocity component with PIV. Experiments in Fluids 8:228-30.

Cenedese, A., and G. Querzoli. 1997. Lagrangian statistics and transilient matrix measurements by PTV in a convective boundary layer. Measurement Science and Technology 8:1553-61.

Dinkelacker, F., M. Schafer, W. Ketterle, and J. Wolfrum. 1992. Determination of the third velocity component with PTA using an intensity graded light sheet. Experiments in Fluids 13:357-59.

Doh, D.H., D.H. Kim, S.H. Choi, S.D. Hong, T. Saga, and T. Kobayashi. 2000. Single-frame 3-D PTV for high speed flows. Experiments in Fluids 29(Suppl.):S85-S98.

Doh, D.H., T.G. Hwang, and T. Saga. 2004. 3D-PTV measurements of the wake of a sphere. Measurement Science and Technology 15:1059-66.

Durst, F., A. Melling, and J.H. Whitelaw. 1976. Principles and practice of laser Doppler anemometry. London: Academic Press.

Gendrich, C.P., M.M. Koochesfahani, and D.G. Nocera. 1997. Molecular tagging velocimetry and other novel applications of a new phosphorescent supramolecule. Experiments in Fluids 23:361-72.

Gharib, M. 1996. Perspective: The experimentalist and the problem of turbulence in the age of supercomputers. Transactions of the ASME, Journal of Fluids Engineering 118:233-42.

Gogineni, S., L. Goss, D. Pestian, and R. Rivir. 1998. Two-color digital PIV employing a single CCD camera. Experiments in Fluids 25:320-28.

Goldstein, R., and D.K. Kreid. 1967. Measurement of laminar flow development in a square duct using a Laser Doppler flowmeter. Journal of Applied Mechanics 34:813-17.

Graham, J.R. 2003. Astrophysical gas dynamics. http://astron.berkeley.edu/~jrg/ay202/node167.html. Accessed November 20, 2005.

Hardalupas, Y., K. Pantelides, and J.H. Whitelaw. 2000. Particle tracking velocimetry in swirl-stabilized burners. Experiments in Fluids 29(Suppl.):S220-S226.

Hart, D.P. 2000. PIV error correction. Experiments in Fluids 29:13-22.

Hinsch, K.D. 1995. Three-dimensional particle velocimetry. Measurement Science and Technology 6:742-53.

Hu, H., T. Saga, T. Koyabashi, N. Taniguchi, and M. Yasuki. 2001. Dual-plane stereoscopic particle image velocimetry: system set-up and its application on a lobed jet mixing flow. Experiments in Fluids 31:277-93.

Johnson, A.E., B. Fletcher, and C.J. Saunders. 1996. Air movement around a worker in a low-speed flow field. The Annals Occupational Hygiene 40(1):57-64.

Kahler, C.J., and J. Kompenhans. 2000. Fundamentals of multiple plane stereo particle image velocimetry. Experiments in Fluids 29(Suppl.):S70-S77.

Koochesfahani, M. 1999. Molecular tagging velocimetry (MTV): Progress and applications. AIAA-99-3786. 30th AIAA Fluid Dynamics Conference, Norfolk, VA, June 28-July 1.

Koochesfahani, M., R. Cohn, and C. MacKinnon. 2000. Simultaneous whole-field measurements of velocity and concentration fields using a combination of MTV and LIF. Measurement Science and Technology 11:1289-1300.

Lawson, N.J., and J. Wu. 1999. Three-dimensional particle image velocimetry: a low-cost 35 mm angular stereoscopic system for liquid flows. Optics and Lasers in Engineering 32:1-19.

Lawson, N.J. 2004. The application of laser measurement techniques to aerospace flows. Proceedings of the IMECHE--Part G: Journal of Aerospace Engineering 218:33-57.

Lebbin, P.A., M. Hosni, B. Jones, B. Beck, C. Lin, and R. Horstman. 2006. Comparison of SPIV measurements for different test room air inlet nozzle configurations. ASHRAE Transactions 112(1):54-59.

Lempert, W.R., N. Jiang, S. Sethuram, and M. Samimy. 2002. Molecular tagging velocimetry measurements in supersonic microjets. American Institute of Aeronautics and Astronautics Journal 40:1065-70.

Lempert, W.R., M. Boehm, N. Jiang, S. Gimelshein and D. Levin. 2003. Comparison of molecular tagging velocimetry data and direct simulation Monte Carlo simulations in supersonic micro jet flows. Experiments in Fluids 34:403-11.

Li, Zhenhai. 1995. A fundamental study of ventilation to improve thermal comfort and indoor air quality. PhD dissertation, University of Illinois at Urbana-Champaign, Urbana, Illinois.

Lin, J.C., and D. Rockwell. 1994. Cinematographic system for high-image-density particle image. Experiments in Fluids 17:110-14.

Loomans, M.G.L.C. 1998. The measurement and simulation of indoor air flow. PhD dissertation, Technical University of Eindhoven, Eindhoven, The Netherlands.

Menon, R.K. 1999. Laser anemometry, Section 29.3. In Measurements, Instrumentation and Sensors Handbook, ed. J.G. Webster. Boca Raton, FL: CRC Press.

Meyers, J.F. 1995. Development of doppler global velocimetry as a flow diagnostics tool. Measurement Science and Technology 6:769-83.

Meyers, J.F., W.L. Joseph, and R.J. Schwartz. 2001. Characterization of measurement error sources in doppler global velocimetry. Measurement Science and Technology 12:357-68.

Miles, R.B., and W.R. Lempert. 1997. Quantitative flow visualization in unseeded flows. Annual Review of Fluid Mechanics 29:285-326.

Miles, R.B., J. Grinstead, R.H. Kohl, and G. Diskin. 2000. The RELIEF flow tagging technique and its application in engine testing facilities and for helium-air mixing studies. Measurement Science and Technology 11:1-10.

Moin, P., and K. Mahesh. 1998. Direct numerical simulation: a tool in turbulence research. Annual Review of Fluid Mechanics 30(1):539-78.

Mueller, T.J. 1996. Flow visualization by direct injection. In Fluid Mechanics Measurements, ed. R. Goldstein. Washington, DC: Taylor & Francis.

Muller, D., B. Muller, and U. Renz. 2001. Three-dimensional particle-streak tracking (PST) velocity measurements of a heat exchanger inlet flow. Experiments in Fluids 30:645-56.

Murakami, S., and S. Kato. 1989. Current status of numerical and experimental methods for analyzing flow field and diffusion field in a room, pp. 36-56. In Building Systems: Room Air and Air Contaminant Distribution. L.L. Christianson, ed. Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.

Nobuyuki, F., K. Toshiaki, Y. Masaji, T. Hiroshi, and W. Isao. 1997. Visualization and color image processing of flow mixing in an air conditioning unit for automobiles. JSME International Journal Series B-Fluids and Thermal Engineering 40(1):101-5.

Pereira, F., M. Gharib, D. Dabiri, and D. Modarress. 2000. Defocussing digital particle image velocity: A 3-component 3-dimensional DPIV measurement technique. Application to bubbly flows. Experiments in Fluids 29(Suppl.):S78-S84.

Pereira, F., and M. Gharib. 2002. Defocusing digital particle image velocimetry and the three-dimensional characterization of two-phase flows. Measurement Science and Technology 13:683-94.

Prasad, A.K., and R.J. Adrian. 1993. Stereoscopic particle image velocimetry applied to liquid flows. Experiments in Fluids 15:49-60.

Prasad, A.K., and K. Jensen. 1995. Scheimpflug stereocamera for particle image velocimetry in liquid flows. Applied Optics 34:7092-99.

Prasad, A.K. 2000. Stereoscopic particle image velocimetry. Experiments in Fluids 29:103-16.

Raffel, M., M. Gharib, O. Ronneberger, and J. Kompenhans. 1995. Feasibility study of three-dimensional PIV by correlating images of particles within parallel light sheet planes. Experiments in Fluids 19:69-77.

Rivir, R., S. Gogineni, D. Trump, L. Goss, and D. Pestian. 1995. High-resolution digital two-color PIV (D2CPIV) and its application to high freestream turbulent flows. Wright Lab, Wright-Patterson AFB, OH. www.stormingmedia.us/96/9642/A964213.html.

Scholzen, F., and A. Moser. 1996. Three-dimensional particle streak velocimetry for room airflows with automatic stereo-photogrammetric image processing. Proceedings of 5th International Conference on Air Distribution in Rooms, ROOMVENT '96, Yokohama, Japan, July 17-19, pp. 555-62.

Stier, B., and M.M. Koochesfahani. 1999. Molecular Tagging Velocimetry (MTV) measurements in gas phase flows. Experiments in Fluids 26:297-304.

Stolz, W., and J. Kohler. 1994. In-plane determination of 3D-velocity vectors using particle tracking anemometry (PTA). Experiments in Fluids 17:105-9.

Sun, Y., and Y. Zhang. 2003. Development of a stereoscopic particle image velocimetry system for full-scale room airflow studies, Part II: Experimental setup. ASHRAE Transactions 109(2):540-48.

Sun, Y., Y. Zhang, L. Zhao, and X. Wang. 2004. An algorithm of stereoscopic particle image velocimetry for full-scale room airflow studies. ASHRAE Transactions 110(1):75-80.

Sun, Y., Y. Zhang, A. Wang, J.L. Topmiller, and J. Bennett. 2005. Experimental characterization of airflows in aircraft cabins. Part I: Experimental system and measurement procedure. ASHRAE Transactions 111(2):45-52.

Tavoularis, S. 2005. Measurement in Fluid Mechanics. New York: Cambridge UP.

Taylor, C.D., R.J. Timko, M.J. Senk, and A. Lusin. 2004. Measurement of airflow in a simulated underground mine environment using an ultrasonic anemometer. SME Transactions 316:201-6.

Towers, D.P., C.E. Towers, C.H. Buckberry, and M. Reeves. 1999. A colour PIV system employing fluorescent particles for two-phase flow measurements. Measurement Science and Technology 10:824-30.

Uemura T., F. Yamamoto, and K. Ohmi. 1989. A high speed algorithm of image analysis for real time measurement of two-dimensional velocity distribution, FED-85, 149:129-34. In Flow Visualization. B. Khalighi, M. Braun, and C. Freitas, eds. New York: American Society of Mechanical Engineers.

Virant, V., and T. Dracos. 1997. 3D PTV and its application on Lagrangian motion. Measurement Science and Technology 8:1539-52.

Webster, J.G. 1999. The measurement, instrumentation and sensors handbook. Boca Raton, FL: CRC Press.

Willert, C.E., and M. Gharib. 1992. Three-dimensional particle imaging with a single camera. Experiments in Fluids 12:353-58.

Willert, C., I. Roehle, R. Schodl, O. Dingel, and T. Seidel. 2002. Application of planar doppler velocimetry within piston engine cylinders. Proceedings of the 11th International Symposium on Applications of Laser Techniques to Fluid Mechanics, Lisbon, Portugal, July 8-11, Paper 38.2, pp. 1-10.

Zhang, J.S. 1991. A fundamental study of two dimensional room ventilation flows under isothermal and non-isothermal conditions. PhD dissertation, University of Illinois at Urbana-Champaign, Urbana, IL.

Zhang, Y. 2005. Indoor Air Quality Engineering. Boca Raton, FL: CRC Press.

Zhang, Y., Y. Sun, A. Wang, J.L. Topmiller, and J. Bennett. 2005. Experimental characterization of airflows in aircraft cabins. Part II: Results and research recommendations. ASHRAE Transactions 111(2):53-59.

Zhao, L.Y. 2000. Measurement and analysis of full-scale room airflow using particle image velocimetry (PIV) techniques. PhD dissertation, University of Illinois at Urbana-Champaign, Urbana, IL.

Zhao, L.Y., Y. Zhang, X. Wang, G.L. Riskowski, and L.L. Christianson. 2001. Measurement of two-dimensional air velocities in a full-scale room using particle image velocimetry. ASHRAE Transactions 107(2):434-44.

Zheng, Q., and J.C. Klewicki. 2000. A fast data reduction algorithm for molecular tagging velocimetry: The decoupled spatial correlation technique. Measurement Science and Technology 11:1282-88.

Yigang Sun, PhD

Member ASHRAE

Yuanhui Zhang, PhD, PE

Member ASHRAE

Yigang Sun is a senior research engineer and Yuanhui Zhang is a professor in the Department of Agricultural and Biological Engineering, University of Illinois at Urbana-Champaign, Urbana, IL.

Received April 2, 2007; accepted August 9, 2007