| It is possible to use the droplet diameter and
velocity data gathered by the phased Doppler particle analyzer [PDPA] to generate measurements of liquid flux
through the sample volume. This method has the advantage of high spatial
resolution. A further advantage is that the spray can be characterized as to
droplet size, velocity and flux density with a single series of measurements
utilizing a single instrument. The accuracy of PDPA mass flux measurements
is typically ±15%. The accuracy of mass flux measurements are, however,
dependent upon numerous factors that lie beyond equipment specifications.
The most influential of all parameters is the optical alignment of the
particular experimental setup. A secondarily important factor is the chosen
measurement volume size, since a prerequisite for successful processing is
that only one particle be present within the measurement volume. The
measurable particle concentration limit depends on particle size. In
general, a much higher concentration of small particles can be tolerated
than of large particles.
The basic issue that must be addressed is
that the cross-sectional area to be used as the reference for counting
particles is dependent upon both the diameter of the particle and its
trajectory through the sample volume. Large particles scatter more light
than small ones. Because detection of the Doppler bursts reference a fixed
trigger threshold, the larger the particle, the larger is the effective
measurement volume. Additionally, the flux measurements must be normalized
for particle velocity and transit time.
The basic algorithm used for PDPA mass flux measurements relies on the
determination of the mean
square burst length for each size class. The mean square burst length is
linearly correlated to the logarithm of particle diameter. The statistical certainty of this
measurement degrades sharply as the number of particles in a particular
class size falls below 100. The method is therefore sensitive to the
typically poorly-populated upper diameter size classes which tend to
represent a relatively large percentage of the total liquid volume sampled
and are therefore more influential in the derived mass flux measurements.
The estimation of the mean square burst length, particularly for the
larger particle sizes, is made more robust by using the measured burst
lengths from all size classes containing more than 100 particles to generate
a curve-fit satisfying the burst|particle size correlation. |