362
M. Weers et al.
wheel (parallel to the axis of rotation) is not taken into account here. With the setup
shown in Fig. 6b an average contact time of 20 μs was determined. During this time
the particle may undergo a deviation into the depth of the wheel which could be
reflected in the results.
The results in Fig. 10b have been obtained with the coaxial setup (position b in
Fig. 6). The COR were ordered according to their revolution rate and the corresponding mean velocity over each range is indicated with a dashed line and a gray area
reflecting one standard deviation. The thick dashed line over the entire velocity range
indicates the expected value for spherical particles based on the literature values for
limestone particles and an aluminum impaction plate. The behavior of the measured
data deviates substantially from the expectation for spherical particles. This observation can be explained based on the results presented in Fig. 9a. At low revolution
rates, the velocity directed into the depth of the blades is larger since also the bouncing angle scatters more. At higher revolution rates, the particles rebound more often
under 90°, thereby reducing the fraction of the non-detected velocity component.
2.4.4 Particle Trajectory in the Classifier at Low and High Particle
Loadings
From the particle velocities shown in Fig. 8b and their angles in relation to the
deflector wheel blades the particle trajectories can be derived. In the beginning, particles with high Stokes numbers are considered where the influence of the airflow
is estimated to be negligible for the approach phase. In an absolute coordinate system, the particle trajectory is rectilinear (in radial inward direction), which will be
transformed into a coordinate system, which rotates with the deflector wheel. In this
rotating system, the trajectories shown in Fig. 11 (left) are obtained for low particle
loadings where the particles can be treated individually.
In Fig. 11 the model predictions (left) are compared with the measurements of
Stender [20] are shown which have been recorded with a high speed camera at a
loading of 1% w . The blades are discernible on the left and the right side in the
pictures and the particle motion is indicated by arrows. For the calculation, particles
with a size of 60 μm were assumed, while the pictures with the high speed camera
were taken for limestone particles with a median diameter x 50,3 of 59.86 μm. On
the left, for each revolution rate three different particle trajectories are labeled which
all start on the circumference but at different positions. The red trajectory starts in
the middle between the blades, while the green trajectory starts at a quarter to the
persuing blade and the blue trajectory begins immediately behind the preceding blade
(experiencing the longest residence time before being hit by the pursuing blade). For
the blue trajectory, the spread of the flight path due to the scattering in velocity and
in approach angle is indicated by the gray area. The fastest particles are indicated
by the dashed line and the slowest particles by the dotted line corresponding to one
standard deviation from the average velocity (blue line).
M. Weers et al.
wheel (parallel to the axis of rotation) is not taken into account here. With the setup
shown in Fig. 6b an average contact time of 20 μs was determined. During this time
the particle may undergo a deviation into the depth of the wheel which could be
reflected in the results.
The results in Fig. 10b have been obtained with the coaxial setup (position b in
Fig. 6). The COR were ordered according to their revolution rate and the corresponding mean velocity over each range is indicated with a dashed line and a gray area
reflecting one standard deviation. The thick dashed line over the entire velocity range
indicates the expected value for spherical particles based on the literature values for
limestone particles and an aluminum impaction plate. The behavior of the measured
data deviates substantially from the expectation for spherical particles. This observation can be explained based on the results presented in Fig. 9a. At low revolution
rates, the velocity directed into the depth of the blades is larger since also the bouncing angle scatters more. At higher revolution rates, the particles rebound more often
under 90°, thereby reducing the fraction of the non-detected velocity component.
2.4.4 Particle Trajectory in the Classifier at Low and High Particle
Loadings
From the particle velocities shown in Fig. 8b and their angles in relation to the
deflector wheel blades the particle trajectories can be derived. In the beginning, particles with high Stokes numbers are considered where the influence of the airflow
is estimated to be negligible for the approach phase. In an absolute coordinate system, the particle trajectory is rectilinear (in radial inward direction), which will be
transformed into a coordinate system, which rotates with the deflector wheel. In this
rotating system, the trajectories shown in Fig. 11 (left) are obtained for low particle
loadings where the particles can be treated individually.
In Fig. 11 the model predictions (left) are compared with the measurements of
Stender [20] are shown which have been recorded with a high speed camera at a
loading of 1% w . The blades are discernible on the left and the right side in the
pictures and the particle motion is indicated by arrows. For the calculation, particles
with a size of 60 μm were assumed, while the pictures with the high speed camera
were taken for limestone particles with a median diameter x 50,3 of 59.86 μm. On
the left, for each revolution rate three different particle trajectories are labeled which
all start on the circumference but at different positions. The red trajectory starts in
the middle between the blades, while the green trajectory starts at a quarter to the
persuing blade and the blue trajectory begins immediately behind the preceding blade
(experiencing the longest residence time before being hit by the pursuing blade). For
the blue trajectory, the spread of the flight path due to the scattering in velocity and
in approach angle is indicated by the gray area. The fastest particles are indicated
by the dashed line and the slowest particles by the dotted line corresponding to one
standard deviation from the average velocity (blue line).