Preliminary wind tunnel simulation study of snow collections gauges

May,

1979

PRELIMINARY WIND TUNNEL SIMULATION STUDY

OF SNOW COLLECTION GAUGES

by

A. A. Haasz and S. Raimondo

UTIAS

Technica1 Note

No.

219

CN ISSN

0082-5263

•

May, 1979

PRELIMINARY WIND TUNNEL SIMULATION STUDY OF SNOW COLLECTION GAUGES

by

A. A. Haasz and S. Raimondo

tJrIAS Teclmi-ca1 Note No. 219 CN ISSN 0082-5263

Acknowledgement

The authors wish to acknowledge the financial support provided

by the Ontario Ministry of the Environment. Furthermore, we thank Dr.

R. B. Caton and Mr. R. Vet of the Ministry for suggesting the project

and their continuing interest in our work.

Abstract

Snow collection performance of three basic precipitation collectors (Sangamo Type A, Aerochem Metric Bucket, Event Bucket) used by the Ontario Ministry of the Environment were experimentally evaluated. There are two basic are as where collection deficiencies may arise during periods of relatively high wind velocities: (1) when snow falls the air flow pattern in the vicinity of the collector may inhibit the capture of some of the snow particles, and (2) af ter the accumulation of some snow in the collector, the entrained flow inside the collector cavity may be sufficient to cause saltation resulting in the escape of same of the snow particles. Full-scale collectors and 1/3 scale models were tested in wind tunnels with snow being simulated by glass beads and mica flakes.

1. 2.

3.

4.

5.

6.

7.

CONrENTS Acknowledgement Abstract Notation INTRODUCTION DIMENSIONAL ANALYSIS

ClIOICE OF SIMULATION P.AR.A.M&rERS FLOW CHARACTERISTICS

4.1

statie Pressure Within the Collectors

4.2

Flow Visualization

4.3

Velo city Flow Field Above Collector Cavity SNOW PARTICLE ESCAPE FROM COLLECTOR CAVITY COLLECTION EFFICIENCY DURING SNOW FALL CONCLUSIONS REFERENCES BIBLIOGRAPHY TABLE FIGURES iv ii iii v 1 1 2 3

3

4

4

5 6 7 8 8

Notation

d partiele diameter

D collector diameter

g gravitation constant

h depth of cavity within collector

H collector height

Re Reynolds number based on collector diameter

Re _x Reynolds number based on X

u flow velo city

u* partiele threshold velocity

ut partiele impact threshold velocity

~ partiele piek-up velocity

~ partiele escape velocity

U_o freestream mean velocity

V_t partiele terminal velocity

X hor i zont al coordinate measured from leading edge of collector

y vertical coordinate measured from top of collector

5 separation bubble thickness above the collectors

6p statie pressure drop within the collector

~ collection efficiency ~ air viscosity ~ dimensionless groups p air density Pp partiele density v

1. INTRODUCTION

The difficulties of obtaining accurate measurements of rainf'all are multiplied many times for the collection of snowf'all data. The Air Resources Branch of' the Ontario Ministry of the Environment presently uses three pre-cipitation collection gauges: the Sangamo Type A, the Aerochem Bucket and the Event Bucket f'or collecting snow to be subjected to stibsequent chemical analysis f'or possible pollutants identif'ication. The three collectors, though different in size and shape, are in the same class of' collectors. During collection, the containers are set out in open spaces (Fig.

1)

and the snow is allowed to fall in and fill the containers. There are two areas where this class of collectors may exhibit def'iciencies during periods of' relatively high wind velocities:

(1)

when snów f'alls, the air flow pattern in the vicinity of the collector may inhibit the capture of' some of' the snow; (2) af ter the accumulation of some snow in the collectors, the entrained flow inside the collector may be suf'f'icient to cause saltation resulting in the escape of some of' the snow. Figure 2 which shows the typical streamlines can help in visualizing the problems.

Our study involved an experiment al investigation into both of' these problems in order to determine their relative importance and provide insight for the construction of' improved collectors. Experiments were perf'ormed using both mica. f'lakes and glass beads to simulate snow. In addition to the three standard collectors, Models l-B and l-BI _{with partitions in the collector} cavity, and Model l-C which allowed for replenishment of air in the collector, were also tested (ref'er to Fig.

3).

2. DIMENSIONAL ANALYSIS .

In order to determine the simulation parameters, criteria f'or choosing scaling f'actors, as well as for providing a means f'or data reduction, dimen-sional analysis using the Buckingham Pi Theorem was perf'ormed. The parameters associated with the collector are: wind velocity Uo, diameter D and height H of the collector, and the deposition height h within the collector. The particle characteristics are: diameter d, density Pp, terminal velocity Vt, and threshold velocity u*. Also there 'are the air densi ty p, the air viscosi ty ~, and the gravitational acceleration g.

Thus there are eleven relevant variables,

which can be related by the eight dimensionless groups:

Tr

l = H/D Tr

2 = h/D -Tr

3 = pUoD/~ (Reynolds number)

114

=

d/D 7T5

=

p/ Pp 7T6

=

_Vt/Uo 7T

7 = U*/UO

7T8

=

_Dg/Uo2 (Froude mmiber)

The first three describe the flow field around and withinthe collectors and the remaining five describe the particle dynamics. Since the particle' s terminal velo city and threshold velo city are f'unctions of the particle diameter and density, such that,

there is no need to simulate

774

and 7T5 independently as long as7T6 and 7T7 are properly simulated. For dynamic similarity between prototype and sma11-scale experiments, all the remaining groups must be the same in both cases. But a quick glance at 7T3 and 71"8 reveals this to be impossible for simulation in air since constant ~eynolds number implies Uo

a

l/~ and constant Froude number implies Uo

a

Dl/2. For guidance in choosing the appropriate nondimensional group to simulate , the equation of motion for· an independ~nt particle can be looked at. The analysis was performed elsewhere (Refs. 1, 2) and the

importance of the Froude number was reinforcedp For proper scaling then

the Froude number should be used.

3. CHOICE OF SIMULATION P~ERS

Snow characteristics required for the selection of the simulation parameters are terminal velocity, threshold velocity for saltation, and wind veloci ties during snowf'all. Mason (Ref. 3) report~ that the terminal veloci-t:ies are in the range 0.3 to 2 mis. This variation is attributed to differ-enoes in particle shapes resulting from varying snowfall, conditions and do not take into account variations due to size distributions. For the threshold velocity, Kind (Ref. 4) reports the impact threshold velocities for uncampacted

snow as roughly 0.15

mis,

whereas Oura et al (Ref. 5) reported threshold

velocities for blowing snow in the range 0.24 to 0.77 ~/s.A distinction must be made between the threshold velocity u* and the impact threshold velocity ui*. If particles from some external source are falling a.:nd impa.cting on a bed of particles then the threshold velocity is lower than if no particles were ;fa11ing.

According to Bagnold (Ref.

6)

the threshold velocity of dry sand is about 0.15 m/ s for 100 lJlD. diameter particles • It can be seen that there is little room for sCaling; therefore, frul-scale experimentshad to be per-formed for the snow escape due to entrainment simulation. Simulation of the full range of snow terminal velocities using ~pheriCa1 glass beads of density 2.42 g/ cm3 requires particles of diameters 70 to 270 Ilm. The glass beOOs innnediately available for experimentation did not allow for the simulation of the larger velocity.

- - - _ .

-Three very narrow si ze range samples (see Fig.

4),

two consisting of

glass beads and one of mica flakes, were prepared using a horizontal

elutri-ator (Infrasizer Ma..rk lIl) which has been developed at the University of

Toronto, Institute for Aerospace Studies (Ref.

7).

The smaller glass bead

sample had a, mea!! diameter of

74

JlID wi'i:;h a stai1.dard d.eviatio:1. of

3

J.ll!l; the

corresponding mean terminal velocity was

0.33

mis

and the threshold velocity

was

0.16

mis.

The large glass bead sample had a

174

lJlll mean diameter wi th a

standard deviation of

13

IJlll.The meall terminal and threshold veloeities were

1.16

mis

a'1d

0.26

mis,

respectively. The mica sample had a mean equivalent

diameter of'

450

!lID and an average thickness of

3.5

JlID. The associated

terminal velocity was

0.24

mis

and no estimate of the threshold velocity is

available.

Fpr the experiments involving measurements of the "critical free stream veloeities" at which snow partieles escape from the collector due to flow entrainment, the threshold velo city for saltation is a very important parameter. Therefore, these experiments were performed with full scale

models (refer to Section

5).

However, in order to evaluate collection

efficiencies during snow fall, proper simulation of the saltation threshold

velocity becomes much less important, and consequently small-scale

experi-ments suitable for tèsting in the UTIAS precipitation wind tunnel were

performed'. Based on the wind veloei ty data compiled by Vet (Ref. 8) (Fig. 5)

and on the termLllal veloeities of the available glass beads and mica, ascale

factor of 1/3 was chosen. Physical diroensions for the full-scale collector

and the small-scale roodels are summarized in Table 1.

4.

FLOW CHARACTERISTICS

The general flow characteristics associated with the collectors could

be intuitively deduced; however, flow visualization experiments with the use

of smoke were performed for the small-scale models in the UTIAS Precipitation

Wind Tu.llnel. Meastl.rements of the statie pressure wi thin the collectors and

velocity profiles al;>ove t·he collectors were also obtained.

4.1

Statie Pressure Within the Collectors

The statie pressure within the collectors could be used as an

indica-tion of the relative,vortex strength since higher the pressure drop fram

the ambient value the faster the rotational flow of the vortex. The

small-scale models were pravided with statie pressure taps alld the pressures were

, measured using a sta..'1dard Betz manometer. Model l-A had pressure taps placed

at various heights along the container wall and experiments revealed no

pressure variation as a.f1..U1ction of height. Thus, in subsequent experiments

only one pressure measurement .was obtained to characterize the vortex strength.

Results were obtained for the various models at six wind tunnel veloeities.

The pressure drop for a particular model praved to vary linearly wi th the

square of the velocity (Fig. 6). Experiments with different h/D ratios for

Model l-A showed no Si~1ificant differences in the ~ measurements. A

signi-ficant difference was however recorded between Model l-C and the remaining

models. This difference was expected and shows that the bleeding of air

into the collectorcavity does indeed aid in the reduction of the y~tex

strength.

4.2 Flow Visualization

The smoke for the flow visualization was produced by v~pourizing

mineral oi1 on a red hot nicronium wire. The vapour was transported from the smoke generator to the mode1s via a smal1 diameter tubing with the aid of a very slow ni trogen gas flow. Flow visua1ization was performed both inside as we11 ·as outside the collectors. Figure 7 shows same of the sketches made of the flow pattern within the collectors.

MOdel 1-A essential1y exhibited one very large vortex with tWQ smaller ones. As the external wind velocity increased,the size of the two smaller vortices decreased and the rotational velocity of the 1arger vortex noticeab1y increased. The same flow pattern was observed with MOdels 2 and 3 except that the smaller vortices were even smaller for the same wind velocity. Interesting flow patterns resulted in MOde1s 1-B and l-C. In Model 1-B at 2

mis

the flow in the downstream chambers resul"ted in severa1 vortices wi th the upper one being the strongest. There was also astrong flow from the rear to the front chal!lber to replenish entrained air at the separation point. In the upstream quadrants a series of very weak vortices was observed. As the external flow velocity was increased, the upper vortices in both the upstream and downstream chambers increa.sed in strength. Model l-C was a modification of the basic design, al10wing air to enter the collector

cavity through slots in the container wa11 in order to replenish the entrained air caused by the wind. At 2

mis

a very slow vort ex vas situated at the collec-tor base; this was caused by the fact that the feeding slots did not extent to the bottom. There was a1so a very slow vortex occupying the front half of the cavity. As the externa1 velocity was increased, this latter vort ex broke up into a number of smaller and much faster vortices occupying a much sma1ler fraction of the cavity.

Since these experiments were conducted for the small-sca1e medels, it is anticipated that for the ful1-sca1e mode1s (a11 except l-C), with the associated Reyno1ds numbers, the flow would be characterized by a single vortex.

4.3 Velocity Flow Field Above Collector Cavity

To estab1ish the size of the separation bubb1e above the collectors, velocity profiles were obtained using a Disa 55Kl hot wire anemometer. For the three basic small-sca1e models, the profiles were taken at the geometric centre (x

= D/2)

above the collectors for wind ve10cities in the 1.2 to 5.1

mis

range. Figure 8 shows two velocity profiles for Model 1-A indicating the presence of separated flow with high turbulence intensity. The develop-ment of the bubb1e (i.e., vertica1 velocity profiles at various x positions) is shown for Model l-A in Fig. 9 for a wind tunnel velocity of 3.85

mis.

Figure 10 shows velocity profiles obtàDned at x

=

D/2

for MOdels 1-A and 3; here the vertica1 dimension y is norma1ized by the thickness 5 of the separa-tion bubble. Using the data in Figs. 9 and 10 (a1so addi tiona1 data for Model 2, and other velocities) the normali~ed separation bubble thiékness

5/x

is plotted as a function of the running Reyno1ds number Rex in Fig. 11. The measured velocity profiles may be used at some future date for computer

simulation of the snow interaction with the flow field in order to obtain theoretica1 predictions of the collector efficiencies.

5 • SNOW PARTICLE ESCAPE FROM COLLECTOR CAVTIY

In order to investigate the snow escape fram the collectors, full-scale experiments were conducted in the University of Toronto Department of Mechanical Engineering low speed wind tunnel (Fig. 12). This tunnel has a cross-section

of 1 ~2 m by '1.8 m and can attain a maximum veloei ty of 13.7 mis. The cross-section was large enough so that the large st collector reduced this area by only 7.5%. The wind tunnel velo city was measured using a pit ot statie tube and an inclined manometer.

The experiments were conducted as follows. The collector to be tested was centred in the wind tunnel (see Fig. 13) and the base adjusted to the

collection height to be tested. A layer, approximately 1 cm thick, of the simulation material was placed in the collectors. The collectors were made of clear acrylic to allow viewing of the material. The wind tunnel was turned on and the velocity increased until a streamer of partieles was seen to skim the surface of the partieles (Fig. 14). This velocity was termed the 'piek-up velocity' ,ul. The velocity was further increased until contin-uous ground activity and partiele escape was detected; the corresponding velocity was termed the 'escape velocity', U2.

The contents of several collectors, after the escape velocity was reached, can be observed in the photographs of Fig. 15, where the areas of partiele remaval are clearly revealed.

The actual amount of material removed, however, depends on the experi-ment running time. The major difference in the simulation characteristics

of mica and glass beads manifests itself the fOllowing way. The mica flakes tend to align themselves with the local air veloeities and form ridge patterns, whUe the glass beads tend to remain as flat surfaces • The partiele removal in Model l-B was primarily in the downstream chambers as would be expected from the vortex flow visualization experiments (Section 4.2). In MOdel l-C, where the entrained flow is replenished via slots in the collector wall, resulting in a weak vortex, the extent of partiele removal was less signi-ficant than in the other models.

The observed values of ul and u2 are plotted in Figs. 16, 17 and 18 for mica, small glass beads and large glass beads, respectively. A general observation fram these figures is that for a given collector, the values of ul and U2 vary linearly with height h. The modifications of Model l-A (MOdel l-B, l-B' and l-C) all performed better than Model l-A, the standard Sangamo collector. The effect of parti tioning the collector into four smaller chambers did have marginal benefits, and this was provided regardless of whether the parti tioning walls were perforated or not. By sub di viding the collector caVity into nine chambers (MOdel l-B') the escape veloeities for large glass beads increased remarkably for a given collection height (Fig. 18). Model l-C aiso showed,a marked improvement in its ability to increase the escape velocity. Its relative performance increased with height h, and it is pos-tulated that this was caused by the fact that the feeding slot area increased with increasing height.

In the case of the Event Bucket (MOdel 3) it was surpr~s~ng that the escape velocity (U2), at the bottom of the collector (h

=

48 cm) was only 6,.4 mis (Fig. 18). A flow Reynolds number dependenee was suspected,there-fore, two other cylindrical collectors with diameters of 10.2 cm and 30.5 cm were tested (Fig. 16). It seems th at the escape velocity at a given height decreases as the collector diameter increases.

The attempts to reduce the results of the various simulation materials to a single curve in terms of the nondimensional groups of Section 2 were

futi1e. Even though a 1inear dependenee on the h/D parameter could be

expected and effects of flow Reyno1ds number detected, a simp1e coinbination of the relevant parameters inc1uding the terminal and thresho1d vela:! ities could not be arrived at on the basis of our present results • It is obvious that in order to gain a better understanding of the partic1e/air interaction phenomenon, further research must be undertaken.

6.

COLLECTION EFFICIENCY DURING SNOW FALL

To determine the co11ection efficiency of the collectors; scaled down experiments were performed in the illIAS Precipitation Wind Tunnel (Fig. 19). The tunnel was modified to attain higher veloeities by reducing the cross-sectional area. The maximum tunnel velocity was thus increased

from 3 to 5

mis.

With a 1inear sc8.1e factor of 3.3, using Fronde sealing,

this would al10w a simulated upper ful1-scale velo city of

.9.1

mis.

A schematic

of the experiment al setup can be seen in Fig. 20 and photographs of a collector

in the tunnel in Fig. 21. The experiments were performed with tl).e smal.1 glass '

beads which simulated snow of 0.6

mis

terminal velo ei ty.

The feed system for the glass beads operated as fo11ows. The glass bead sample, 100 grams per run, was contained in a cylindrical hopper above the tunnel, and the beads emerged through the conical base via a 1.5 mm

diameter orifi ce at a mass rate of 0.17

gis.

This extreme1y slow: m.as:> rate

was required to ensure that the glass bea4s were fal1ing independent1y of each other wi th a minimum of aerodynaJDic interference. The beads then f e11 through a short 1.9 cm diameter pipe before impinging on an a.ng1ed flat plate

to produce . a 1ine source perpendicular to the air flow. The beads rebounding

at different heights, in combination with some turbulence and size segregation effects, contributed to a spreading of the partiele stream, providing a quasi-uniform area densi ty bead flow over an area sufficient1y large to cover the opening of the large st collector.

The mode1s were mounted on a smal.1 platform (15 cm in diameter) and the

heights adjusted such that the collector opening was at the same level for 'al1

mode1s. The platform could be moved upstream or downstream to inter cept the partiele stream. To obtain the co11ection efficiency, two values are required:

(1) the amount of the .100 grams co11ected in the collectors, and (2) the amount

of partieles that would have traversed the same collection area hadthe collector not been there. To obtain the latter, the normalization factor, a number of smaJ.l trays (1.27 cm high, 1.27 cm wide and 7.62 cm long) were used to cover the platform which was p1aced at the same level as the collector openings would be located. The contents of the trays, covering a suitab1e area, were weighed to obtain an average weight per unit area. Experiments

were then performed,for the various collectors and efficiencies were calculated.

It must be noted that the normal.i~ation factor had to be evaluated for each

tUnnel velocity used. Furthermore, to ensure rE:lpeatabi1ity, all experiments were performed a minimum of three ti:m.es and the results proved to be repeatab1e to within a few percent.

The collection efficiencies were p10tted as a function of wind tunnel veloei ty and Reyno1ds number based on the scaled-mode1 collector diameter

(Fig. 22). As the tunnel velocity is increased, that is, the particles are coming at a shallower angle with respect to the horizontal, the collection efficiency decreases drastically. This c~~ be explained in two ways: (1) because of the higher velocity the local streamlines near the separation point are tilted upward at a greate~ angle, and (2) the vertical component

of the airflow can exceed the terminal velocities of the partieles While the greater wind velocity quickly carries the partieles past the collection area. The collection efficiency for Model l-A is higher than those for

Models 2 or 3 for the same wind velocity, and this difference can be attributed to flow Reynolds number differences. Model l-B did not behave very d ifferently from Model l-A. However, the collection efficiency of Model l-C was very

difficult to evaluate since there was large scatter in the measured efficiencies and values could not be reproduced.

The efficiencies calculated could be expected to be somewhat higher than the full-scale values since the model Reynolds mmlbers were .lower. Tur-bulence could also be expected to effect the collection efficiencies. Again, further research is required regarding the evaluation of collection

efficien-cies during snaw fall.

7. CONCLUSI,ONS

1. The effect of air entrainment and consequent vort ex formation in the collector cavity on the snow particles present in the collector was investigated experimentally for full-scale collector models at various wind velocities. Snow was simulated with glass beOOs and mica flakes. Particle "piek-up" and "escape" velocities were measured as a function of internal collector height. It was found that both of these velocities increased with increasing height. Also, by suitably bleeding air into the collector cavity to replenish the entrained air, the vort ex strength was reduced and consequently higher piek-up/escape velocities were measured. The use of subdivisions within the collectors resulted in marginal improve-ments, however, for full-scale applications this may cause snow bridging problems.

2. The collection efficiencies of the class of collectors tested for falling snow decreased rapidly wi th increasing wind veloeities. Thi s effect is probably caused by a combination of the following: (1) the modified flow field above the collector cavi ty causes the particles to travel downstream suffici~ntly far such that they are prevented from entering the collector opening, and (2) some of the particles which actually enter the cavity may be suspended in the entrained vortex causing them to escape.

1. Etkin, B. Goering, P. I I 2. Raimondo, S. Haasz, A. A. 3. Mason, B. J. 4. Kind, R. J. 5. Oura, H. Ishida, T. Kobayashi, D. Kobayashi, S. Yamada, T. 6. Bagnold, R. A. 7. Raimondo, S. Haasz, A. A. Etkin, B. 8. Vet, B. 1. Bernard, M. 2. Gerde1', R. W. stram, G. H. 3. Jairel1, R. L. 4 • Mellor, M.

5.

Oura, H. REFERENCES

"Air Curtain· Walls and Roof's - Dynamic Structures" , Proceedings of' Conference on Architectural Aero-dynamics, Royal Society of' London, 1970.

"Single and Dual Air Curtain Jets Used as Protection Against Precipitation", Ul'IAS Report No. 227, June 1978.

"The Physics of' Cl ouds " , (2nd Ed.), Clarenion Press, Oxf'ord, 1971.

"A Critical Examination of' the Requirements f'or Model Simulation of' Wind-Induced Erosion/Deposition Phenomenp. Such as Snow Drif'ting", Atmospheric

Environment, Vol. 10, 1976, p. 219.

"Studies on Blowing Snow II", Physics of' Snow and Ice, Part 2, Inst. Low Temp. Sci., Sapporo, 1967.

"The Movement of' Desert Sand", Proc. Royal Society A, Vol. 137, 1936.

"The Development of' a Horizontal Elutriator ... The Inf'rasizer MK lIl", UTIAS Report No. 235, 1979.

Ontario Ministry of' Environment, Air Resources Branch (Private COIl1lllunication).

BIBLIOGRAPHY

"Weather Bureau' s Mountain Snowf'all Works", Civi1 Engineering, Vol. 9, No. 3, March 1939.

"Wind Tunnel Studies With Scale Model Simulated Snow", General Assemb1y of' Helsinki, International Association of' Scientif'ic Hydrology. Inter. Union

of' Geology and Geophysics, Publ. No. 54, 1960. "An Impraved Recording Gauge f'or Blowing Snow" , Water Resources Research, Vol •. 11, 1975, pp. 674.

"Gauging Antarctic Drif't Snow", Antarctic Meteorology, Pergamon Press, New Yor~ 1960.

"Studies in Blowing Snow I", Physics of' Snow and Ice, Part 2, Inst. Lew Temp. Sci., Sapporo, 1967.

Table 1. Collector Dimensions* Ful1-Scale Model Outer Collectors Diameter I-A 20.3 1-B 20.3 l-C 20.3 2 29.7 3 44.5+

*

All dimensions in cm + At top of collectors Inner Diameter 19·7 19.7 15.9 28..5 .

,

44.0+

Sca1ed Down Model

Outer Inner

Height Di.ameter Dia.xooter

39.4 6.4 5.5 39.4 6.4 5.5 40.0 6.4 4.4 26.0 8.9 8.3 47.0 13.5 + 12.7 + Height 11.4 11.4 12.0 7.6 12.7

FIGURE I. Field set-up of event bucket.

c

Model i-A (Sangamo)

Ir

203

c"!l

Model

i-BI

E u E u E u o tÓ C\I

I_

0 = 29.7 cm ..

I

\

Model 2 (Aerochem Bucket)

1_0

= 20.3

cm~1

Model l-B

,

E u

FIGURE 3. Schematic diagrams of collectors tested.

\

0=44.5 cm Model 3 (Event Bucket) 1

2

=

15.9

c~1

Model

1-C

~I

7

(b)

_{cl

FIGURE 4. Snow simulation materials: (a) mica, (b) small glass beads

(71-77~m),

30.0

-

(/) ... 20.0

E

-

MAXIMUM

"0 Q) 10.0

&

Cf) 8.0 "0 c:

i

6.0 ~ 0 4.0 0 ~O~

____

~

__

~~~~~

__

~

__

~

____

~

____

~~

__

~

______

~. 02 80 ~ _99.9 0/0 Under - Velocity

FIGURE 5. Cumulative frequency distribution of daily wind speeds while snowing.

• Model I-A (h/ORfI.7)

20 • Model I-A (h/OIWI.O)

... Model I-A (h/O Rf 0.5)

o Model 2 16 I> Model 3 o Model I-C

-

_cP

₁₂

-a..

<] ₈ 4

o

4 8 12 16 20 24 U~ (m/s)2

U

_o

=

2

mIs

Uo

=

5

mIs

Model I-A

vu-~Q

01]

o

Model 1- B

~

o

? I

0/

o

1

Model 1- C

-

_E

o

1.0

-0.5

1.0

'0

_\

\

o \

\

\

\

o \

,

~

\

~

\

\

o

U

_o

=

5.1

mIs

• Uo

=

2.0 mIs

- - Mean Velocity - - - Turbulence lntenaity

2.0

3.0

4.0

5.0

Mean

Velocity

U

(mis)

FIGURE 8. Typical rnean velocity and turbulence intensity profiles measured above small scale collector Model - lA.

2.0

1.5

-E

o

->-

1.0

Uo

=

3.85 mIs

LO

2.0

Mean Ve

locity

• x

=

0.150

x

x

=

0.50

o

x

=

1.00

o

x

=

1.50

.. x

=

1.850

3.0

U (

mIs)

4.0

1.0

0.9

0.8

0.7

0.6

«>0.5

...

>-0:4

Q3

0.2

0.1

Model

Uo(m/s)

•

I-A

2

0

_I-A

_5.1

Á

3

2

r:.

₃

_5.1

x

=

0/2

~

yl

U

~

0.1 0.2

0.3 0.4

0·5 0.6

0.7 0.8 0.9

U/Uo

FIGURE 10. Normalized mean velocity profiles above small scale collector

models at x

=

D(2.

x ... ,1

1.1

ï • 1.01- • Model I-A )( Model

2

o Model 3 0.9

08

ro

I

•

Q7

8

2.3

X

= (Re x )O.165. 0.6

•

)( o )( 0.5

•

•

Q41 I I I I ~ ~ 1,000 2,000 5,000 10,000 20,000 30,000

Reynolds Number

Rex

u

• ••• : •• ~ ••• ' ; : ; : . "0. ~.: : . . . : •• ; .: . : .... .

(a)

r°-1

( b )

FIGURE 14. Schematic diagrams depicting concepts of (a) pickup, and

(b) escape veloeities.

--WIND

DIRECTION

FIGURE 15. Particle patterns within full scale collectors af ter escape

velocity has been reached for mica (left hand photos) and

14

_•

_{Model I-A (0-20.3 cm)} )( Model I-B

-

... _{Model I-C} ~ c _{0-10.2 cm} E 12 0 _{0-30.5 cm}

-

=:J-10

>--

u 0 ~ 8 ~ :::s ₆ ~ u ₀

-iï: 4 0 0 0 t

•

ti) "-E Ol =:J

>

-'u

0 ~ Cl) ~ 0 u ti) w 4 0 ",

14r

₁₄

o

Model

I-A

x

Model 1- B

121-

_{• Model}

_I-C

₁₂

--

UI

~

10

"-

_E

₁₀

E

-

_-

_y

-

_'"

:J

_::J

8

₈

~ ~

-

0

-

0

~6

0

~

6

Cl)

>

a.

Cl) ::J

4

g-

₄

~ 0 0 U)

.-0..

_LU

2

₂

o

10

20

30

h

(cm)

40

50

o

10

₂₀

₃₀

h

(cm)

40

50

.,'

14

12

-

_~

IQ

E

-

_:::s

₈

~

+-.-

_u

.Q

6

~

~

4

.::tC. U

.-

_e..

2

I

o

10

~

I

20

• Model

x

Model

o

Model

Á

Model

o

Model

I

30

h

(cm)

I-A (Sangamo)

I-B

1- BI

14

I-C 12

3

(Event Bucket)

IQ

-

f/)

8

"-E

~

:::s

6

~ +-(,)

o

4

Q)

>

Q)

0.2

0 (,) f/)

w

I

I

L

40

50

0

10

2Q

30

h

(cm)

FIGURE 18. Pickup and escape velocities for large glass beads.

Splash Plate

Wind Direction

..

W

Hopper

Tunnel Roof

Tunnel Floor

FIGURE 20. Schematic of wind tunnel setup for collection efficiency experiments.

FIGURE 21. Small-scale

model

of collector installed in UTIAS Precipitation

Wind Tunnel.

100

• Model I-A (Sangamo)

-

~

• Model 2 (Aerochem Bucket)

0

-

_~

_{• Mode I 3 (Event Bucket)}

80

_{+ Model I-B}

~ o

Model

I-C

0 c: Q) 0

₆₀

'Po 'Po W

c:

₄₀

0

-

0 Q)

-

₀

20

U

oL---~---~---~---~~---~~

1.0

2.0

3.0

4.0

5.0

-

_~

100

'

o

-~ o c: Q)

o

80

'Po

60

'Po W c:

.2

40

-

o

Q)

o

u

20

o

Wind Tunnel Velocity

Uo

(mis)

10,000

20,000

Reynolds Number

30,000

Re

40,000

~

lIrIAS Technical Note No. 219

Institute for Aerospace studies, University of Toronto (lIrIAS)

4925 Dufferin Street, Downsview, Ontario, Canada, M;3H 5T6

Haasz, A. A. J RaiJoondo, S. 27 pages 22 figures 1 table

1. Sncw collection gauges 2. Glass beads 3. Mica flakes 4. Saltation 5. Entrainment

I. Haasz, A. A., Raimondo, S. Il. urIAS Technical Note No. 219

~

Snow collectlon performance of three basic precipitatlon collectors {Sangamo Type A, Aerochem lI.etric Bucket, Event Bucket} used by tbe Ontario lIJ.nistry of the Environment were experimentally

evalu:lted. ~here are two basic areas where collection deficlencies may arise during periode

of' relatively high '1ind veloelties: (1) wben snow'falls the air fiO',o( pattern in the vicinity

of tbe collector rr.ay inhib'!.t tbe capt"l.U'e of same of the snO"i'I partleles, and (2) aftar the

accumulatlon of sorne snow in thc collector J the entra.ineà flmf inside the collector cavity may

be sufficlent. to cause saltatlon resUl.ting in the escape of some of the sno'W. particles • Full ..

scale collectors and 1/3 scale models were tested in wind tunnels witb sncw being s1mulated by

glass beads and mica flakes.

Available copies of th is report are limited. Return this card to UTIAS, if you require a copy.

urIAS Technical Note No. 219

Institute for Aerospace studies, University of Toronto (urIAS)

4925 Dufferin Street, Downsview, Ontario, Canada, M;3H 5T6

Haasz, A. A., Raimondo, S. 27 pages 22 figures 1 table

1. Snow collection gauges 2. Glass beads 3. Mica flakes 4. Saltation 5. Entrainment

I. Haasz, A. A., Raimondo, S. Ir. urIAS Technical Note No. 219

~

Snow collect!on performance of three basic prec!pi tation collectors (SangaJJO Type A, Aerochem

Metric Bucket, Event Bucket) used by tbe Ontario ~1inistry of tbe Environment were experimentally evaluated. There are two basic areas where collection deficiencies rnay arise during periods

of relatively high wind velocities: (1) wber. snow fails tbe air flow pattern in tbe vic1nity

of tbe collector lT.a.y inhibit tbe capture of some of tbe SIlOW particles, and (2) af ter tbe

accumulation of some snow in the collector, the entra.ined flow inside the collector cavi ty may

be sufficient to cause saltation resulting in the escape of same of the snow partieles • Full ..

scale collectors and 1/3 sca.le models were tested in wind tunnels w1 th snow be1ng simul.ated by

glaas beads and mica fl.akes.

Available copies of th is report are limited: Return th is card to UTIAS, if you require a copy.

lIrIAS Technical Note No. 219

Institute for Aerospace studies, University of Toronto (urIAS)

4925 Dufferin Street, Downsviev. Ontario, Canada, M;3H 5T6

Haasz, A. A., Raimondo, S. 27 pale. 22 figure. 1 table

1. Snow collection gauges 2. Glasa beada 3. Mica flakes 4. Saltation 5. Entrai~nt

r. Haasz, A. A., Raimondo, S. Il. urIAS Technical Note No. 219

~

Snow col.lection performance of three basic precipitatlor.. collectors (Sangamo Type A, Aerochem

Metric Bucket, Event Bucket) used by tbe Ontario I'á.niotry of tbe Envirorunent were experimentally

evaluated. There are two basic areu where collectlon deflclencles may a.rise during periods

or relatively high wind velocities: (1) wben sncw fails tbe air flow pattern in tbe vicinity

of tbe collector may inhibit tbe capture of some of tbe sncw particles, and (2) after tbe accumulatlon of same snow in the collector, the entrained now inside the collector cavi ty may

be sufficient to cause saltation resulting in tbe escape of sanc of tbe snow particles • Full

-scale collectors and 1/3 -scale lIIOdels were tested in wind tunnels wi tb snow being s1mul .. ted by

glass besds and mica flakes.

Available copies of this report are limited. Return this card to UTIAS, if you require 11 copy.

lIrIAS Technical Note No. 219

Institute for Aerospace Studies, University of Toronto (urIAS)

4925 Dufferin Street, Downsview, Ontario, Canada, M3H 5T6

Haasz, A. A., Raimondo, S. 27 pages 22 figures 1 table

1. Snow collection gauges 2. Glass beada 3. Mica flakea 4. Saltation 5. Entrainment

I. Haasz, A. A., Raimondo, S. Ir. urIAS Technical Note No. 219

~

Snow collection performance of tbree basic precipitation collectors (Sangamo Type A, Aerocbem

Metric Bucket, Event Bucket) used by tbe Ontario Ministry of tbe Environment were experimentally

evsluated. There are two basic areas where collectio:l defic1encies may &.rise during periods

of relatively high wind velocities: (1) wben snow fails tbe air flow pattern in tbe vicinity

of tbe collector may inhibit tbe capture of sane of the sncw particles, and (2) after tbe

accumulation of same snow in the collector, the entrained !"lew inside the collector cavi ty may

be sufficient to cause saltation resulting in tbe escape of some of tbe snow particles • Full

-sca.le collectors and 1/3 scale modele were tested in wind tunnels wi th snow being simulated by

glass beads and mic .. flakes.