EnviRobot – Air Purification Testing

Crosby Laboratory

University of Maine

Orono, ME 04469

May 8, 2013

By: Emmanuel Marsh-Sachs

Kyle Staples

Josh Stubbs

Keith Pearson

Rebecca Hanks

Brian Farnsworth

Table of Contents

Introduction ................................................................................................. 3

Objectives ...................................................................................................... 3

Apparatus...................................................................................................... 3

Equipment ...................................................................................... 3

Instrumentation ............................................................................. 6

Theory ............................................................................................................ 7

Fluid Mechanics Calculations: ........................................................ 7

Measuring Flow in Ducts: ............................................................... 7

Air Filter Pressure Drop: ................................................................. 9

Uncertainty ................................................................................................... 9

Procedure .................................................................................................. 10

Results ......................................................................................................... 11

Conclusions ............................................................................................... 16

Introduction The goal of the EnviRobot capstone design project is to create several robotic attachment

systems that promote a healthy environment. These systems are designed to fit on a single

mobile robotic platform so only one may be used at a time. Three systems have been designed;

one to plant seeds, one to purify air, and one to acquire data about the environment. This

experiment involves testing the air purification system to ensure the air filter housing is

appropriate to process the airflow rate through the filter achieving maximum purification.

Objectives 1. Take velocity measurements across a cross-section of the duct at small increments,

starting from the duct wall and moving inward, mapping out the velocity profile. The

velocity measurements are to be taken in the straight portion of the duct.

2. Take these velocity measurements utilizing both a dirty filter and a clean filter.

3. Take static pressure differential measurements, using a manometer, for both a clean and

dirty filter. Compare the pressure drop through the clean filter with the manufacturer’s

posted value.

Apparatus

Equipment The apparatus that is to be used in the air purification experiment can be seen in Figure 1.

The volumetric flow rate of air through the duct is provided by the fan pictured on the left side of

the duct. The fan that was used is a DC blower which is to be controlled by sending a voltage

input to provide the desired flow rate through the duct. The inlet of the duct is circular to match

the outlet area of the fan. This circular cross-section expands into a rectangular cross-section for

the portion of the duct where the filter is contained. This is done to decrease the air velocity into

the required range of effectiveness for the filter. There is a straight portion of the rectangular

duct before the air filter to allow the flow, after expansion, to return to ordinary flow before

traveling through the filter. The length of this straight portion is equal to four diameters of the

duct inlet. After the air passes through the filter, it flows through a four inch straight duct, before

being contracted through a circular duct. A diagram of the experimental set-up can be seen below

in Figure 1.

Figure 1: 2-D Diagram of the Air Purification Experiment

The cross-sectional dimensions of the assembled filters are 8 in. x 8 in. This includes the

plywood encasing that is surrounding the filters. The actual filter median itself is only 5.5 in. x 6

in. Since the cross-sectional area of the duct is not equal to the cross-sectional area of the filter

median, spacer board was used to fill in the gap. This is illustrated in Figure 2.

Figure 2: Cross-Sectional Area of the Filter Assembled into the Duct

Duct Dimensions:

Duct Inlet 3 in. diameter

Expansion Duct Length 12 in.

Straight Duct Dimensions 8 in. x 8in.

Straight Duct Pre-Filter Length 12 in.

Pressure Tap Location Pre-Filter 2 in. before filter

Air Filter Dimensions 5.5 in. x 6 in.

Air Filter Depth 5.875 in.

Straight Duct Post-Filter Length 4 in.

Pressure Tap Location Post Filter 2 in. after filter

Duct Outlet 3 in. diameter

Contraction Duct Length 12 in.

Rubber Tubing

Two pieces of rubber tubing of diameter to accept a Pitot tube end and cut to a length of about 2

feet.

Rubber Stoppers

Nine rubbers stoppers, ½ in. diameter, were used to plug unused pressure taps.

Weatherproof Stripping

One roll of 3/8 in. diameter weatherproof stripping was used to seal air leaks around the holes

that were used during data collection.

Fan

Jabsco Model: 35515-0010

Fan Outlet 3 in. diameter

Fan Voltage 12 VDC

The fan airflow rate is described by the curve shown below in Figure 2.

Figure 3: Performance Curve for the fan. Displays Airflow vs. Current and RPM

Instrumentation

Hot-Wire Anemometer

Serial Number: Q641740 Model Number: HHF 2005 HW

Range: 40 to 3940 ft/min

Uncertainty: ±(10% + 1 ft/min)

Inclined Manometer

No Serial Number Listed

Range: 0 to 100 psi

Maximum Operating Temperature: 150°F

Uncertainty: ±.01 in. H20

Pitot Tubes

Two Pitot tubes were used in combination with the inclined manometer in order to obtain the

static pressure drop through the filter.

Ruler

C-Thru Metric

Uncertainty: ±1/16 in.

Theory

Fluid Mechanics Calculations: Velocity measurements are taken at each of the specified locations in the duct according to the

schematic in Figure 3. These velocity measurements must be converted into a volumetric flow

rate to ensure that the flow rate is within the manufacturer’s specified range for the filter. The

velocity measurements are first averaged in order to get two flow velocity values, one for before

the filter and one for after the filter. These values are then converted to a volumetric flow rate by

using equation 1 below.

Equation 1

Where:

Q = Volumetric Flow Rate (ft^3/min)

Vavg = Average Velocity (ft/min)

A = Filter Cross-Sectional Area (ft^2)

L = Filter Length (ft)

W = Filter Width (ft)

Measuring Flow in Ducts: In order to properly determine the volumetric flow through the duct, a method must be used that

encompasses the low flow rates seen at the duct wall as well as the higher flow rates in the

middle of the duct. The American Society of Heating, Refrigerating and Air Conditioning

Engineers (ASHRAE) recommends using the log-Tchebychev method to measure the air

velocity at a number of points at the same cross section of the duct and average the values. The

standard recommends points at which to measure the velocity. Those points are displayed in

Figure 4 below.

Holes are drilled through the duct in accordance to ASHRAE standards and sealed in order to

allow the placement of an anemometer probe into the duct at desired locations to take velocity

measurements for velocity profile mapping. Five holes are drilled along the top of the duct on

one cross-section before the filter and five holes are drilled on one cross section after the filter.

Since the height of the anemometer probe can be varied in the duct, these holes will allow

measurements to be taken at all of the required locations.

Figure 4: Velocity measurement locations inside the cross section of the duct

The velocity measurements taken at these points are necessary to create velocity profile graphs

of the flow. These profiles are plotted on a surface plot that shows a three-dimensional model of

flow through the duct. Measurements are taken along horizontal and vertical cross-sections of the

duct. These respective cross sections are shown below in Figure 5.

Air Filter Pressure Drop: The static pressure drop through the filter is measured by using an inclined manometer. This

pressure drop can be explained as a dissipation of energy as the air passes through the fibers of

the filter. The flow has to change direction in order to get past the fiber and then change again to

resume the original flow direction. This change in direction results in a momentum loss in the

airflow, which causes the pressure drop.

Uncertainty The uncertainty in the calculated value of the volumetric flow rate must be found in order to

accurately compare the values. With the relationship between the measured velocity and area of

the filter known, the Kline McClintock technique can be used to calculate the uncertainty. The

general form of the equation can be found in Equation 2.

[(

)

(

)

(

)

]

Equation 2

Where R is the result of the independent variables xn, wR is the uncertainty in the result and wn is

the uncertainty in the independent variables. The uncertainty in Q, the volumetric flow rate can

be found in Equation 3:

[(

)

(

)

(

)

]

Equation 3

The uncertainty in any measurement is the measure of how well a measurement can be taken

both repeatedly and accurately. The uncertainty in measuring velocity can be found from

OMEGA, the manufacturer of the hot wire anemometer. For the model used they indicate an

Figure 5: Horizontal and Vertical Cross-Sections of the Duct

uncertainty of 10% of the measurement taken plus the least significant digit. For the uncertainty

calculations, all of the measured velocity values have been averaged, and this value was applied

to the uncertainty equation.

The uncertainty for the ruler used to measure the cross sectional area of the filter has been

reported above as:

The partial differential equations are calculated for each contributing variable for use in the

uncertainty equation. The partial differential equation of Q with respect to the average velocity

is:

The partial differential equation of Q with respect to the length of the filter is:

The partial differential equation of Q with respect to the width of the filter is:

Plugging these partial differential equations into Equation 3 the uncertainty in the volumetric

flow rate can be calculated as:

The uncertainty in the measurement pressure drop through the filter is a direct result of the

inclined manometer used.

Procedure Preparation:

1. Soil filter by utilizing wood dust and other air particulates in the wood shop at Crosby

Lab.

Velocity Profile Modeling:

1. Place the anemometer probe shaft alongside the duct. Raise the probe to the first height

location from the bottom of the duct where a velocity measurement is to be taken

according to the log-Tchebychev diagram (.498 in.). Place a piece of tape on the

anemometer probe shaft where it is barely visible on top of the duct.

2. Repeat step one for each of the other four height locations in the measurement schematic.

3. Place the anemometer probe into the first tap location, down into the first height location

in the duct and turn on the anemometer.

4. Turn on the fan and record a velocity measurement.

5. Move the anemometer to the each of the other four heights in this tap location and take a

velocity measurement at each.

6. Repeat steps 3-5 at each of the other tap locations, including the other four on the pre-

filter side of the duct and the 5 tap locations after the filter.

7. Replace the clean filter with the dirty filter and repeat this process.

Static Pressure Drop Testing:

1. Place one Pitot tube in the first drilled hole before the filter and place another Pitot tube

in the first drilled hole after the filter. Make sure both are pointed in the direction of the

flow and oriented in a fashion so that the static pressure is taken rather than the total

pressure.

2. Seal the Pitot tubes in the duct with weatherproof striping.

3. Connect the ends of the Pitot tubes into the inclined manometer using the rubber tubing.

Make sure the Pitot tube before the filter is connected to the high-pressure end.

4. Turn on the fan and take a pressure measurement reading from the inclined manometer.

Record this measurement as the static pressure drop through the filter at this location.

5. Repeat steps 1- 4 at each of the drilled holes before and after the filter. There are 5

measurement locations in total.

6. Turn off the fan and replace the clean filter with the dirty one from preparation

7. Repeat steps 1-5 for the dirty filter.

Results In this section, surface plots will display a three-dimensional graph of flow velocity throughout

the duct. The x-axis corresponds to the length of the cross-section of the duct and the y-axis

corresponds to the height of the cross-section of the duct. The four plots display the velocity

profile for flow before the clean filter, after the clean filter, before the dirty filter, and after the

dirty filter. The velocity profile for flow before the clean filter can be seen below in Figure 6.

Figure 6: 3-D Surface Plot of Flow Velocity Through the Duct, Clean Filter, Pre-Filter

It can be seen that the peak velocity is measured in the bottom corner of the ductwork at high

values of x. The flow velocity also seems to be higher in the duct corners. The velocity in the

center of the duct is generally lower.

The velocity profile for flow after the clean filter can be seen below in Figure 7.

Figure 7: 3-D Surface Plot of Flow Velocity Through the Duct, Clean Filter, Post-Filter

The velocity profile exhibits an expected shape for airflow through a rectangular duct. The flow

has a higher velocity near the center of the duct and it is slower near the walls of the duct.

The velocity profile for flow before the dirty filter can be seen below in Figure 8.

Figure 8: 3-D Surface Plot of Flow Velocity Through the Duct, Dirty Filter, Pre-Filter

It can be seen that the peak velocity is measured in the top corner of the ductwork at high values

of x. The flow velocity also seems to be higher in the duct corners. The velocity in the center of

the duct is generally lower.

The velocity profile for flow after the dirty filter can be seen below in Figure 9.

Figure 9: 3-D Surface Plot of Flow Velocity Through the Duct, Dirty Filter, Post Filter

The velocity profile exhibits an expected shape for airflow through a rectangular duct. The flow

has a higher velocity near the center of the duct and it is slower near the walls of the duct. The

profile has a slightly more irregular surface than that of the profile for flow after the clean filter.

Overall the velocity surface plots exhibited some similar characteristics and some distinct

characteristics. One observation that was made was that the velocity was generally higher near

the duct walls and especially near the corners of the duct for flow before the filter. This is most

likely due to the fact that the cross-sectional area of the filter median was smaller than the cross-

sectional area of the duct. Plywood was used to fill in the gap between the two areas. This caused

the flow toward the walls to travel faster in order to travel through the filter median, which was

in the center of the duct.

The surface plots for flow after the filter show a more expected pattern for airflow through a

rectangular filter. The fastest flow is near the center of the duct, whereas the flow near the duct

walls has a lower velocity. In this case the no-slip condition is preserved. The surface plots for

the dirty filter had a more irregular shape than the surface plots for the clean filter. We believe

that this is because the air is either bouncing back off of the dust particles trapped in the filter

fibers or being redirected around these particles.

Volumetric Flow Data

Volumetric Flow Rate (CFM)

Pre-Filter Post-Filter

Clean Filter 30.72 26.8

Dirty Filter 36.33 26.3

Static Pressure Drop Data

Static Pressure Drop across the Clean Filter

y-position (in.) 2 6

x-position (in.) Static Pressure Drop (in. H2O)

0.488 0.48 0.48

1.88 0.46 0.46

4 0.45 0.45

6.12 0.44 0.44

7.512 0.44 0.44

Static Pressure Drop across the Dirty Filter

y-position (in.) 2 6

x-position (in.) Static Pressure Drop (in. H2O)

0.488 0.44 0.44

1.88 0.44 0.44

4 0.44 0.44

6.12 0.44 0.44

7.512 0.44 0.44

The static pressure drop across the dirty filter was measured to be lower than the static pressure

drop across the clean filter. This was most likely caused by the dirty filter not being soiled

enough to actually increase the pressure drop through the filter. The pressure drop would then be

expected to be the same for both filters. The discrepancy between these values can be explained

by the uncertainty in the inclined manometer used which is 0.01 in. H2O.

Conclusions After analyzing the results, a few conclusions can be made. First, it is optimal to use a filter

whose median’s cross-sectional area that is equal in size to the duct cross-sectional area. This is

preferred in an air purification system because it makes the most effective use of the space inside

the duct, since there is more surface area for the filter fibers. This also makes it very difficult to

assemble the filter into the duct without obstructing the flow, which is undesirable.

The method used to soil the dirty filter did not result in a higher pressure drop through the filter.

With no additional data regarding the filter’s cleanliness, no conclusions can be drawn from the

pressure drop data collected.

Another possible contribution to the irregular velocity profile is the length of the straight portion

of duct before the filter. ASHRAE recommends a straight portion of length equal to 8 to 10

diameters of the expansion inlet diameter, before passing air through the filter. Our ductwork

utilized a straight duct that was only 4 diameters. In conclusion it is wise to follow ASHRAE

standards while designing an air purification system.