EPA-AA-EOD/ES-83/1
Technical Report
Humidity Measurement Comparison Tests
Dew Point Hygrometer
vs.
Wet Bulb Psychrometer
Sherman D. Funk
January 1983
Notice
This Report does not necessarily represent final EPA
decisions or positions. It is intended to present technical
analysis of the issue using data which are currently
available. The purpose in the release of such reports is to
facilitate the exchange of technical information and to inform
the public of technical developments which may form the basis
for a final EPA decision, position or regulatory action.
This document has been approved under the EPA peer review
process and is available for internal and external distribution.
Engineering Staff
Engineering Operations Division
Office of Mobile Sources
Office of Air, Noise, and Radiation
U.S. Environmental Protection Agency
2565 Plymouth Road
Ann Arbor, Michigan 48105
-------
Abstract
A series of comparison tests were made to characterize and
quantify causes of differences in results between the dew point
hygrometer anC the wet bulb psychrometer methods of measuring
humidity. The EPA Laboratory is implementing the dew point
hygrometer method in the Light Duty test cells. Tests were
necessary in order to identify any potential impact on test
results. This study provides supportive data to address the
test procedure comparability aspects of EPCA.
The results of these tests indicate that the pyschrometer
specific humidity indication will average approximately 5 gR/lb
higher than the dew point hygrometer. The major causes of this
error are the wick water temperature and ventilating air
velocity.
Introduction (Background)
The Energy and Policy Conservation Act of 1975 and letters
dated February 1980 from the EPA Administrator to Ford and GM
establishes policy that mandates the EPA laboratory to use
procedures for testing vehicles for fuel economy that are the
same as those utilized for the 1975 model year. Provisions
were made by Congress to allow test procedure changes that
would improve the test accuracy and reduce variability of
emission and fuel economy measurements as long as there is no
significant impact on Corporate Average Fuel Economy (CAFE).
However, the sensitive nature of any such change necessitates a
quantification of the results of the action.
The EPA laboratory has been in the process of developing
and installing new ambient monitoring instruments in the Light
Duty test cells that will be interfaced to the real time
Laboratory Computer System (LCS) . One of these instruments,
the dew point hygrometer will replace the wet bulb psychrometer
for measuring humidity. Studies performed and reported in
December 1980, data samples taken from official tests in 1982,
and recent direct comparison tests indicate differences in the
results of these two measurement methods. The comparisons
showed the wet bulb psychrometer humidity results to be
approximately 5 grains higher than the dew point results. The
study of 1980 indicated that the major differences were caused
by the wick water temperature and low air velocity of the
psychrometer. In order to provide additional confirmatory and
supplemental data that relate to these measurement differences,
another series of comparison tests was designed and performed.
This paper reports the results of that study.
-------
Theory of Operation
The wet^a-fld dry bulb psychrometer and the condensation type
dew point indicators are two distinctly different methods of
measuring humidity.
The basic~form of a wet and dry bulb psychrometer uses two
thermometers. The bulb of one thermometer is covered with a
moistened wick and is called the wet bulb. The bulb of the
other is left bare and is referred to as the dry bulb. The
accuracy of humidity measurements requires these thermometers
to be closely matched. The evaporation of water from the
moistened wick of the wet bulb thermometer produces a lowering
in temperature. The thermometer may be ventilated by a sling
or forced air circulation. Ventilation should be provided to
the wick at a minimum velocity of 900 feet per minute. (See
Reference 4-NBS Circular 512) .
Using the reading of the two thermometers and the
barometric pressure, the humidity, specific and/or relative,
can be determined. Specific humidity is calculated by
determining the saturation pressure of water vapor at the wet
bulb temperature by a least squares fit of Keenan and Keyes
steam table and the partial pressure of water vapor at the dry
bulb temperature. These are determined by equations as
recommended by Wexler and Greenspan in the NES paper "Vapor
Pressure Equation for Water in the Range of 0 to 100°C" and by
Eric Zellin in "Procedures for Calculating Humidity" (See
Reference) .
The condensation dew point method of measuring humidity is
a fundamental technique by which a surface is cooled until a
dew layer is formed on that surface. The temperature of the
surface at that point is called dew point temperature. The dew
point temperature is defined as that temperature to which water
vapor must be reduced to obtain saturation vapor pressure or a
relative humidity of 100 percent. The unit discussed in this
paper is an optical condensation type dew point meter. With
this unit a sample of air is drawn across a mirror surface at a
flow rate of 2.0 SCFH. The mirror is cooled by thermoelectric
heat pump action until a dew layer is formed by the
condensation of moisture from the sample air. A beam of light
from an LED is directed at the mirror which reflects the light
to a photo cell. The photo cell detects the light and produces
a corresponding current. When the dew layer is formed, the
reflectance of light will sharply decrease causing a decrease
in photo current. The system is designed to control the dew
layer in a feedback contol loop by heating or cooling of the
mirror surface. This heating or cooling is referenced to a 40
-------
percent (40%) decrease in photo current level from a clean dry
mirror condition. The temperature of the mirror surface is
measured by-~a platinium RTD thermometer with a specified
accuracy of_i 0.4°F, is displayed on a digital meter and is
read by the-- Laboratory Computer System. This reading is the
dew point temperature and is used to calculate the specific
humidity from the partial pressure of water vapor at that
temperature and the barometric pressure.
For the data analysis of this experiment, the dew point has
been used to represent the correct humidity. The dew point was
verified by an ice bath closed loop coil technique that
produces a dew point of 32°F.
Test Plan
A series of 32 test sequences was developed for various
combinations of conditions to characterize the differences
between the wet bulb and the dew point humidity measurement
methods. Psychrometer air velocity, wick exposure, water
temperature and level, and humidity levels were varied in the
test plan. The test set up consisted of a Sargent Welch
psychrometer frame, a flexible hose, and adapter chamber, the
laboratory ventilation system with an adjustable damper, a
mercury thermometer (certified accurate and readable to _+
0.2°F), a cotton wick, a mercury thermometer (readable to
0.5°F) matched to the certified version, a type J thermocouple
and recorder, a Thermo Systems Model 1650 heater wire
anemometer, and a General Eastern Model 1200 APS dew point
meter. The equipment was set up in the Gas Blending room (See
Figure 1).
Simultaneous humidity measurements were made using the two
methods under the combinations of conditions as shown on Table
C. Sock up means the wet bulb was entirely covered in the
tube. Sock down means the sock was moved down on the
thermometer 3/4" from the top of the tube. Test conditions
included the reservoir full, and 1/2 full, the wick water
temperature at ambient and within _+2.0 degrees F from the wet
bulb temperature, and with low (450 FPM) , medium (700 and 900
FPM) and high (1300 and 1500 FPM) air velocities of the
psychrometer. The points where the wick water was below the
wet bulb temperature were later deleted from the data.
The readings were taken by two operators. The readings
made by SF were taken with a magnified jewelers glass.
Measurements were also taken of four room humidity settings
using ambient water temperature, air velocity at 900 FPM,
reservoir full, and the sock completely covering the
thermometer inside the tube. These specific humidity levels
were approximately 80, 70, 50 and 35 grains of water per pound
of dry air. The first set of tests were made on 11/30/82 and
12/1/82 and consisted of six readings each. A set of
confirmatory tests were run on 12/1/82 and 12/2/82 and
consisted of three readings for each condition. The room
humidity level comparisons were run on 12/3/82 and 12/9/82.
All readings taken were approximately 2 minutes apart in each
sequence.
-------
Discussion (Summary of Test Results)
The attached tables show the results of the tests. Table A
shows the te|t plan overview and includes the average humidity
at each setTdf indicated conditions. Table B shows the average
effect of each parameter. The average was determined from all
the pairwise differences involving the parameter. The original
and confirmatory tests indicate the following:
Increasing the air velocity from 450 to 1500 feet per
minute in steps of 300 feet per minute decrease the
measurement difference at both the cold and ambient water
temperature an average of 1.29 gr/lb.
Decreasing the water temperature from ambient to the wet
bulb temp or up to 2° above decreases the differences at
the air velocities tested an average of 3.55 gr/lb.
With all other conditions at optimum (sock up, standard
humidity level, and full reservior), increasing the air
velocity to the recommended level or decreasing the water
temperature to the wet bulb temperature) separately will
cause a decrease in the error but will not eliminate it
completely until all three conditions are optimized.
As shown in Table B, total bias caused by low velocity,
ambient water, low reservoir and sock not completely
covering the bulb in the tube can cause the wet bulb to
read about 1 grains/lb. higher than the dew point. The
typical bias would be at the lower air velocity, ambient
temperature water, and a half full reservior. At these
conditions the psychrometer reading would average 5.2
grain/lb. higher.
Although the test did not address this point, type J
thermocouple accuracy tolerance can introduce an additional
+ 2.0 grain error. Since the wet bulb sensitivity is +2.5
gr./lb. per +1°F, an accuracy spec of +.75°F on Type J
could cause a 2.0 gr./lb. error. Likewise, a j+0.4°F
technician integration error would introduce a +1.0 gr./lb.
error.
As Table C shows the majority of these tests were run on
three consecutive days with room humidity level checks
being made on 12/9/82. The majority of the original tests
were read by Steve Pfeiffer while the majority of the
confirmatories were read by Sherm Funk. It appears that
after a plot analysis and a later confirmation test the wet
bulb data on lines 1, 8, and 9 were mis-read low by 1°.
These points were adjusted on Table A and shown in
parenthesis, but were deleted from the calculated values of
Table B and the plots on Figure 2.
-------
One of the difficult parts of this experiment was to
maintain the_temperature in the reservior to recommended levels
(+2.0°F from ""the wet bulb temperature). Crushed ice was mixed
with the watEr and as the ice melted the water temperature
would rise. ~ -It was necessary to take six readings two minutes
apart during the time the water temperature was in the four
degree window. It was found that if more ice was used the
temperature rise was considerably slower. NBS recommends that
the water temperature be maintained at or slightly above the
wet bulb temperature, (See Reference, NBS circ. 512).
Some points were read where the wick water temperature was
below the wet bulb temperature. After the data were plotted,
the results showed that at these conditions the psychrometer
read lower than the dew point meter. If the dew point meter is
considered the accurate instrument, the results agree with the
NBS statements. This phenomena was taken into account and
these points were deleted from our reported data.
Conclusions/Recommendations
This experiment demonstrates and quantifies that the
typical conditions of our psychrometer (low air velocity,
ambient temperature water and less than full reservior) cause a
positive bias of about 5 grains. If the sock is down or a
contaminated wick condition is added, the bias increases to
about 7 grains. Furthermore, the thermometer calibration and
reading error could increase the bias to 10 grain/lb. or more.
These tests explain the causes for the levels of error that
have been seen in previous comparisons and reaffirms the
humidity report of December 1980.
The dew point method is a very accurate and reliable
technique for measuring humidity. The dewpoint hygrometer
units installed in this laboratory use the fundamental
method of condensation, and incorporates such measures as
an alarm for excessive contamination, automatic correction
circuitry, a highly accurate platinium RTD thermometer for
surface temperature measurement, and a mirror surface
self-cleaning capability. These features eliminate the
inherent characteristics of the wet bulb psychrometer that
can result in a biased humidity measurement. In addition,
the use of a dew point temperature is a more direct means
of calculating humidity as opposed to wet bulb and dry bulb.
Recommendations
It is recommended that the dew point meter be implemented
as the method to measure humidity and that they replace the
wet bulb/dry bulb psychrometer and the Esterline Angus
recorders in the Light Duty test cells.
It is recommended that the Laboratory Computer System (LCS)
be used to collect and process the output data from these
units and that the data be used to set room humidity levels
and to calculate NOx correction factor for official
emission tests.
-------
References
EPA Admiu_istrator's Letters to General Motors and Ford,
Subject: bImpact of Test Procedure Changes on Corporate
Average Tuel Economy (CAFE), February 19, 1980.
EPA Technical Report, Test Cell Humidity Investigation
Report, Sherman D. Funk, April 17, 1980
EPA Technical Report, Assessment of Test Cell Humidity
Measurement and Control, EPA-AA-EOD-80-13, Sherman D. Funk,
December 1980.
NBS Circular 512, Methods of Measuring Humidity and Testing
Hygrometers, A. Wexler, W.G. Brombacher, Sept. 28, 1951.
Vapor Pressure Equation for Water in The Range of 0 to
100°C. A Wexler and L. Greenspan, NBS, Feb. 19, 1971.
EPA Memo, Comparison of Humidity Correction Factor
Procedures, Eric Zellin, May 31, 1973.
EPA Internal Paper, A Procedure for Calculating Humidity,
Eric Zellin, July 1975.
Instruction Manual Dew Point Meter 1200 APS, General
Eastern Corp., June 1979.
EPA Internal Document, Technical References and Procedures
for Calculating Humidity, Eric Zellin, January 1983.
-------
Attachments
Table A - -- ^Humidity Differences Relative to Variable
^Parameters (Wet Bulb - Dew Point)
Table B - Analysis of Paired Humidity Differences (Wet Bulb
- Dew Point)
Table C - Wet Bulb vs. Dew Point Humidity Tests, Tables and
Data
Figure 1 - Humidity Measurement Comparison Test Set Up
(Room 305)
Figure 2 - Factors Affective Psychrometer Accuracy
-------
i i
! i
TABLE A:
HUMIDITY DIFFERENCES
RELATIVE TO VARIABLE PARAMETERS
(WET BULB - DEW POINT) (GR/LB)
AIR VEL
FPM ,
SOCK UP 2
RES lj FULL
ORIGINAL •
N = 6-1 5
CONFIRM e
N = f 7
SOCK UP 6
RES. FULtf
ORIGINAL; ,
N = 6 11
CONFIRM ::
N = 3
!•'•
SOCK i;
\ DOWN l<3
RES. FULL;
ORIGINAL, .
N = 6
HUMIDITY.
ORIGINAL . .
N = 6
§78.6,.
69.8
52.4
30.2
f
?9
JC,
- •__
TWP
i
450 i 700
1
+ 1.6 !
T"
-0, +2°F.
3
4
900 : 1300
+ 0.4
! + 0.2
j
+ 0.9
+ 0.2
+ 1.4 ;
5
1500
+ 0.04
J
- 0.3 ! + 0.3
. . ! ; i
1 ;
1
!
;
:
i r
- —.._----. — .. . — -.
* NOTE: DATA IN PARENTHESES ARE COR
- ONE DEGREE READING ERROR OF
75 + 2°F'
6789
450 700 900 1300
!
! i
+5.0 . + 4.0
t
+5.4 +4.6 i+3.6 +3.0
+ 1.4 + 0.6
1500
+ 2.6 .
- 0.1
(+ 6.0)* |(+ 5.2)* ;(+ 1.5)'
+ 4.6 +4.5 ; + 4.3 '. + 3.0
+8.0 ! + 6.4 i
+ 4.8
1 '
i
I
1
1 !
,
: + 2.2 ;
: +2.6
; + 1.9 i
+ 1.8
ECTED FOR A POSSIBLE i
WET BULB i
i A.
1 i
"."". '."."~1.'_. .TT... " .
.....
-------
TABLE B - ANALYSIS OF PAIRED DIFFERENCES
PARAMETER
CHANGED
RELATED PAIRED DIFFERENCES
AVERAGE
SHIFT
GR/LB
DECREASE H2O
TEMP (75 - WB)
INCREASE AIR
VELOCITY BY
500 fpm
RESERVOIR
FILLED
(1/2F - FULL)
SOCK POSITION
DOWN - UP
(-3.4*, -3.6*, -4.4, -3.2, -4.0, -2.7)
(-1.2, -1.0, -1.4, -1.7, -0.3, -1.5, -1.6, -1.6)
(-0.5, -0.2, -0.4, -0.7, 0.0)
(-3.4, -2.2, -1.8)
-3.55 '' f
I >«r
-1.29
-0.36
-2.50
EXAMPLE:
* -3.4
* -3.6
= 1.6 - 5.0 (See Table A) @ 450 fpm
.4 - 4.0 @ 900 fpm, etc,
TOTAL BIAS LOW VEL, AMB H2O, 1/2 FULL, SOCK DOWN
TYPICAL BIAS LOW VEL, AMB H2O, 1/2 FULL, SOCK UP
TYPE J T/C ERROR
MANUAL INTEGRATION
ERROR
MAXIMUM BIAS
-7.7
-5.2
+ 2.0
+1.0
10.7
-------
Jl-bo
/?-/
/5-/
/5- /
/J-/
- A
19;, // -
"i1 /?
"^ />'
r //I - Q
li
J» I'
il
3
3
*
OPid
Zt'fT.
SP
s e
^-p—
s? .
5P
5F
6f
SF
sr
Sf
Sf3
se
#£3.
UP
UP
UP
ur
UP
up
ur
OP
UP
UP
UP
oP
uf>
up
UP •
UP
FOIL
Fun.
UP
0 o\jJi\)
up
UP
UP
UP
w
7V'
7?'
7V-
IS" ;
n? \
.0 i )3<5<9
DA'S US&'f
577
?3.l>
Sj.fc
7*.y .
?y. o
s^Y
^o./
0 ,
WO
*><>>. o &O-P
. 7
50. S
^o.^
^./
5 :
^•7 I ^-i
bTT : 5V-7
-0.5
71-. 3
563
-»•••/ _•.
^t < r**
» ! (3)
I" ! (?)
-------
-------
13
1 I*.282 »
^.FIGURE. 2 .-..FACTORS AFFECTING PSYCHROMETER ACCURACY
200 400 ; 600 800 1000 1200 1400 1600
•:: I :.•:•';.'•::! PSYCHROMETER AIR'VELOCITY fFPNO ' :-'l •
1800 2000
20 Squarii to I ho I:u'h
------- |