..

laboratory )
AN INEXPENSIVE AND QUICK
FLUID MECHANICS EXPERIMENT
J.T. R YAN, R.K. W OOD, P.J. CRICKMORE
University ofAlberta
Edmonton, Alberta, Canada T6G 2G6
T
he fir st laboratory course in che mical engi-
neering at the University of Alberta is taught
to about seventy students in the first term of
t heir j unior year, and its primary objective is to
improve and develop the st udent s' writing skills.
They are re quired to write three reports. The first
two repor ts are each about ten pages in length and
deal with technical material which is fami liar to t he
student. Each of the first two experiments is per-
formed and t he reports written within a period of
two wee ks, and the corrected re ports are returned
to the st udents within another t wo weeks. The En-
gli sh construction and t he presentation of t he re-
port s are brut ally cri t icized by the academi c staff
responsible for t he course. The students then do a
st anda r d engi neering experiment and write t heir
third (hopefully r eadable) report.
The key to this approach is to present simple and
short experiments. Typicall y, the first two experi-
ments should each t ake less than t en minutes t o
complete. Quick exper ime nts have t he advantage of
conveying t o t he students that even though t he tech-
nical aspects are easy, descri bing them in clear, un-
derstandable English is often very difficul t .
For t he last six years t he first experi ment we have
used has been a computer simulat ion of a simple
concept that t he students should be able to under-
st and. Us ually, t he si mulated experiment is based
on a fundamental principle t hat was taught to t he
st udents in t he previous semest er, i.e., the vapor
pressure of water as a function of temperature . The
st udents run the program, specifying t he t empera-
ture, with the simulat or re t urning a slight ly inaccu-
rat e value of t he vapor pressure. Th e students are
asked to compare the si mulated vapor pressures
with t hose in t he st eam tables and those calculated
Copyr ight ChE Di vision ofASEE 1993
140
from a published correlation of vapor pressure and
te mperature. They then write a re port about the
simulated experiment.
Experience has taught us that both of the short
experiments should not be simulations si nce, if t hey
are, the st udents will write a simulated second r e-
por t. Their attitude seems to be, "If you don't t ake
t he exper iment seriously, why should we take t he
report ser ious ly?" Clearly, t his is not the objective of
t he course. So, we make t he second exper iment a
real experiment-but st ill quick.
One experiment that is simple and fas t is t he
filling and blowdown of a tank of air. In spite of
its simplicity, t his experiment is surprisingly rich in
its technical cont ent, involving ideal gas ther-
modynamics, unsteady state mat erial bala nces, and
simple fluid mechanics. All of t hese subj ects have
been covere d in t he preceding t erm or are being
taken concurrently with this course. A further
advant age is that t he complexity of the data analy-
sis can be adjusted t o accommodate t he t echnical
skills of the students.
EXPERIMENTAL EQUIPMENT AND PROCEDURE
Th e experiment consist s of two stages: 1) filling
the tank with air at about 90 psig and 70F, and 2)
emptyi ng the tank by venting t he compressed air to
the atmosphere. The equipment (shown in Figure 1)
J. T. Ryan is a professor and registered engineer. He has taught thermo-
dynamics, fluid mechanics, and process design for more than twenty-five
years.
R. K. Wood is professor of chemical engineering at the University of
Alberta. His recent teaching responsibilities, in addi tion to the introduc-
tory laboratory course, have involved process analysis, optimization,
dynamic modeling, and simulation. His research is concerned with digital
simulation of the dynamic and control behavior of process systems and
the computer control of distillat ion columns.
P. J. Crickmore is an associate professor of chemical engineering at the
University of Alberta. He received his BSc (Chemical Engineering), MSc
(Mathematics), and PhD (Chemical Engineering) degrees from Queen's
University, Kingston. Research areas include oil sands, coal and envi-
ronmental sampling, and remediation.
Chemical Engineering Education
cons ists of a modified 30-pound propane bottle and
a manifold mount ed at the top of t he bottl e. The
manifold has four nozzles, each of which is isolated
wit h a quick-acting ball valve. St andard 1/2-inch
copper tube and fittings are used for t he manifold.
The nozzles are brass plugs which have hol es dri lled
to diameters ranging fr om 1/ 16 to 7/64 inch. These
plugs are soldere d into the outlet of 1/2-inch unions
which are located above each of the ball valves.
While it is not required because of the limited air
supply pressur e, a relief valve is installed on the
tank for educational purposes. A cheaper alterna-
tive would be t o incorporat e t he r elief valve in t he
manifold. High-pressure air is supplied from t he
building service air through a flexibl e hose. Another
quick-acting ball valve is installed on t he tank at
t he hose connect ion.
The instrumentation consists of a fast response
t hermocouple , a pressure transducer, and associ-
ated signal conditioni ng equipment. The thermo-
couple is installed t hrough t he t ank wall. Since one
of t he objectives of t he experiment is to finish the
experi ment quickly, t he dat a-l ogging was done with
a microcomputer using an OPTO 22 int erface.
The experi mental procedure is straightforwar d:
With the pressure in the tank at atmospheric, close
valves (Vi - V4). Start logging the pressure and tem-
perature using a one-second sampling time. Open V5.
(2)
When it is combine d with the definit ions of hand u
for an ideal gas , t his equat ion can be us ed to solve
all the cases appr opri ate to this experiment . The
simplest case occurs if t he incoming kinetic energy
and the init ial mass of the syst em is neglected. Then
hi == u
2
(3)
If we assume constant heat capaciti es, then
T
2
==kT
j
(4)
141
Qcv+Lmj {hj+(t ~ +gZj}
= Lme{he +(t )v; +gZe }+W
cv
+ID2 {h2+(t ~ +gZ2 }
-ml{hI+(t )vi +gZI} ( 1)
When we apply t his equation to the filli ng of a tank,
neglecting the heat t ransfer, work, pot ential energy
changes, an d all kinet ic energy t erms except for t he
input, we get
FILLING THE TANK
This part of t he experime nt focuses pr imarily on
t he r modynamics. As it is usuall y present ed in ther-
modynamics texts, t he t heory for filli ng a tank with
an ideal gas is correct-but t he maj or assumption is
wrong. The st andar d assumption is that t he proces s
is adiabatic. When t he experiment is actually per-
formed, however, t he domin ant influe nce is the heat
t r ans fer, not t he thermodynamics. But , t he st udents
believe the textbooks and t heir professors-who also
believe t he textbooks!
Van Wylen and Sonntag' !' give the following equa-
t ion as the appropriate for m of the first law of ther-
modyn amics over a control volume for a uniform
state, uniform flow process:
The tank wi ll reach the supply pressure in about ten
seconds . Close V5 and open one or more valves in the
manifold. Th e pressure in the tank will reduce to
near atmospheric in about one minute, depending on
whi ch oalueis) is opened. At this time the data-log-
ging can be stopped or the experiment repeated.
The cost of t he equi pment is small (less than $100,
not counting t he instrumentation and microcom-
put er ), and the total shop time r equired for machin-
ing and wel ding is less than two hours. Our advice
is to purchase a new propa ne bottle an d have t he
propane relief valve removed upon purchase since
t his valve is extremely difficul t to remove . An enti re
class can perform t he experi ment in about two days
by using sche duled 15-minute t ime slots, so the in-
str umentation and dat a-logging computer can usu-
ally be borrowed.
High pressure
air supply
#30 propane
bottle
Figure 1.
V
1
<,
V V
4
Unions
3
E8
8l\ .
1/2" Quick acting
Cu pipe ball valves
Spring 1993
TABLE 1
Comparison of Exper imental and Calculated
Adiabatic, Final Air Temperatures
dm .
CIt = - m
N
(6)
Since t he volume of the system is constant and the
temperature is nearly constant, t he material equa-
tion for the air in the tank is
dm _ VCY dP (7)
CIt = RT <It
The mass flow rate of air through the nozzlets) is
m N =PNAN v N (8)
Provided that the air pressure in the t ank is above
the critical pressure required for sonic velocit y, the
velocity of the air t hrough the nozzle is equal to
v
N
= (kRT
N
) I/ 2 (9)
Th e different ial equation for the unsteady st at e ma-
terial balance reduce s to
v; dP _ ( )112
RT <It = - PNA N kRT
N
(10)
The trick is to convert, in a simple way, the ther-
modynamic variables evaluated at nozzl e conditions
Chemical Engi neeri ng Education
(5)
100F
315F
305F
355F
Approximate Final
Air Temperature
Met hod of
Det ermination
Experiment
Equation 4
Equation 2 (corrected for initial mass)
Equation 2
T _ p ( n- l j / n
To = Po
If the decompr ession is isothermal, n is equal t o
one. The pr ocess is isentropic when n =k. Th e value
of n is found from the slope of the line t hr ough the
measured pressure and temperature data when plot-
ted on log-log paper. Many st udent s use a regr es-
sion program t o est imate n from their data; how-
ever, we require the T versus P plot for educat ional
reasons. Using the equipment and procedure de-
scribed earlier, the typical exper imental value of n
was 1.04 0.005. Even though the temperature of
the air in the tank drops by approximately 85F, the
students conclude that the decompression process is
better approximated by an isothermal process than
by an ise ntropic one. This conclusion simplifies the
mathematics of the material balance. A more accu-
rate analysis is possible but is not worthwhile given
the intent of the course.
After the st udents have established t hat t he de-
compression process is approximately isothermal,
t he unsteady state material balance equation fol-
lows easily
where k == C
p
/ c. .
A better approximation of the final t emperature
can be made, a priori , by correcting for t he initial
mass of air in the tank. A further refinement can be
mad e, after t he experi ment, by incl uding an esti-
mat e of the incomi ng kinetic energy. Supposedly,
t he theory can be expanded or refined-however,
our objective is to introduce the st udent s to a simple
laboratory experiment and t o the difficulty of writ-
ing a technical r eport on such a simple experiment.
Table 1 shows a comparison of the measured final
temperature and those calculated from Eq . (2),
cor rected for both the initial mass and kinetic en-
ergy of the incoming air but still assuming no heat
transfer. The point to be mad e from t his table is
that all of the calculations are si mply wrong and
differ from the measurement by a minimum of ap-
pr oximat ely 200F.
As soon as the st udent s see the difference between
the theory taught in thermodynamics lectures and
t he results of t he experiment, they question the ex-
peri ment . There is sufficient time to do multiple
runs, but t hey find that the repeated experiments
produce virtually t he same results. The students
are faced wit h an experiment which has precision
but , in t heir minds, is of dubious accuracy. To ex-
plain the difference between the experiment and
t he temperature calculated from Eq. (4), many stu-
dent s do all the corrections indicated in Table 1.
They work hard to prove that the theory is right and
the experiment is wrong. Ultimately, they realize
Eqs. (2) and (4) are simply not t rue and are based on
a bad assumption, and finally they concl ude that
the process is not adiabatic and t hat the heat trans-
fer is t he domin ant effect. A typical value of the
heat trans fer is 14 - 18 Btu/cycle.
VENTING THE TANK
At this point the students are convi nced t hat ex-
peri ment s are worthwhile, but they are somewhat
ske ptical of t heory. Venting of t he tank is designed
to resolve this conflict for them. The venting of air
illust r at es t he polytropic decompression of an ideal
gas and an unsteady state material balance.
The relations hi p bet ween the te mperature and the
pressure of air in the tank, as it empties, must be
established before t he material balance is attempted.
Analysis of t he dat a is relatively simple and is cov-
ered in most introductory courses in engi neering
t he rmo dynamics. The decompression of t he ai r in
t he tank is taken to be a polytropic pr ocess. The
appropriate equation in terms of t he measured vari-
ables is
142
Time (sec)
Figure 2.
CONCLUSION
The experiment described in this paper is ideal
when the experiment requirements are a quick t urn-
around time, inexpensive equi pment, and flexible
technical content.
tion, shown below, is simple; however, it is valid
only until t he critical pr essure ratio is reached.
fn ( == - KI t (15)
Another tack would be to regard this equation
merely as t he basi s of a correlating equation. The
equation would be the same as Eq. (15) but have a
leading coefficient of K
2
, as
( 16)
The students are required to plot the ratio of the
measured pressure to the maximum pressure ver-
sus time on semi-logarithmic paper. They then com-
pare the slope of the line determined by t he data
and Eq. (14) and that predicted by Eq. (13).
A typical set of data taken by the staff, a regres-
sion line based on Eq. (16), and the theoretical pre-
diction from the solut ion of Eq. (15) are shown in
Figure 2. A nozzle with an internal diameter of 3/
32-inch was used. The empirical value of was
foun d to be 0.028 based on ten runs with a total of
462 data points. The value of'K, predicted by theory,
at the average temperature, was 0.027.
About 90% of the st udent s find that the difference
in the slope of the lines between theory and dat a is
5% or less. This difference is not st at ist ically signifi-
cant given the inaccuracies in measurement of the
nozzle diameters and the volume of the tank and
manifold. Surprisingly, the ot her 10% of students,
who predict differences of up to 200%, make the
mistake of using the wrong nozzlets ) or recording
the nozzle diameter(s) incorrectly. None of the stu-
dent s have difficulty with the mathematics, t hough
some think that Eqs. (14) and (16) apply even whe n
the velocity in the nozzle is subsonic.
NOMENCLATURE
A Area ; fe
h ent halpy; Btu/lb
K, cons t ant ; defined by Eq. ( 14)
cons tant; defined by Eq. ( 16)
ke kinetic energy; Btu/lb
m mass flow rate; lb/s
m mass of syste m; lb
n polytropic exponent
P pressure; psi a
Q rate of heat transfer; BTU/cycle
Cont inued on page 149.
70

60

50 40 30
- - Empirical K2 = 0.028
20 10
Limit for
sonic velocity
0.2
0.3
0.1
o
0.25
0.4
P
Po
0.5
0.6
to t hose measured in t he tank. The theory required
for t his transformation is full y develope d in most
mechanical engineering thermodynamics texts,
though not in many chemical engineeri ng texts. Our
students do not deal with compressible flow in lec-
ture courses unt il after the experiment. This prob-
lem is solved by simply stating that the thermody-
namic variables at t he nozzle can be evaluated at
tank conditions by applying a correction factor. Stu-
dent s seem to like correction factors. The two rela-
tionshi ps, shown below, are derived by Holman.!"
TN= T k 1 ( 11)
P = (12)
N k + 1
When the change in mass inventory is equated to
t he mass flow rate out of the nozz lers), the differen-
tial equation for t he pressure in the tank is
dP = _ AN(kRT)1/2(_ 2_ )( 2r
k
+-\>J p (13)
dt - V
cv
k + 1
Since t he students previously established that the
absolute temperature in t he tank is approximately
a constant, they can now write t his differential equa-
tion in a short form as
dP = - KIP (14)
dt -
where K
1
is the constant term in Eq. (13). The solu-
1.01 _-------- ----- - ------,
0.9
0.8 - - Theory K1=0.027
0.7
Spring 1993 143
Fluid Mechanics Experiment
Continued from page 143.
R gas constant; Bt ull b(R)
t time; s
u internal energy; Bt ull b
v velocity; ftis
V volume; ft3
W power input; BTU/cycle
p density; lb/ft"
Subscripts
cv control volume
e exiting air
incomi ng ai r
N nozzle
o t ime zero - valve opening
1 initial state
2 final st ate
REFERENCES
1. Van Wylen an d Sonntag, Fundamentals of Classical Ther-
modynamics, 3rd ed., SI Ver sion, J ohn Wiley, New York
(1985)
2. Holman, J.P., Thermodynamics, 4th ed., McGraw-Hili , New
York (1988) 0
REVIEW: Mass Transfer
Continued from page 126
st rongly recommended.
Chapter 17, on sorption processes, discusses fixed
bed adsorption and ion exchange. The presentat ion
on why loading and elution in ion exchange are not
symmetrical is particularly clear and easy to under-
stand. In general, t he authors assume that the reader
is familiar with these separation processes. Readers
who are not (par t icularl y elect rodialysis ) will find
these chapters difficult , but readers familiar with
the processes will gain deeper insight.
A third part of the book st arts with Chapter 18,
whi ch compares the M-S, Ficki an, and irreversibl e
thermodynamics approaches to mass trans fer. This
is a very enlight ening chapter, and sophisticate d
readers should re ad it following Chapter 2 or 3.
Chapt er 19 cites references. A rather complete list
of symbols starts on page 160. I found myself refer-
ring t o t hi s li st often and wish it were in a more
prominent location.
Th e fourth part of t he book consists of t hirty-six
wor ked exercises (pages 163 to 238) which consider
some very interesting and chall enging problems. Al-
though the solut ions are not polished, t hey are cer-
tainly sufficient to show how to attack the problems.
A major problem with this book is highlighted in
the Guidelines to the Reader on page 11: "This t ext
was written to accompany overhead transparencies
Spring 1993
The 1993 (maroon) revised printing of the
CHEMICAL RE A CTOR OMNIBOOK
is now out,
and it still costs $24.
Order at your bookstore,
or FAX your order and card number to
OSU Bookstores, Corvallis, OR at 503-737-3398
in a course on mult icomponent mass trans fer. So
the Figures are quite important. " Unfort unately,
many readers will not pay enough attention to this
section and will find reading t he book difficult until
they have learned t he proper way t o read it . Also,
since t he figures are hand drawn, t he reader needs
to learn how t o decipher t he authors ' script. The
inclusion of equat ion numbers would be useful. Some
of t he examples are confusing si nce t he problem
statements are not clear (e.g., Figure 6.2) and dat a
or formulas are sli pped into the solutions wit h little
explanation te.g., Figures 3.7 and 5.5). Stat ements
such as "Qualitat ively the reasoning shoul d be clear,"
(page 91) will unintenti onally demoti vat e readers
who are struggling, and they shoul d be removed.
The basic ideas of t he M-S approach are not sum-
mari zed until pages 64 and 65. A much earlier expo-
sition of t his would he lp many readers. Also, since
t he authors assume considerable fami liarity with
mass transfer, Chapter 18 coul d appear earlier in
the text. If a second edition is planned, the authors
could ai d readers by correcting t hese probl ems. One
hopes that t he authors will make t his effort si nce
t he book presents a very important t opic in a way
which will be accessibl e to most readers.
Where can t his book be used in t he curriculum?
The book is a curious mix of sophist icat ion (M-S
t heory and challengi ng problems) and of approxi ma-
ti ons (difference solutions and overly si mplified t her-
modynamics). Because of t he subject mat ter and t he
assumed high degree of knowledge in mass t rans fer
and separations, t his text is appropriate at the gr adu-
ate level. However, t he approximations and some
lack of rigor may cause difficult ies. It book would be
a very useful supplement in a graduate-level course,
part icul arly if journal articles are used in most of
the course. It is also a very good source of problems
and exa mples for a graduate-level course.
Finally, for pract icing professionals who mi ssed
the M-S t heory in their formal education, t his book
would be very useful for self study. Wesselingh and
Kri shna will st imulate and frust rate, but the reader
will never be bored. 0
149