Analysis of A Flooded Heat Exchanger: Laboratory
Analysis of A Flooded Heat Exchanger: Laboratory
Analysis of A Flooded Heat Exchanger: Laboratory
Analysis of a Flooded
Heat Exchanger
U
Department of Chemical Engineering • Lehigh University, Bethlehem, PA 18015
nderstanding heat-transfer phenomena is a vital part temperature measurements). Steam pressures are measured at
of all fields of engineering education. Exposure to the supply header and inside the shell of the heater. Flowrates
the theory and practice of heat transfer is provided of the process water, cooling water, and steam condensate are
in all undergraduate chemical engineering departments. Most measured by the old reliable “bucket and stopwatch” method.
departments also have heat-transfer experiments to enhance The experiment is started up following the sequence of
understanding of basic concepts. establishing cooling water, then starting process water cir-
The Chemical Processing Laboratory at Lehigh Univer- culation, and finally slowly opening the steam valve. For a
sity has several heat-transfer experiments, some operating fixed process water flowrate, the steam valve is opened until
at steady state conditions and some operating in dynamic the steam pressure in the shell side of the heater is about 15
batch mode. Experimental data are collected and analyzed to psig. At steady-state conditions, the experimental data are
calculate heat-transfer rates, differential temperature driving used to calculate heat-transfer rates on both sides of both Figure 1. Flowsheet.
forces, and overall heat-transfer coefficients under various heat exchangers. In theory, these four duties should all be the
conditions (Reynolds Numbers). same at steady state, but experimental measurement errors liquid condensate to exit into the drain. So the heat-transfer The steam pressure in the shell is at the full supply pressure.
One of our most effective experiments has two pilot-plant- and small heat losses to the atmosphere produce results that area exposed to condensing vapor is the entire heat-transfer The liquid level is adjusted to alter the heat-transfer rate by
scale tube-in-shell heat exchangers operating in series as do not match exactly. Some data reconciliation is required area, which is the total number of tubes times the surface area changing liquid level and hence the heat-transfer area. Thus
shown in Figure 1. This experiment is used for both steady- before overall heat-transfer coefficients can be calculated. The per tube (π D L). in this configuration, differential temperature (ΔT) is fixed.
state and dynamic experiments, as discussed in detail in Luy- dependence of the coefficients on the flowrate of process water In this situation, the heat-transfer rate is changed by chang- The heat-transfer area changes.
ben, Tuzla, and Bader[1] in the two senior chemical processing through the two heat exchangers is determined in a series of ing the position of the upstream steam valve, which changes The flooded design has the advantage that a smaller heat-
laboratory courses. The students have taken a course in heat appropriate experiments. the pressure (and temperature) on the steam side of the heat transfer area is required because of the larger ΔT driving
transfer in the first semester of the junior year. exchanger. Thus different heat-transfer rates are achieved force. The steam temperature in the shell is the temperature
Flooded Heat Exchangers by altering the differential temperature driving force (ΔTLM) of saturated steam at the full steam supply pressure. It is
Process water is pumped from a tank into the tubes of the
first heat exchanger, which is heated by steam on the shell Our experimental heater normally operates with the shell between the hot condensing steam in the shell and the colder not reduced because of pressure drop over a control valve.
side. Condensate leaves through a steam trap, discharging completely full of vapor. As the steam condenses, it flows out process water in the tubes. Heat-transfer area is fixed. The Another advantage is the smaller valve required in the liquid
into a drain. The process water then flows into the tubes of the bottom of the shell into the steam trap, which only permits differential temperature driving force between the steam and line compared to a larger valve on the vapor line.
a downstream heat exchanger, which is cooled by cooling the cooler process water changes. The flooded design has some dynamic control disadvan-
water. The process water flows back into the feed tank. All Aaron Fink graduated with honors from Lehigh University in 2014 with a
An alternative configuration is sometimes used in industry. tages. Because it takes some time to change the liquid holdup
three flows are controlled by control valves. The steam valve B.S. in chemical engineering. Since graduating he has started an engi- Instead of having a large steam valve upstream of the heat in the shell, there can be an undesirable dynamic lag in a
is “air-to-open” and is located in the steam line upstream of neering position at W.R. Grace in their Manufacturing Leadership Program. exchanger with gas flowing through it, a smaller control valve control loop that is using heat input to control some variable.
the heater. The other two valves are “air-to-close.” William L. Luyben has taught at Lehigh for 47 years in the areas of is located on the exit liquid condensate line with liquid flow- The discussion above has considered a steam-heated heat
ing through it. There is no steam trap (or condensate drum
process design and process control. He is the author of 14 books and
Stream temperatures are measured at the inlet and outlet more than 280 papers. exchanger. The same situation occurs when a process vapor
conditions of all heat exchangers, using redundant dial and with its liquid level held by a control valve in the liquid exit is being condensed by a cooling medium. Flooded condens-
thermocouple sensors (to demonstrate the uncertainties in © Copyright ChE Division of ASEE 2015 line). Liquid condensate collects in the bottom of the shell. ers are sometimes used in distillation columns and reactors.
88 Chemical Engineering Education Vol. 49, No. 2, Spring 2015 89
References 2 and 3 give information about flooded The liquid condensate backs up into the heater because the pres- kg/s). So the sensible heat removed from the liquid in the Simultaneous Nonlinear Equations
heat exchangers. sure at the trap is 14.7 psia and the pressure of the steam inside the bottom of the heater can be calculated. The known variables presented in the previous section are
heater is lower than atmospheric. Enough hydraulic head (ρh ) must Q L = Fcond C p ( TS − Tcond ) QV, QL, FP, Fcond, CP, AV, AL, TS, Tcond, TPin, and TPout. The five
(2)
Lehigh Flooded Heat Exchanger exist for the liquid to leave the system. unknown variables are T1, T2, UV, UL, and QL1. The overall
= ( 0.012 kg / s ) ( 4.18 kJ / kg − C˚) ( 98.00 − 38.9 ˚C )
The heater is mounted vertically with 54 tubes (14.1 In the numerical example presented below, the elevation of the heat-transfer coefficients in the vapor and liquid spaces (UV
inches in length and 0.25 inches in outside diameter). bottom of the heater above the trap is 20 inches and the liquid level = 2.964 kW and UL) are unknown, but we would expect UV to be much
The process water inlet and outlet are located at the in the sight glass is 8 inches above the base of the heater. The total larger because condensing steam should have a large film
Some of this heat is removed as the process water flows
bottom of the vessel. There are two tube passes, which liquid height between the trap exit and the top of the liquid in the coefficient compared to stagnant liquid.
up through half of the tubes (QL1) and some as it flows down
means that the process water first flows up through 27 shell of the heater is 28 inches. Therefore the absolute pressure The five equations describing the system are listed below.
through the other half (QL2). We do not know either of these
tubes and then down through 27 tubes. Figure 2 gives inside the heater shell is
heat-transfer rates, but we know they total QL. Q V = FP CP ( T2 − T1 ) (8)
a sketch of the system during flooded operation. 28
PS = 14.7 psia − ft ( 62.3 lb / ft 3 ) ( ft 2 /144 in 2 ) = 13.69 psia (1) The total heat transferred is calculated from the measured
To achieve flooded conditions, we start from normal 12
process water flowrate and its measured inlet and outlet
steady-state conditions with the process water valve
At this pressure, the saturation temperature of steam is 208.4 ˚F temperatures. ( TS − T2 ) − ( TS − T1 )
opened wide and the steam valve positioned (about 50% Q V = U V AV (9)
(TS = 98.00 ˚C). Therefore, steam is condensing at 98.00 ˚C and
open) to achieve 15 psig in the shell. Then the steam Q Total = FP C p ( TPout − TPin ) ( 3) ln ( TS − T2 )
the temperature of the liquid at the top of the condensate inside the
valve opening is gradually reduced. This results in the T −T
heater is 98.00 ˚C. This pressure cannot be measured because the = (1.31 kg / s ) ( 4.18 kJ / kg − C˚) ( 36.9 − 30.0 ˚C ) ( S 1)
steam pressure in the heater gradually dropping. At
some point the pressure gauge shows zero psig. Further pressure gauge does not display vacuum pressures. = 37.78 kW
reduction of the steam valve opening produces a liquid The temperature of the condensate leaving the trap is measured
level in the sight glass installed on the side of the heater. (38.9 ˚C) and the flowrate of the condensate is also measured (0.012 ( TS − T1 ) − ( Tcond − TPin )
The heat transferred from the condensing steam vapor in the Q L 1 = U L ( 0.5 ) A L (10)
ln ( TS − T1 )
upper part of the heater above the pool of liquid condensate
T −T
must be the difference between the total and that transferred ( cond Pin )
to the liquid.
Q Total = Q V + Q L (4)
Q V = 37.78 − 2.964 = 34.82 kW ( TS − T2 ) − ( Tcond − TPout )
Q L − Q L 1 = U L ( 0.5 ) A L (11)
ln ( TS − T2 )
T −T
( cond Pout )
The total heat-transfer area is calculated from the physical
dimensions of the 54 tubes. Q L 1 = FP CP ( T1 − TPin ) (12)
A Total = ( 54 tubes) π ( 0.25 inch ) ( 0.0254 m/inch ) (14.1 in )
2
% calc variables kW kW
t1 := 305K t2 := 307K qliq1 := 1kW
W u liq := 1.5 u vap := 0.4
qtot=fp*cp*(tpout - tpin);qliq=fs*cp*(ts-tcond);qvap=qtot-qliq; 2 2
mK mK
% Initial guesses of 5 unknowns:
Given
xo=[30.6 36.9 1.5 .4 1]’;
options=optimset(‘MaxFunEvals’,1000);
⎡ (ts − t1) − (tcond − tpin)⎤
aliq
qliq1 (
fp ⋅ cp⋅ t1 − tpin ) q liq1 u liq⋅
2 ⎢
⋅ ⎢
⎡ (ts − t1) ⎤ ⎥
⎥
% Use fsolve to solve 5 nonlinear algebraic equations
⎢ ln⎢ (t ⎥ ⎥
[x,fval,exitflag,parameters]=fsolve(@floodfunc,xo,options); ⎣ ⎣ cond − tpin)⎦ ⎦
t1=x(1,1),t2=x(2,1),uvap=x(3,1),uliq=x(4,1),qliq1=x(5,1),
⎡ (ts − t2) − (ts − t1)⎤
qtot,qvap,qliq qvap (
fp ⋅ cp⋅ t2 − t1 ) q vap u vap ⋅ avap ⋅ ⎢ ⎥
⎡ (ts − t2)⎤ ⎥
⎢
⎢ ln⎢ (t − t )⎥ ⎥
⎣ ⎣ s 1 ⎦ ⎦
% function “floodfunc.m” ⎡ (ts − t2) − (tcond − tpout )⎤
aliq
% Use fsolve to find solution of 5 nonlinear algebraic equations qliq − q liq1 u liq⋅ ⋅ ⎢ ⎥
2 ⎢ ⎡ (ts − t2) ⎤ ⎥
function f=floodfunc(x) ⎢ ln⎢ (t ⎥ ⎥
% Unknowns are t1, t2, uvap, uliq, qliq1 ⎣ ⎣ cond − tpout )⎦ ⎦
t1=x(1,1);t2=x(2,1);uvap=x(3,1);uliq=x(4,1);qliq1=x(5,1); Find Block
% fixed variables
⎛⎜ t1 ⎞⎟
fp=1.31;fs=0.012;cp=4.18;tpin=30;tpout=36.9;ts=98.06;tcond=38.9;
⎜ t2 ⎟
atot=0.386;aliq=0.219;avap=0.167; ⎜ ⎟
% calc variables ⎜ u vap ⎟ := find ( t1 , t2 , u vap , u liq, q liq1)
qtot=fp*cp*(tpout - tpin);qliq=fs*cp*(ts-tcond);qvap=qtot-qliq; ⎜ u ⎟
⎜ liq ⎟
% 5 equations ⎜ q liq1 ⎟
f(1,1)=qvap-fp*cp*(t2-t1); ⎝ ⎠
f(2,1)=qvap-uvap*avap*((ts-t2)-(ts-t1))/log((ts-t2)/(ts-t1));
Solutions
f(3,1)=qliq1-uliq*aliq*0.5*((ts-t1)-(tcond-tpin))/log((ts-t1)/(tcond-tpin)); kW
f(4,1)=qliq-qliq1-uliq*aliq*0.5*((ts-t2 )-(tcond-tpout))/log((ts-t2)/(tcond- t1 = 303.489K uvap = 3.233 qliq1 = 1.857kW
W
2
tpout)); m ⋅K
f(5,1)=qliq1-fp*cp*(t1-tpin); kW
t2 = 309.847K u liq = 0.585
2
m ⋅K
92 Chemical Engineering Education Vol. 49, No. 2, Spring 2015 93
1
of convective heat transfer in a realistically complex process tion (kW)
unit in which phase changes are occurring. T1 – temperature of process stream as it leaves the top of the
liquid flooded section and enters the vapor-heated section
Nomenclature (˚C)
T2 – temperature of process stream as it leaves the top vapor-
AL – total heat-transfer area covered by liquid (m2) heated section and enters the liquid flooded section (˚C)
ATotal – total heat-transfer area of 54 tubes (m2) Tcond – condensate temperature leaving shell side of heater (˚C)
AV – total heat-transfer area exposed to steam (m2) TPin – temperature of process stream into heater (˚C)
Cp – heat capacity of water (kJ kg-1 K-1) TPout – temperature of process stream out of heater (˚C)
Fcond – flowrate of condensate and steam (kg/s) TS – saturation temperature of steam at PS (˚C)
FP – flowrate of process stream (kg/s) UL – overall heat-transfer coefficient in lower region where
PS – steam pressure in shell (psia) liquid covers tubes (kW K-1 m-2)
QL – total energy transferred from condensate in lower section UV – overall heat-transfer coefficient in upper vapor region
(kW) where steam is condensing (kW K-1 m-2)
QL1 – energy transferred from condensate into process stream
as it flows up from the process inlet to the vapor/liquid References
interface in the lower section (kW)
1. Luyben, W.L., K. Tuzla, and P.D. Bader, “Combined Steady-State
QL2 – energy transferred from condensate into process stream And Dynamic Heat Exchanger Experiment,” Chem. Eng. Ed., 43(1),
as it flows down from the vapor/liquid interface to the 39 (2009)
process exit (kW) 2. Buckley, P.S. “Historical Perspective,” Ch. 2 in Practical Distillation
QTotal – total energy transferred into process stream (kW) Control (1992) van Nostrand-Reinhold, Florence, KY
QV – energy transferred from condensing steam in upper sec- 3. <www.separationprocesses.com/Operations/POT_Chp03b.htm> p