The Effect of Longitudinal Heat Conduction on

G. D. B A H N K E
Lieutenant, Chilean Navy,
Graduate Student,
Mechanical Engineering,
U. S. Naval Postgraduate School,
Monterey, Calif.
C. P. HOWARD
Associate Professor of
Mechanical Engineering,
U. S. Naval Postgraduate School,
Monterey, Calif.
Periodic-Flow Heat Exchanger Performance
A numerical finite-difference method of calculating the effectiveness for the periodic-flow
type heat exchanger accounting for the effect of longitudinal heal conduction in the
direction of fluid flow is presented. The method considers th-e metal stream in crossflow
with each of the gas streams as two separate but dependent heat exchangers. To ac-
commodate the large number of divisions necessary for accuracy and extrapolation to
zero element area, use was made of a general purpose digital computer. The valves of
the effectiveness thus obtained are good to four significant figures while those values for
the conduction effect are good to three significant figures. The exchanger effectiveness
and conduction effect have been evaluated over the following range of dimensionless
parameters.

l.o > Cmiu/Cm„ z 0.90

1.0 ^ Cr/Cm-m ^ CO
1.0 NTU0 ^ 100
1.0 Z (hA)* > 0.25
1.0 > -4s* ^ 0.25
0.01 sc X ^ 0.32

Introduction fluid flow, as would be found in heat exchangers constructed with

IN THE analysis of the performance of periodic-flow
heat exchangers, the differential equations and boundary condi-
tions which result when considering the most general case are
sufficiently complex to preclude a general analytic solution. Thus
the results of work which are found in the literature [1, 2, 3, 4, 5]1
can be classified by certain simplifying assumptions. The two
most general categories are transient and steady-state behavior.
Transient behavior, of course, concerns the response of the heat
exchanger when any of the regular periodic conditions associated
with steady-state operation are changed. Subdivisions of these
two categories can then be classified as special cases.
This paper is concerned with the special case of steady-state
operation of the periodic-flow heat exchanger in which the thermal
conductivity of the exchanger material is finite in the direction of
1 Numbers in brackets designate References at end of paper.
Contributed by the Gas Turbine Power Division and presented at
the Aviation and Space, Hydraulic, and Gas Turbine Conference and
ProductsShow, Los Angeles, Calif., March 3 - 7 , 1 9 6 3 , of THE AMERICAN
SOCIETY OF MECHANICAL ENGINEERS. Manuscript received at
A S M E Headquarters, December 4, 1962. Paper No. 6 3 — A H G T - 1 6 .
continuous flow channels such as tubes, plate-fin, and finned sur-
faces. The introduction of this finite thermal conductivity re-
quires that the cross-sectional area of the material through which
heat conduction can occur also be introduced as a variable. Thus
it is found that the effectiveness of the heat exchanger is a func-
tion of six dimensionless parameters rather than the usual four.
To cover the entire range of possible application for these six
parameters would require a rather large volume of tables or
curves; therefore, the results presented here have been condensed
to cover the range of interest which would be encountered in gas-
turbine regenerator applications. However, the method pre-
sented is not restricted in range of the parameters.
Of the more recent material found in the literature, two of these
provide the bounds of the present work. Lambertson [6] covers
the special case of zero thermal conductivity with finite rotor
speed or "blow time," while Hahnemann [7] covers the special
case of finite thermal conductivity with infinite rotor speed or
"zero blow time." Hahnemann's work, which was done in
England in 1948, remained classified until late 1955 and still does
not seem to be generally available, hence, his resulting analytical
solution is given.

Nomenclature
A = heat-transfer area on side ing to subscript, Btu/hr,
designated by subscript, deg F
sq ft c = specific heat of fluid (con-
Asz = solid area available for stant pressure) or matrix
longitudinal heat con- material depending on
duction on the side of subscript, Btu/lb, deg F
Cmax, sq ft h = unit conductance for ther-
/I s„ = solid area available for mal convective heat
longitudinal heat con- transfer, Btu/hr, deg F,
duction on the side of sq ft
sq ft k = unit thermal conductiv-
ity, Btu/hr, sq ft, deg
,4 s = total solid area available F/ft
for longitudinal heat L = total length of the matrix
conduction, sq ft in the direction of fluid
C = heat capacity rate {We) of flow, ft
fluid or matrix accord- N = number of subdivisions ac-
cording to subscript
T = temperature of matrix, deg
F
Txd.j) = subscripted matrix tem-
perature on the side of
Cmai, deg F
T„<f,o) = subscripted matrix tem-
perature on the side of
Cmra, deg F
t = temperature of fluid, deg F
txd.jt = subscripted fluid tempera-
ture on the side of CID0,,
deg F
t„(/,„) = subscripted fluid tempera-
ture on the side of Cmm,
deg F
(Continued on next page)

Journal of Engineering for Power APRIL 1964 / 105

Copyright © 1964 by ASME

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 09/29/2013 Terms of Use: http://asme.org/terms

 Method (^)j A rn _ [i , i
JTJ7 Z i J av g — jy Ul(i,j) — ii(i+l,j)l (11)
The idealizations and boundary conditions assumed in the
derivation of the governing differential equations for the periodic- where ATV-g represents the mean temperature difference between
flow heat exchanger are as follows: the fluid and the matrix element. For a small enough element the
(а) The thermal conductivity of the heat exchanger matrix is arithmetic mean temperature difference may be assumed valid, so
zero in the direction of "metal flow," finite in the direction of that
fluid flow, and infinite in the third direction which is normal to
both the fluid and metal flow. A'/'ave = — Ui<(.|) + /*<i+l. /)] — — \'l'z«.i) + Tz(i,j+1)].
(б) Thermal properties of the two fluids and the matrix ma-
terial are constant with temperature and time. The fluid and matrix temperatures are subscripted to indicate the
(c) No mixing of the fluids occurs either as a result of direct average temperatures across the inlet and outlet of the element
leakage or carry over. for simple unmixed crossflow.
(d) The convective heat-transfer coefficients and thus the con- Considering the energy balance for an element, the energy
vective conductances between the fluids and the matrix are con- transferred to the element by convection plus energy transferred
stant with flow length. to and from the element by conduction must equal the energy
(e) The fluids pass in counterflow directions. stored in the element:
( / ) Entering fluid temperatures are uniform over the flow inlet
cross sections and constant with time.
(g) There is no convective heat exchange at the fluid entering
or exit cross sections.
(/i) Regular periodic conditions are established for all matrix
elements; i.e., steady-state operation.
Fig. 1 is a representation of such an axial flow and radial flow
type rotary regenerator. A typical element from each fluid side
is shown in the schematic representation of Fig. 2. At this point
it is possible to take several different approaches to the solution
of the problem. One approach is to derive the partial differential TYFLCAL. ELEMEN

equations by using the rate equations in conjunction with an

energy balance as was done by Hahnemann. With the further
assumption of infinite rotor speed, Hahnemann was able to find ILLUSTRATIVE MATRIX ARRANGEMENT FOR AXIAL FLO*

a closed form analytical solution. Another approach, and the

one which will be used here, was proposed by Dusinberre [8].
With this method a numerical analysis technique is applied
directly by subdividing the rotor into a finite number of elements
and writing the energy balance relations using the rate equations.
Repeated use of these finite difference equations in some orderly
iterative scheme eventually results in a particular solution. As
the number of elements chosen increases, the closer a solution will
approach that for the differential equations. This was the
method used by Lambertson but with the assumption that the
thermal conductivity of the rotor material in the direction of
fluid flow was zero.
If the elements are considered to be fixed in space, then each
element can be regarded as a crossflow heat exchanger with a gas TYPICAL ELEMENT

stream and a metal stream. For an element on the side of Cmax,

the heat-transfer rate by convection is equal to the rate of change ILLUSTRATIVE MATRIX ARRANGEMENT FOR RADIAL FLO*

in enthalpy of the fluid across the element, Fig. 1

-Nomenclature-
W = mass flow rate of fluid Dimensionless Parameters fer units on side desig-
(lb/hr) or matrix (lb, E = heat exchanger effective- nated by subscript
rev/hr), according to ness, ratio of actual to the
subscript NTU NTU
thermodynamically maxi- ° = " [ t t W ] =
Ci, Ci, . mum possible heat-trans- overall number of trans-
Di, D 2 ,. fer rate fer units
= const, dimensionless
Ei, E?,. E, no conduction .As* = ^Isn/Asj = conduction area
Fi, Ft, . AE — E, conduction ratio
E E, no conduction
Subscripts X„ = — conduction pa-
= conduction effect t'miniJ
avg = average min/Cmax = capacity rate ratio of fluid rameter on side of Cmin
i = inlet streams
X, = kAsx
——— = conduction pa-
min = minimum magnitude Cr/Cmin = capacity rate ratio of matrix tv mnT Lj
max = maximum magnitude to minimum fluid capacity rameter on side of Cmax
n = side of Cm;„ rate kAs
X =
o = outlet, or overall in the case of (hA)* = (hA)J(hA), = conduc- CmiuE
NTUo tance ratio total conduction parame-
X = side of Cmax NTU = (hA)/C = number of trans- ter

106 / APRIL 1964 Transactions of the A S M E

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 09/29/2013 Terms of Use: http://asme.org/terms

 Solving for the outlet fluid temperature for elements on the side of
N, Cmin from equation (7) gives

AsxNr '"(/+1.0) = Fil„V,0) + Fi[T„(/,g) + T„(/,g+l)] (12)

+ " [ ( J ' x t f - l • ;') + ?'x(i-l.J+lO — (.Tx(i.j) + T'ld'.j+l))]
which holds for all elements on the side of C m i „ .
As^V,. There are then eight finite difference equations (3), (4), (5), (6),
" 2 L N T H ' . H * ) — I+I.FL + J X ( I + I , / + D ) ]
(9), (10), (11), and (12) with sixteen dimensionless coefficients.
1

Q Besides the four usual dimensionless parameters needed for the

= (Tx(i.j+i) — T xi i,j)) (2) case of no conduction in the direction of fluid flow,

Cmin/Cmax = capacity rate ratio of fluid streams

B y using equations (1) and (2) the outlet fluid and metal tem-
Cr/Cmio = capacity rate ratio of rotor matrix to minimum
peratures for an element on the side of C'max ma}' be solved for in
fluid
terms of the remaining temperatures and the results expressed in
equations with dimensionless coefficients. (/iA)* = conductance ratio
Since the longitudinal heat conduction is zero at the ends, the
NTUo = NTU. = overall number of trans-
energy balance equation is modified from equation (2) for ele- " [l + (hA)*]
ments of the first and last row. Three expressions for the matrix fer units,
outlet temperature are thus obtained:
two additional dimensionless groupings are necessary in these co-
First r o w elements efficients to take the conduction into account. One of these
TX(t.j+1) = D dxti.i) — D-jTx{i,j) + De[Tx^,j) + ?':c(2,;+i)] (3) groupings will be the ratio of the solid cross-sectional area available
for conduction on the side of Cmm to the solid cross-sectional area
Middle row elements available for conduction on the side of C m a l ,

Tx(i.j+\) = Dilxii.j) — DiTxU.j) As* = As„/Asx, conduction area ratio.

+ DATzii-i.j) + TW-i.z+i) + T z(i+ i,,') + Tx(i+i,y+i)] (4) The second dimensionless group will be a conduction parameter
Last r o w elements
which can be defined as

jfxCA'r.y+l) = Dstx(.Xr,j) — DiTx(Kr,j) kAsx

Xx = = conduction parameter on side of C max>
+ Ds[Tx{Nr-l.i) + 2'j (l v,_i,/ + i)] (5) Cma^L

Bj r solving for the fluid outlet temperature on the side of C'max in kAs.
X n — ——~r = conduction parameter on side of Cm
terms of the matrix outlet temperature, from equation (1) only CmiuL
one expression is obtained which will hold for all elements on the
side of CnmK,

lx(i+i,j) = Ditx{i,j> + D„[T,<i.» + jPxd'.j'+i)] (6) kAs

X = total conduction parameter,
C/minL K " + a7*] =
Similarly, for an element on the side of Cm-m the rate equation is
with As the total solid area available for conduction
- ATave = -Jj— [<»</+!.S> — in</.S>] (7)
NnNr /Is = Asx + As,,.

with This later form offers the advantage that for a given value of the
conduction parameter, the resulting heat exchanger effectiveness
A'Z'ovg — 0 [{T nUl0 ) + T„(/,a+1))] — — [<„(/,„) + <„</+!,0)]. will not be greatly affected b y small changes in As*. The expres-
sions for the dimensionless coefficients in terms of the dimension-
less parameters are given in the Appendix and their derivation
Energy balance:
can be obtained from reference [9].
Cmii From the schematic representation shown in Fig. 2, it should be
2LN„
Sec OF CM
As,,iVr WN.
— (7'iif/_i,S) + r„ ( / _i,„ + I ) )] — A; [(2',|</+1,(|) + T„lf+l,g+l))
2LNn

Cr
— (Tn(f.o) + 5r»(/.,+i))] — [{T,i(/.») — 2'„(/,„ + i);] (8)

By the same method the three expressions for the outlet matrix
temperatures may be obtained by using equations (7) and (8).
1

J
First r o w elements
i
'/\i(A'r,0+l) = Fetn(Nr.u) ~ FjTn^'r,a) + [(A'r-1,o) + T„Wr-l,g+l) ]
(9)
WSk T,

r
Ml T«i-
1
T« i.fi tn !.«">
Middle row elements
-
11
- 1
i ° u
Tn(f.o+1) = fV„(/lrt — FlT„(f,g) tt
+ Fz[TnU_i,g) - f TnV-l.g+l) + JVz+l.j) + 0+1)] (10) V>1
i
Last r o w elements

jf"(i.o+i) = Fdn(i.e) ~ FnT„{\,g) + + Tntj.j+i)] (11) Fig. 2 Schematic representation of a periodic-flow heat exchanger

Journal of Engineering for Power APRIL 1964 / 107

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 09/29/2013 Terms of Use: http://asme.org/terms

 noted that the left edge of the regenerator is physically the same
as the right edge and, therefore, the matrix inlet temperatures for
elements of the first column on the side of Cma* are the same as
the outlet temperatures for corresponding elements of the last
column on the side of Cmin. Expressed mathematically,
= T„(/.ffn+U) / = + 1
where i can take the values of 1, 2, 3, . . . N,. This is referred to
as the reversal condition.
For simplicity of calculation, a temperature scale can be used
for which the fluid entrance temperature is zero on the side of
Cm\n and unit}' on the side of CmBx,
tnn.o) = 0 g = 1, 2, 3, . . . Nn
ixci./) = 1.0 j = 1, 2, 3, . . . Nx
From the finite difference equations it is seen that in order to
solve for the matrix outlet temperature for a particular element,
it is necessary to know the matrix temperature of the next ele-
ment. An estimate of these temperatures can be made by deter-
mining the temperature distribution for the assumption of zero
conduction in the direction of fluid flow since the next element
temperatures are not needed (i.e., the coefficients D3, Ds, F3, and
Ft are zero for no conduction). For the calculation procedure
then, a matrix temperature distribution is assumed on the left
edge and the problem is solved for the no conduction case to ob-
tain the initial estimate of matrix temperatures for the problem
with conduction. This also provides the no conduction effective-
ness necessary for comparison purposes. Now with this initial
estimate of all element temperatures, the calculations are made for
the conduction case beginning with the first element, first column
on the side of Cm„ and working down the column. When the
first column of elements is completed, the second is calculated
and so on until the outlet temperatures of all elements on the
side of Cmax are determined. In practice it was found that con-
vergence was more rapid if two passes per column were made be-
fore proceeding to the next one. The reason for this, as men-
tioned previously is that the equations for the outlet tempera-
tures contain the matrix temperatures of the next element. At the
seal, represented by the double line in Fig. 2,
- x(i,ATx+l) = T„(/.i)
since the matrix outlet temperatures for elements of the last
column on the side of are physically the same as the inlet
temperatures of the corresponding elements on the side of Cm,„.
The same calculation procedure is then applied to the side of Cmi„.
If the temperature distribution assumed on the left or starting
edge were correct it would then be duplicated on the right or
finishing edge (i.e., the reversal condition would be fulfilled). If,
however, this is not the case, the resulting temperatures on the
right side are now used as starting values on the left side and the
calculations repeated. After each pass (i.e., the complete calcula-
tion of a temperature distribution) an energy balance is made;
and before a solution is accepted for a particular set of parameters,
the heat balance error together with the reversal condition has to
be fulfilled to a specified accuracy. The heat balance error was
determined from the difference in enthalpy change for the fluid
on the side of C m a l to Cm-,„ compared to the enthalpy change on
the side of Cmai)
Cnmxitxi txo) Cmin(lno tni)
Cmax(^ii tx o)
Cmin
[<ii(iVr+l,(j)]avg
Cmax
l -
l - UxWr+1, j)
For the results presented here this error was established at
0.0004 percent for a maximum absolute value. It was found that
the heat balance error criterion was far more stringent than for
108 / A P R I L 1964
Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 09/29/2013 Terms of Use: http://asme.org/terms
the same magnitude error required for the reversal conditions.
The effectiveness of a heat exchanger is defined as the ratio of
the amount of energy transferred to the maximum thermody-
namically possible,
Cmax(^zi (^no tni)
E = = [L OVr+l.ff) Jnvg.
Cmiu(txi - tni) ' (<ii - tni)
The values of the effectiveness thus found by this method are
good to five significant figures. However, this would only apply
for the case of a certain number of subdivisions. If the number
of subdivisions is increased and the calculations are repeated, a
different value of the effectiveness is obtained. A plot of effective-
ness versus element area will reveal a linear relationship as the
element area approaches zero [6, 9]. Thus to arrive at the given
tabulated values, the solutions were first generated for N r = N x
= N„ = 12, recomputed for Nr = Nx = Nn = 16 and finally
extrapolated to zero element area. This method reduced the
accuracy of the extrapolated effectiveness to four significant
figures.
To best illustrate the influence of conduction on heat exchanger
performance, the "conduction effect" is defined as the ratio of the
difference between the effectiveness with no conduction and with
conduction to the no conduction effectiveness,
AE E, no conduction E, conduction
=
IT E, no conduction
and correlated with the conduction parameter X. The values of
the conduction effect thus obtained are then good to three sig-
nificant figures.
To accommodate the large number of calculations required
use was made of a general purpose digital computer, and the
limits on accuracy were dictated by computer time available for
calculation rather than by the method employed. With the
current availability of digital computers, such a method as pre-
sented here could be considered as a closed form solution and in-
troduced as a subroutine in a general analysis of gas-turbine power
plant behavior.
Results
Before explaining the results shown in Figs. 3, 4, and 5 and
Tables 1, 2, and 3, it is best to review the influence of the various
parameters on the periodic-flow heat exchanger effectiveness for
the case of no conduction in the direction of fluid flow.
For fixed values of Cmm/Cmm Cr/Cmin, and (hA)*t the effective-
ness will increase with increasing NTU 0 and will asymptotically
approach unity.
For fixed values of Cmin/Cmax, NTU 0 , and (hA)*, the effective-
ness will increase with increasing Cr/Cm in and will asymptotically
approach some value determined b}' Cmm/Cmand NTU 0 which
is less than unity; such that at C r /C min infinite,
1 - exp
[ ( C I - O H
E =
.. Cmin
1 - exp I NTU„J
^max |
and if C m i n /C mol is unity, the expression becomes
E _ NTUo
I + NTU„"
For fixed values of Cr/Crain, NTUo and (hA)*, the effectiveness
will increase with decreasing Cmift/Cmm. The percent change is
largest in the lower NTU 0 region and increases with larger
]avg values of C,/Cmm.
For fixed values of Cml„/Cm„, Cr/Cmia, and NTU 0 , the effec-
tiveness will generally decrease with decreasing values of (hA)*,
however, for large values of NTU 0 and Cmin/C„mx near unity the
reverse occurs. At small values of NTU 0 regardless of
Transactions of the A S M E
 .004 5 8 . 0 2 3 4 5 B J 2 3 4 2 3:4
COCOCTION PARAMETER, X COKOUCTIOM PARAMETER, X

Fig. 3 Fig. 5

The effectiveness will decrease asymptotically to some value

IS
Fig. 4
the effect will be to reduce the effectiveness. In all cases the
effect is largest for small values of Cr/Cmm. For (liA)* in the
range of 0.25 to 1.0, the effect is small regardless of the values of
the other parameters.
The foregoing effects occur whether conduction is present or
not. The addition of conduction simply superimposes its effect
on the heat exchanger with the others.
For all other parameters held constant, an increase in the con-
duction parameter X will produce a decrease in the effectiveness.
compatible with the values of the other parameters. However, for
a design or operational point of view, this limiting value is not as
important as determining the deviation from the ideal or no
conduction effectiveness which could be expected for changing
values of the characteristic parameters. Figs. 3, 4, and 5 offer a
qualitative picture of this deviation in the form of the conduction
effect plotted versus the conduction parameter for various com-
binations of the other parameters. Tables 1, 2, and 3 provide
quantitative values for the conduction effect over a slightly
wider range of the parameters and also provide the means for
interpolation to cover specific cases.
One of the more interesting effects occurs for a constant value
of the conduction parameter and increasing NTUo. ^Tule an
increase in NTU0 increases the effectiveness, the increase for the
conduction case is less than for the no conduction case and,
therefore, the conduction effect will increase. This behavior will
continue until the no conduction effectiveness approaches unity
and then the conduction effect will decrease with a further in-
crease in NTUo (see Fig. 5).
It was found 'that the conduction area ratio As* had only a
small influence in the range of 0.25 to 1.0. Also the influence of
As* was negligible when varied ± 3 0 percent with respect to
{liA)*. Hence the values which were computed and given here
were obtained with As* = (7iA)*.
Lastly it is seen that when Cr/Cmm reaches a value of 10, the re-
sults are practically the same for C,/Cmin infinite. Thus for
values of Cr/Cmi„ ^ 10, the Hahnemann analytical solution can
be used (see Appendix).
A final observation on these results shows the prediction of
London [6] that the conduction effect would vary directly as the
conduction parameter,
provides a very good engineering approximation for NTUo > 10
and X < 0.1.

Journal of Engineering for Power APRIL 19 6 4 / 109

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 09/29/2013 Terms of Use: http://asme.org/terms

 1 CM
o 1 ro
O 1
o «o o ro st in m o -c -o r- r»- oo oo CD OO o- o- o»
1
1
II 1 1
• 1 1
oo 1 1
< 1 1
1
<3 •o a CD st o ro o CM o st h- ro o o m CM m CM «O st
• • o-
• ist•co
• t•— » I-c• oto Io I CM
• 1 in« o~
1 • I ro
• in • Iro•N OO O O O

< 1 •
jf-ocoooooooo—>—CMCMCMcgro Jt-OCOCOOO-OOOOO———»-'-———CM
X 1 o St 00 CD O O O O O O — — — CM CM CM CM CM

h- 1
1 1
O
Ui 1
u. 1
LL UJ CO
UJ UJ CD o
V. O
z UJ o o CM ^r ^ ui in in in in o >o-o-o o >o «o «o >o-o r-r«- st st in in m m in in -o <3 -o >o >o -o -o «o -o -o
o o O rg ^t^i/iiDLDi/iirxxjo-o-ooor^-r-N-f^r- CM

1
• 1
o H- 1
u u 1
o o UJ 1 o UJ 1
z o u. o LL
o u. St LL Jt
o m UJ O o UJ o
H N z o ^-{MCMCMCMrorororororororororororororo.* II z o CM CM CM CM ro ro ro ro ro ro ro ro ro ro ro M ro M ro CM CM CM CM ro ro ro ro ro ro ro ro ro ro ro ro ro ro ro
< Z o z o
UJ 1 1
Z X H 1 z >- 1
o u 1 u o 1
=> 1 1
o cc a
<
z O z C M o z C M
>-x o O o o
o u u
o ———————————————rg CM CM CM 0
h- 1 1
1 1 1
O 1 1 1 1
5 1 1 1 1
1 1 incMomo—roin-oN-o-oi—"-cMin-o>ON-N-
1 jj—ouico—roui-ocoo—ro.inn—N")Ln-oo 1 & «0 f*- r>- CD CO 00 CO 00 CO O" o> & a o- o o o- o>
1 o ^•ONSNCDCDmcDCOOOOOO-OOOOO 1 o
o 1 • •••••••••••••••••a* o 1
o o o 1 o
• 1
^ 1 II HH I
N X 1
X 1 <
< 1 z 1
Z 1 u
u •c CD ro o o ^ ro CMro st in o> o- ro x CD st en inV.m
V o CD ro m oo c w o ^ in jt oo co-o ro oo CM ^ in oozr^ UJ o
z UJ • o
o O«-O ^ <> FSL I N 00 O O — C M FO J T ^ H-CD CO CD
z
I ^••or^N-oococDaDcoooooooooooo o 1
o 1 1
11 11
< 1 < 1
OOOOOOOOOOOOOOOOOOOO a i o OOOOOOOOOOOOOOOOOOOO OIO OOOOOOOOOOOOOOOOOOOO
m 13 CQ| 3 I I I » • I I • I I • I • i I I • i I I
f CM ro St l/l <C N- CO o CM st >o 00 o o o o o o I 1 X | V- P-NfO^Ul-ONfflfrONtf-OflOOOOOOO
— — • - — — C M : * O C C O O « 1z
'-in -J 1

CM o o— ro »- coo—o- in ro J* co st o CM
a ro " O CM st st CM 0 0 Jt 0 » CD * O M 0 CM RO O M O-
•• • I• • I• II• t• I • III• I
«o o CM * m -o r- oo oo o o — » - st in <o <o oo
— — — — — — — — — C M CM CM CM CM CM CM CM CM

O X CM ro ro O st CM m CM «o O ^ ro 00 CM st 00 ro ro
ro m O O" ro O ^ CO CM CO ro ao CM m o - o CM r«-CM
I I I I • I T • • • I I I I I • • • I I
a -c t*- oo o o — C M CMTO ro a st o r - oo oo » - J J - O C D O O O O — C M C M F O F O R O J'INM-O-O ^ O CD O-O O o O CM C
MCMC
M ro ^ J

— — CD <3St MO»-O- a 0> -O O CO rofw & EGRO— o m oo o — M ro CD O oo o R - O oo ^ o J J o co CM

«o oo in —o o st r- o CM O oro o o m o st o R - 0 " 0 CM-O O CM & -O OO CM St <3 0 0 O O st F - C D >C

CM ro ^ mm o-o-o-o r-r«-ooco co ao O'-CM CM jf CM ro m m m -o >0 <0 -o r- N- r- 00 o- o o > o CM ?o m m m >0 ~o -o -o -o h- r«- r^ 00 00 ao co o

CM C
MCM ro ro ro ro ro st st st a a m^<3 r«- N- O - CM CM CM ro ro ro ro ro ro ro ^ st st st m m m m <3 —CMCMCMCMrorororororororo^ st st st X m
"X « I

— — »-•— CM CM CM CM CM CM CM RO st * a <3 — CM CM CM CM CM CM CM ro ro "O — — — —— — C M CM CM CM CM CM CM CM CM

O
s I I
st o o- m o a h- o ro m o> CM m oo o st—•o o- >o
st o <31— co oo o o o o o o o^«- ro ro ro m
CD CM

^ R - E G CM CM st
— I I

m CM-o m CM CM—-o Jt ro oo CM ro o ro <o K CM mooom-orO"Crorororom""omrocMCM^^

a <—co o jt CM N o* o- m k h- m—^ n Jt oo m
CM^o o -o ro ro o ro o- m st -o
mm « C M
o o N- CD N. o -o o——r- o- o oo m st—«- oo * •t< i ii• ii• • • ti• • • ii •
" O- CD «o 00 CM o a «o m CM^^ CO ^ ro JT o ro -o
<0 O "O O rornN- O- O—CM ro ^ m -O O C
M ro ro co ro—m o-—ro * m «oo o o»- CM m -o N- r» o> o * CM r- o ro m «o r- co o CM CM ro o CD OO >
jy>o-or^r-r^N-r>-cDcocococDcocoooooo ^•Os-N-NCOCOCD<DCDCOCD(>0-O00l>0>0 ^-Or^r-COCDCOCOCOCDO'O'OOOOOOO'O

I
I I
OOOOOOOOOOOOOOOOOOOO <1
O I O O O O O O O O O O O O O O O O O O O O O O i o OOOOOOOOOOOOOOOOOOOO
0) 3 CMTO m >o N CD «> O C M J» <i 00 O o o o o O B
x5I 3 IM
t- »-C •rO
• Jin-O^OOOOCM^>OCDOOOOOO
• I• « ttI« tI• • M I• •
X
< i z
— CMRO^LTLOH-COOOCM^OCDOOOOOO
———F—CMJ-«OOOOO < '-'-'-——CMJ-OCDOO < I Z, -'-'-CM^OCOOO
_t I —m —m
jI

110/A PRIL 1 964 Transactions of the A S M E

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 09/29/2013 Terms of Use: http://asme.org/terms

 a-jroj—oouo-frco
r>» <o <o co <o ro J T o> eg
0 « 0 I • • • • •
P-—ro-smr-cocoo

I O C M O — C O C O —•OF*-
oromcMor^-sf-o
J R>- co O- O CM J - -O CD jrr^-cco-o—CN^J I N H - co o O-•— CM CM
6 o 9 o g g o 00
U-lMCO-OO-^o

I
UJ co

CM .a-in m Is-a-o CM c\j a - m in-o N - u aa C

M j ^ m o >o CM ^ jf in m «o o «o-o

— LU o moro.*r~-ocot>«o
•O
•fO•
OC•D•OC
•M
•^•
f m-o
••
—C
MCMC
M rO ^T in •O h- — CM <M CM T O ^ JR JT —C
MCMC
MCM ro ro ro ro R - CM CM CM CM FO FO FO FO

~ ~ I

— ——— CMIM") — —— IMCMCM

•OCM«- -OOCOCMinh-
J'Oh-^COOOOO-O
•II• I• • •t

I
. .1 I
I I I I I I
<I <I < I
O I o ooooooooo O I o ooooooooo ooooooooo Q | O ooooooooo
CO I 3 CO i 3 CO I 3
X I V- X 11-
< I z < I z
-J I -J I

r-—rnjinr^cocoo

JOtDOO—CvjNl^f J F W C C ^ ^ O - C M CM
jfrf^-COOOO ———

LU CD
O
JJ
CM ro JT in M R>-co o CM CM a zx M in <O R^ co co o o TlTun-O-O^-h- eg a a in m >o o -o-o
» 1
1
o 1 — O I
o
o a.
u. a
•n UJ o
a in O — CM CM CM FO RO & 3 S
ii z o — CM CM CM CM ro FO ro FO
- CM CM CM M IS-
z o
1
i- 1
o a 1
-v 1
a: o
o z C M
o O
o
- — — — CM W O — — — — — IMCMCM <_>
1
1 1
1 1
1 1
1 1 I I
1
1
1 o
o 1
o 1 o
1 i
1 1 i i
II 1 i
X
<t
1 i
1
2: 1
o
V. o
z UJ
o
31
u 1
1
I 1 1
I I < 1
< I O 1 oOOOOOOOOO
O I o ooooooooo O I O O O O O O O O O O
CD | Z3 o OOOOOOOOO
CO I ZD CO | =7
X | H- X | H- 2E 1
< 1 z
< Iz < IZ -1 1
I

Journal of Engineering for Power APRIL 1 964 / 111

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 09/29/2013 Terms of Use: http://asme.org/terms

 mcMCMoro—roincD«oocM<o© , cMr-cocuin
mom—-o-o—inm-sr*-—eooaojfjnnco
m o O J » - » m o m CD ro m h-m ^ ^ UT—r^-vcoininajiri I N - O H - O co O CM M R « - O RO J F «O R - O - O oo F - *
• •••»i»0««*(**ttt** • •••••••• •••••••••••••••••••a
•o a CM a imn o «c r»- h- co eo ® co o o o o
<© a cm jt jt r^ co co aj

OCMJ»OO.*——^oocMcMorooino«of^
o >0—m aoo—ro Jt "Or-f- coaof- m.*—
J'OMDOOOOOOO-'— « • • i• • • « i• • •i• • • i• • •
jfON-coaooooooooooooaoo
jr^r-cooo——

(M rn # * in m m m in in o o -o o <3 -c -o-o >o cgro^-^mininininintnin'O'O'Oininininin

CM ro Jt Jt in in -o -o m

—CMCMCMCMCMrorororororororororororocM —CMCMCMCMCMrOfOrorOrOmrOWirOfOCMCMCMCM
P>J CM <M CM RO R<L RO CM

I I
I I I \
ro o cd ro o o o—cm ro a jt a jt ro m o h- N.-c
jti in• oi r- f- n- oo co co co co oo oo oo co n- in * ro
i• i • i• • i• • • i• • it •

Q|o ooooooooooooooooooo <I a I o O O O O O O O O Q O O O O O O O O O O O

O I o ooooooooo CD I 3 •••••••••••••••«•••#
CO I ZD x I h- ~cMrojrm-or*-cooocM.*>ocooaaooo
< IZ <-—«-—>-CM J" <OCDOO
-J I —OJ

cs

^ o r- ^ o •— m r~ o -o in o -o in o m CM «O O
;yinr-cNm—r^cNf-oo—^r-o—i/ror-
i • i • i• • • i• •• i i• • « • •
•oocM-^mo-or-r—cocooowo————
—OJ C M CM OJ

^•or^-cooooo—cMCMCMmrojt.»min Jj- •or- COO O O O O——CMCMCMrOrOrOfO

rg fn Jt ^ m m <o-o fw r-co oo co o — ^ CM r o ^ I N M M - O O - O - O - O R - ^ N - c o co ^ O
CM ro # in in m mo -o -o <3 -o r»- r- h- h- r-

CM C M — C M C O O O ( D J T D - I N A O C O H - R O I N C O C M ^ > O
^ o ^ N. o CM ^ m f» co—m m r»-o-o o — m o CM N—o-«o h--o ro r-r» «0 fO fr CO Jt
J T •— j t R - c o O — C M M a
i n o F - co A O — - o
•• • • • • • • • i• • • • « < • i •
ro ro ro ro ro jt jt Jt Jt a <3 «o r- o
— C M CM CM CM -cMCMCMCMfOfororomromj* j-d-^ininin —cMCMcMCMromroforomrorororo.*:**.*
s: h~ l
i_> u |
3 l
o CM —oo m o in o - o - c o rn o jt t^ co
co—CM Jt in in -o h- f- r- oo o o- o o — o o
— '— — »-•-«- ^-CMCMCMCMCMCMfOm Jt Jt <3 — — » - C M CM CM CM CM CM CM CM CM «- — ^ — — — C M C M C M C M C M ^ -

CM co n-CM r—^ n. o—in oo o C M JT «>• o o es

^ • M - O H - R - A O CD CD C D O o o o o o o o o o
— — — — — — — — CMCMCMFO

Jt CM o CM N. r-——co oo o in K—rn—CM CM o O —CM o CM o Jt—o o o m (MTO om i«- o— o

—co-ocor^o^-—cMfMco——or^-ommopo r^-ooorocor-—rofoo—or^fn^-—inr-
h-o n - — ^ «o CD o — C M ro in-o ^ r^-—fn ^ j co om^ocM^-ooooo—ro^jfincoooo
j-o-or-r>-r^r-cDcocDcoaocDcr)coooo-oo -^•Of-r^CCOOCOCOCOOOOOOOOOOO'

I O O O O O O O O O Q O O O O O O O O O O O ) l o ooooooooooooooooooo ooooooooooooooooooo
1 I 3 —CM ro Jt in O h- CO O O CM a O CD o o o o o •• I • • IIItt• t< • • • II •
X< II zK
CD I 1 3
-f^rOjm-OSCOO-OtNJOCDOQOOOO : 11- —CM^ <J COO C
M ro ^ if)<31- CO O O CM Jt O CO O O OO O
— — — — CM^-OOOOO tIZ ——CMrf-OCOO
—m

112 / APRIL 1964 Transactions of the A S M E

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 09/29/2013 Terms of Use: http://asme.org/terms

 y i/lNtooo —m^in-of-coh-

jf N- co o—to in -o st O CD COO -~CM»Oro ^ -O CD CO O O « - » - • -

st'O aatDOOOOO

eg ro jy S in -O CO O— CM RO M «R> «o N. T - R^- CM to.*.* mm-0 m in

- CM CM CM CM tO St in -o — CM CM CM CM M J O A ^ - CM CM CM O J M RO M CM
—CMCMCMCMFOTOTOCM

"——•——1—CM'OtO

I I

II —>—<

I
I I
I <I « I
O | o ooooooooo a I o I o ooooooooo
CD | 3 o o o o o o o o o
O O O O O O O O O O O ! O
cc I Z3 CD I
ta I D
a: I X I a: I »-
<• I z <I « |z
-J I

<N

•oocMjinr-ajoo •OOfMrf S ' O S N N

om-rcor--®—ai-o
JOMXJOO—— rff-O-O — CMCOCO.*
^or-ajo—«!ir>r- s •© T*- co W- CM ro ro
jJ^MOOQOOO

oo—in st o o st in CM
•oomocMP-oo.*
i i • i • • t•i
CM ro st * in o r— r«- r*- CMro-sjji/vno-oo csiro^T .*mm<omin

— CM CM CM CM TO MR^ — C M C M C M C V O F O st St — CM CM CM CM to to RO CM CM CM CM CM to fO tO CM
— — |

——«— — CMOfjO a

TO O F - CNMN CM CM CMN-
• • • • • • • •ro
•
X M O R - ^ CD CD H -

II —— I

I I
I I
I I I I
< I < I
O I o ooooooooo ooooooooo •Li
o
co
|o
i
ooooooooo o |o ooooooooo
C3 |
XI
3
to| 3
X H- r - c M t o ^ i n o o o o
< I z < *
-J I

Journal of Engineering for Power APRIL 19 6 4 / 113

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 09/29/2013 Terms of Use: http://asme.org/terms

 — oro'cmco-a-—in CMCD®->O—INROIN-
rocM-n-omfo—om
i i ••i i •i • (•••••it*
o CM M
•O R - R^ R - •O o c% ro a *o r- N-O •O cj cm ro jt >0 -O <3 m

J-Oh-COCOOOOO j-cr^-coajo-ooco ^-o h-CD CO O O O-CD

cMfo^t ^mminin^ cv ^ a in m m ^ ro CM ro * a m m m jr ro

mcoocooo-o^r-
^•orom^co«oo—
— CM CM CM CM CM CM C M — -C
M CM C
M CM CM CM CM— — CM CM CM CM CM <M C M —

i i
o o i
\ => l

I I

< I
O O O O O O O O O a i o ooooooooo ooooooooo
o | 3

ro -o co o»—ro in r- ^or^-ooo—CMroro .3-Oh-COOO —CM —

CMro^^in-OCOO.— CM ro ^t a in O N- r^- r- CM "o J - J T M -o >o >o M

•——(MiNCMro Jjm-o — CM CM CM CM ro ro a ro -C
MCM CM CMO
' ro ro C
M

o z CM
o o
—————CMroro u
0 — — — — — CMCM—
1
1 1
1 1
I I 1 1
OrOfO—f-CMrOO>»- 1
rom<or-r*-oro-ocM 1
i i •i i * • • • 1 o
——CM o 1
o- 1 o
1
1 1
II 1
X 1
< 1
X 1
u
o
z UJ
o
Z
o 1
I

1 1
ooooooooo < 1 ooooooooo
I • I II I•I t 9 1 O O O O O O O O O

»-CMro^moooo
—CM^tO L L
< 1
-J 1

114 / APRIL 1964 Transactions of the AS M E

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 09/29/2013 Terms of Use: http://asme.org/terms

 •OOCMNI.^OOOI/1

sto P - R - OD o CM ro in jj-or-oocoo—— — J O M U C D O O O O j-ostoajooooa

- I

CM RO * st st in in 1/1 st CM ro st st in in st ro

— C M CM CM ro ro Jt in «-CM C
M CM ro ro ro eg — CMCMCMCMCMCMCM— — CM CM CM CM CM CM — —
-i |

- — — CM CM CM

|| —

I I
<I
O O O O O O O O O O I o ooooooooo ooooooooo >Io ooooooooo
en i => J13 •«••••••»
: j»- — CMro^rmoooo
I IZ —CM^O

-Q
CO
A
a
H

00 O — St IM o O u»CO
oirmot-j—coaj-o
•i i i • i t • *
mo—cM-»r-o—ro •<3 o CM ro st R - CU O O OotMro-yor-p-f- •oo CMWI * m >o-om
—————CSJCM

ro in P - oo co o CM st o Tor-cocoo- •»«or-cooooooo .^or^cocooooco

C M R O * -3" St O P - C O O
CM'O * a in in o o in CMro st * * mmin* cMro*

rocOCMJTO^CMOC
——CMCMCMrO^inO —CMCMCMrororocM — CM CM CM CM CM CM C M — — CM CM CM CM ( M CM C M —

— i — I

r-r-rocMP-inoror-
P-OCMrOfOJTCMCDCM
• •••••lit
— —— CMCMRO

I I

O O P - O - » O O — O
CM
•
CD P - — O in J* C D ©
• • • • • • « «
—cor^roo*cooo
mor-cocooooo

I
I I
11 < I
OIo O O O O O O O O O
O I O O O O O O O O O O O | O ooooooooo ooooooooo
CD | 3 CD I 3 I • I I I • •• •
X I H- —CMroj»inoooo
< I Z —CM^O
—J I

Journal of Engineering for Power APRIL 19 6 4 / 115

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 09/29/2013 Terms of Use: http://asme.org/terms

 Summary and Conclusions2 C l = JiSiiE. N X U o
i +
1 1 1 L = NTU„[l + (/U)*] —
Cmax (7iA)*J 2iVr| 1
The numerical method presented is completely general for the
problem of steady-state performance of the periodic-flow heat Cmin Cr N. Cr N,l
c2 E, =
exchanger with finite thermal conductivity in the direction of Cra Cmin N r Cmin Nr
fluid flow. The ranges of the governing parameters which have
been covered are Cm l N r
As* A'r
Ci =
Cmax 1 -4s* 2 E s = X 1 + As* 2
1.0 ^ Cmin/Cmax ^ 0 . 9
C, = Ci + (1 + Ci)(C 2 + 2 C 3 ) E, = Ei + (1 +El)(E1 + 2E3)

1.0 ^ C r / C m i n < CO c6 = Cl + (1 + C,)(C 2 + C,) Ei = El + (1 + Ei)(E2 + E3)

- - § 2 Ex
1.0 sC NTUo $ 100 Fi =
E<

1.0 ^ (hA)* > 0.25 Cl + (1 + C,)(2C 3 - C2) Ei + (1 + E>)(2E3 - Ei)

Z>2 = F2 =
c< Et
1.0 ^ 4 b * J: 0.25
(1 + Ci)C3 (1 + EJE*
Di = F, =
c, Et
0 . 0 1 ^ X $ 0.32
1 - Cl 1 - ffi
A = F< =
which is the range of interest for gas-turbine regenerator applica- 1 + Cl 1 + Ei
tion. The method itself is not restricted in range.
The values of the effectiveness reported are good to four sig- Cl Et
A = F„ =
nificant figures while those for the conduction effect are good to Ex
1 + c, 1 +
three significant figures. This reliability was obtained by extrapo-
lation of the results for a finite number of subdivisions to an 2E\
infinite number of elements. I)e -
C6
It was found that the effects of (hA)* and As* on the heat ex-
changer effectiveness were sufficiently small that the results Ci + (1 + Ci)(C 3 - C2) Ex + (1 4- Ei)(E, - E2)
could be presented graphically over a wide range of conditions •D; = F-, =
E5
to provide a rapid means of observing the influence of thermal
conduction. (1 + Ci)C3 (1 + Ei)E3
De =
It was also found that for values of Cr/Cmm ^ 10, the analytical E5
solution of Hahnemann for Cr/Cmm = can be used, thus an
upper bound on this finite difference method has been established.
the Hahnemann solution for Cr/Cmm infinite [7].
In the Nomenclature of this paper, the variables used by

References Hahnemann can be expressed as

1 H. Hausen, "Accomplished Calculations of Heat Exchange in 1

Regenerators," MAP Reports and Translations No. 312, November (hA)*'
1, 1946.
2 B. H. Schultz, "Regenerators With Longitudinal Heat Conduc-
Wi Cmin)
tion," I M E - A S M E General Discussion on Heat Transfer, London,
England, 1951, pp. 440-443.
3 W. M. Kays and A. L. London, Compact Heat Exchangers, Wo Cmaxj
The National Press, 1955.
4 H. H. Sogin and K. Hassan, " A Design Manual for Regenera- A.- = [1 + (M)*]NTU0
tive Heat Exchangers of the Rotary Type," W A D C Technical Report
55-13.
5 J. E. Coppage and A. L. London, " T h e Periodic Flow Re- and = v [1 + (/iA)*]NTU0.
X
generator," TRANS. ASME, vol. 75, 1953, pp. 779-7S7.
6 T. J. Lambertson, "Performance Factors of a Periodic Flow
Heat Exchanger," TRANS. ASME, vol. 80, 1958, pp. 586-592. The following abbreviations are then defined
7 H. W. Hahnemann, "Approximate Calculation of Thermal
Ratios in Heat Exchangers Including Heat Conduction in Direction
5i
5j = A . . ( f , - x )
of Flow," National Gas Turbine Establishment Memorandum No.
' \ Ifo /
M36, 1948.

5 1 <^1
= - [ $ < (
8 G. M. Dusinberre, Heat-Transfer Calculations by Finite
W
Differences, International Textbook Company, 1961. \(<r + 1) +
9 G. D. Bahnke, "Effect of Longitudinal Heat Conduction on Wo J
Rotary Regenerators," MS thesis, U.S. Naval Postgraduate School,
Monterey, Calif., 1962. RV
oA,$i
0 - IF„,

APPENDIX Ss SA Ai
2 6 27
The following is a list of the dimensionless coefficients for the
finite difference equations (3), (4), (5), (6), (9), (10), (11), and (12)
V _ A
in terms of the governing variables:
9 3

2 The opinions and assertions of this paper are not to be construed

as official or reflecting the official views of the United States or
Chilean Navy Departments.
cos" 1
(Vp3»/)
116 / APRIL 1964 Transactions of the A S M E

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 09/29/2013 Terms of Use: http://asme.org/terms

 2 v_ /-P cos -'" - -5.
3 3

- 2 v_ /P- cos ( -<P + -1T) - -

5.
333

T3 _/- cos
= -2 vp ( -'" - -1T) - -5.
::I 3 3
The effectiveness is then computed as

E = I - P (I +u
I
:J {r.S(I
uAiw,
+ u) [
en - e"

+ ( -;:;-
T2 ) 2
( ell - en) + -;:;-
T3 )( 2
(en - e rt ) ]

+ -ucI>,
7•• 2 (
u 2 -W·'
W.
- 1) [ crt - eTi + ,', (era - en )
~
T.

'where tbe quantity

P = (1 - en)(e r • - eTl) + -T. (1 - ert)(e Ti - en)

T2

+ -T. (1 - crJ)(e rt - crt) + 1 {T.

-- 3 [e rt - era
~a TVi UAiWi
1 +u W.

+ 7." (
T! ) 2
(era - en) + -;:;-( 1'3 ) 2
(e Ti - crt) ]

+ ( -T.T3 ) 2
(en - ert)eTl ] - -T.
UW i
2 [
ert - eTi + -T.
1".
(e Ti - en)

Wi
ur. 2 -

+ -T.r3 (en - ert ) + --W.- [ er1(c rt - e")

Wi

+ -~
T.
(crJ - en)e rt + -~
T.
(en - ert)erJ ]} .

Journal 01 Engineering lor Power APRIL 1964 / 117

Downloaded From: http://gasturbinespower.asmedigitalcollection.asme.org/ on 09/29/2013 Terms of Use: http://asme.org/terms