AD-A113 355 VON KARMAN INST FOR FLUID DYNAMICS RHOO-SAINT-GENESE--ETC F/S 13/1THERMAL ENERGY STORAGE IN PHASE CHANGE MATERIAL.(U)MAR 62 P WHITE, J BUCHLIN AFSR-81-0119

UNCLASSIFIED EOARD-TR-B2-6 ML

i lllflflflflllIflEEEEllEilEEEElEEEEEEEEEEllIEEEEEEEIIIEEEE

EEEEEEhEh

III''III '11111 I 111IN H

IL2 . 4 • III

AD A ll 3 355 RE .IN'STRMTIONSBEFORE COMPLE'rI::; EORM

.eort Nux'be r 2. Govt Accession No .Recipient's Catalog Nuniibr

EOARD -TR-8 2- 8 __

4. Title (and Subtitlc) 5. Type of Report & Period Co-'re,.

Final Scientific Report

Thermal energy storage in phase 1 Feb 81 - 31 Jan 82change material 6. Performing Crg. Report Nu:r;erI

7. Aut.hur (s) 8. Contract or Granr N-,ber

P. ,dITE & J-M. BUCHLIN Grant AFOSR 01-0119

9. Performing Organization Name and Address 10. Program Element, Project, Task

von Karman Institute for Fluid Dynamics Area & Work Unit Nu;-Ubers

Chaussee de Waterloo, 72 P.E. 61102FB-1640 Rhode Saint Gen~se, Belgium Proj/Task 2301/Dl

W.U. 120

11. Controlling Office N-me and Address 12. Report Date

European Office of Aerospace Research and March 1932Devel opment/LNT March_1982

Box 14 13. Nuznber of PagesFPO New York 09510 S1

14. Monitoring Agency Name and Address 15.European Office of Aerospace Research and

Development/LNTBox 14FPO New York 09510

16. & 17. Distribution Statement

Approved fur public release; distribution unlimited.

18. Supplementary Notes

19. Key Words

Latent heat storage;Matrix heat exchanger;Solar energy application

$0. AbstractThe present study deals with an experimental investigation of low temperature ther-mal storage based on macroencapsulation of Phase Change Material (PC,).The storage performance capabilities of capsule bed, tube bank and tubular single-pass heat exchanger are compared. The tests are conducted on the VKI Solar Utility I

Network (SUN) which is a closed loop facility designed to study air heating system.An original data acquisition chain based on two conversing mircroprocessors isdeveloped to carry out mass flow, pressure drop and temperature measurements. Theexperimental results are interpreted on the basis of comparison with numericalpredictions and they allow to draw the following conclusions.Each type of matrix has its own range of operation for practical application butfrom a heat transfer standpoint, the PCM capsule packing unit is strongly recom-mended. It is suggested to extend this investigation to the effect of Reynoldsnumber to find optimum range for therro-mechanical efficiency. k°e i

8 ffciecy 14)

82 04 12 142

EOARD "TR-8 2-B 9

This report has been reviewed by the EOARD Information Office and is

releasable to the National Technical Information Service (NTIS). At NTIS

it will be releasable to the general public, including foreign nations.

This technical report has been reviewed and is approved for publication.

WINSTON K. PENDLETONLt Colonel, USAF

Chief Scientist

GORDON L. HERMANNLt Colonel, USAFDeputy Commander

Acoession For

NTIS GRA&IDTIC TABUnannouncedJustificatio

ByDistribution/

Availability CodesAvail and/or

Dist Special T0ISPECTED

1A, ....

UK -; 11,'!11'

A c know 1 edg-emo'~n t

Thainks to the United !rtates Air Force and the taxplayers

of the United Ftater, for financingI thii year of stuly, in Eurone.

1 will strive to apply ry' traininR to the benefit of mny country.

iable of Conte,.cs

I. Introduction ........................................ 1

II.Heview of Literature ................................ 2

III . Theory ............................................ 4

IV. Comparison of Different Fysters.................... 11

V. Fxrerimental System ................................ 13

VI. Measurement Fyste ................................ 14

VII. Fxperimental rrogram and Results ................. 15

A. Tube Cross Flow ............................... 17

B. Packed Fed .................................... Ll

C. Tube Parallel Flow ............................ 59

VIII. Comparison of Uumerical and Experimental Result.67

IX. Efficiency Considerations ......................... 70

X. Conclusions and Recommendations .................... 71

XI. References ........................................ 72

Appendix I: Photographs of Experimental Program ....... 73

Appendix II: Diagram of S.U.N. facility ............... 79

U/

I: Introducti on

lm.nino you have a recnhouse in Helgium. Fore yn,:

have a natural solar collector that doesn't cost :.rou an:,thi.:;-,

but you still ;,end thouoards of dcl!r!; a year on heatin

fuel. WI-a't's more, and this is the really frustratin- art,

on the few beautiful days there are, you have to oT',cn win-

dows to let out the hot air durini the day, a n then heat

with oil during the night. You think this is crazy, which

it is, and you decide to capture some of that heat you're

letting out during the day, and use it at night.

So what do you do? There are several nossibilities.

You can put some rocks under your greenhouse, and then blow

the hot air through durinc the da:. and blov cold air throu2h

the hot rocks durinr the night. Or you can fill old coke

bottles with nraffin, or some other phase change material,

and run the hot air throug-h them. Th-it way you have addi-

tional heat capacity in the same volume, due to the heat of

fusion of the material. Alsro the storage will be at a more

or less constant temperature, the temperature of fusion, while

it is being loaded and unloaded with heat. Another thing you

night do would be to put some plastic tubes under the ,reenhouse,

put the phase change material in between the tubes, and then

blow air through the tubes. The hot or cold air will then

give or take up heat as it nasses. Finally, you could put

a phase change material in tubes, and blow the air through

the tubes. There are other nossibilities, and the different

geometries and encapsulation methods are limited only by

one's imaginiation.

What would be good to know is how to make a heat storar,e

system that, over the lifetime of th- -rf-enhouse, will

maximize the amount of money saved by acting, instead of

continuing to throw out the heat cantured diirinr the day.

For the storage system you'd 3ike to know how to get

the most heat in and out in the shortest amount of time;

P

that will corrennond to the minimum amoint of electric ener,:r

needed for 1lowin- the air. in fact for a ntorage system, one

can name a criterion, a coefficient of n-rformance, 5uzt like

a heat numn has, which tells how m:;ch heat is saved fcr a P ven

quantity of work, or pumninr energy.

The goal of this renort is to use the results of an ex-

per4mental investiration three heat exchinaer storage units,

in order to decide which is best. the advantages and disad-

vantages of the encapsulation methods and geoetries will be

discussed; together with thin, a theoretical model and its

relation to the exrerirentally observed results will be pre-

sented.

II: Review of Literature

For his doctoral thesis at the von Karrman Institute,

Theunissen(S) has done a review of the state of the art in

heat exchanger modeling and the chemical processes which are

involved in heat-of-fusion storage systems. As the present

work is comolementary to the Theunissen thesis, a short re-

view of some additional articles is presented here.

The world expert for phase change materials is Dr.

Maria Telkes. She first suggested the idea of using Glauber's

salts, paraffin, and eutectic salts for storing heat in the

1950's . In Solar Engineering (I) she notes in a review of

the costs and benefits of various storage systems, that the

people who work with phase change materials should be physical

chemists, since nany of the problems associated with these

materials are chemical problers. Examples of this are the

separation of the liquid from the solid in hydrated salts,

and supercooling during the resolidification of the meltel

material. She also points out the need for attention to be

paid to the method of encapsulation of the material.

In the Journal of Solar Engineering(2) a numerical sim-

ulation is presented which is similar to the one done in the

present program. That is, the model consists of enerpy balances

both on the heat transfer fluid and on the solid nh'Ie,

which receives the heat. In the trade nuLl Aiclti n, .olar

Age (3,1) passlve latent heat stora.-Ye systms arc discuissei

as are hyhrid syztens and the need for low-enerwy-conumin

in another article by'~'ek ) th., need' tr Consider

3yntoms which uso electric resi stance hartinr- durinr off-

peak utility periods, and the need to consider the encapsulation

are discussed. Fmith et a]. (0) note that the liriting

factor in heat exchanger performance of storage systems is

the resistance to heat transfer in the solid nhase. In

modeling the microscopic phenomena of the phase change heat

exchang'e, they stress the imnortance of natural convection in

determining the shane of the melting front and in thus in-

fluencing the rate of heat transfer.

The heat transfer coefficient, h , models all the heat

transfer from the fluid to the solid phase, or vice versa.

This is always based on empirical correlations, and thus may

be a dominant cause of error in any modeling of heat exchangers.

In this brief sample of the literature on the latent-heat

storage subject, there are three probles which are ccntinually

mentioned. First, there is the need to model mathematically

the performance of the heat storage unit. Second, there is

the need to devise low-cost encansulation methods. Finally

there is the need to consider the rurpin requirerents for

any heat exchanger, and to relate these to its thermal per-

formance.

The goal of the present work is to address each of these

three issues, based on the experimental program of the snring

of 1981.

14

III: 'Pheor_'v

We have the fir;t la w of therm odynam ' , which says

that, for a r,;-ven device,

Rate of heat :.:ate of heat Rate of heatpoing in + Production = roin,- out

For the air poing into a heat accumulatcr, one can

consider the rate of heat goinrr in to be:

U intA 3C p,air n Convection of hot air in

The rate of heat going out equals

NV UintA3 Cnar T out Convection of hot air out

+ h1 (Tair (y)- T PC(y)) Heat given to solidfrom air

+ h2 A(T (Y)- a (y)) Heat given to exterior

2 4 air ambient through walls of storage

un it

Where A,= Heat exchange surface area

A 2= Accumulator cross-sectional area

A 3= Accumulator interstitial area

A4= External surface area of acumulator

The rate of heat production, in the case of a machine

derrading electrical or mechanical energy equals:

e , V air C p,air dTair /dt

If we put all these terms into an equation accordin, to

the first law of thermodynamics, we have:

airtintA 3 ep,air T. + .V .C dT /dt=ar int3parTin hair air p , (air air

eairUin tA3C p,airTou t I I air (y -T PC"..: ( ) h2A air y -T ambient (y)

Now if we take this equation and div, d( al ltrmrt

the volume of the accumulator, we h.ave (rpmembrinr thiit

volume is eq'ual to leni-th times cro-;s-sectional are-):

Q . tA -C - . I; An C, ,a T ou ' i ir c, i d'. r

air in 3 p,ar in ai irmnt 3 a out air n a ,r air+

A 2 L A L A, L dt

hI A 1 (T air(y)-T (Y)) h2 A ? air(V)-Tambient(.))+-

A2L A 2L

From this eauation, we let

S = A 4/A L

Ss= A 3 / A 2= V air /A2L

a,= A 1/A2 L

h,= heat transfer coefficient between air and solid insidethe heat accumulator

h 2= heat transfer coefficient between air inside and outsideof the heat accumulator

Rewriting the energy balance with the new notation,

we have:

airintsC rairkTin,air Tout,air) + v aC p,airdTair /dt

L

h as(T a ir (y)-TpcM (y)) + h2 Sc ( air (v)-T ambient(y))

Letting the length of our accumulator become infinitesimal,

we have:

ea rUintsCp,air dTair/dy +Eveair C,air air

h a s(T air (y)-T pc(y) + h2 c(T air (y)-Tambient(Y))

or a complete statement for the energy balance on the air

Lg

Similarly, we have an energy balance for the PC'.1, that

is that all the heat it receives results from heat given un

by the air:

PCMVPCMCPPC! dT c/dt = h 1 ' air PC,.,,PC '.i 1 (y)-T (C )

Now we begin to make assumntions. First of all, let us

say that the external losses and the accumulation of heat in

the air are negligible when conpared to the convection terms

in the energy balance. The energy balance then becomes:

-,ir dT / dy = h as(T a (y)-TP,( )Parint s p,air 'air 1 s ir CM ) )

together with the equation above it on this page.

The next assumption is that the TpcI is constant with

time and space, at the temperature of fusion of the PC',

while the fusion is occuring.

Thus we obtain:

dTair(Y) h a s

- - dy

Tair fusion PairUintesCp,air

Remembering the definitions of a s andes, we have:

dT air(Y) hlA

dyTair (Y)-Tfusion A1 LairUintCp,ai r

and we see that the multiplier of dy equals:

hiA

mC .L

p,air

where m is the air mass flow-rate.

Zow integratin r both :ido s of the e-nerry :larceA , I:e, v:

a r 6 ,2 Yr

F1l a r y-:,)- uinairuon

= "tn( /

i fi o ' r

whe re h we idreie ats:

air fua

From thich de rivo e a osrc te£loir ~~

= exyr (-/t y/L)T. - T

'in fusion

where St* is defined as hiA I

DjaL "

Physically this dimensionless rroup the Stanton num!-. r

gives us the ratio of the hea t trsnsfered in the hea t accu;ulator

to the heat convected throurgh the accumulator.

From this derivation we can construct the follewinc 7r.rh:

--

TC : sIn, rf th1s <rarh at r" . .

-I fr --r t/ e-rVP!v exp( r2

-, .... / ,) ) _,,L

o th' vat, _ r te - ,r !n a .

or na l v-lut Vy the time the a (-a l a r, r -co7-.mo etel,'

t r aver !, wh.le fo.r a Ft of 2, the red temncraturp

will decay by a factor of (ite) ti en ts cri ,al value

by. the t 1i-p the accu.rmul tor s, t.r ,, rn~d. The r a h beo6

illustrates the s n ca n tr:e of tl: eoo-y !Ia-.

TrAs, -

.~5tU10

The hither the Stanton nunl. er is, the hir,}:'r the heat,

transfer will be, and the hi.gher the rate of reduction of

air tenrerat,re will be as th accu-.ulqtor is traversed.

This haa 2 imnortant consenuences. First, if one is interested

in storinr heat rapidly, one wants as largne a temnerature

reduction as possible. Secondly, one must consider a closed

loop solar heatin,: systen, includi.i-a a solar collector and

a heat stnra -e unit. Thp hih:r th- tpnreritur- goin,- into

t I" P r en r r t .

.~ ~~~~ ~~ ~~~ w.. .. ,: . '.,-, ,n'T

h- ot : ! .. -. r

,:n~ n i'. - -r ,:" .- f-,' t r v0w o"h et a . ... , , " r ,. t'; r,"

th'r:~ r -!

Dia;,rri of Collector-Ftra'e Loo'

One may have noticed by nr." th".t not ,:c; ha- 'een sq:11

about h i , the heat transfer coefficient. This ;s a

which can be addrossed( only by .e-.ns of erririvii correlaton!.

W' kno; that orii-inally., the Stantcon numrer locally en-vilf]

h / ai C U1nr r ,.lr mnt

h /k X k/ C / ar nVt' r,air "" " n .

Fn it rem-ains to find the Nussr,2t nunber by means, of e-riri-al

correloItions. For n nocked bed, we have:

Nu= (1+1.5( I-E ))(2+.C)', 1 0557 le'Pr') Pr 1 /7RV 1+2.h,(rr 2 3 -1) -

While for the tube cross flow,

Nu= .3 e 6

and for the tube rar. l el flow, in the turbulent rei-r,

N:u= 0.22 Pr Ro"R

imom ...... A

Ly re arranrn in -- the ter ! - n the ener y ral;Lnce e'I'lat cn7r

we arrive Fit the folloiinr :

d7'a/ir(y/L) = P. i irC airtint

-, ' r nt,V:.,. ...... h1 :"

-~~~ -- a_____ - a rn C ;j • 1-£( t ,C , pr'

i air r,a-r int s v r ,

These 2 equations and their discretization form the sulu.ject of

the work of Ouestois (9).

A brief summary of the numerical approach follows. We have

a backward time, backward space discretization scheme:

dT/dt = T( i ,j )-T( i-I ,. )/C for i=1 to n

d /d(y/l,) = T(i,j) - T(i,j-1)/ (y/L) for j= 2 to m

This is straiphtforward; the only part which requires a bit

of thought is what to introduce as the heet capacity of the

PC!' when the fusion is occuring. The ar.proa- h chosen is first

to use a delta function, of finite width to express a "heat

capacity of fusion". Normally this would be, with a precise

fusio. tem-erature;

TThis is first modelled by havinr7 a finite-width spike centered

on the fusion temperature;

C.

l"rav,., T )

ani then by a smoother di -tr ibuti.on of heat caracity abo-'it t ..0

te-.perature of !usion, as n'n on the tor of tho next. .

s# su 1 aon therffo ro icr i tI1 car'-c 0 f

the -PC*, as the fusion temperatu.re s a nnroachod, an-I tho

Phpnomonon of the to .noerature re'naining constant during

fusion is simulated.

IV: Comnarson of Diffrnt 'tor'-ke F[rstens

To cc pare the different methods of heat storage, it is

i7-ortant to consider: a water storage, a rock bed storage,

a paraffi.n rhnse change storage, an! a CaC 2 CH 0 system;p 2all these systems are comnrared for air as the solar-heated

fluid.

Water Rock bed Paraffin CaCI 2('2 0

1. Pipes 1. Rocks 1. Paraffin 1. CaCI 6' 0First 2. Storage 2. Air blower 2. Encarsu- 2. FncarsuationCosts tank 3. Air ducts lation 3" Ducting

3. Heat 4. Walls to 3. Ducting 4. Blowerexchanger hold rocks . Blower

4. Blower5. Water

pump6. Ducting

for air

Volumerequired 1. Water 1. Volume of volume of PC'.' + Fncapsu]ation

Volume rock bed2. Heat

Exchangervolume

Operation Tn all cases, a function of the heat transfer coefficient,costs the pressure loss across the storr,, and the voluie flow

rate of the air.

Fror. thi:; q1,a1 it. at'v- , evaluation ,: r o i % nunti, tit, ivr

conmparison of volurue requirfrmnt s for di*'feront :tor'.e -'doz

en . set stor'-ie canacitv, of' 5CO mni a,;nules, c,* n 1 ti ,-v'-

erature c n'.: n-e f'rc. I tc to 5 O .

;.o- water, C = 'I w /&"- 02 = . l .,o

.hus the volume nf water reou red1 .huln b

5X 1 /(1. n:1 x 0 )=1 0, ir 1 ui = 11) I c.,

Yor rocks, C .9 Kjoul/G--C , 2502 KG/. 3 and we

will assume a void fraction of 0.4. One cubic meter of

storare will then contain 0.6.. 3 of rock, or 1500 KG.

We then have:

5 X 10 8/(900 X 20)= 2.8 X 10 K71

2.8 X 10 /(1.5 X 10 3)= 1 .5 of rock bed

For Paraffin, C = 2000 JOUL''F/KG- 0 C and we see thatP

1 8 4i5 X 10 /(2000 X 20) = 1.25 X 10 KGr are required. Therefore

the mass of paraffin required will be lower than that of water

or rocks, but one must remember that encarsulation is required,

which will have void space as well as volume itself.

For CaC12 6!2 0 , the heat, sensible, of solid and liquid

phases is 2000 Joule/KG- 0 C, the heat of fusion is 170000 joule/KG

the temperature of fusion is 29 0 C , the solid density is

1.5 KG/Liter, and the liquid density is 1.14 KG/Liter. We have:

5 X 10 8/ (200C X 2000 + 170000) = 2.), X 10 K-

2.4 x i3 / 1500 = 1.6 %1

As with paraffin, there is encapsulation and void space required

but there is still a significant volume reduction over rocks

or water.

Together with less volume, there is a smaller bed through

which to pumn the hot air from the solar collector, and this

will result in lower blowinr, or r-npinr costs, since there is

less r,:;sure loss with the: smullcr accuimu] tor.

13

As mfntioned before, three heat accumulator.s, or throe

methods of encansulation, were tested. The corCncotr are:

tube cross flow, tube pirIllel flow, ind a raeked bed. T

are lia,rammpd below.

t "z

/*

TOP VIEW~ ."........ - _

Tube Parallel Flow Tube Cross Flow Pac'ked Bed

In addition to ther, ,al characteristics, such as the stanton

number, there are several other means by which the aceum~ulators ;

may be compared, as tabulated below:

Tube TubeCross Flow, Packe'cd Ped Parallel Flow

Heat exchange2.. 2Surface area 18.1 14 I .3 I' 5. .

Void fract-ion 0.59 o.4 0.35

Vol.. PC',l 0,.2 6 0.5 O.CVol. total

7.17 0.11 0.03Vol. total

I{eat Capacity/:1,3 7 e~ j u e1 e a o l 8 e a o l(150C to 350C) 77leaol15 eaol18 eaue

Total Volume 333 Liters 175 Liters 210 Liters

Pressure drop 8 mm Fl2o0 .5 mm F20 3 m E20

Inadiin othralcarceisisscha testdo

From a first examinitivn of these data for the three

accumulators, one can see some obvious tradeoffs. For exam-

ple the tube parallel flow, while havinig the highest heit

capacity per unit volume, also has the least area for heat

exchange. The volume of encansulation is also low, which

results in a low first cost, but this advantage is offset 1y

the low heat exchange surface area, which results in higher

operating costs due t- greater volumes of air being pumned

for the given amount of beat transfer to take place.

Looking at the data for the three accumulators shows that

the packed bed gives the best compromise between heat exchanre

surface area and volume of encapsulation material. Ancther

factor, which must not be overlooked,is flow distribution.

In the course of the experimental program , it was found that

in the case of the tube cross flow and the tube parallel flow,

all the air flow tended to go down one side of the storage

unit due to an elbow and a broad angled diffuser just upstream

of the storage unit. The packed bed, on the other hand had the

advantage of evenly distributing the flow of air about itself.

VI: Measurement System

The measurement system consisted of two measurement chains.

The more elaborate is the thermocouple temperature measurement

chain. This starts with copper-constantan thermocouples, and

is shown in the diapram below:

The Mek II microprocessor commands the syste, and the

multiplexer scans 48 channels of thermocouple measurements,

which are then digitized by the analog-to-diritil cn-vrtor.

Th hexadecimal numb-r , an avera'e r-f 2 ianurr(-nts tJ'r :n

sequentially in tio,:, is trannf-rr-d from the meror- nf the

'ekTI microroce ssr to the Pet romodore comuter. There

a softwar- prrr'i' in 1, iin, thl ca!flration curv- for th!

thermocot'Ir es m ahes thp final conversion from the milIivolt

reading of the '"i It i rloexer to the te'rature data whic h are

required.

The second .neasurenent chain is the fw .measurement

chain. Thic is rirri' an orifice nlate fiow-meter, with

a water manometer used to measure the nressure drop across it.

For this measurement, a previously obtained calibrattion curve

was used to determine the air mass flow rate.

VII: Exnerimental Program and Resilts

The experimental program was designed to test 2 ideas.

First of all, the validity of the numerical model needed to

be established. Second, the assumptions 7oing into the model

needed to be tested. The most dangerous assumption is that

the flow of heat into the accumulator is evenly distributed

at each point in the cross-section perr wndicular to the flow.

It is a one-dimensional model. In each of the three accum-

ulators tested, the accumulators were instrumented with thermo-

couples in the geometric center of the cross-section perren-

dicular to the flow, and at severalpoints along the lcnrth of

the accumulators. This enabled a visualization of the evolution

of the air and phase change material temperatures along the

length of the bed. At three levels in the accumulators thermo-

couples were put at different points around the cross-section

in order to test the uniformity of air and phase change material

temperatures.

A resum4 of the experiments done follows. The primary

goal was to do one complet charge and one complete discharge

of each of the three accumulators. This was achieved. A lar;-

amount of time went into instrumenting the accumulators, and

then finding the "bugs" in the system and correcting the!.

There were 14 tests altogether, which are tabulated on

the next ,o'e, together with a list of other r,-r.aratory ictiviti-n.

Test

Number )ate Title Comment;, Purrose

2/I/01-3/14/81 Prenaration, calibration of thprrocounles

1 3/14/81 Charge of First charge, comnlete system workecTube CrossFlow unit

2 3/1G/61 Soe leakaFe of liquid P'. occurei;flow distribution problems

3 3/19/bI Discharqe Test terminated before comrleteof TCF unit resolidification had occured

4 3/19/81 Charg7e of Foam was added for better flowTCF unit distribution if diffuser

5 3/24/81 Dischazge 8-hour discharge, still notof TCF unit sufficient to totally resolidify

6 4/(/81 Charge of More liouid PCM leakage noticed;

TCF unit also leakage in ducting and damrerswas discovered

7 4/10/81 Discharge of Comnlete discharge achieved; re-TCF unit solidification in more than 15 hours

8 4/14/81 Charge ofTCF unit

S4/14/81-5/12/81 a. Preparation of thermocouplesfor packed bed and tube parallel flo'-b. Resolution of air leakage problemc. Mixture of PC', for packed bedand tube parallel flow at Solvay

chemical comnanyd: Filling packed bed and tubeparallel flow units with PCM

9 5/12/31 Charge of Sure of leak-proof air system;TCF unit air flow measured downstream of

accumulator, as well as upstream

10 5/13/81 Discharge of Problems with data acquisitionTCF unit during test

11 5/14/81 Charge of Some leakage of PCIPacked bed

12 5/15/81 Discharge of No problems, comrlete discharge

Packed bed

13 5/19/81 Charge of 1o PCI leakage, very long timeTube parallel renuired for PC' to meltflow unit

14 7/1/31 Tube parallel 2 days renuired for ex-erimentflow discharge

17

Graphs of the exnerimental results follow: tney are

groupea in the tree categories: tube cross flow, ricked

bed, and tube parallel flow.

A. Tube Cross Flow:

The tests on the tube cross flow unit served to give: exre-

rience with the experimental apoaratus, but due to the fact

that there was marked leakage in the nase chane material,

they are not conclusive. Additional nhase chan~e materia]

was not available, and developing a system for sealing the

tubes at either end would have constituted a complete Pro.ject

in itself. The problem lay in degradation of the g]ue at

either end of the tubes, where Plastic cans were fixed.

In addition, since the air flow rate going into the accu-

mulator was less than what was measured at the orifice plate

upstream, due to air leaks in the damner system, the convective

heat input was not known with sufficient Precision to warrant

a comnarison with the numerical simulation. A comrarison

of the numerical results and the experimental results for Test

number 1 showed that the phase chan7e in the exreriments proceeded

much more slowly than the computer simulation predicted. This

was because the mass flow rate which was used as input for the

simulation was that which was measured at the orifice plate

upstream of the accumulator, and did not take into account the

air leakage which eccured between the orifice plate and the

accumulator.

These limitations not withstandinr, some meaningful

information can be drawn from the experiments on the tube cross

flow unit. Graphs 1 and 2 show the evolutions of the air anI

phase change material temperaturps. These have the characteristic

plateau of fusion. Also as expected, the air temperature at

each point is higher than the PC'l temperature, until steady state,

when the whole accumulator is at the influent air temperature,

is reached. The time for reachinr steady state, at 600 C, was

7 hours.

The other information that can be drawn reFards the eienness

of the flow distribution, and the thermal Performance as repeated

charges and discharges were Performed. The second charge, as

shown in graphs 3 and 4, shows alr,!ady that the nretty rlateaus

A.____________________________________________

18

of the first charge are less flat than in the first test.

Also the time for charring is about 4 hours, as versus 7

hours in Test 1. This should be due both to the fact that

some PC'I had already leahed, and the fact that the PCM ma, not

have been comnletely solidified when the test was started.

The second condition would mean that some of the latent heat

of fusion remained in the bed at the test's start.

Granhs 5,6, and 7 address the nuestion of air temnerature

distribution about the cross sections of the accumulator.

These show that in each cross section, the temperature in

the center of the tube on the side (curves 8,11, and 13)

were higher than the temperatures at the 2 ends of the central

tube. This can be attributed to a higher heat capacity of

the greater amount of plastic at the ends of the tubes. This

higher heat capacity will result in a lower rate of temperature

increase for a given heat input. In addition the air flow

rate should be somewhat higher at points 8,11, and 13 than at

points 7,9,10,12,14, and 15, due to less resistance to the flow.

The slightly higher flow rate would again result in a higher

rate of temperature increase for the points exposed to the

faster flow.

Graphs 8,9, and 10 from test number 4 show that the sawe

pattern of temperature distributions about the cross-sections

remains, even with the additon of foam to even out the flow

going into the accumulator. Thus one can conclude that the

addition of the foam resulted in additional pressure loss which

was not justified by an improvement in thermal performance.

Tests number 6 and 7 were wade to see the temperature evolution

during a complete discharge. Graphs 11, 12, and 13 show the

inlet and outlet temperatures, their difference in time, and the

integral of the terperatuire difference, respectively. The

corresponding curves for the discharge are shown in graphs 14,

15, and 16. Graph 15 shows the phenomenon of su-ercooling ,

appearing in the dip in the temperature difference and subsequent

rise. Both graphs of the integrals of the temperature differences

should be multiplied by the same mass flow rate to arrive at the

amount of heat accumulated or discharged. rince the interral is

not multiplied by nC in either case, the nur-rital va]uc-

19

can he enmared wi- see thart fo7r th'! - r, it is 7 1 r5

w,-rvas for the s cliar~'e t 3., v~ 1 ii. Ts on - th ird

of the heatt msust be Icc t t hro'iq~. I r rle rar!h 17 sho)'.7 wel I

how thr' P(." tesnrer-itur- fol lnote a; tes-rernat 'r;e drn7 thel

course o r a cha.r~r- or' the storag7'e urit, ardl lrar- 1 ' hosthe fa'ce

s an - 2 c uir ves fo r q ( i s c-ar ,- . r q 1 0 , P20, Q~v n 'P1 s n w

that the "",I tervt irttiirre aro? 7rre or I ess cons tort abou,.t t" ',

cress-sections ,Iur,-n;- the di sr-rrge.

helocal rlin.c--oson o f ri-nercool - r- can -11so b-e seen f ro7

th e o ra'hs . Thi s is a decrease of' the ternprature bePlow the

fusion temneraturc, follocwed3 by a shorr rise to the- fusion ter-

perature as resolidification occurs, and then a decrea-e in tem~-

perature to that of the influent air. The sunercooling7 poses

noroblems for the latent-heat-of-fusion systems because: 1 ) All the

heat is released in a short tir.e when the chrystpallaZat~on occurs,

thus causing diff iculties in con~trollin:g the solidif ication anl

2 ) Duri ng the sunercoolinF-, air is beinr, pumped at an expense

in electric energy, for no benefit in thermal energy transfer.

orarhtTAIR (C)70 .0

* 4

4,4 -t0 4,CNo4

lp44,4,-

to,40 E--

wo Lo

t% wto44,D

4,4,4 t-40 04,4 0JJLo t%

0 *I0 to

4 , .00

4,40tq Cto, L

0 * ,pip, 400,04,4,G

to LUJto t

A,40

040

* u MAS 40.0.5 G/E

TUEC4,4,WCAGIGMDCC262

21

,¢ranh

ATPCM (C)

60.0

50.0-

t'-

- 4.. LUCDC

40.0

00 c

* L40.0 '

A0 o

CL.

44

,i

* c

' ' D T - -8

to

0.0, 7200LO/=.05 1440 0 100.S2EC2 "

4* --"

4

DAE4-4-'

20.0 MAS TINE0. (SEC)E4'A

B C

:- TUEPERTUEA FL!ICARNCIN OFE TC.C,.-IE .

- *. *.. ,- - LU ..

<TAIP(C)-' 00

-4 1" ccIf

1

r.

J

ikl LIL.'IL

/

, /E

/ /

// 0

/ C

-7 EF CP.! CH/41 -4 DD '.L2

I4 *- I'

* "5,*- ,' -- -.

e:. LU

'_',,-- ,,_,- -* * - - - -

F2 1

lranh'

K' f~o

1* -7

77

im iI

qrar

-~ ~ ~ I .0' r~'

*~ Ei

I F E.,

rL E

L -T

25

STA~F rr

if

/O*~- Iifs ipIoWe /o- tII 13 v

T 1 l

IJ

L- .7 . ' rE A

2o

Irraph7

4, pm.TA I R( C7O.0~

Lt~~6 .0 Z

If L .L /

~ / A E: C

,/ /1 T

DAT -,- IJ

1- SS - -

U- E CP z -- 1 ;-f.H C",

*OME

I

TAIR (C)I 65.01

"* 55.0

* 'S. Cz ct7

t4 0

tot050

* :

1 o

E_

LLJ

5.00.

44'S BA E 4- 3-

"f:r, z ,TE FL'S( 05 ECSE

t .- .3

I: I25.

'S LjL

BATEI'24-03-81,.41~~~~~'0 'SMS F00 05 /E

TUB~ES CROSS FLOW, CHARGING MODE. CACL2 61H20

~~ *.

r~ ~ ~ ~ 7 r, n --. ..!

..

TAIR (C)65.0

'44

454'o40 14 5d5.0-

totP (* L-C

4 0,S.

to

I.4 -4

0 *

.44425 4.0

44.4to4

.444

44

0.420014040260 . 90 .

44'4AUR A FUNCTIO OFi~' TIME

.44EPRAUE4FFRITPOIT IDL EE

OAE 24403-8

44 AS LW .5 G7E

TUE ROSFO, HP[G44E AC262

29

.4. TAR (C) ranh 10

65.0

to 0

.~~.35.0 -

'A 1*

-. Q4

to 25.0

'4 LU

05.* 0 u70 u1400260.

t4

V'4

.05 1<tS0'A1A TUEtRS Lb1 HRGN lD.CC262

CL

.0

': 3'

" CC /0Gr -t T

60 .0 , (

44

I-

S LUJ

4g.

* '4

14 L. j

//

I. F.Ei1 R

- 8 1

DE I .05 k, .

. - - " F L41

//

,*1 -...-

'E.T AP ... k'S/

*1 tU CGS L0I AL26f

r" 31

Graph 12DTAIR(C)

4, - 40.0

30.0U-)

to0 .0

- 20.0

L.I

44o..o i

2:: TEllPS (SEC )0 .0 - -- - , i .0.0 3760.0 7320 0 11280.0 1 504. 0,

I DIFFERENCE BETWIEEN [INE ADOULTi_ ITADO7E

I, AIR TEMPERATURES (DTAIR)

• 2 I DATE 9-04-81

ff::lMASS FLOW' .05 KG/SEC.__,TUBES CRO.SS FLOV,' CHARGING HODE. CACL-2-6H-2O

444 . ..4. ...... .. . ..

32

Grarh 13

C-

.51±t1 NhTAFD) CGC

.24E05

.0.

_% I4 IF1" 4 04e-0.

* ,,.0TIM INER7O OFDFEPIEBTEE k

9. 4.

OU. LET A . E R R TDAIR 4DT VRS

Lu/-''DATE .3-04-8

MASS FLW .0- /T-H aROS FL . CHR.r)HOE CA

I., 33

-r TAIRKOC

6 •

15 R

-j..j. 4-j

.2i-

*. . ., ,-, [ RF1

- -0 102 .0 212 010". 31 0500 0B D

I ...-" a

-. - - T :

..V ,, .. 4. :. r,.l' --- :2 - ]"

fV E Z "'

5* .--C -- i, 2 _

.' 5.- - rCsn .--, -.

.I F

C.i4--- FDD-.',!T -,

DTAI (C) Grarnh 15

50.0~

-$ 40.0

30.0-

LU

0 0 20.0-

LU

ow0 WU

* .4 HE( SE

* . p.0

* 0.0 0.00 0 tO 0 2 0L

DIFFERENCE BETW4EEil OUTLET AND INLET

o AIR TEIIPERATURES (UTAIP)

L. DATEz004-

_ AIR FLOW=0 .O'K-d PER HECONO

TU2,ES COSS FLOV 0UVtCHARGF, CACL2-UJ

.3E IQ UNT(DTA IF *UT (C',8EC)

-i

t - 0

-0. F.a

32 "1w05-0 0C )0.

.0 10000 .0 21600 .0 02400 .0 43200 .0500.0 !80.

~~ T I INTECRAT I N OF I FFEFENICE BETVEEN INLET ANDOU TLET- A IR TE HIPEP;"%TUlESK I H'T 1AIHFU'.0T) VE FUS TI i

DATE=O-04-8 I:: w AIR FLOW=0 .05BNG PER SECOND

TLI8EG C08 FL0OV .0 1 SCHARGE. CACL2-6H2'C

.,C.

jt-j

..-.. ~9..

C.. .. J

A9 -- (

F-j ET7

/ .7

-f (C)

*T I

I IL)4

'4MW ..

,. .,.,

* a ai . '

• a ;

I a ... 4.._, , , -

" .: ' 1I" ;"

a .. , [-

V '..,, - -L"--. .. :' 2 _'H ;;- ° , .. .

>4,, 39

Grah,2

TCQ(7C

4

4'

4,

to .0-. E

S.S.,4I TW H C-A

* 240

Grrh21

TtiC'F( C)

1404

,01

o i4.

I44 I-- IiS C 1

~ L 'lTT

B. Packed Ped:

Graphs 22 throurh 27 show the inlet and ou tlet

temperautres, their difference, and the intecral cf that

difference as a function of tire, in the sam° fashion as the

granhs for the ch'arre and discharre for thie turb cross flow.

Here the numerical values )f th- :r.t-r-rin are c!o7er to each

other in the chlarge - nd di, ch-irr e, -it ;.7- X 1" and 14.4P

0, resnectively.

Granhs 28 and ?0 show the air ard I'" te-er..tures at

the top, middle, and bottom of the accumulator-, and. (Irart's

30,31, and 32 show the PC'.! temteratures at cross-sections at

the inlet, middle, and outlet of the accunulator. It is evi-

dent that the temnerature is very unifor . across the cross-

sections, and, as r .ntioned before, this is a marked advan-

taige of the racked bed. In the discharpe of the packed bed

there is again the problem of sunercoolina (Granhs 33-3),

but by havin7 the encansulation small, and thus makinr the

exit from supercooling occur in a more dispersed fashion,

the effect is smooth-d out in the air outlet temneratire

readings. Also, the air temperature is made nore uniform

about the cross-section as the bed is traversed(iraphs 34-3,3),during the charge of the unit.

Granb. 22

600 AIR (C)

-~ 50 .0

L L/

L LJ ? o .

- 4'.

2. D E=ia" - I

L4 L

143

Gran ri31

500ADTAIR (C)

*20,0I

~~0.00. 720. 140-. 10 0280.

DI.E.NC BEa.FN NE NDOTE

* L AI 20.0AE .7 PE EC I

---- PAKDBDC-~kG(,'3MDE A~2621

L L L

;rnh 24

IN'I1DAIR4DT) (C-4,SEC).472E±0t

0 t4

.. 41 9 +0

'N* . to._0.4 CL

.40.

0. 7.0014CO. . 60 2 0.

-A.4 C0TI NERTO FDFFPNEBTEN L7N

OULTTHE.4%4E. N Tr R-,D' VRI U RE

(I4 L 4 L L

S 45

T~ro'rGrivih 25

60 .0

-'-9 .4.450.0

A 9 -A

.4 L 40 .0

Till'

f' 17~ r)j:

E!~ ri FF'E.9

D. E

HAS' TLU R E EICV ~ ~ ~ ~ ~ ~ Jr[ '7[ D!-- 1 %P: i *liIi.~~~"* I*

Graph 26

DTAIR (C)

.4 o

:0.4

40.0

Lu30.0-.

14..,

.--

000

.4

Lua10 .0 " - _,-'

~.4.4

!TIM "(SEC)0 . 0 io

.'4.0 7 ,2,,, .3 14466 .7 2170C0 .0o ,8.9"n.3.,.,,. 466 . 4400.0

44

i 2. DIFFERENCE BETVIEEN OUTLET AND INLET

-"- AIR TEMPERATURES (DTM[R)

.4

~DATE 81-5-15

'"_,_ MASS FLOV' RATE .07 KC;/ , L:C-

1 A4

.4

* 4. (*f

Granh 27

INT(DTAIR4DT) (C'SEC)*:" .448OE+O

' I"

VI'1.

'44

STIE269E E

0. 4

V-, .7E 0 .0 722%.64466. '10C 0 2G9 3 1C)C 7 434

',4

D T ,

too MAS FLWSATE.7 G/E

* I **I* I14e,

' " 6_ L k7 43. k

Ls L

.7 E 0 .0 7 ,.,.64 6 . 1 ,3 . 8#. i1 6 .T IE NEGA[O F 1FEECEBTWE ILT4N

.11 OU L TT4E A U E T H iD . [':D ))V R U ] '4 ., *o-,DAT481 8-14 I

- MASFLWRAE .0")iE::4 *.

PAKE BD ISHRG CCL-64

n l *il I U III4 I

~r-a nh 2F

TI (CD

t4. 60Q-. . . . . . . . . . .

t4/

/ 7P

A T 7,

JL'I

u i

A I LTI , T = I I

ranh 29

'TFCIM (C)

60 *g

Cls 4 7-

I, -

24 'D C

LiA!

Cr) TE 4

J / G P 1*~T 4-. 1:

50c

-,ratd- 3C(

T00' TFCH (C

6" .

o. LiJ

C,

TL I Fr E

ID 30'- I

II-

hI-.

CL.

TFK7 F= 14 ,i(j

e 'I! T- .0 ~

V 7HC

I 'ir-anh 323

- IIV

U/

-

D4 ii T E

AIR4 F!(,E' .0' f PE E'

P~ A/'lD'7Il1! CL C

rah33

40 C TCHC

ftI I E0

40 0

lr*R TI" A U L i-Ci C

* I ,.- -' r'rT

HA S .C!;.' A'.,.0

ftP A C "EE '. * . G

';ranh

605.0

TPCH( C )

60 .0

:IL

"

A~4 .0~

CLL

ID

A711

JJr

" - 40,C0

DAT I ,

1.I "" .. + l 2 _/ j14J6:'C J <

4-. e- --m", L0 , RA E CI 1 . EA,' '!

[ (

P, .I ELI EEL [i I7*P ' I'A7Q-IiZ

*155

tranh .35

TFCI,'1( C)

A.LI

DA T2 r'1r-1ft- FLO ,'P..

PA-TJ - T

Granh 3(;

TAIPK C)

60 .0

3 0.II

.2,

~ Cr7

-~~ 40.

. .. j

.4

L C, P 7-0

'+ Li

A- " i " -~

, i. ...

- , ___..._...Z

*4* " . 1 - - -.---

A..-

• '-' DATE 81 -Yi-1,5

: @ tl~~.",'z7_ FLO., .',",?ITE .07 U,§,,'E7_____P.A.]IE[ EBED U*--C-H.'"_' lr \I '" I--

.. . .....L i " b ... ."..' .2 .

J ranh 37STA IR(

4 t4

.4

*4

44

-AlI

*1 t

A".

4,.',

LU ,.

A E -'5 S ' 1 lJ CJ

CL

P4 E D 11- F

G rah 3T

S A TAP C)60 ,0A:

444

I L2

• # LU c.i ,_ ,;ii/ T

4--. -o 1 .r -,'D 0i _: . ._

o ~~~ A T7 E 81 ,5-1

]M.,,S:,b FLI-O" R, 7TE 0n G , CP' ,E D.E ' ED -0. 1LI3 ITKIK ]lI

59

C. Tube Parallel FIonw:

Here again the inlet and outlet air temneratures in

the center of the accumulator, their difference, an ,! the

intepral of the difference are shown on Glranhs 3 to 4,

This can be comrared with confidence to the data for the

packed bed charge, since we are confident of the mass flo-,:

rates. We determint the averige rate of eneryv storare as

follows: rtPacked Ped: nC "T d t/t

nairJA c harge

=.07 kr/sec X i00o8 Joule/Kf-°c XL72,000 'C-sec123000 see

= 1450 Joule/sec

Tube Parallel flow

.095 Kg/sec X 1008 Joule/Kg-°C

X 883000OC-sec / 70000 sec

= 1200 joule/sec

Thus we can see that the packed bed stores heat at a rate

1.2 times that at which the tube parallel flow stores heat.

Further indications of the thermal nerformance of the

tube parallel flow unit are shown in graphs L and 4-,. These

show temperature evolutions at different heights in the ac-

cumulator. Graphs 4h and 45 show the inlet and outlet air

temperature distributions about the cross-section of the

accumulator. These last 2 praphs show the surprising

result that the air temperatures are close to each other as

the accumulator is entered, and there is a larger spread amonr

them at the exit from the accumulator. This may be due to

irregular melting of the PCM about the accumulator, as well as

motion of the melted PCM from one part of the accumulator to

another. A complete discharge was performed on the tube parallel

flow u~it, and the noteworthy aspect of that test was that it

took almost 48 hours for the resolidification to occur, because

6f an air inlet temperature of 180C.

Qraph 39

ThTA R'65 .0

4--L

A-T I

I* IILE -

Now ...- ~

Grainh hO)

DTAIR (C)50.0

40.0

,21 w0 o

-- U

L3

t* 20.0co

=', TU'E( E

to t - . 7 00LiLI

Lii

10.0-r

0.0 I4400.0 29800.0 43200s .0 57600. 720.

T DIFFERENCE BETW.EEN INLET AND COUTLET

".. --- AIR TEM1PERATURES <DTAIR)

t, J3 DATE:5-19-8 [

p

t, MASS FLOW .0gb KGiSEC

4.

k-a~~ .....'.. Il I.. . ... . ... U E PL F ..... .. i .... ll ...

.88E±C 4 ItT(DTAlIF4.,BT) (C, SFC)

.76E+34-

.530E±0

0to

..

.353E±06-

b 0

t4 ~.1 77F+06

,.0.

00.

00 =TIME INTEGRATION OF DIFFERE1,CE BET\,.EEN INLET ANDi

0 *=OUTLET TEMPERTFETIN(TIRD) EUSIN

DATE=5-19-8 IMA~SS FLOVd .095 KG/SECTUBE-" PARALLEL FLO'd .OCIAARGE. CACL2-SH20

A Tt 1R(

r

I 69.) ~ I

* , -- -

--- 6. .7- . .............

' I ,,* 6 - --i

L.LI!

aT1 1I ,:_

I I F C ; l"17

"" AT i' -1 E 1... D , -- 3,,', F n l-' ,q , 7 :. 0 , , i'""' ,.* ~~~D T' E .51 '3 .... , H- AS .. . . .F. L L c.'..

' " L ,J "[ UE~T E F 'i F : , ,L L E L F L O IT ."C4- H : '

...

aL

I.,'

=, I ~ l .. . . . . . . . . .

ilK

63~~ Q1,- E- T 1l(ZErah

0 7,

Ia U-Cl._l -JF T I

*l TE[I!

D. :E= 5 IS'_ 'R"3 FL I,,

ft L 'L

TO4P7F'T,

Granrh 'I4

~0 .0* .*

L

T- C.7-j, 1

1- '1C-F ,E 1L-jE:

DA,. -*-9 ov L Y=,: '

TaEP4R4LE FT - CL

C. i TVEGBE'

Granh 45

A-I

TE IF" -D

- 3'1.I AS' FL0 I

A-T

As' ;1F1'R LC

VIII: Comnarison of Numerical and Fx jerimental Results

Graphs h6 and 47 show the numerical curves and the

experimental points for the racked bed unit. Graph 4C show5

that for the charge, the numerical simulation rroceeds faster

than the experimentil data. Again for the discharge, in Grarh

47, tie numerical c'rves proceed more abruptly than exreriment.

This difference between the numerical and experimentil results

is probably due to the resistance to heat transfer of the FCM

in the solid phase. such a resistance, neglected in the

numerical simulation, would indeed result in a slower rate of

discharging and charginr. Thus if this added resistance ,ere

taken into account, the experimental and numerical results

would be in better agreement.

TE M PERaTURIF- r,Gc- m C. prh

t- SORTIE EERIM ENTAL

5'- 0 MIL

35--

3000 + + 4- +

+ TrEMPS (S)

2-5" -,Fo l (,,,o'>040 4 'ZO (1

60- TEM PERATURE

DE L'ftIR (C)

55~o rLERE EXPERIMENTAL0 MILIEU _5-0 NUMERIQUE 0 t

3o - + +

TEMPS (S)0 "1 15 ~ 20 (" Ic

TE t'PER,]TURE 7Granh IrT 7

MCP (*c)+ SORTIC EP~RMUA

GO - C~~~ MILIEU S)P-jr~JANUMERIQUE'

300

104-

TEMPS (5)0 1 0

0 410 LO 30 4 0

TEMPE PtITUPE

DE L'F91IRQ"C)

0 MILIEUI- NUMERIcRUR

50-

40

0+0

30 -00+ 0 +0

+

TENP1S Ns

I( I-Icf- 7c 30

TO

IX: Efficiency Considerations

The mechanical energy renuired to store the heat in each

accumulator can be written as 4' X V = Power = . ,h s --as

that the power required to maintain a volume flow of V across a

pressure difference of AP, is enual to the rroduct of the two.

Lookin at this in terms of wor!', it car be said thit W in the

amount of work done in moving a volume of fluid V across a

nressure dror 61. Thus we have the amount of work required for

our accumulators by measuring the pressure drop accross them.

The thermal enerry transfer is ;C . 6T, whereAT is theD, a3r

difference between the inlet and outlet air temperatures, and

r is the mass flow rate. We then find Q = fvC A T. Now ifpwe want a measure of the amount of heat transferred in relation

to the amount of work required to transfer that heat, we can

devide Q by W. This is simply C AT/AP. This then is readilypavailable data from the experiments, and it has been found that

in every case it is on the order of 50 to 100. This high

apparent efficiency though must be considered to be divided

first by 2, since there is a charge and a discharge of the

accumulator in any application where the heat is used. Then it

muss .gain be divided by 3, f'or the overall efficiency of

the clectric system. This brings it down bv almost an order

of magnitude, to 8 to 16 already., down from 50 to 100.

Together with this idea of efficiency, a criterion for

stopping the charge or discharge should be developed. That is,

the W remains more or less constant during the charge and dis-

charge while Q decreases with time due to AT decreasing with

time. Thus the efficiency is decreasing throughout the experi-

ment and after a certain point, oil or electricity could be used

more advantageously than by pumping low temperature air in ani

out of a storage unit.

X: CONCLUSIONS A::D RECO:,MONDATIOU!S

The experimental study shows that from a heat transfer

standpoint the packed bed is the best method of encapsulation.

It has a negligible pressure loss, large heat exchange sur-

face area, a self distribution of the flow and an avoidance

of the problems of supercooling. On the other hand, the geo-

metry which was worst in heat transfer characteristics, the

tube parallel flow, was the only one of the three that had no

PCM leakage problem, while the packed bed,though the containers

were sealed with silicon rubber on the caps, did have leakage

problems. Thus the problem is one of feasibility and cost of

construction as well as heat transfer.

Further studies of such storage units should include an

operation over a broad range of Reynolds numbers, to find the

optimum range for thermo-mechanical efficiency. Also,

experiments should be done which include solar 2ollector w:

inlet temperature varying with time, instead of the fixed

inlet temperature which was used in the past experiments.

72

References:

1. Telkes, IKaria, "Feat of fusion syctems for solar h-atinand coolin-" , Solar IEnrineerinrz, eptembhr, 1977.

2. tGross, IR.J. , et al. "i~u erical Sinulation of Dual-'led iLThermal Energy Ctorare S'ystems", Journal of Solar ener' °yEngineering, Vol. 102, P. 287, ,ovember, 1 .

3. Eissenberg, David, and 7.Wyman, Charles, "What's in "tore PorPhaLse Change?" , Solar Age, 'lay, 1CQ130, P. 13.

h. Riordan, Michael, "Thermal Storage: A Rtsic Guile to the Ptateof the Art", Solar Age, Aril, 1978, P. 10.

5. Telkes, Maria, "Thermal Lner y Storage in Salt Hydrates-,Solar Energy Materials, 2, 1980, n-s. 381-393.

6. Smith, R. N. et al. "Heat Exhcanger Performance in LatentHeat Thermal Energy Storage", Journal of Solar EnergyEnrineering, Vol. 102, lay, 1980, pps. 112-118.

7. questois, Remi,"'Numerical Study of Melting Phenomena inPhase-change material Heat Fxchangers", Senior thesis,Free University of Brussels, Faculty of Engineering, 1981.

8. Theunissen, Paul-Herve,"Studies of Heat Accumula-iors", in-ternal reports of the von Karman Institute for Fluid Dynamics,Rhode-St.-Gen~se, Belgium.

73

Arendix IPhotographs of Exnerirnental Program

74~

if"

C u

0 14

YA-.

'I'L

(As

~~{. L;

Aga:

C.-

rI

Al

Photo 5: Filling elements OfPacked B~ed Unit

- A

Appendix TI

Diagram Of S.U.N. Facility

79

7pe

1N- ---a Pane

2 Leat sqcamuTlator3 Refri -er'int for di sch-.r-, e-xrcr r--rt,,

4 Fl1ectr-.c re.istancc, heater for charg-i n,, exro-rircnts

5Air Blower6Freun-atic Purr

II

' w'