Inlet Thermal Stratification_flexible Manifold

8/6/2019 Inlet Thermal Stratification_flexible Manifold

1/6

SOLAR 93THE 1993AMERICAN SOLAR ENERGY SOCIETYANNUAL CONFERENCE

Washington, DCApril 22028,1993

Editors:S. M. BurleyM. E. Arden

American Solar Energy SocietyU.S. Section of the International Solar Energy Society2400 Central Avenue, Suite G-lBoulder, CO 80301Printed on recycled paper


2/6

TANK STRATIFICATION WITH A FLEXIBLE MANIFOLDD.E. Adams and J.H. DavidsonSolar Energy Applications LaboratoryColorado State UniversityFort Collins, CO 80523 USA

Use of a flexible, porous manifold to increase the level ofstorage tank stratification in domestic solar water heatingsystems is studied in a 372-liter storage tank. The initialtank temperature profile, inlet temperature, and test durationare varied in three testing schemes. Flow rate is 0.07 l/s.Stratification level is quantified by vertical temperatureprofiles and a new dimensionless mix number based on theenergy in the storage tank weighted by vertical location.The mix number ranges from 0 to 1, with 0 representing aperfectly stratified tank and 1 representing a fully mixedtank. Results show that under operating conditions typicalof direct, constant flow rate solar systems, an orlonmanifold is 48 percent more effective than a conventionaldrop-tube at achieving stratification.

Thermal stratification in solar storage tanks is a criticalfactor in the design of effective water heating systems.Methods of increasing stratification include: operating withflow rates low enough to turn over the tank only once a day(single-pass), isothermal operation, and/or use of astratification enhancing distribution manifold. Manifoldshave the advantage over the other options of providing tankstratification without requiring modifications in systemoperation.Manifolds can be made of either rigid porous tubes orflexible porous fabrics (l-6). Design of these devices isbased on matching the pressure gradients of the manifoldfluid and the tank fluid to prevent inflow or outflow fromthe manifold until the fluid returning Erom the collectorreaches the location in the tank where tank temperatureequals mm fluid temperature. Because rigid manifolds usevertical resistance elements to match pressures, they aredifficult, if not impossible, to design to operate effectively

over a range of temperatures and flow rates. Flexiblemanifolds adapt to different operating conditions becausepressure gradients in the tank and the manifold are matchedcontinuously by variations in the cross-sectional area of themanifold.The purpose of this study is to determine the level of tankstratification that can be maintained in a direct solar systemusing conventional flow rates and a flexible, fabricmanifold. Stratification is characterized by verticaltemperature profiles and a new mixing number based on theheight weighted energy in the tank. Performance of theflexible manifold is compared to that of a conventionaldrop tube inlet.

The experimental facility includes an insulated, plastic 372liter water storage tank (UA = 2.7 W/K), a 310 liter electricwater heater used to simulate collector return water, and acold mains water supply. Tank temperatures are measuredwith 19 T-type thermocouples mounted in a thermocoupletree. Inlet water temperatures are measured with athermocouple inserted in the pipe just upstream of the inlet.A turbine flow meter is used to measure the volumetricflow rate of the water entering the storage tank.The conventional inlet is a vertical 2.54cm diameter PVCtube which delivers water to the top of the tank. Theflexible manifold, shown in Fig. 1, is composed of a knitorlon sleeve clamped to a modified inlet tube. The fabric isalso attached to a weight which rests on the bottom of thestorage tank so that the manifold does not float to the topof the tank as air bubbles attach to the fabric. The verticalmomentum of the incoming fluid is reduced by forcing thewater from holes drilled around the circumference of a,

277


3/6

plugged drop tube. Flow from the tank into the manifold isnot possible because the manifold collapses until thedifference between the tank and manifold pressures is 10.Once the manifold fluid reaches the tank depth at which itsdensity equals the tank fluid density, the fabric expandsallowing flow from the manifold to the tank.

MOMENTUM _DIFFUSER 4.lNLETSECTION73mm DIA

FLEXIBLEMANIFOLD I 1030mm

I Time (minutes) 1o- 10

Tinlet (c)1 w . I5010 -20 40

20-30 3030-40 3040-50 4050-60 5060-70 4070- 0 3080-90 40 L

1163mm 4. RESULTS

Fig. 1. Flexible manifold design3. TESTINGThree testing schemes are used to evaluate the two inletdesigns. In each test, as water enters the storage tank fromthe inlet at the top of the tank, water is drained from thebottom of the tank at the same fixed flow rate (0.07 l/sbased on conventional flow rates of 0.01-0.02 kg/s per m2of collector area). Tank temperature profiles, inlet watertemperature, and flow rate are recorded at l-minuteintervals.In Scheme I, the upper half of the tank is filled with hot(50-55C) water and the bottom half is ftiled with cold (15-20C) water. Water is then delivered to the tank at aconstant intermediate temperature (30C). The length of thetest is 48 minutes which is sufficient time for the coldwater to be removed from the tank (assuming no mixingoccurs). In Schemes II and III, the storage tank is initiallyfilled with 15-20C water and test duration is 90 minutes,the time necessary to turn over the tank. In Scheme II,water is input at 5OC. In Scheme III, temperature of theinlet water is varied every 10 minutes as shown in Table 1.

In a preliminary study of 13 synthetic fabrics, a 7.3 cmdiameter, orlon manifold was determined to be the mosteffective at achieving and maintaining stratification (6). Ingeneral, materials which perform most effectively areloosely-knit synthetic fabrics which stretch easily in onedirection and maintain physical integrity even after longexposure to high temperature water (7,8). Stratificationlevels in the storage tank equipped with this best flexiblemanifold are compared to stratification levels in the sametank equipped with a conventional drop-tube..4.1-e Profw

Normalized tank temperature profiles obtained at 8-minuteintervals during Scheme I tests are plotted in Figs. 2(a) and(b), for the conventional inlet and flexible manifold,respectively. The ordinate is the normalized tank height,the distance from the bottom of the tank (Y) divided by thetotal tank height (H). The abscissa is the normalized tanktemperature defined as the local temperature (T) minus theinitial minimum tank temperature (T,) divided by themaximum initial tank temperature difference (Th-Tc). Thethick solid line (final-str) represents the theoreticaltemperature profile that would exist at the end of the test ifno mixing occurred. This ideal case is numericallypredicted using a plug flow model with no mixing. Thesimulated tank is initially made up of isothermal disks ofvolume and temperature consistent with the experimentalconditions. Losses to the surroundings are taken intoaccount. The thinner solid line (final-mix) is determinedtheoretically by assuming that any time water enters thetank, the entire tank mixes completely. The mass weightedaverage temperature of the experimental tank at thebeginning of a test is used as the initial condition for themixed tank model.

278


4/6

0.81

0.0 O-2 O-4 0.6 0.8 1.0T - T,

TEMPERATURE -Th-Tc60

0.0 0.2 0.4 0.6 0.8 1.0T - T,

TEMPERATURE -Th-Tc

cb)Fig. 2. Scheme I tank temperature profiles(a) conventional inlet (b) flexible manifold( E -0 min., n -8 min., n -16 min.,

A -24 min., n -32 min., z -4Omin.,* -fmal(exp), - -final&r), - -fi.nal(mix)).

As shown in Fig. 2(a), in the ideally stratified tank, waterin the upper half of the tank remains at the initialtemperature (minus losses to the surroundings), and waterin the lower half of the tank is at the temperature of theincoming fluid. In both the simulated fully-mixed tank

1.0

0.8*Ix5 0.62E! 0.4

0.2

1.c

0.8*lx

k 0.6CDEix 0.4

0.2

0.01 3 O-2 0.4 0.6 0.8 1.0TEMPERATURE (C)

0.6 0.8TEMPERATURE (C)

(a)

(b)Fig. 3. Scheme II tank temperature profiles (a)conventional inlet (b) flexible manifold.( a -0 min., + -10 min., n 20 min.,

4 -30 min., I 40 min., 0 -50 min.,A -60 min., 4 -70 min., I-80 min.,

m -fmal(exp), - -final(str), - -final(and the conventional tank, the tank is isothermal at the endof the test; however, since in the actual tank, mixing occursonly in the top half of the tank, total energy stored in theconventional tank is greater than that predicted for a fully-mixed tank. This result points out the fallacy of basinglevel of mixing on only the slope of the temperatureprofile. As shown in Fig. 2(b), use of the flexible

279


5/6

manifold significantly reduces mixing. At the end of thetest, water temperatures in the upper half of the tank areonly slightly lower than those predicted by the stratifiedtank model and in the lower half of the tank, measuredwater temperatures are only slightly greater than in an idealtank.Tank temperature profiles for Scheme II tests are plotted inFig. 3. Temperature is plotted at lo-minute intervals as afunction of normalized vertical position. In both these testsand Scheme III tests, temperatures are not normalized sincethere are not constant hot and cold bounding tanktemperatures. Inspection of the tank temperature profileafter the first 10 minutes reveals that the temperature at thebottom of the tank is increased when using theconventional inlet. This temperature rise indicates thatsome mixing occurs throughout the entire tank. Duringthis same time, the flexible manifold restricts mixing to thetop third of the tank.Tank temperature profiles obtained under Scheme III areplotted in Fig. 4. As in Scheme II, mixing occursthroughout the conventional tank after only 10 minutes andat the end of the 90 minute test, the tank is nearlyisothermal. In contrast, as shown in Fig. 4(b), use of theflexible manifold restricts mixing and in the lower portionof the tank, the final temperature profile is nearly identicalto that predicted for a stratified tank. In the upper part ofthe tank, measured temperatures are as much as 10C lessthan in the ideal case, but are significantly higher than inthe conventional tank.

A new quantitative measure of tank stratification is basedon the energy in the storage tank weighted by verticallocation. The mix number is,

MIX#=M*trMUP)(MdrLX) (1)where M is the first moment of energy given by,

for a tank of height H, with n isothermal nodes. Thedistance measured from the bottom of the tank to the centerof node i is yi, and Et = pIc,,,V,T,. hemix number hasa value of zero for a tank with a measured moment ofenergy (Mexp) equal to that predicted by the fully stratifiedtank model (M&. Mix number equals one if theexperimental moment of energy equals the moment ofenergy predicted by the fully mixed tank model (Mmix).

,0.6

0.4=

0.2'

oo-l 0.0 o-2 o-4 0.6 0.8 1.0

TEMPERATURE (C)GO

l.Ol-J--0.84Xc -(.6

-2 O-4 0.6 0.8 1.0TEMPERATURE (C)

0Fig. 4. Scheme III tank temperature profiles(a) conventional inlet (b) flexible manifold( E -0 min., + -10 min., n 20 min.,

0 -30 min., I-40 min., Cl 50 min.,A -60 min., A -70 in., I -80 min.,

* -fmal(exp), I -final@), - -fmal(mix)).

Mix numbers for the conventional drop-tube inlet and theflexible manifold are compared in Table 2. As expectedfrom tank temperature profiles, the mix number associatedwith the flexible manifold is much less (closer to theperfectly stratified value of zero) than the mix numbercalculated for the tank using the conventional drop-tube.

280


6/6

The use of the flexible manifold improves stratificationunder each of the three testing schemes compared to theconventional drop tube inlet. Under Scheme III, with therealistic conditions of variable inlet temperature associatedwith variable insolation for constant flow rate systems, theflexible manifold reduces mixing by 48 percent compared tothe conventional inlet.

Thesis, Civil Engineering, Colorado State University,November 1992.(7) DuPont, Comparative Heat Resistance of Fibers,Technical Bulletin X-56,1956.(8) DuPont, Properties of DuPont Industrial FilamentYarns, Technical Bulletin X-272, 1988.

2 MIX NUMB=Inlet Type Scheme Scheme SchemeI II m

Conventional ,620 .556 ,737Drop-TubeFlexible ,161 ,401 .3831 Manifold 1 I I 1

A new mix number based on height weighted energy givesan accurate indication of thermal stratification in solarstorage tanks. Mix numbers obtained in a 372-liter tankindicate that a knit orlon flexible manifold is 48 percentmore effective than a conventional drop-tube at achievingtank stratification.

6. REFERENCES(1) Loehrke, R-I., Holtzer, J.C., and Gari, H-N.,Stratification Enhancement in Liquid Thermal StorageTanks, Journal of Energy, Vol. 3, pp. 129-130, 1979.(2) Sharp, M-K., Loehrke, R-I., Stratified Thermal Storagein Residential Solar Energy Applications, Journal ofEnergy, Vol. 3, No. 2, pp. 106-113, 1979.(3) Gari, H.N., and Loehrke, R-I., A Controlled BuoyantJet for Enhancing Stratification in a Liquid Storage Tank,.Journal of Fluids Engineering, Vol. 104, pp. 475-481,1982.(4) Fanney, A.H., Klein, S-A., Thermal PerformanceComparisons for Solar Hot Water Systems Subjected toVarious Collector and Heat Exchanger Flow Rates, uEnergy, Vol. 40, pp. 1-12, 1988.(5) Davidson, J-H., Carlson, W-T., Duff, W-S., Impact ofComponent Selection and Operation on Thermal Ratings ofDrain-Back Solar Water Heaters, Journal of Smmineerine, Vol. 114, No. 4, pp. 219-226, 1992.(6) Adams, D.E., Design of a Flexible StratificationManifold for Solar Water Heating Systems, Masters

281

Documents

Inlet Thermal Stratification_flexible Manifold