Geysering Effects

7/24/2019 Geysering Effects

1/6

Proceedings NZ GeothermalWorkshop 2002

AREVIEW OF GEYSERING FLOWS

X.LU A.WATSON

Geothermal Institute, The University of Auckland, Auckland,

SUMMARY -Thispaper reviews geysering theories developed by some previous investigators. Thereview covers natural geysers and geysering phenomena in engineering equipment. Important aspects withregard to flow processes are discussed, such as boiling, flashing, geometry, and heat transfer mode alongthe channel. The mechanism of geysering wells is summarized, based on measurements at TeAroha, New Zealand. Thisstudy shows that neither a heat source nor chambers in the flow passage arenecessary to produce a geysering flow.

1. INTRODUCTION

The word geyser is originally derived fromanoldIcelandic verb, which means to erupt. Itrefers to natural phenomena observed ingeothermally active areas in which hot water andsteam intermittently ejects or erupts into theatmosphere. Flows with the same intermittentcharacter are found in some engineeringequipment, and in this paper geysering has a

broad meaning and applies to phenomena thatoccurin nature and in engineering equipment.

Natural geysers are rare and unique phenomenafound where there is surface geothermal activity.Many ofthem are tourism attractions and need to

be preserved. Arecent study showed that geysersmight exist on Saturns giant moon Titan and

Neptunes moon Triton (Lorenz, leading tothe research of geysering becoming morefascinating.

Fluid flow in engineering equipment may bedesigned tobeinthesteadystateorinsomeofuniform transient,howevergeyseringcanoccurin certain conditions and result in problems.Geyseringisinthecategoryofatwo-phaseflowinstability,andexamplesarefoundinthemissileindustry (rocket engines), nuclear industry,

petroleumindustry,and ingeothermalwells. Inrocketenginesforexample,thefuelsupplymust

besteady,andgeyseringinthefuelsupplylineshascausedproblems(Murphy,1965).

Naturalgeysershaveattractedscientificattentionfor about 190years (Rinehart, but inengineering equipmentsuchflowshaveonlybeenstudiedinrecenttimes. Itissignificant thatrecentlytherehas beennocrossfertilizationofideasbetween earth science andengineering.

The aim of this paper is to briefly reviewknowledge about the mechanics of geyseringflows, especiallyin engineering equipment. Some

important points emerge and are discussed. Afterthat, a geysering model isdescribed

the transient two-phaseflowpoint ofview.

2. REVIEW OF LITERATURE ONNATURALGEYSERS

Natural geysers have some comrnoncharacteristics. They all discharge liquid andvapour to the atmosphere to various heights for a

period of time. Then they become quiet for acertain time until the next eruption occurs. This

periodic process a geysering cycle. Eachgeyser has its unique performance. The height oferuption may vary fiom less than a meter to ahundred meters or more. The interval of theeruptions may be a few minutes, days, weeks, oreven months. Some geysers erupt in a singledischarge while others erupt inseveral successivedischarges. The period and activity of a geysermay change withtime for some reason. Generally,the variables responsible for the geysering

characteristicshave not been isolated.

AllenandDay (1935)reviewedtheworkup tothattime. TheyreportedBunsenstheory,whichwasbasedonboilingbeginningapproximatelyatthemiddle depth of the channel.However, histheory did not satisfactorily explain theintermittent nature of the geysers andhow thewaterwas heated to boiling. SomeresearchersincludingAllen andDay(1935)who reviewedmanyearlierpapers,supported a newideathatthe

boilingdidnot takeplaceinthegeyserchannelbutat a lowerplace, wherethetemperaturewas

higher.Thisidealedtotheconceptthattherewasanunderground chamber at thebottom of thechannel,wherewaterwasheatedtoboilingbefore

being discharged and the chamber refilled,periodically.

Many researchers (Allen and day, 1935;Andersonet al., 1978; Rinehart, 1980) accepted the conceptthat it is the boiling of water that drives theeruption ofmost of the geysers. However, someinvestigators thought that gases might play someroles in the eruptions. Rinehart (1980) suggestedthat some gases such as C02 might play an

important role because the behavior ofgeysersdischarging water with dissolved gas wasmarkedly different to that of a steam-activated,hotwater geyser.

267


2/6

Many investigators have conceptualized thesubsurface plumbing system to explain thegeysering process, especially the intermittenteruptions. Allen and Day(1935) concluded thatthere are three essential elements of a geyser: aheat source, a water source and a chamber with avery or tortuous channel above. Based ontemperature-depth curves of some geysers atYellowstone National Park, they thought that heatsource is magmatic and is transported by steam.The water source is supplied by the inflow ofcould water fiom cavities. Theinflowofcold water is not constant but is greatestafter eruption and then decreases the nexteruption.

Steinberg et al. (1981) developed a typicalconceptual model as shown in Fig.1 . Theis connected to the surface by a narrow channel.Two feed points were assumed. One is for thedeep inflow of geothermal water, and another isfor the shallow inflow of cold ground water. Thismodel was used to derive their theory of thegeyser process.

and vapour bubbles rose into the channel withoutcollapsing. It was also shown that intermittentovefflow occurs prior to eruption when boiling

begins and the channel is nearly full.

Steinberg et al. (1981) built more complexmodels, the first using water. It was found that,after eruption, the inflow rate of the cold waterincreased very rapidly and that the majority of thevapour condensed when cold water entered thechamber. As the pressure increased duringchannel filling, the cold water inflow ratedecreased. Continuous heating caused the pressureand temperature in the chamber to increase, andwhen the water became saturated boiling anderuption followed. In their first model the ratio ofchannel length (L) to the channel diameter (D)was = = Their secondmodel was smaller, with of 40and used fieon-113 as the fluid. As aresult of these experiments they suggested thatliquid superheating was a possible importantmechanism for initiating geysering. Their thirdmodel was used to investigate the influence of themechanical pulses on the geysering period. Thedata showed that high mechanical stressingreduced the degree of superheating and causedmore fiequent geyser eruptions.

-.

-

(a) c o l d w a t e r b e a r i n g h o r i z o n ; ( b ) l o wp e r m e a b i l i t y z o n e ; ( c ) in f low o fc o l d

wat er ; ( d ) i n f l o w o f h o t w ater

Fig. 1:Conceptual model of a geyser(from

Steinbergetal.,1981)

Some laboratory models have been built toinvestigate the mechanismofnatural geyser. One

advantage of laboratory models is that the mainparameters such as the pressure and temperaturein the chamber, and the inflow rate of the coldwater can be measured. Representative oflaboratory investigations are those of Allen andDay (1935); Forrester and Thune (1942);Anderson et al. (1978); Steinberg et al.(1981);and Saptadji (1995). Although each model had adifferent configuration and dimensions, all

basically consists-

-

-

-

a chamber placed at the bottoma tube or tube-like channelan inflow of subcooled fluid into the heatingarea and the channela heat sourceat the chamber

Saptadji (1995) carried out an experimental studyin the Geothermal Institute at the UniversityofAuckland. Her laboratory models were based onthe similar ideas as those of Anderson et al.(1978) and Steinberg et al. and revealed

the flow regimes in the channel of the model andhow they changed. In the experiment, thechamber was filled with water and thechannel was filled to a certain level. As thechamber was heated, convection occurred in thechamber and the temperature increased. Severalminutes later, smallbubbles started to form. Athigher bubble generation rates with bigger bubblessomeofthem separated fiom the wall and rose upthe channel. The water level in the channelcontinued to rise at a faster rate resulting anoverflow. After a few minutes, a large bubbleappeared to enter the channel but collapsed at the

bottom of the channel .and only small bubblescontinued rising to the surface. This occurredseveral times, with bubble collapse at increasinglyhigher position in the channel. Finally, vigorous

boiling occurred in the chamber, a large vapourbubble rose into the channel without collapsingfollowed quickly by other large bullet shaped

bubbles (Taylor bubbles) that filled almost thewhole cross sectional area of the channel forcingthe water above out of the channel in an eruption.

Anderson et al.(1978) observed that an eruptionwas initiated once boiling occurred in the chamber

268

I


3/6

3. REVIEW OF ENGINEERINGLITERATUREONGEYSERING

3.1 Geysering in Propellant Feed Systems

Geysering problems have arisen in the design ofpropellant feed systems for liquid fuelled rocketmotors in missiles (Murphy, which

typically use long lines to connect the fuel(propellant) tank to the engine. Since the

propellants are cryogenic they are heated in thefeed line by the atmosphere during missile fueling

before launch. Geysering during thisperiod mayempty the feed line of liquid and the liquid refillcan seriously damage the circuit. The fuel isdangerous if it is allowed to leak. Murphy (1965)carried out an experimental study to investigategeysering in vertical tubes in the form of

(ie a tube with closed lower endopening into a reservoir at the top) with ratiosranging from 1.5 to 30. Murphy found that the

ratio is the most significant parametercontrolling the occurrence of geysering. The heatflux appeared to have minor effect.

Heatflux Coolerfluid fluid

Fig.2: Geyseringinpropellantfeed systems

According to Murphy the mechanism of this geysering can be interpreted as follows(Fig.2). When the heat input through the tubewall, it warms the liquid adjacent to the wall. Thedensity ofthis liquid decreases and the liquid risesupward as seen in(A).At the same time the coolerliquid from the top liquid tank (reservoir)

descending down the center of the tube forming aconvection. The warm liquid rising adjacent to thewall forms a boundary layer, which grows inthickness from the bottom of the tube to the top.After a period of time, the thickness of the

boundary layer is sufficientlyincreased and blocksthe downward flow of the cooler liquid, hencestops the convection. Continuous heating throughthe wall causes a further rise in temperature of the

fluid until it becomes saturated and begins toboil at (B). Bubbles are formedon the tubewall, and then detach and rise upward due to the

buoyancy. They coalesce and form a large bubble

(Taylor bubble) as shown in (C). The formation ofthe bubbles reduces the pressure below themwhere more bubbles form in the saturated liquid.This reaction causes the vapour to formso

269

rapidly and violently that it expels the liquidupward fromthe tube in an eruption.

-.-

3.2Geysering in Nuclear Heating Reactors

In the last decade, some concerns have been raisedabout the possibility of geysering during startup

from low pressure and low flow conditions innatural circulation nuclear reactors. Geyseringcauses safety problems as well as unwantedunsteady operation. Jiang et al. (1995) observedthat, during a geysering cycle, especially when thevapor condenses, an energetic pressure wave iscreated in valves and other components in thesystem. Very strong mechanical vibrations of thewhole system have been observed. Fig.3 showsthe schematic diagram of the geysering in a 5reactor riser. The terminology is confusing. Jianget al. (1995) refer to one of their observed

phenomena as flashing but in comparison to

natural geysering, it would be better calledunstable two-phase flow. What they refer to asflashing instability is what the geothermalcommunity as geysering. Paniagua et al.(1999) investigate a similar phenomenon, and for

both sets of authors this can be summarized asfollows, using Fig 3, which shows a vertical tube(called a riser) with a smaller diameter heatedsection at the bottom and a vertical liquid flowentering. For the tube section shown the ratio of

= 50.

The section shown is part of a liquid circuit.

corresponds to upward flow of liquid.Because of the continuous heating at the bottom,subcooled boiling occurs in the heated section

The generated vapour bubbles flowupward into the tube, which still contains sub-cooledwaterso they condense The fluidtemperature in the riser gradually increases due tothe continuous condensation of the vapour. Thisunstable two-phase flow process finallymakes thefluid temperature at the top of the riser reachsaturation. Flashing occurs at the top part of theriser because the upward flow causes thehydrostatic pressure to decrease.This will result

in a decrease of the hydrostatic pressure below theboiling point and cause the flashing to migratedownward This rapid bubble formationeventually leads to geysering. At the same time,the inletmass flow rate increases significantly dueto the increase ofthe systemdriving head. Asthissubcooled inflow enters the heated section,boilingdisappears and cold water (with certain degree ofsubcooling) enters the riser. Flashing graduallydisappears and only single liquid upward flowexists in the riser after which another similarcyclestarts.


4/6

Inflow

range 90 at bottom to70 at the top.When fully opened, the wells discharge like ageyser. investigation (Michels et al.,1993) showed that the major component of thewater in the well is bicarbonate, hence is themain reason causing geysering.

-.-

These wells offer a very good opportunityto studythe geysering process since they have highratio and simple geometry.

4.DISCUSSION

From the above review of the geyseringmechanism, geysering can be referred to asintermittent boiling or flashing process. Butthisdoes not mean that intermittentboiling or flashing

3.3Gas-driven Geysering Wells

We consider geothermal wells as engineeringequipment here. Rinehart (1980) listed manytypical examples of geysering wells in hisbookand noted that geysering wells discharging at lessthan the boiling point of water were reported incounties such as Iceland, Russia, France, andSlovakia.

The Crystal Geyser at Green River, Utah is onesuch well. It is a gassy cool water geysering well,activated by the evolution ofcarbon dioxide. Thewater temperature in the pipe before eruption is

only about 15 This geyser maintains a fairlyconstant pattern. It ejects a stream of water to50m for about 5 to 10 minutes. Immediately afterthe eruption the water in the well is about 8 m

below the surface, and about 3 hours after theeruption, it overflows with and hissingin the pipe. The overflow increases in vigorreaching its maximum height in only 4 to 5seconds after which the flow stops and the waterlevel declines in the well. Many otheractivated geysering wells are found in USA. Bothof these wells are located in the regions wheresubterranean carbon dioxide and helium exist in

high concentrations.

At TeAroha,North Island of New Zealand, thereare three geysering wells, namely the MokenaGeyser, the Wilson Street Bore, and the DomainTrust Bore. The water temperatures in these wells

4.1HeatTransfer Mode

It is important to point out that in both naturalgeyser models and nuclear reactor models(Fig.3), boiling first occurs at the lower heating

parts. The formed bubbles flow upward, then enterinto the tube-like channel (or tube) and collapsedue to the subcooled liquid in the channel. As the

bubbles condense, heat is transferred to thesurrounding liquid in the channel. Hence theliquid temperature in the channel becomes higherand higher, creating conditions that flashing in thechannel. Bubble rise and condensation is the most

process for transferring heat upwards

because natural convection is inhibitedby theratio and thermal conduction is much too slow.

4.2GeometryandFlow Regime

It is found that not all intermittent boiling orflashing will result in geysering. For a tubechannel, the ratio of the tube length to thediameter is a very important factor asconcluded by Murphy (1965). The review in this

paper shows that geysering has occurred overrations ranging from30 to 110. This is because,when the intermittent boiling or flashing occurs, a

very long and narrow tube is most likely togenerate a slug flow regime with the formation ofTaylorbubbles that expel the overlaying liquid outof the tube.Itis also clear that a chamber is notindispensable to geysering.

270


5/6

5. MECHANISM OFGEYSERINGINWELLS

Measurements on the Wilson Street Bore at TeAroha have been presented elsewhere by

Nurkamal et (2001) and Luand Watson (2001).

In 1999, Iman Nurkamal carried out a well test inWilson Street Bore (WSB) at Te Aroha andsupplied some useful data of thisgeysering well. Based on these data, a conceptualmodel of a geysering well wasgenerated by a detailed interpretation of thegeysering mechanism in WSB (Lu and Watson,2001). The well is 70m deep, is cased todiameter over almost its full depth, and watercontaining dissolved enters close to the

bottom. At high wellheadpressures the well flowssteadily. It is considered that a steady mass flow

rate of solution enters the well from the formationover the full range of possible wellhead pressures,

but that below a certain wellhead pressure theflow becomes transient partway up the well. Thegeysering mechanism can be summarized as seenin Fig.4. State A is thestart of the geysering cyclewhen the water just reaches the top of the well andoverflows. Just before overflow, the solution inthe well is in thermodynamic and chemicalequilibrium. As soon as it overflows, theequilibrium is broken due to the decrease ofhydrostatic pressure that causes flashing - gasto be released from the new solution at depth. As

the bubbles increase in the well, the hydrostatic -pressure decreases at different depths resulting inmore disequilibrium in the solution.This causesmore gas to come out of solution. From Fig.4A to C, the flash point is going deeper, downtowards the bottom.This allows more solution atthe bottom to release gas as bubbles. At Fig,4B the bubble growth is very fast due tothis rapid

a4

FlashPoint .

.

A

chain reaction along the well. The upper part ofthe wellis filled by Taylor bubbles, which ejectwater in front of them producing eruptions at thewellhead. Because the inflow at the bottom of thewell isso slow (about0.025 it cannot supplyenough solution to release gas and maintainthe slug flow in the upper part of the well. Thiscan be seenfrom Fig. 4C to E that the reference

elementRE is below the flash point curve. Thedecrease of gas supply reduces the void fractioninthe well and as a result the water level falls.Meanwhile, the inflow at the bottom contributestothe increase of hydrostaticpressure at every depth,which prevents new gas coming out fiomsolution. Hence, the degassing process stops, withonly the remaining gas in the well causingunsteady upward two-phase flow and making thewater level fluctuate. At Fig, 4D,most of theremaining gas in the well has escaped and the voidfraction is small. From to E, the waterlevel rises at almost constant speed it reaches

the wellhead at E. There is no new gascoming out fiom the solution because thehydrostatic pressure increases at all depths belowwater level. After anothercycle starts.

numerical simulationhasbeen carriedout and the results showed a reasonable cyclicwell discharge supporting this interpretation.Further field measurements are being planned to

justifythis theory.

6.CONCLUSION

The review of the geysering mechanism showsthat the theories developed by the previousinvestigators can not be applied directly to the

geysering wells. It is thesupersaturated solution that causes theintermittent flashing instead of superheatedwater.A heat source is not indispensable in a geyseringwell. It is found that there is a transient two-phase

Water Level

B C D

Figure 4: geysering cycle inWilsonStreetBore

27

E


6/6

flow in the well, and that eruptions occur viaTaylor bubbles. Transient two-phase flow studiesare very importantin understanding the behaviorof gas-driven geysering wells and may bedominant in the study ofboiling water geysers.

7.ACKNOWLEDGEMENTS

The authors would like to acknowledge theassistance of Dr Alexander Gorine, HonoraryResearch Fellow at the Geothermal Institute, forthe discussions ofthis work.

8.REF'ERENCES

Allen, E. T. and Day, A. L. (1935). Hot Springsofthe Yellowstone National ParkInstitution of Washington, PublicationNo. 466,525

Anderson, L.W., Anderegg, J. W., and Lawler, J.

E. (1978) Model Geysers.American Journal ofScience, 278,

Forrester, J. D., and Thune, H. W. (1942). AModel Geyser. Science. Vol. 95,No. 2460,206.

Jiang,S. Y., Yao, M. S., Bo, J. H., and Wu,S. R.(1995). Experimental Simulation Study on Start-up of the 5 Nuclear Heating Reactor.

Nuclear Engineering and Design, Vol. 158, 11123.

Lorenz, R. D. (2002). ThermodynamicsGeysers: Application to Titan. Vol. 156,Issuel, March 2002,

Lu,X. and Watson, A. (2001). Interpretationofthe Measurements of a Geysering Well at Te

Aroha. Proceedings of NZ Engineering and .Technology Hamilton, 133-136.

Michels, D. E., D. andHochstein, M.P. (1993). Dischargeof Fluids at TeAroha Proceedings Geothermal Workshop1993.Auckland,New Zealand, 21-27

Murphy, D. W. (1965). An ExperimentalInvestigation of Geysering in Vertical Tubes.

Advance Cryogenic Engineering, Vol.359.

Nwkamal,I. (1999). The Fluid Mechanics of aGeysering Well in Ta Aroha.Master's Thesis,

Department of Mechanical Engineering,University ofAuckland

I, Lu,X., Watson, A and Gorine, A.V.(2001). Flow measurements in the bore of acarbon-dioxide driven geysering well,Proc.

NZ Geothermal Workshop, Auckland, NewZealand, 189-193,2001

Paniagua, Rohatgi, U. S., Prasad, V. (1999).Modeling of Thermal Hydraulic Instabilities inSingle Heated Loop During StartupTransients.Nuclear Engineering and Design, Vol.193,207-226.

Rinehart, J. S. (1980). Geysers and GeothermalEnergy.Springer-Verlag,New York.

Saptadji, N. M. (1995) Modeling ofGeysers.Thesis, Department of Engineering Science,University of Auckland.

Steinberg, G. S., A. G. and Steinberg,A. S. (1981). Geyser Process: Its Theory,Modeling, and Field Experiment. ModernGeology, 67-86

272

Geysering Effects

Documents

Transcript of Geysering Effects