Click here to load reader

Prototype of a thermally driven heat pump based on integrated Organic Rankine Cycles (ORC)

  • View
    220

  • Download
    3

Embed Size (px)

Text of Prototype of a thermally driven heat pump based on integrated Organic Rankine Cycles (ORC)

  • mtzer

    s wthe se se, fRC

    ery ines aboue to prs. Betthat a

    an oil-free high speed dynamic compressor-expander assemblyrotating on refrigerant vapor bearings. Unfortunately earlierattempts of the latter failed because of the lack of appropriatematerials and because of the problem of the depletion of the ozonelinked to the CFC refrigerants, which were, at the time, the onlynon-ammable and non toxic refrigerant candidates for such high

    technical feasibility of a thermally driven heat pump based ondouble Rankine cycle using a high speed oil-free Compressor-Turbine Unit (CTU).

    2. ORCeORC high speed concept

    An ORCeORC system includes an ORC power cycle drivinga reversed Rankine heat pump cycle (see Fig. 1). The condenser andthe subcooler, if introduced, are common to both cycles. This

    * Corresponding author. Tel.: 41 21 693 67 29; fax: 41 21 693 35 02.

    Contents lists available at

    Ener

    elsevier .com/locate/energy

    Energy 41 (2012) 10e17E-mail address: [email protected] (J. Demierre).upgrade of the renewable heat from the environment. Thermallydriven heat pumps can use various fuels including wood pellets ornatural gas and are usually based on absorption heat pump cyclesor on a combination of a power cycle driving a compression heatpump cycle or a combination of both. One concept of the secondtype has been proposed by Strong [1] and is based on the use of anORC power cycle driving a reversed ORC heat pump cycle, bothusing the same uid. Such a concept can be made using a scrollexpander and a scroll compressor usually lubricated with oil or by

    one concept of miniature high speed centrifugal compressordirectly driven by a high speed electric motor rotating on refrig-erant gas bearings [2,3] opens the way to such devices with orwithout electric motor.

    The aim of this paper is to present the development of a ther-mally driven heat pump prototype based on a double ORC. Adescription of the principle is given at Section 2. The applicationthat is studied in this work is residential heating for small buildings.The goal of developing such a prototype is to demonstrate theSupercritical evaporatorGas bearingsResidential heatingOil-free

    1. Introduction

    Single combustion in boilers is a vThe higher cost of fuels and concernput more and more political pressurfor most building heating purposelinked to the use of heat pumps0360-5442/$ e see front matter 2011 Elsevier Ltd.doi:10.1016/j.energy.2011.08.049obtained with an in-house supercritical evaporator simulation program and measurements made on theDTC is presented. The design steps of the compressor-turbine are briey presented. The compressor-turbine unit has been balanced and tested, with air, at speeds up to 140,000 rpm.

    2011 Elsevier Ltd. All rights reserved.

    fcient heating process.t pollution are likely toevent the use of boilerster alternatives are allllow the recovery and

    temperature cycles. Progress in materials and the emergence ofnew uids, to substitute the CFCs with a reasonably high temper-ature chemical stability and/or acceptance, allow the reconsidera-tion of these ORCeORC concepts. At present, HFC-134a or R600 arekey working uid candidates, together with new low GlobalWarming Potential (GWP) refrigerants being in developmentwithin the same range of pressures. The recent demonstration ofThermally driven heat pumpORCheat exchanger for the supercritical evaporation is the double tube coil (DTC). A rst experimental setuphas been built to test the pump and the supercritical evaporator. A comparison between the resultsKeywords:system are the compressor-turbine unit, the supercritical evaporator and the pump. The selected type ofPrototype of a thermally driven heat puCycles (ORC)

    J. Demierre*, S. Henchoz, D. FavratEcole Polytechnique Fdrale de Lausanne, LENI-IGM-STI, Station 9, 1015 Lausanne, Swi

    a r t i c l e i n f o

    Article history:Received 7 October 2010Received in revised form24 August 2011Accepted 28 August 2011Available online 25 September 2011

    a b s t r a c t

    The concept studied in thiheat at the condenser) andcycle, both cycles using thdirectly coupled on the samadvantage of being oil-fredevelopment of an ORCeO

    journal homepage: www.All rights reserved.p based on integrated Organic Rankine

    land

    ork is a low power ORCeORC heat pump system (providing about 20 kWat is composed of an ORC power cycle driving a reversed ORC heat pumpame uid. The centrifugal compressor and the radial in-ow turbine arehaft rotating on self-acting refrigerant vapor bearings. The system has theully hermetic and with low maintenance costs. The paper presents theprototype, with HFC-134a as working uid. The main critical parts of theSciVerse ScienceDirect

    gy

  • Nomenclature

    Latin lettersDh0i fuel lower heating value (J/kg)_E mechanical power (W)_Q heat rate (W)_m mass ow rate (kg/s)

    Nu hfl, Nusselt number ()

    Pr cpml, Prandtl number ()

    Ref Cfm=r

    , Reynolds number based on ()A heat transfer area (m2)

    l thermal conductivity (W/(m K))m dynamic viscosity (Pa s)

    4cross sectional areaperimeter

    , hydraulic diameter (m)

    r density (kg/m3)

    Indices1 condenser (PHX) outlet, refrigerant2 evaporator (DTC) inlet, refrigerant3 evaporator (DTC) outlet, refrigerant4 condenser (PHX) inlet, refrigerantC1 condenser (PHX) inlet, cooling waterC2 condenser (PHX) outlet, cooling waterF fuelfumes fumes resulting from the combustion

    J. Demierre et al. / Energy 41 (2012) 10e17 11C ow velocity (m/s)cp isobaric specic heat (J/(kg K))COP Coefcient of Performance, ()

    f dPdL

    f

    12rC2

    , friction factor ()

    h heat transfer coefcient (W/(m2 K))system works between three main temperature levels and istherefore similar to an absorption heat pump. The thermally drivenheat pump dealt with in this paper uses a supercritical evaporatorat the high temperature of the topping ORC and a geothermal probeat the low temperature evaporator of the heat pump cycle. Heat issupplied to the house heating system at medium temperature bya condenser. The studied concept is a relatively low heat rate

    k roughness (m)L tube length (m)nsc compressor specic speed ()nst turbine specic speed ()P pressure (Pa)T temperature (C)U overall heat transfer coefcient (W/(m2 K))u standard uncertainty

    Greek lettersU error on the overall heat transfer coefcient between

    simulation and measurements ()hc compressor isentropic efciency ()ht turbine isentropic efciency ()

    condenser

    supercritical evaporator

    evaporator

    compressoturbine uni(CTU)

    Fig. 1. Schematic owsheet and T-s diagramH1 evaporator (DTC) inlet, hot thermal oilH2 evaporator (DTC) outlet, hot thermal oillm logarithmic meanmeas measuredpump power cycle pumpR refrigerantsimu predicted by simulationW watersystem (about 20 kW thermal at the condenser in the case treatedhere) with a one-stage centrifugal compressor and a one-stageradial in-ow turbine. The compressor and the turbine aredirectly coupled on the same shaft rotating on self-acting refrig-erant vapor bearings. This allows the system to be oil-free, fullyhermetic and with low maintenance costs in spite of the morecomplex circuitry. Because of the characteristics of dynamic

    Superscipts the entity is received by the system from the outsidee the entity is delivered by the system to the outside

    AcronymsCTU Compressor-Turbine UnitDTC Double Tube Coil heat exchangerEU European UnionGWP Global Warming PotentialORC Organic Rankine CyclePHX Plate Heat eXchanger

    r-t

    Entropy

    Tem

    pera

    ture

    of a simple ORCeORC heat pump unit.

  • is the heat supplied by the fumes resulting from the methanecombustion and _E

    pump is the electrical power consumed by the

    4. Prototype layout

    The layout of the prototype is shown in Fig. 4. For this rstprototype, instead of having a unique condenser for both cycles, itwas decided to have one condenser for each cycle for an easiercontrol of the system. This gives the possibility to test the turbineand the compressor more independently, since, with this layout,the turbine outlet pressure and the compressor outlet pressure canbe different. Two on/off valves enable the connexion of bothcondensers whenwewish to simulate a unique condenser. The CTUhousing is connected to the low pressure of the heat pump cycleand a valve enables to regulate the pressure in the gas bearings. Theturbine bypass and the valves at the turbine inlet and outlet allowto start and heat up the cycle without having the working uid toow through the turbine. This allows to start the turbine only whenthe refrigerant vapor at the inlet is entirely dry. In fact, at highspeed, any droplet would damage the CTU rotor. A vertical tubewith a large section (about 150 mm diameter) after the heat pumpcycle evaporator acts as a separator to avoid droplets at the inlet of

    nergORCeORC pump (which is the only electrical power consumed bythe ORCeORC). _E

    pump is divided by 0.56 to take into account for the

    efciency of a modern combined cycle power plant that producesthis electrical power, in order to dene a COPwhich is strictly basedon the conversion of the fuel (here, methane) into heat. For a rstcompressors and turbines a low density uid is preferred andrefrigerant HFC-134a has been selected for the rst prototype. It ischemically stable at relatively high temperatures (at least up to180 C).

    3. Preliminary design

    A preliminary designwas made based on the results of previousstudies [4,5]. An ORCeORC system design and optimization toolhad been developed. This computer tool consists of a modeldeveloped on a commercial owsheeting software, Belsim-Vali [6]that is linked to an in-house energy integration tool [7] and anin-house multiobjective optimization tool [8]. The three pieces ofsoftware are linked using an interface, OSMOSE, developed in ourlaboratory.

    The ORCeORC system design and optimization tool enables thecalculation of the optimal pressure and temperature levels, of theoptimal refrigerant mass ow rates, as well as of the optimalcompressor and turbine preliminary design (rotational speed anddiameter of the wheels). A polynomial approximation of thecorrelation of Rohlik (1968) [9] is used to model the turbine isen-tropic efciency ht. This relation gives the maximum efciency thatcan be achieved for a xed turbine specic speed nst. Thecompressor isentropic efciency hc is modeled using a similarmethod. A correlation (adapted from Balje 1981 [9,10]) that relatesthe maximum efciency to the compressor specic speed nsc isapproximated. Pressure drops in the heat exchangers and pipes,and heat losses are neglected. The expansion in the valve isconsidered to be isenthalpic. The pump isentropic efciency isconsidered to be constant and equal to 0.5. The losses related tothe compressor-turbine shaft are calculated using the model givenin [2,3].

    The ORCeORC heat pump studied in this work is a system forresidential heating purpose. The system has to be designed toproduce hot water at 60 C (water initial temperature is 10 C) andto heat up water for oor heating from 30 C to 35 C. The totalheating power (hot water production oor heating) is about20 kW. The hot heat source is the combustion gases resulting froma stoichiometric combustion of methane that are cooled down tothe lowest temperature as possible (for example to 4 C in wintertime of our calculations). The cold heat source is the brine froma geothermal probe (that is cooled down from 4 C to 0 C in ourcalculations).

    A two-objective optimization is done, using the conditions givenabove, with the system COP (coefcient of performance) as the rstobjective and the compressor-turbine unit (CTU) rotational speedas the second objective. The CTU rotational speed is chosen as thesecond objective, because it becomes critically high for such lowpower compressor and turbine. The COP of the system is dened asfollows:

    COP _QW

    _Qfumes

    _Epump=0:56

    (1)

    _QW is the heat supplied to thewater (hot water and heating), _Q

    fumes

    J. Demierre et al. / E12approximation, _Qfumes is evaluated as follows [11]:_Qfumesy _mFDh

    0i (2)

    _mF is the mass ow of the fuel (methane) and Dh0i is the lowerheating value of the fuel (based on the standard state,P0 1.01325 bar and T0 25 C). The Pareto curve resulting fromthe optimization, using the system COP and the rotational speedof the CTU as objectives, is shown in Fig. 2. Each point corre-sponds to a particular design solution. It appears that the COPincreases with the CTU rotational speed to reach a maximumvalue of about 1.7 at a speed of 250,000 rpm. In theory, theoptimal compressor and turbine rotational speed is even higherthan 250,000 rpm, but the losses linked to the shaft (gas bearinglosses and windage losses) increase with the increase in rota-tional speed. This result shows the importance of using a bearingtechnology that enables to reach high speeds and so, justies theuse of gas bearings.

    The optimal values of some main design parameters as a func-tion of the COP are plotted in Fig. 3. It appears that the optimaltemperature, of almost all solutions, is about 180 C, that corre-sponds to the upper limit set in our calculations to ensure anadequate chemical stability of HFC-134a. Following those results,the values of the preliminary design parameters for the ORCeORCprototype were determined as given in the Table 1.1.4 1.6 1.80.5

    1

    1.5

    2

    2.5x 105

    COP []

    CTU

    rota

    tiona

    l spe

    ed [r

    pm]

    Fig. 2. Pareto curve of the optimization with the COP and the CTU rotational speed asobjectives.

    y 41 (2012) 10e17the compressor.

  • 0.06

    0.08el

    dia

    met

    er [m

    ]turbinecompressor

    J. Demierre et al. / Energy 41 (2012) 10e17 131.4 1.6 1.8

    0.02

    0.04

    COP []

    whe

    65

    70

    [bar]5. Selected equipment

    5.1. Supercritical evaporator

    The supercritical evaporator has to work under relatively severeconditions. In fact, the supercritical evaporator has to be designedto reach refrigerant pressures up to 70 bar, to heat up the refrig-erant, in the worst case, from 15 C to 180 C and to work witha temperature of thermal oil (Syltherm 800) up to about 200 C.

    (see Fig. 5). A DTC heat exchanger is a spiral tube-in-tube heat ex-changer allowing a counter-current operation with high tempera-

    1.4 1.6 1.840

    45

    50

    55

    60

    COP []

    supe

    rcrit

    ical

    eva

    p. p

    ress

    ure

    1.2 1.4 1.6 1.8160

    165

    170

    175

    180

    COP []

    turb

    ine

    inle

    t tem

    pera

    ture

    [C]

    Fig. 3. Design parameters regarding the COP.

    Table 1Values of the preliminary design parameters.

    Supercritical evaporation pressure 65 barTurbine inlet temperature 180 CCondensation temperature 40 CHeat pump evaporation temperature 5 CTurbine wheel diameter 18 mmCompressor wheel diameter 20 mmture differences and high pressures.

    5.2. Pump of the power cycleSeveral types of heat exchanger have been reviewed (detailsare given in [12]). The selected type is the double tube coil (DTC)

    Fig. 4. ORCeORC prototype layout.The pump has to operate with a pressure difference going up to55 bar. Another constraint is to be oil-free, therefore withoutcontact between the liquid refrigerant and the lubricating oil.Consequently, a diaphragm pump has been selected.

    Fig. 5. Double tube coil (DTC) heat exchanger.

  • 6. First test rig

    A rst test rig was built to test the critical tests equipment (thesupercritical evaporator and the diaphragm pump) before theconstruction of the complete prototype. Several tests were per-formed to characterize the diaphragm pump and the DTC heatexchanger in supercritical conditions.

    6.1. Layout

    The rst test rig approximately corresponds to the power cyclewithout the turbine (see Fig. 6). The turbine is replaced by amanualregulating valve (noted V in Fig. 6). R134a is evaporated in the DTCwith the thermal oil (Syltherm) coming from a regulated electricboiler. The refrigerant condenses in the plate heat exchanger (PHX)with cold water at about 9 C at the inlet. The diaphragm pump (P)is connected directly at the outlet of the condenser (plate heatexchanger) with a check valve (c) downstream. The mass ow of

    relief valve (s1) is connected from the high pressure side to the low

    J. Demierre et al. / Energ14pressure side of the cycle. It opens when the high pressure exceeds70 bar. Another pressure relief valve (s2) opens when the lowpressure exceeds 17 bar. The pressure limits of 70 bar and 17 barcorrespond to the maximum allowable operating pressures,respectively, at the outlet and inlet of the diaphragm pump. Apressure gauge (PVin) is placed downstream of the DTC to check thelevel of the high pressure when the manual regulating valve (V) isactivated.

    6.2. Test results

    6.2.1. Supercritical evaporatorThe tested DTC has the following specications:

    Heat transfer area: 0.23 m2 inner tube outer diameter: 1/2 inch inner tube wall thickness: 0.049 inchthe refrigerant is measured with a Coriolis ow meter (M). Thetemperature of R134a is measured with thermocouples (T1, T2, T3and T4) at the inlet and outlet of each heat exchanger. The refrig-erant pressure is measured at three points with piezoresistivepressure sensors (P1, P2 and P3). Thewater temperature is measuredwith thermocouples (TC1 and TC2) at inlet and outlet of thecondenser and the thermal oil temperature is also measured withthermocouples (TH1 and TH2) at each side of the DTC. A pressureFig. 6. DTC heat exchanger test rig (ORC in which the turbine is replaced by thevalve V). outer tube outer diameter: 1 inch outer tube wall thickness: 0.083 inch material: stainless steel estimated tube roughness k : 50 mm tube thermal conductivity l: 15.2 W/(mK)

    Tests were performed with different values of R134a mass ow_mR, thermal oil temperature at the DTC inlet TH1 and R134a pres-sure at the DTC inlet P2. The tested values were approximately:

    _mR(kg/s): 0.05, 0.08, 0.1, 0.13 TH1 (C): 50, 120, 180, 220 P2 (bar): 45, 55, 65

    The volume ow rate of thermal oil is approximately the samefor all tests at about 4m3/h. The R134a temperature at the DTC inletT2 varies between 15 C and 50 C depending on the test points. Bycombining the different values given to _mR, TH1 and P2, measure-ments were done for 41 operating points.

    The heat rate measured varies between 0.5 kW and 30 kW. Theoverall heat transfer coefcient U calculated on the basis of themeasurements varies between 360 W/(m2K) and 1020 W/(m2K).The overall heat transfer coefcient is evaluated as follows:

    U _QR

    A$DTlm(3)

    where _QR is the heat rate given to the refrigerant, A is the heattransfer area of the DTC and DTlm is the log-mean temperaturedifference. The log-mean temperature difference is calculated asfollows:

    DTlm DTH2 DTH1lnDTH2=DTH1

    (4)where

    DTH1 TH1 T3 and DTH2 TH2 T2 (5)The experimental results have been used to validate the in-houseDTC supercritical evaporator simulation tool [12] that was usedfor the design of the supercritical evaporation unit of the ORCeORCprototype. The heat transfer correlations implemented in this toolare correlations commonly used for single-phase ows, howeverthe authors have not found in the literature any study of theirvalidity when applied to supercritical evaporation of R134a. Severalcorrelations for the evaluation of the Nusselt number Nu and thefriction factor f are implemented in the simulation tool (see Table 2for Nu and Table 3 for f). Five heat transfer correlations have beenconsidered:

    Gnielinski with DarcyeWiesbach friction factor (G DeW) Gnielinski with Gnielinski friction factor (G G) Gnielinski with Filonenko friction factor (G F) PetukhoveKirillovePopov with Filonenko friction factor(PeKeP F)

    DittuseBoelter (DeB)

    The 41 operating points, that have been tested, were simulated.The simulations for each operating point were done for all possiblecombinations of heat transfer correlations (25 combinations). Theerror on the overall heat transfer coefcient U is calculated asfollows:

    Usimu Umeas

    y 41 (2012) 10e17U Umeas (6)

  • shows the calculated relative uncertainties which, for most oper-ating points are less than 1%. However in same cases, the relativeuncertainty exceeds 5%.

    6.2.2. Pump of the power cycleThe required rise in pressure (about 55 bar) has been reached

    with the diaphragm pump. However, the pump nominal mass owis slightly too high compared to the requiredmass ow of R134a forour application. Since it is a volumetric pump, the delivered mass

    4 PeKePeF G DeW 6.285 DeB G DeW 3.2

    Table 2Heat transfer correlations implemented in the DTC model (if is given in the Table 3).

    Correlation Range of application

    Gnielinski

    Nu f8Ref 1000Pr

    1 12:7f8

    rPr2=3 1

    2300 < Ref < 5 1060.5 < Pr < 2000

    Petukhov-Kirillov-Popov

    Nu f8RefPr

    1:07 12:7f8

    rPr2=3 1

    104 < Ref < 5 1060.5 < Pr < 2000

    Dittus-BoelterNu 0:023Re4=5f Prn Ref > 104with n 0.4 for heating 0.7 Pr 160and n 0.3 for cooling L/f > 10

    J. Demierre et al. / Energwhere Usimu and Umeas are the overall heat transfer coefcientscomparing simulation andmeasurements. The overall heat transfercoefcient error U averaged on the 41 tested operating points isgiven at Table 4 for each combination of heat transfer correlations.The average value of U is also shown in Fig. 7, as well as theminimum and maximum. The results show that the best predictionis obtained when the correlation of DittuseBoelter is used for bothR134a and the Syltherm thermal oil. The average of the error U isless than 2% and the maximum error for the 41 tested operatingpoints is less than 3%. This means that the DTC model with theDittuseBoelter correlation is accurate enough to design thesupercritical evaporator for such application. Following the results,it was decided to use three DTC heat exchangers (in parallel), witha heat transfer area of 0.23 m2 each, for the supercritical evapora-tion of R134a in the ORCeORC prototype.

    An uncertainty analysis is performed for the measured overallheat transfer coefcient Umeas. The standard uncertainty for thedifferent measurements is:

    Temperature (thermocouples type K): uT 0.2 K Pressure (piezo-resistive): uP 0.2 bar Mass ow rate (Coriolis force): u _m 0:2% _m

    The uncertainty of the overall heat transfer coefcient is calcu-lated using the law of propagation of uncertainties as described in[13]. The overall heat transfer coefcient standard uncertainty uU is

    expressed, in our case:

    Table 3Friction factor correlations implemented in the DTC model.

    Correlation Application

    Gnielinskif 0:79lnRef 1:642 smooth pipeFilonenkof 1:82log10Ref 1:642 smooth pipeDarcyeWiesbach(Churchill empirical formula)

    f 8 8Ref

    12 1A B3=2

    1=12where

    A 2:457$ln 1 7Ref0:9 0:27$k

    f

    16B 370 530

    Ref16

    rough pipeu2U vUvT2

    2u2T

    vUvT3

    2u2T

    vUvTH1

    2u2T

    vUvTH2

    2u2T

    vUvP2

    2u2P

    vUvP3

    2u2P

    vUv _mR

    2u2_m (7)

    This expression is evaluated numerically for each of the 41 oper-ating points. The standard uncertainty of the measured overall heattransfer coefcient is different for each operating point. Fig. 8

    6 G DeW G G 4.377 G G G G 3.678 G F G G 3.669 PeKeP F G G 4.9610 DeB G G 2.1611 G DeW G F 4.3712 G G G F 3.6513 G F G F 3.6614 PeKeP F G F 4.9215 DeB G F 2.1416 G DeW PeKeP F 4.3117 G G PeKeP F 3.5918 G F PeKeP F 3.619 PeKeP F PeKeP F 4.7820 DeB PeKeP F 2.1221 G DeW DeB 3.9122 G G DeB 3.2723 G F DeB 3.2824 PeKeP F DeB 4.1925 DeB DeB 1.82Table 4Tested correlation combinations and corresponding average errors on the overallheat transfer coefcient U.

    Model nb. Syltherm corr. R134a corr. U aver. (%)

    1 G DeW G DeW 5.442 G G G DeW 4.753 G F G DeW 4.72

    y 41 (2012) 10e17 15ow is more or less proportional to the rotational speed. Therefore,the rotational speed of the pump was too low compared to itsnominal speed and induced high torque peaks. Therefore, aftera few minutes of operation, the temperature of the electric motorbecomes too high. A way to solve the problem is to add a bypassbetween the outlet and inlet of the pump, in order to increase themass ow of the latter without changing the mass ow in the restof the system. This is a solution for the rst experiments, but in thefuture a new diaphragm pump which is better adapted to therequired mass ow will be acquired.

    7. Compressor-turbine unit design

    The compressor-turbine unit (see Fig. 9) was designed on thebasis of an electrically driven oil-free compressor (about 3 kW)described in [2,3]. The electric motor was replaced by a turbine.

    The turbine shape was designed in two steps using the ConceptsNREC [14] software. At rst, a preliminary design was made usingthe 1-D simulation tool. This step consists of determining the

  • 1 2 3 4 5 6 7 8 9 10 11 12 10

    1

    2

    3

    4

    5

    6

    7

    8

    9

    Mod

    Erro

    r on

    the

    calc

    ulat

    ed o

    vera

    ll he

    at tr

    ansf

    er c

    oeffi

    cien

    t (%)

    nt

    J. Demierre et al. / Energ16Fig. 7. Average, minimal and maximal error on the overall heat transfer coefcie

    7

    %)optimal geometry of the inlet and outlet of each element (volute,nozzle, interspace, rotor and diffuser). The second step is the 3-Ddesign of the different elements. To evaluate the performancesand check the ow, CFD calculations have been performed.

    1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 20

    1

    2

    3

    4

    5

    6

    Operating

    Rel

    ativ

    e sta

    ndar

    d un

    certa

    inty

    of U

    (

    Fig. 8. Calculated relative uncertainties of the measured overall hea

    Fig. 9. Compressor-turbine unit without the compressor and turbine stator partsduring the balancing (compressor on the left and turbine on the right.).3 14 15 16 17 18 19 20 21 22 23 24 25el nb.

    U for the 25 tested correlation combinations (for model nb. refer to Table 4).averageminmax

    y 41 (2012) 10e17The different elements of the turbine side were designed withthe main constraint of having the clearances between the rotor andthe static parts as small as possible. On this prototype the clear-ances between the compressor and turbine wheels and the staticparts is about 0.1 mm. Finite element analysis was done to checkthe thermal and mechanical stresses and the displacements due tothe high rotational speed (up to 240,000 rpm), the high pressure(up to 70 bar) and the high temperature differences (about 0 C onthe compressor side and 180 C on the turbine side). The CTU hasbeen balanced and tested with air at rotational speeds up to140,000 rpm.

    8. R134a GWP concerns

    The selected working uid, R134a, has a Global WarmingPotential (GWP) equal to 1300, which is quite high. Since 1 January2011, EU Member States do no longer grant EC type-approval ornational type-approval for vehicles equipped with an air-conditioning device using R134a [15]. Concerning stationaryequipment, as in our case, the EU regulation requires to clearlyindicate on the product that it contains such a refrigerant and itsquantity [16]. As described in Section 2, the system is fullyhermetic, since there is no sealing on moving parts. In case oftoughening of the regulation, R1234yf (GWP 4) could be used forthis application instead of R134a. Preliminary designs and

    1 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 point nb.

    t transfer coefcient for each of the 41 tested operating points.

  • simulations of ORCeORC systems with a heating capacity of 40 kW(twice bigger than what is presented in this paper) using R1234yfor R134a have been performed by the authors [17]. This preliminaryanalysis tends to indicate that the COP would be in a similar rangefor both uids and that the design parameter values would besimilar. At present, there is no reason to exclude R134a for thisapplication. Moreover, the use of R134a for this prototype isconvenient, because the material and equipment compatibility ofthis refrigerant is well documented as compared to R1234yf.Furthermore, tests of the supercritical evaporator have been donewith R134a for reasons of costs and availability.

    9. Conclusion

    The preliminary design of a prototype of an ORCeORC thermallydriven heat pump for residential application (about 20 kW heatingpower) is presented. The chosen diameter of the compressor andturbine wheels are respectively 20 mm and 18 mm and theirnominal speed of rotation is about 200,000 rpm. A counter-currentdouble tube coil supercritical evaporator is selected because of itsrobust design. The detailed layout for the ORCeORC prototype isgiven. A rst test rig has been built to test the supercritical evap-orator and the diaphragm pump of the power cycle which are twocritical components. The results obtained with an in-house super-critical evaporator model has been compared to the measurements

    behavior at rotational speeds at least up to 140,000 rpm.

    Solutions AG in Herzogenbuchsee (BE, Switzerland) for the supplyof the gas bearings.

    References

    [1] Strong DTG, Development of a directly red domestic heat pump, Ph.D. Thesis,University of Oxford, UK 1980.

    [2] Schiffmann J, Favrat D. Design, experimental investigation and multi-objectiveoptimisation of a small-scale radial compressor for heat pump applications.Energy 2010;35(1):436e50.

    [3] Schiffmann J, Favrat D. Experimental investigation of a direct driven radialcompressor for domestic heat pumps. International Journal of Refrigeration2009;32(8):1918e28.

    [4] Demierre J, Favrat D, Low power ORC-ORC systems for heat pump applications.In: 9th International IEA Heat Pump Conference, Zrich, Switzerland, 2008.

    [5] Demierre J, Favrat D. Thermally driven residential heat pumps based onintegrated Organic Rankine Cycles (ORC). In: Smart energy strategies; 2008.

    [6] Belsim, Vali IV, http://www.belsim.com 2006.[7] Marchal F, Kalitventzeff B. Process integration: selection of the optimal utility

    system. Computers & Chemical Engineering 1998;22:S149e56.[8] Molyneaux A, Leyland G, Favrat D. Environomic multi-objective optimisation

    of a district heating network considering centralized and decentralized heatpumps. Energy 2010;35(2):751e8.

    [9] Whiteld A, Baines NC. Design of radial turbomachines. Harlow, UK: LongmanScientic & Technical; 1990.

    [10] Balje OE. Turbomachines: a guide to design, selection and theory. New York,USA: John Wiley & Sons; 1981.

    [11] Borel L, Favrat D. Thermodynamics and energy systems analysis. Lausanne,Switzerland: EPFL Press; 2010.

    [12] Henchoz S. Study of a supercritical evaporator for an ORC-ORC thermallydriven heat pump (student semester project). Switzerland: Tech. Rep., Labo-ratory for Industrial Energy Systems (LENI), Ecole polytechnique fdrale deLausanne; 2008.

    J. Demierre et al. / Energy 41 (2012) 10e17 17Acknowledgments

    The authors would like to thank the Swiss Federal Ofce ofEnergy (OFEN) for its nancial support and Fischer Engineeringdone on the DTC heat exchanger. The error between the predictedoverall heat transfer coefcient and the measurements is less than3%. The tests have shown that the diaphragm pump is able toprovide the required pressure rise (about 55 bar). Finally, thecompressor-turbine unit has been tested with air and shows a good[13] Lira I. Evaluating the measurement uncertainty: fundamentals and practicalguidance. Taylor & Francis; 2002.

    [14] Concepts NREC. AGILE engineering design system, http://www.conceptsnrec.com; 2007.

    [15] European Parliament, Council. Directive 2006/40/EC of the European parlia-ment and of the council of 17 May 2006 relating to emissions from airconditioning systems in motor vehicles and amending council directive 70/156/EEC (text with EEA relevance), http://eur-lex.europa.eu; 2006.

    [16] European Parliament, Council. Regulation (EC) No 842/2006 of the Europeanparliament and of the council of 17 May 2006 on certain uorinated green-house gases (text with EEA relevance), http://eur-lex.europa.eu; 2006.

    [17] Demierre J, Favrat D, Theoretical study of thermally driven heat pumps based ondouble Organic Rankine Cycle: working uid comparison and off-design simu-lation, In: 10th International IEA Heat pump Conference, Tokyo, Japan, 2011.

    Prototype of a thermally driven heat pump based on integrated Organic Rankine Cycles (ORC)1. Introduction2. ORCORC high speed concept3. Preliminary design4. Prototype layout5. Selected equipment5.1. Supercritical evaporator5.2. Pump of the power cycle

    6. First test rig6.1. Layout6.2. Test results6.2.1. Supercritical evaporator6.2.2. Pump of the power cycle

    7. Compressor-turbine unit design8. R134a GWP concerns9. ConclusionAcknowledgmentsReferences