8
Prototype of a thermally driven heat pump based on integrated Organic Rankine Cycles (ORC) J. Demierre * , S. Henchoz, D. Favrat Ecole Polytechnique Fédérale de Lausanne, LENI-IGM-STI, Station 9, 1015 Lausanne, Switzerland article info Article history: Received 7 October 2010 Received in revised form 24 August 2011 Accepted 28 August 2011 Available online 25 September 2011 Keywords: Thermally driven heat pump ORC Supercritical evaporator Gas bearings Residential heating Oil-free abstract The concept studied in this work is a low power ORCeORC heat pump system (providing about 20 kW heat at the condenser) and that is composed of an ORC power cycle driving a reversed ORC heat pump cycle, both cycles using the same uid. The centrifugal compressor and the radial in-ow turbine are directly coupled on the same shaft rotating on self-acting refrigerant vapor bearings. The system has the advantage of being oil-free, fully hermetic and with low maintenance costs. The paper presents the development of an ORCeORC prototype, with HFC-134a as working uid. The main critical parts of the system are the compressor-turbine unit, the supercritical evaporator and the pump. The selected type of heat exchanger for the supercritical evaporation is the double tube coil (DTC). A rst experimental setup has been built to test the pump and the supercritical evaporator. A comparison between the results obtained with an in-house supercritical evaporator simulation program and measurements made on the DTC 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. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction Single combustion in boilers is a very inefcient heating process. The higher cost of fuels and concerns about pollution are likely to put more and more political pressure to prevent the use of boilers for most building heating purposes. Better alternatives are all linked to the use of heat pumps that allow the recovery and upgrade of the renewable heat from the environment. Thermally driven heat pumps can use various fuels including wood pellets or natural gas and are usually based on absorption heat pump cycles or on a combination of a power cycle driving a compression heat pump cycle or a combination of both. One concept of the second type has been proposed by Strong [1] and is based on the use of an ORC power cycle driving a reversed ORC heat pump cycle, both using the same uid. Such a concept can be made using a scroll expander and a scroll compressor usually lubricated with oil or by an oil-free high speed dynamic compressor-expander assembly rotating on refrigerant vapor bearings. Unfortunately earlier attempts of the latter failed because of the lack of appropriate materials and because of the problem of the depletion of the ozone linked to the CFC refrigerants, which were, at the time, the only non-ammable and non toxic refrigerant candidates for such high temperature cycles. Progress in materials and the emergence of new 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 are key working uid candidates, together with new low Global Warming Potential (GWP) refrigerants being in development within the same range of pressures. The recent demonstration of one concept of miniature high speed centrifugal compressor directly driven by a high speed electric motor rotating on refrig- erant gas bearings [2,3] opens the way to such devices with or without 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. A description of the principle is given at Section 2. The application that is studied in this work is residential heating for small buildings. The goal of developing such a prototype is to demonstrate the technical feasibility of a thermally driven heat pump based on double 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 driving a reversed Rankine heat pump cycle (see Fig. 1). The condenser and the subcooler, if introduced, are common to both cycles. This * Corresponding author. Tel.: þ41 21 693 67 29; fax: þ 41 21 693 35 02. E-mail address: jonathan.demierre@ep.ch (J. Demierre). Contents lists available at SciVerse ScienceDirect Energy journal homepage: www.elsevier.com/locate/energy 0360-5442/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.energy.2011.08.049 Energy 41 (2012) 10e17

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

Embed Size (px)

Citation preview

Page 1: Prototype of a thermally driven heat pump based on integrated Organic Rankine Cycles (ORC)

at SciVerse ScienceDirect

Energy 41 (2012) 10e17

Contents lists available

Energy

journal homepage: www.elsevier .com/locate/energy

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

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

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

Keywords:Thermally driven heat pumpORCSupercritical evaporatorGas bearingsResidential heatingOil-free

* Corresponding author. Tel.: þ41 21 693 67 29; faxE-mail address: [email protected] (J. Dem

0360-5442/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.energy.2011.08.049

a b s t r a c t

The concept studied in this work is a low power ORCeORC heat pump system (providing about 20 kWheat at the condenser) and that is composed of an ORC power cycle driving a reversed ORC heat pumpcycle, both cycles using the same fluid. The centrifugal compressor and the radial in-flow turbine aredirectly coupled on the same shaft rotating on self-acting refrigerant vapor bearings. The system has theadvantage of being oil-free, fully hermetic and with low maintenance costs. The paper presents thedevelopment of an ORCeORC prototype, with HFC-134a as working fluid. The main critical parts of thesystem are the compressor-turbine unit, the supercritical evaporator and the pump. The selected type ofheat exchanger for the supercritical evaporation is the double tube coil (DTC). A first experimental setuphas been built to test the pump and the supercritical evaporator. A comparison between the resultsobtained with an in-house supercritical evaporator simulation program and measurements made on theDTC is presented. The design steps of the compressor-turbine are briefly 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.

1. Introduction

Single combustion in boilers is a very inefficient heating process.The higher cost of fuels and concerns about pollution are likely toput more and more political pressure to prevent the use of boilersfor most building heating purposes. Better alternatives are alllinked to the use of heat pumps that allow the recovery andupgrade 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 fluid. Such a concept can be made using a scrollexpander and a scroll compressor usually lubricated with oil or byan 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-flammable and non toxic refrigerant candidates for such high

: þ 41 21 693 35 02.ierre).

All rights reserved.

temperature cycles. Progress in materials and the emergence ofnew fluids, 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 fluid candidates, together with new low GlobalWarming Potential (GWP) refrigerants being in developmentwithin the same range of pressures. The recent demonstration ofone 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 thetechnical 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

Page 2: Prototype of a thermally driven heat pump based on integrated Organic Rankine Cycles (ORC)

Nomenclature

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

Nu ¼ hfl, Nusselt number (�)

Pr ¼ cpml, Prandtl number (�)

Ref ¼ Cfm=r

, Reynolds number based on ɸ (�)

A heat transfer area (m2)C flow velocity (m/s)cp isobaric specific heat (J/(kg K))COP Coefficient of Performance, (�)

f ¼ �dPdL

f

12rC2

, friction factor (�)

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

Greek letters˛U error on the overall heat transfer coefficient between

simulation and measurements (�)hc compressor isentropic efficiency (�)ht turbine isentropic efficiency (�)

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

ɸ ¼ 4�cross 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 combustionH1 evaporator (DTC) inlet, hot thermal oilH2 evaporator (DTC) outlet, hot thermal oillm logarithmic meanmeas measuredpump power cycle pumpR refrigerantsimu predicted by simulationW water

Supersciptsþ 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

J. Demierre et al. / Energy 41 (2012) 10e17 11

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

condenser

supercritical evaporator

evaporator

compressoturbine uni(CTU)

Fig. 1. Schematic flowsheet and T-s diagram

system (about 20 kW thermal at the condenser in the case treatedhere) with a one-stage centrifugal compressor and a one-stageradial in-flow 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

r-t

Entropy

Tem

pera

ture

of a simple ORCeORC heat pump unit.

Page 3: Prototype of a thermally driven heat pump based on integrated Organic Rankine Cycles (ORC)

2.5x 10

5

rpm

]

J. Demierre et al. / Energy 41 (2012) 10e1712

compressors and turbines a low density fluid is preferred andrefrigerant HFC-134a has been selected for the first prototype. It ischemically stable at relatively high temperatures (at least up to180 �C).

1.4 1.6 1.80.5

1

1.5

2

COP [−]

CT

U r

otat

iona

l spe

ed [

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

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 flowsheeting 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 flow 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 efficiency ht. This relation gives the maximum efficiency thatcan be achieved for a fixed turbine specific speed nst. Thecompressor isentropic efficiency hc is modeled using a similarmethod. A correlation (adapted from Balje 1981 [9,10]) that relatesthe maximum efficiency to the compressor specific 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 efficiency 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 floor heating from 30 �C to 35 �C. The totalheating power (hot water production þ floor 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 (coefficient of performance) as the firstobjective 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 defined asfollows:

COP ¼_Q�W

_Qþfumes þ

�_Eþpump=0:56

� (1)

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

þfumes

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

þpump is the electrical power consumed by the

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

þpump is divided by 0.56 to take into account for the

efficiency of a modern combined cycle power plant that producesthis electrical power, in order to define a COPwhich is strictly basedon the conversion of the fuel (here, methane) into heat. For a firstapproximation, _Q

þfumes is evaluated as follows [11]:

_Qþfumesy _mFDh

0i (2)

_mF is the mass flow of the fuel (methane) and Dh0i is the lowerheating value of the fuel (based on the standard state,P0 ¼ 1.01325 bar and T0 ¼ 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, justifies 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.

4. Prototype layout

The layout of the prototype is shown in Fig. 4. For this firstprototype, 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 fluid toflow 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 ofthe compressor.

Page 4: Prototype of a thermally driven heat pump based on integrated Organic Rankine Cycles (ORC)

1.4 1.6 1.8

0.02

0.04

0.06

0.08

COP [−]

whe

el d

iam

eter

[m

]turbine

compressor

1.4 1.6 1.840

45

50

55

60

65

70

COP [−]

supe

rcri

tical

eva

p. p

ress

ure

[bar

]

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.

Fig. 4. ORCeORC prototype layout.

J. Demierre et al. / Energy 41 (2012) 10e17 13

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.

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 mm

Several types of heat exchanger have been reviewed (detailsare given in [12]). The selected type is “the double tube coil” (DTC)(see Fig. 5). A DTC heat exchanger is a spiral tube-in-tube heat ex-changer allowing a counter-current operation with high tempera-ture differences and high pressures.

5.2. Pump of the power cycle

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.

Page 5: Prototype of a thermally driven heat pump based on integrated Organic Rankine Cycles (ORC)

J. Demierre et al. / Energy 41 (2012) 10e1714

6. First test rig

A first 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 first 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 flow ofthe refrigerant is measured with a Coriolis flow 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 pressurerelief valve (s1) is connected from the high pressure side to the lowpressure 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 specifications:

� Heat transfer area: 0.23 m2

� inner tube outer diameter: 1/2 inch� inner tube wall thickness: 0.049 inch

Fig. 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 flow_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 flow 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 coefficient U calculated on the basis of themeasurements varies between 360 W/(m2K) and 1020 W/(m2K).The overall heat transfer coefficient is evaluated as follows:

U ¼_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 ¼ DTH2� DTH1

ln�DTH2

=DTH1

� (4)

where

DTH1¼ TH1

� T3 and DTH2¼ 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 flows, 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 coefficient ˛U is calculated asfollows:

˛U ¼ Usimu � Umeas

Umeas(6)

Page 6: Prototype of a thermally driven heat pump based on integrated Organic Rankine Cycles (ORC)

Table 4Tested correlation combinations and corresponding average errors on the overallheat transfer coefficient ˛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 4.724 PeKePeF G DeW 6.285 DeB G DeW 3.26 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 1.82

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

Correlation Range of application

Gnielinski

Nu ¼f8ðRef � 1000ÞPr

1þ 12:7

ffiffiffif8

rðPr2=3 � 1Þ

2300 < Ref < 5� 106

0.5 < Pr < 2000

Petukhov-Kirillov-Popov

Nu ¼f8RefPr

1:07þ 12:7

ffiffiffif8

rðPr2=3 � 1Þ

104 < Ref < 5� 106

0.5 < Pr < 2000

Dittus-BoelterNu ¼ 0:023Re4=5f Prn Ref > 104

with n ¼ 0.4 for heating 0.7 � Pr � 160and n ¼ 0.3 for cooling L/f > 10

J. Demierre et al. / Energy 41 (2012) 10e17 15

where Usimu and Umeas are the overall heat transfer coefficientscomparing simulation andmeasurements. The overall heat transfercoefficient 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 coefficient Umeas. The standard uncertainty for thedifferent measurements is:

� Temperature (thermocouples type K): uT ¼ �0.2 K� Pressure (piezo-resistive): uP ¼ �0.2 bar� Mass flow rate (Coriolis force): u _m ¼ �0:2%� _m

The uncertainty of the overall heat transfer coefficient is calcu-lated using the “law of propagation of uncertainties” as described in[13]. The overall heat transfer coefficient standard uncertainty uU isexpressed, in our case:

Table 3Friction factor correlations implemented in the DTC model.

Correlation Application

Gnielinskif ¼ ð0:79lnRef � 1:64Þ�2 smooth pipeFilonenkof ¼ ð1:82log10Ref � 1:64Þ�2 smooth pipeDarcyeWiesbach(Churchill empirical formula)

f ¼ 8½ð 8Ref

Þ12 þ 1

ðAþ BÞ3=2�1=12

where

A ¼ ½2:457$ln 1

ð 7RefÞ0:9

þ 0:27$kf

�16B ¼ ½370 530

Ref�16

rough pipe

u2U ¼�vUvT2

�2

u2T þ�vUvT3

�2

u2T þ�

vUvTH1

�2

u2T þ�

vUvTH2

�2

u2T

þ�vUvP2

�2u2P þ

�vUvP3

�2u2P þ

�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 coefficient is different for each operating point. Fig. 8shows 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 flowis slightly too high compared to the requiredmass flow of R134a forour application. Since it is a volumetric pump, the delivered massflow 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 flow of the latter without changing the mass flow in the restof the system. This is a solution for the first experiments, but in thefuture a new diaphragm pump which is better adapted to therequired mass flow 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 first, a preliminary design was made usingthe 1-D simulation tool. This step consists of determining the

Page 7: Prototype of a thermally driven heat pump based on integrated Organic Rankine Cycles (ORC)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 250

1

2

3

4

5

6

7

8

9

Model nb.

Err

or o

n th

e ca

lcul

ated

ove

rall

hea

t tra

nsfe

r co

effi

cien

t (%

)

averageminmax

Fig. 7. Average, minimal and maximal error on the overall heat transfer coefficient ˛U for the 25 tested correlation combinations (for “model nb.” refer to Table 4).

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 410

1

2

3

4

5

6

7

Operating point nb.

Rel

ativ

e st

anda

rd u

ncer

tain

ty o

f U

(%

)

Fig. 8. Calculated relative uncertainties of the measured overall heat transfer coefficient for each of the 41 tested operating points.

J. Demierre et al. / Energy 41 (2012) 10e1716

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 flow, CFD calculations have been performed.

Fig. 9. Compressor-turbine unit without the compressor and turbine stator partsduring the balancing (compressor on the left and turbine on the right.).

The 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 fluid, 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 ¼ 4) could be used forthis application instead of R134a. Preliminary designs and

Page 8: Prototype of a thermally driven heat pump based on integrated Organic Rankine Cycles (ORC)

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

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 fluids 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 first 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 measurementsdone on the DTC heat exchanger. The error between the predictedoverall heat transfer coefficient 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 goodbehavior at rotational speeds at least up to 140,000 rpm.

Acknowledgments

The authors would like to thank the Swiss Federal Office ofEnergy (OFEN) for its financial support and Fischer Engineering

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

References

[1] Strong DTG, Development of a directly fired 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, Zürich, 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] Maréchal 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] Whitfield A, Baines NC. Design of radial turbomachines. Harlow, UK: LongmanScientific & 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 fédérale deLausanne; 2008.

[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 fluorinated 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 fluid comparison and off-design simu-lation, In: 10th International IEA Heat pump Conference, Tokyo, Japan, 2011.