Dynamic performance estimation of small-scale solar cogenerationwith an organic Rankine cycle using a scroll expanderB. Tw...
Rankine cycle. Such a cogeneration unit should utilise the highthermal efficiency of solar thermal collectors to provide wa...
a simple non-regenerative ORC with a scroll expander to producean isentropic efficiency as high as 83%. It is demonstrated ...
_Wloss À _Qex þ _Qsu À _Qamb ¼ 0: (9)3.2. Parameter estimationThis section describes the ORC test setup used to determinev...
also assumed that heat exchangers are designed and sized appro-priately so that sufficient heat exchange is attained to pro...
4.5. ORC pumpThe pump is not independently modelled as a component, butthe power required to maintain the cycle pressure d...
performance from tests, and the condition that scroll exit temper-ature is greater than Thw þ DTpinch to aid water heating...
Table 9 shows some key effects of perturbing the former twoparameters for the month of December. Generally, _morg affectst...
6. ConclusionsA scroll expander model was calibrated and implemented intoa larger dynamic model of a solar thermal cogener...
[19] Geoscience Australia, Geodesy e Astronomical Information (2005), Availableonline: http://www.ga.gov.au/bin/astro/sunr...
Upcoming SlideShare
Loading in …5
×

Ciclo de rankine con regeneración

1,060 views

Published on

Published in: Business, Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
1,060
On SlideShare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
6
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Ciclo de rankine con regeneración

  1. 1. Dynamic performance estimation of small-scale solar cogenerationwith an organic Rankine cycle using a scroll expanderB. Twomey*, P.A. Jacobs, H. GurgenciQueensland Geothermal Energy Centre of Excellence, The University of Queensland, Level 5, Bldg 45 Mansergh Shaw, Staff House Road, St. Lucia,Brisbane, Qld 4072, Australiah i g h l i g h t s< A scroll expander is modelled and parameter values are estimated experimentally.< The scroll expander shows a maximum isentropic efficiency of 59%.< A dynamic model of a low-cost solar thermal cogeneration system is presented.< The cycle thermal efficiency is 3.47%.< 1710 kWh electricity is produced per year as a by-product of heating water.a r t i c l e i n f oArticle history:Received 11 December 2011Accepted 28 June 2012Available online 13 December 2012Keywords:Organic Rankine cycleDynamic modellingScroll expanderSolar cogenerationa b s t r a c tSmall-scale solar thermal cogeneration shows promise as an effective way to get increased benefit out ofa given solar availability, since it does not waste potential during summer after the water capacity isheated. In this paper a scroll expander is tested in a small organic Rankine cycle (ORC) and used tocalibrate a static expander model. Validation of the scroll expander model shows agreement generallywithin 10% for the shaft power, 5% for the rotational speed and 6 K for the exhaust temperature, withsome outliers at very low pressure ratios. This calibrated model is then incorporated into a largerdynamic model of a solar thermal cogeneration system, designed for some larger dwelling unit or smallcommercial establishment that requires a larger volume of hot water. An annual simulation is conductedusing a collector area of 50 m2, and the scroll expander shows a maximum isentropic efficiency of 59%while the ORC efficiency is 3.47%. The total energy produced is 1710 kWh and the hot water available ison average 2540 L/day. The maximum instantaneous power that can be produced by the system is 676 W,and it is possible to shift the time period that the system is producing power to match the peak demandperiod by adjusting the solar store volume.Ó 2012 Elsevier Ltd. All rights reserved.1. IntroductionAmidst uncertain economic times and increasing electricitycosts, many home and small business owners are looking to investin renewable energy, and especially solar energy, as a stablesupplement. There are a number of emerging systems, but solarpower is typically generated by one of two methods. The first isphotovoltaic (PV) solar capture that converts solar irradiancedirectly to electricity and typically has an efficiency between 12 and20%. Solar thermal captures the irradiance as thermal energy andhas a heating efficiency generally about 60%, so it is the betterchoice for pure heating applications. To generate electricity, theheating efficiency needs to be multiplied by a power conversionefficiency and the latter is low unless a high degree of concentra-tion is used. The result is that for small installations, low-concentration solar thermal is generally best for water and spaceheating and PV is more effective to generate electricity.Additionally, solar thermal panels are more effective during thesummer months than the winter, but the amount of hot waterrequired does not change much, so there is much wasted potentialduring the summer when the capacity of water is heated. A solarthermal power system may be competitive in applications thatneed both electricity and heat, such as process heat for a brewery orfood processing plant, or water heating for a larger dwelling unit.The goal of this study is to estimate the performance andcomment on the feasibility of a small-scale solar thermal cogene-ration unit that uses cheap and available equipment, with a focuson the scroll expander being used to extract energy in an organic* Corresponding author. Tel.: þ61 7 3365 7164.E-mail address: b.twomey@uq.edu.au (B. Twomey).Contents lists available at SciVerse ScienceDirectApplied Thermal Engineeringjournal homepage: www.elsevier.com/locate/apthermeng1359-4311/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.applthermaleng.2012.06.054Applied Thermal Engineering 51 (2013) 1307e1316
  2. 2. Rankine cycle. Such a cogeneration unit should utilise the highthermal efficiency of solar thermal collectors to provide waterheating, while generating electricity during the off-peak periodswhen the sun’s potential would otherwise be wasted. Fig. 1 showsone such system configuration which is the focus of this paper.A solar thermal cycle heats a fluid in the solar store vessel, whichis connected to a heat exchanger that is the evaporator for an ORC.Hot water is produced when the ORC condenser cools the workingfluid using a continuous stream of town supply water. The maincharacteristics of the system are the following: Solar collectors: 50 m2evacuated tube type Solar thermal cycle: Pressurised and well insulated closed-loopcycle heating 800 L solar store of water with circulation pumpand controller ORC: Expansion of R134a through scroll expander, with evap-orator utilising the solar store, and condenser utilisinga constant stream of town supply water.Water is pumped through the solar thermal cycle to heat thesolar store. When the store reaches a temperature, Tst,max, deter-mined by the optimum conditions of the ORC, the refrigerant cyclepump is started. R134a heats up by passing through the solar storeheat exchanger, is expanded through the scroll expander, iscondensed in the lower heat exchanger, then returns through thepump, completing the cycle. When the solar store drops toa temperature, Tst,min defined by the ORC conditions, the refrigerantcycle is stopped. The cycle may be started again if there is stillradiant emittance, but when the sun sets, the solar thermal cyclewill be available until the temperature drops to Tst,min, then deac-tivated until the next day.This study builds on prior literature, which has mostly focused onstatic solar models, and presents a daily dynamic model and annualsimulation of a solar thermal cogeneration system. The model givesan estimation of the total power produced and hot water delivered,as well as providing cycle efficiency, solar utilisation and expanderefficiency for different parts of the year. Positive outcomes wouldindicate that the addition of low-cost ORC components to standardsolar thermal systems might be financially worthwhile and hasbenefits which make it comparable to PV systems.2. Recent ORC literatureIn the first international conference on ORC power systems heldin Delft, ORC 2011, there was some focus on solar thermal ORCs.Presentations on large- and small-scale systems, pilot plants,combined heating and power and scroll expanders show agreementthat there is potential in very small-scale ORCs for cogeneration.Oudkerk et al. [1] presented an evaluation of an ORC-based micro-CHP system, less than 50 kWe, involving a hermetic scroll expander.Of the fluids tested, R245fa was deemed as most suitable, and a scrollexpander isentropic efficiency of 71% was achieved. The testers wereleft desiring a more suitable expander with a higher maximum inlettemperature and volume ratio in order to maximise the system.Kosmadakis et al. [2] provided a comparison of a novel double-stageexpansion in a 2.5 kWe solar ORC. They found that the maximumcycle efficiency increased from 4.3% using single-stage expansionwithout enhancements, up to 9.8% using a double-expander system.A simulation model of an experimental small-scale ORC cogeneratorwas presented on poster by Clemente et al. [3]. This model consid-ered both scroll expander and piston-type expander, and deter-mined a maximum expected scroll expander efficiency near to 65%and a maximum cycle efficiency without recuperator of 7%. They alsonoted that if the expansion ratio is lower than the best-efficiencyone, which corresponds to the built-in volume ratio, the expanderefficiency decreases rather quickly. Mikielewicz et al. [4] discussedtheir experiences from operation of different expansion devices indomestic micro ORCs. Three expansion devices were compared:a rotating plate expander, scroll expander, and pneumatic engine.They found that the pneumatic devices performed better with effi-ciencies near 61e82% to the scroll expander at near 29e52%.A performance and design optimisation study by Quoilin et al.[5] based on low-cost solar ORCs provides modelling results forsolar thermal electricity generation for remote off-grid areas ofdeveloping countries. An overall steady-state electrical efficiency ofup to 8% was achieved at a nominal working point, and it waspointed out that a dynamic model is needed to evaluate the yearlyenergy output. A comparison of working fluids showed that thebest performer was Solkatherm SES36, with R245fa also beinga good performer. A more complex solar Rankine cycle is presentedby Bao et al. [6]. This configuration consists of two collectors, twoexpanders, a regenerator and an internal heat exchanger, andutilises a zeotropic mixture Isopentane/R245fa at multiple massfractions to attempt to optimise the thermal efficiency. The thermalefficiency of this system was found to be significantly higher thana single-stage system using pure Isopentane or R244fa.Qiu et al. [7] give an overview based on market research ofa number of expansion devices for micro-CHP systems usingorganic Rankine cycles and conclude that scroll expanders and vaneexpanders are good choices for systems running at about 1e10 kW.The authors also comment that scroll expanders can be scaleddown very well. Schimpf et al. [8] present a model of a solar assistedcombined heat pumpeorganic Rankine cycle-system. The analysisshowed that it seems plausible that the additional costs of an ORCand advanced controls could be covered by the electricity gener-ated. The authors note that the application of a combined system ismore suited to larger dwelling units like hotels, senior citizen’shomes and multiple dwelling houses.A solar organic Rankine cycle similar to that presented in thispaper is discussed by Saitoh et al. [9]. This paper focuses on theelectricity generation using a scroll expander during the summer,and achieves a Rankine cycle efficiency of 11%. Using the workingfluid R113, a scroll expander efficiency of 63% was achieved. Man-olakos et al. [10] provide a design and experimental results froma 2 kWe solar ORC system using evacuated tube collectors, scrollexpanders and R134a as a working fluid. The system obtaineda maximum overall efficiency of 4%. Oliveira [11] provides anexample layout of a combined heat and power solar thermal systemwhich utilises a Rankine cycle to produce work from solar heat, andrecovers useful heat from the condenser. Mathias et al. [12] useEvacuated tubecollectorsSolar storagevesselScroll expanderCondenserPumpWshDynamic model boundaryWQinWTo hot waterstorageCity supply waterQoutmFig. 1. Solar ORC system configuration.B. Twomey et al. / Applied Thermal Engineering 51 (2013) 1307e13161308
  3. 3. a simple non-regenerative ORC with a scroll expander to producean isentropic efficiency as high as 83%. It is demonstrated that thehighest isentropic efficiency is attained when both the specificvolume ratio of the fluid undergoing expansion is matched to thebuilt-in volume ratio of the expander, and when the volume flowrate is matched to the rotational speed. An accompanying analyticmodel showcases the improvement of having two expanders inseries, increasing the overall pressure ratio as high as 14.23, to givean overall system efficiency of 7.7%.3. Scroll expanderA scroll expander has been identified as an appropriate devicefor small-scale solar thermal cogeneration because it is compactand inexpensive, low-power to suit small-scale scenarios, andgenerally shows high efficiency and has few moving parts. Thissection presents a model of a scroll expander suitable for incor-poration into a solar thermal ORC dynamic model, along withparameter identification and experimental validation.3.1. ModelThe scroll expander model used in this study is a semi-empiricalmodel developed by Lemort et al. [13] and the key processdescriptions and equations are included in this paper forcompleteness. Fig. 2 depicts the processing of the fluid as it flowsthrough the scroll expander in a number of hypothetical stages.(a) Adiabatic supply pressure drop (su e su,1): This process accountsfor all pressure losses between the expander inlet and thesuction chamber. It is approximated as an isentropic flowthrough a converging nozzle as per Eq. (1) with throat area, Asu,computed by parameter optimisation._m ¼Asuvthr;suffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2Àhsu À hthr;suÁq(1)(b) Isobaric supply cooling-down (su,1 e su,2): Heat exchange isobserved between the fluid entering the suction chamber anda fictitious isothermal envelope, which is a lumped variablerepresenting the expander scrolls and shell. Supply heattransfer is given by:_Qsu ¼ _m$Àhsu;1 À hsu;2Á¼241 À eÀAUsu_m$cp35 _m$cp$ÀTsu;1 À TwÁ;(2)where supply heat transfer coefficient AUsu is assumed to vary withthe mass flow rate according to:AUsu ¼ AUsu;n$_m_mn; (3)and AUsu,n is a parameter to identify.(c) Internal leakage (su,2 e ex,2): Fluid that leaks inside themachine undergoes no useful expansion and is a significantcontributor to loss in the expander. All leakage paths arelumped into a hypothetical area Aleak, which is computed bycomparison to flow through an isentropic nozzle of said throatarea:_mleak ¼Aleakvthr;leakffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2Àhsu;2 À hthr;leakÁq: (4)The inlet pressure is Psu,2 and the outlet pressure is themaximum between Pex,2 and Pcrit, leak. The critical pressure iscomputed by considering the vapour as a perfect gas.(d) Adiabatic and reversible expansion to the adapted pressure(su,2 e in): The adapted pressure is related to the built-involume ratio of the expander rv, which is parameter thatmust be identified. The relationship is:vin ¼ rv$vsu;2: (5)(e) Adiabatic expansion at constant machine volume (ad e ex,2): Theunder- and over-expansion losses occur at this stage when theadapted pressure is higher or lower, respectively, than thesystem pressure at the expander outlet. In order to equalisethese pressures, the model assumes that some fluid flows outof or into the discharge chamber instantaneously when theexpansion chamber opens onto the discharge line.(f) Isobaric exhaust cooling-down or heating-up (ex,1 e ex): Heat isexchanged between the fluid exiting the expansion chamberand the isothermal envelope. This is modelled similarly as inEq. (2).The internal power produced is a combination of the suctionpower, expansion power and discharge power:_Win ¼ _minÂÀhsu;2 À hinÁþ vinÀPin À Pex;2ÁÃ: (6)Mechanical losses are lumped into a mechanical loss torqueparameter sloss, which is a parameter to identify. The expandershaft power is then defined by:_Wsh ¼ _Win À2$p$N$sloss60: (7)To complete the heat balance over the expander, a global heattransfer coefficient, AUamb, between the isothermal envelope andthe ambient is used to compute ambient losses:_Qamb ¼ AUambðTw À TambÞ: (8)Then, assuming that mechanical losses are directly added to theisothermal envelope, the heat balance is completed:fictitious isothermal envelopes=cte V=cteWinsusu,1 su,2MQsu QexMin inMleakMex,2ex,1exWloss QambFig. 2. Conceptual scheme of the expander model by Lemort et al. [13]. From left toright across the figure, there are three regions: the supply (su), internal (in), andexhaust (ex). The ‘cte’ notation in the figure indicates that the variable describedremains constant through the process.B. Twomey et al. / Applied Thermal Engineering 51 (2013) 1307e1316 1309
  4. 4. _Wloss À _Qex þ _Qsu À _Qamb ¼ 0: (9)3.2. Parameter estimationThis section describes the ORC test setup used to determinevalues for the scroll expander parameters which were mentionedin the previous section. The facility used was the QGECE low-pressure organic Rankine cycle at the University of Queensland.The cycle is depicted in Fig. 3. A series of steady-state experimentswere conducted on the laboratory ORC by varying the pump speed,load on the expander, and evaporator temperature, and averagingthe results for each case over a defined time period. The primaryrecorded variables were expander shaft power, rotational speed,and exhaust temperature. A complete record of the results isavailable from the University of Queensland online technicalrepository [14].The simulation software DYMOLA (Dynamic Modeling Labora-tory) [15] was used to model the cycle including the scroll expanderequations in Section 3.1. Dymola is a complete modelling andsimulation tool for complex dynamic systems in various engi-neering disciplines. It includes accurate fluid models and somebasic libraries for thermal cycle modelling which have beenimproved and adapted for organic Rankine cycle and solar thermalmodelling.The scroll expander parameter values were estimated using the‘Calibrate Steady State’ function in the ‘Design’ library included withDymola. This calibrator works by minimising an error function basedon the sum of squares of the weighted differences between eachsteady-state output variable and the corresponding model variable.The inputs were selected based on their ability to be controlledrelatively independently from each other in the experimental rig,and on how close they match with the model inputs. They werechosen to be mass flow rate _m, pressure ratio of inlet to outlet rp, andthe inlet temperature Tsu. Three control values were selected for eachinput, based on the operational limits of the test rig, to generatea series of 27 test cases. The values are summarised in Table 1.The outputs selected were the shaft work Wsh, rotational speedN and outlet temperature Tex because they could be easilymeasured in the test rig and the model. The weighting applied tothese variables was the inverse of the maximum recorded valuesquared, to match up with Dymola’s calibration scheme. Table 2presents a summary of the outputs and their weights. The esti-mated parameter values are shown in Table 3. The aforementionedDymola error function returned 3 ¼ 0.0335. To put this inperspective, Fig. 4 shows the difference between model outputsand experimental outputs. The error was consistently high for caseswith a pressure ratio of 1.3, mostly between 13% and 20%. Apartfrom at this pressure ratio, the error was generally below 8%.4. System modellingIn this section, a dynamic model of the cogeneration systemcontained within the scope boundary in Fig. 1 is developed. Themodel should be used to fulfil the following objectives: Estimate the daily and annual power generation of the system Estimate the volume of hot water available to the establish-ment due to solar capture Determine the maximum temperature observed in the solarthermal cycle to limit boiling Assess the compatibility of a scroll expander with this type ofsystem Observe sensitivity to a change of collector area, solar storevolume and ORC mass flow rate.The model is based on one 24-h day cycle. In order to attainannual results, a model for each month is given parameters basedon the average of all days in that month. Specific details are given inthe following subsections, and month-specific parameters given inTable 4.Within the system model, it is assumed that the piping in thesolar cycle is insulated according to AS 3500.4 Section 8.2 [16], tothe effect that heat loss from the pipes is negligible. The powerrequired by the solar circulation pump and controller are neglectedbecause they are small compared to the solar thermal energy, andsince they are used for heating, should not be included in electricalefficiency. Also, local heating and cooling effects, including thermalstratification of storage vessels, are neglected for simplification ofanalysis. This is a conservative approach since stratification effectshave a positive influence on the efficiency of solar systems [17]. It isGBldg. aux. waterElectric elementheaterOil returnReceiverElectrical resistornetworkMTpθpθpθpθpθpθθθθθCoriolisflow meterqFig. 3. Diagram of the QGECE low-pressure ORC used for testing the performance of the scroll expander. Note that q represents temperature sensors.B. Twomey et al. / Applied Thermal Engineering 51 (2013) 1307e13161310
  5. 5. also assumed that heat exchangers are designed and sized appro-priately so that sufficient heat exchange is attained to producea heat exchange pinch point of DTpinch. Finally, it is assumed that thescroll expander can be attached to a generator and synchronised toan electrical grid with minimal inefficiencies.The system model was created in the Dymola environment,using the Modelica language. Separate modules for each the solarfluid cycle, the scroll expander and the cooling heat exchanger weredeveloped and connected within a testbed module.4.1. Solar insolationThe solar insolation is a direct input into the solar fluid cyclemodel, and is measured in W/m2. The value of irradiance followsa half sine curve as in Eq. (10)S ¼ 0; 0 t triseS ¼ Smax$sinpðtetriseÞtsetetrise; trise t tsetS ¼ 0; tset t 24:(10)where t is time in hours after midnight. The peak solar irradiance,Smax, is defined such that the total daily solar exposure, Ex, equalsthe value given by the Bureau of Meteorology [18] for the average ofall days of a particular month in Brisbane, Australia, going back to1990. The sun rise and set times are of the 15th day of every month,as given by Geoscience Australia [19] for 2010. The parameters usedfor each month are shown in Table 4.4.2. CollectorsThe collectors are modelled by the total collector efficiency hcol,which is a lumped parameter that describes the ratio of heat energyaddition to solar insolation incident on the collector, accounting fordifferent sources of losses. The total collector efficiency is describedby characteristic curves given by Ecofys [17], as shown in Fig. 5. Theaverage daily ambient temperature is given by the Bureau ofMeteorology [18], and shown in Table 4.4.3. Solar fluid cycleThe collectors, pipes, pump, and solar store are modelledtogether in a single lumped control volume, defined as the solarfluid cycle. This is to distinguish it from the hot water component ofthe system. Heat energy from the collectors is directly injected intothis control volume, and it is assumed that the solar storage vesselrepresents an appropriate average of the energy held in all of thesolar cycle components. The cycle is defined by the parametersshown in Table 5.The temperature of the solar store, Tst, can be calculated at anytime using equations of state such as those included in fluidpackages in Dymola, and require two state variables, chosen for thiscase as specific internal energy, ust, and store pressure, pst. Storepressure is a constant given in Table 5 and specific internal energy iscalculated as per Eq. (11).u ¼Ustrst$Vst(11)The store internal energy, Ust, is calculated using an energybalance over the store taking into account solar gain, heat transferto the ORC and standing heat losses:dUstdt¼ _Qsol À _Quhx À _Qstand: (12)Heat transfer to the ORC, _Quhx, is given by the scroll expandermodel. Solar gain is given by the product of insolation, collectorarea and collector efficiency, as in Eq. (13)._Qsol ¼ S$Acol$hcol (13)Standing heat loss from the solar store is modelled as a rateproportional to the temperature difference between the tank andambient, using values given in the industry guide by Ecofys [17] forproperly insulated tanks, as described in Eq. (14). The heat loss rate,AUst, is given in Table 5._Qstand ¼ AUst$ðTst À TambÞ (14)An output connector is added to the model with the value of Tstto be fed directly to the scroll expander flow boundary model andmass flow switch. The total solar exposure is also defined in themodel:dEsoldt¼ S$Acol: (15)4.4. Condenser heat exchangerThe heat exchanger at the expander outlet has two purposes: tocondense the ORC fluid and to heat the continuous supply of townwater to a sufficiently high temperature. It is modelled usinga simple heat balance for counter-flow heat exchangers, as in Eqs.(16) and (17).Qlhx ¼ _morg$Dhorg (16)Qlhx ¼ _mprc$Dhprc (17)All inlet and outlet temperatures are set according to theconditions, as shown in Table 6.The mass flow rate of the process stream, _mprc, is calculated andis dependent on the organic stream temperature as in Eq. (17).Table 2Scroll expander test outputs.Variable Weighting_Wsh (552)À2N (1257)À2Tex (385)À2Table 3Scroll expander parameters.Parameter Description ValueMedium Fluid medium R134aAUamb Overall heat transfer coefficient to ambient 0.5 W/KAUsu,n Supply overall heat transfer coefficient 35 W/KAUex,n Exhaust overall heat transfer coefficient 0.1 W/KAsu,thr Supply nozzle throat area 50eÀ6 m2Aleak Internal leakage area 4.04eÀ6 m2rv Built-in volume ratio 1.57Vs,exp Swept volume in expander mode 53.1eÀ6 m3sloss Mechanical loss torque 0.94 N mTable 1Scroll expander test inputs.Variable Input 1 Input 2 Input 3_m 0.05 kg/s 0.06 kg/s 0.07 kg/srp 1.3 1.5 1.7Tsu 80 C 100 C 120 CB. Twomey et al. / Applied Thermal Engineering 51 (2013) 1307e1316 1311
  6. 6. 4.5. ORC pumpThe pump is not independently modelled as a component, butthe power required to maintain the cycle pressure difference andmass flow rate is calculated and included in efficiency and energycalculations._Wpump ¼_morg$Àhpump;out À hpump;inÁhpump(18)Enthalpy hpump,in is estimated using the Dymola equation ofstate function with the two state variables being ORC condenserpressure, p1 and bubble enthalpy for that pressure (also calculatedby Dymola). Assuming the pump behaves isentropically, hpump,outis estimated by the same Dymola function, using state variablesof ORC evaporator pressure, p2, and bubble entropy at thecondenser pressure. The parameters p1, p2 and hpump are given inTable 7.4.6. Dynamic simulationThe dynamic simulation module incorporates the solar fluidcycle, scroll expander, cooling heat exchanger, flow boundary andinsolation models, as well as temperature and heat flow sensorsand switches. Fig. 6 shows how the components are connected. TheTst output signal is used to turn the scroll expander mass flow onand off. The mass flow rate is set to _mn when Tst À DTpinch reachesTst,max and set to zero if it falls below Tst,min. These temperatureswere chosen based on the range which produced the best expanderFig. 4. Check of scroll expander modelling against individual experimental test measurements.Table 4Environmental inputs.Month Ex (kWh/m2/day) Smax (W/m2) Sun rise/set (24 h) Tamb (C)January 6.8 780 0506/1848 25.3February 6.1 736 0531/1833 24.9March 5.6 715 0548/1806 23.5April 4.7 645 0605/1732 20.8May 3.8 553 0621/1708 17.9June 3.4 514 0636/1701 15.3July 3.7 551 0637/1710 14.3August 4.5 636 0619/1726 15.3September 5.6 738 0546/1741 18.1October 6.1 754 0513/1755 20.6November 6.7 782 0449/1816 22.7December 6.9 781 0447/1839 24.2Fig. 5. Characteristic efficiency curve for evacuated tube collector [17].B. Twomey et al. / Applied Thermal Engineering 51 (2013) 1307e13161312
  7. 7. performance from tests, and the condition that scroll exit temper-ature is greater than Thw þ DTpinch to aid water heating. Thenominal ORC mass flow rate is the highest value that will allowa single continuous ORC period during the day without the need toswitch off and let the solar store recharge. The mass flow rate couldbe adjusted seasonally in order to optimise the system. Thedynamic test bed parameters are shown in Table 7.4.7. System metricsSome key thermal power variables are defined here which areuseful to judge the effectiveness of the system. The net powerproduced in this cycle is defined by:dEcycdt¼ _Wsh À _Wpump: (19)The cycle’s thermal efficiency, hth, is the ratio of net workproduced to the heat input:hth ¼_Wsh À _Wpump_Quhx: (20)The second-law thermal efficiency, hii, is an indicator of howclose the thermal efficiency is to the highest value permissibledepending on the temperature range (Carnot efficiency), as inEq. (21):hii ¼hth1 ÀÀTlhx;min=Tuhx;minÁ : (21)The scroll expander effectiveness is generally measured usingisentropic efficiency. This is the ratio of actual power produced tothe ideal power that would be produced if there was no entropyincrease for the same pressure ratio:hs ¼_Wsh_m$Àhin À hout;sÁ; (22)where isentropic enthalpy, hout,s, is calculated using the Dymolaenthalpy function which requires two state variable inputs, beingthe outlet pressure and the inlet entropy.5. System performanceThis section aims to determine the performance of the scrollexpander, find the total energy generated by the system annually,and discern the effects on the system of varying key parameterssuch as ORC mass flow rate, collector area, and solar store volume.Nominal hot water temperature is imposed: Thw ¼ 65 C.The temperature of the solar store is the most important vari-able for which the performance of the ORC is dependent. Fig. 7shows the solar store temperature over one day, for the monthsof December, February, April and June. December and June are themonths with the highest and lowest solar exposure, respectively,as in Table 4. August and October are omitted because they arevery similar to April and February. The maximum expectedtemperature is 130 C, which confirms that the minimum workingpressure to prevent boiling in the solar thermal cycle using water is278 kPa.The scroll expander performance as a unit can be quantified byits isentropic efficiency, as described in Section 4.7. The expanderperforms at between 57% and 59% isentropic efficiency for allmonths of the year. This is a respectable efficiency for a powerexpander, however, isentropic efficiencies of up to 83% have beenreported in the literature [1,3,12]. Greater isentropic, and there-fore system, efficiency might be achieved by using a similar sizedscroll expander that has been purpose-built for expansion.Modifications to inlet and exit nozzles as in Mathias et al. [12]and improved lubrication as in Wang et al. [20] might helptoward this goal.The ORC performance is quantified by the peak expander power,_Wsh;max, rotational speed, Nmax, start and stop time, ton and toff,daily net energy produced, Ecyc, and total daily volume of hot waterproduced, Vprc. Table 8 shows the daily values for these variables forthe average day of each month. Refer to Fig. 9 for a visualisation ofthe shaft power evolution over one day. It follows that the totalannual energy produced by this system is 1710 kWh. At a standardrate for Brisbane of 22.7 c/kWh [21], this might equate to an annualvalue of $388.17. For the additional equipment required to form theORC (the pump, scroll expander and piping), this system wouldlikely pay for itself over its lifetime. With scroll expanderimprovements, the annual energy produced could significantlyincrease.The cycle efficiencies, as described in Section 4.7, do not varysignificantly over the course of the year. Thermal efficiency, hth,remains fairly constant at 3.47%, and the second-law efficiency, hii,varies over the course of each day between 15.7% and 23.2%. Fig. 8shows the daily trend for December, which has a similar shape toother months.The second-law efficiency is low, and there is potential forimprovement.It is evident that the pure electrical efficiency of this configu-ration is lower than that for PV. However, the low efficiency iscompensated for by the small costs associated with the additionalrequired equipment.5.1. Parameter sensitivityThe three parameters most sensitive to change are ORC massflow rate, _morg, collector area, Acol, and solar store volume, Vst.Table 5Solar fluid cycle parameters.Parameter Description ValueAcol Evacuated tube collector area 50 m2Vst Combined volume of solar thermal cycle 800 Lpst Pressure of solar cycle loop 300 kPaAUst Standing heat loss coefficient of store 2 W/KTable 6Cooling heat exchanger temperatures.Heat exchanger passage Temperature designationTorg,in TexTorg,out TsatTprc,in TambTprc,out ThwTable 7Dynamic testbed parameters.Parameter Description Value_mn Scroll expander nominal mass flow rate 0.07 kg/sTst,max Temperature maximum for expander activation 120 CTst,min Temperature minimum for expander deactivation 85 CDTpinch Heat exchange pinch point temperature difference 5 Cp1 ORC condenser pressure 850 kPap2 ORC evaporator pressure 1500 kPahpump ORC pump efficiency 0.6B. Twomey et al. / Applied Thermal Engineering 51 (2013) 1307e1316 1313
  8. 8. Table 9 shows some key effects of perturbing the former twoparameters for the month of December. Generally, _morg affectsthe ORC efficiency and is limited directly by solar store volume.Acol scales the entire system but is restricted by the available roofarea. For both of these parameters, an increase in value givesa modest increase in cycle energy, while only the collector areagreatly affects the hot water volume. These parameters could beadjusted to tailor the system to particular energy and waterneeds.The third parameter, Vst, has the most interesting effect.Increasing or decreasing it causes the performance curves to shiftbackward or forward in the time domain, as can be observed inFig. 9. Volume Vst can therefore be adjusted to match the electricitygeneration timeframe to the peak power usage, if appropriate. Thisis advantageous because generally the value of electricity is higherduring peak, and peak demand does not usually coincide with peaksolar irradiation. The drawback is that a larger and more costlystorage tank is needed, with better insulation in order to adhere toindustry standards. However, there will be less thermal mixing ina larger tank which could cause more distinct thermal stratification,benefiting the solar system’s efficiency.Fig. 6. Flow diagram of the dynamic simulation module.Fig. 7. Solar store temperature.Table 8System performance for the average day in each month.Month _Wsh;max(W)Nmax(rpm)ton(24 h)toff(24 h)Ecyc(kWh/day)Vprc(L/day)hsol (%)January 676 1050 0841 1844 6.05 3270 51.0February 663 1030 0908 1820 5.46 2950 51.4March 660 1024 0923 1751 4.99 2700 51.3April 660 1024 0946 1659 4.17 2270 51.2May 660 1024 1011 1605 3.34 1820 50.9June 660 1024 1031 1546 2.96 1610 50.3July 660 1024 1027 1607 3.20 1750 50.1August 660 1024 1001 1648 3.90 2120 50.1September 660 1024 0918 1732 4.86 2630 49.9October 664 1032 0847 1747 5.34 2890 50.4November 674 1047 0823 1812 5.89 3190 50.5December 675 1049 0824 1833 6.11 3300 50.8B. Twomey et al. / Applied Thermal Engineering 51 (2013) 1307e13161314
  9. 9. 6. ConclusionsA scroll expander model was calibrated and implemented intoa larger dynamic model of a solar thermal cogeneration systemusing evacuated tube collectors suited to some larger dwelling unitor small commercial establishment. Using a model presented byLemort et al. [13], the scroll expander calibration shows agreementgenerally within 10% for the shaft power, 5% for the rotational speedand 6 K for the exhaust temperature, with some outliers at verylow pressure ratios.When implemented into the dynamic simulation program, witha collector area of 50 m2and simulated over a year: The expander shows a maximum isentropic efficiency of 59% The ORC first-law thermal efficiency is 3.47% and second-lawefficiency varies between 15.7% and 23.2% The total energy produced is 1710 kWh and the total hot wateravailable is on average 2540 L/day The maximum instantaneous power observed is 676 W.Additionally, it was noted that it is possible to shift the timeperiod that the system is producing power to match the peakdemand period by adjusting the solar store volume.The expander isentropic efficiency is quite good, and could befurther improved by mechanical modification as noted in theliterature, such as inlet and exit nozzle improvements andimproved lubrication. The electrical efficiency metrics aregenerally quite low, and the total energy produced is notremarkable, but there is potential for improvement and theenergy produced is significant enough to potentially cover thelow costs of the ORC equipment over its lifetime. Also, thecapability to shift electricity production to match demand couldbe very advantageous in the solar energy market, where peakelectricity demand does not usually coincide with the day’s peaksolar irradiation.References[1] J.F. Oudkerk, S. Quoilin, V. Lemort, Evaluation of an ORC-based micro-CHPsystem involving a hermetic scroll expander, in: ORC (Ed.), First InternationalSeminar on ORC Power Systems, TU Delft, Netherlands, 2011.http://www.orc2011.nl.[2] G. Kosmadakis, D. Manolakos, G. Papadakis, Investigating the double-stageexpansion in a solar ORC, in: ORC (Ed.), First International Seminar on ORCPower Systems, TU Delft, Netherlands, 2011.http://www.orc2011.nl.[3] S. Clemente, D. Micheli, M. Reini, R. Taccani, Simulation model of anexperimental small scale ORC cogenerator, in: ORC (Ed.), First InternationalSeminar on ORC Power Systems, TU Delft, Netherlands, 2011.http://www.orc2011.nl.[4] D. Mikielewicz, J. Mikielewicz, J. Wajs, E. Ihnatowicz, Experiences fromoperation of different expansion devices in domestic micro ORC, in: ORC (Ed.),First International Seminar on ORC Power Systems, TU Delft, Netherlands,2011.http://www.orc2011.nl.[5] S. Quoilin, M. Orosz, H. Hemond, V. Lemort, Performance and design optimi-zation of a low-cost solar organic Rankine cycle for remote power generation,Solar Energy 85 (2011) 955e966.[6] J.J. Bao, L. Zhao, W.Z. Zhang, A novel auto-cascade low-temperature solarRankine cycle system for power generation, Solar Energy 85 (2011) 2710e2719.[7] G. Qiu, H. Liu, S. Riffat, Expanders for micro-CHP systems with organic Rankinecycle, Applied Thermal Engineering 31 (2011) 3301e3307.[8] S. Schimpf, K. Uitz, R. Span, Simulation of a solar assisted combined heatpump-organic Rankine cycle-system, in: World Renewable Energy Congress2011, Sweden.[9] T. Saitoh, N. Yamada, S. Wakashima, Solar Rankine cycle system usingscroll expander, Journal of Environment and Engineering 2 (2007)708e719.[10] D. Manolakos, G. Papadakis, S. Kyritsis, K. Bouzianas, Experimental evaluationof an autonomous low-temperature solar Rankine cycle system for reverseosmosis desalination, Desalination 203 (2007) 366e374. 5th Conference onDesalination Strategies in South Mediterranean Countries (EuroMed 2006),Montpellier, France, May 21e25, 2006.[11] A.C. Oliveira, A new look at the long-term performance of general solarthermal systems, Solar Energy 81 (2007) 1361e1368.[12] J.A. Mathias, J.R. Johnston Jr., J. Cao, D.K. Priedeman, R.N. Christensen, Exper-imental testing of gerotor and scroll expanders used in, and energetic andexergetic modeling of, an organic Rankine cycle, Journal of Energy ResourcesTechnology e Transactions of the ASME 131 (2009) 1e9.[13] V. Lemort, S. Quoilin, C. Cuevas, J. Lebrun, Testing and modeling a scrollexpander integrated into an organic Rankine cycle, Applied Thermal Engi-neering 29 (2009) 3094e3102.[14] B. Twomey, Experimental test results from QGECE laboratory small-scaleorganic Rankine cycle using a scroll expander, Technical Report 2012/03,School of Mechanical Engineering, University of Queensland, 2012, Availableonline: http://espace.library.uq.edu.au.[15] Dassault Systemes, Multi-engineering Modeling and Simulation (2012),Available online: http://www.3ds.com/products/catia/portfolio/dymola(accessed 29.02.12).[16] Standards Australia, AS 3500.4 Plumbing and Drainage e Heated WaterServices (2003).[17] Ecofys, Planning and Installing Solar Thermal Systems: a Guide for Installers,Architects and Engineers, Earthscan, 2005.[18] Bureau of Meteorology, Monthly Mean Daily Global Exposure (2011),Available online: http://www.bom.gov.au/climate/data/index.shtml (accessed24.11.11).Fig. 9. Vst parameter sweep.Fig. 8. Cycle efficiency for December. The y-axis is capped at 0.30, just above theCarnot efficiency.Table 9Parameter sensitivity.Parameter _Wsh;peak (W) Ecyc (kWh/day) Vprc (L) hsol (%)Reference 675 6.11 3300 50.8_morg ¼ 0:06 kg/s 561 5.44 3189 49.0_morg ¼ 0:08 kg/s 798 6.58 3356 51.8Acol ¼ 40 m2660 4.93 2682 51.6Acol ¼ 60 m2719 7.06 3799 48.7B. Twomey et al. / Applied Thermal Engineering 51 (2013) 1307e1316 1315
  10. 10. [19] Geoscience Australia, Geodesy e Astronomical Information (2005), Availableonline: http://www.ga.gov.au/bin/astro/sunrisenset (accessed 24.11.11).[20] H. Wang, R.B. Peterson, T. Herron, Experimental performance of a compliantscroll expander for an organic Rankine cycle, Proceedings of the Institutionof Mechanical Engineers Part A-Journal of Power and Energy 223 (2009)863e872.[21] Origin Energy, Electricity Tariffs (QLD) (2011), Available online: http://www.originenergy.com.au/2087/Electricity-tariffs-QLD (accessed 28.02.12).GlossaryA: area, m2c: specific heat, J/(kg K)E: energy, J, kWhh: specific enthalpy, J/(kg K)_m: mass flow rate, kg/m2N: speed, rev/minp: pressure, PaQ: heat energy, J_Q: heat power, Wr: ratio, eS: solar insolation, W/m2T: temperature, Ct: time, sU: heat transfer coefficient, W/(m2K)U: internal energy, Ju: specific internal energy, J/kgV: volume, m3, Lv: specific volume, m3/kg_W: power, WEx: daily solar exposure, kWh/m2/dayGreek symbolsD: difference, ee: error, eh: efficiency, er: density, kg/m3s: torque, NmSubscriptsamb: ambientcol: collectorcrit: criticalcyc: cycle netex: expander exhausthp: heating and powerhw: hot waterii: second lawin: expander internalleak: expander leaklhx: lower heat exchangermax: maximum/peakmin: minimumn: nominaloff: ORC stopon: ORC startorg: organic fluid streamp: constant pressurepinch: heat exchanger pinch pointprc: process hot water streamrise: sun risesat: saturationset: sun setsh: shaftsol: solarst: solar storestand: standing heat losssu: expander supplyth: thermalthr: throatuhx: upper heat exchangerv: volume propertyw: isothermal envelopeB. Twomey et al. / Applied Thermal Engineering 51 (2013) 1307e13161316

×