1-s2.0-S0038092X13002788-main

laen

all

nol

ilan

Received 13 May 2013; received in revised form 15 July 2013; accepted 16 July 2013Available online 15 August 2013

2013 Elsevier Ltd. All rights reserved.

aected by dispatchability issues. The growing share ofelectricity generated with these technologies demands forstorage capability in order to ensure continued and reliable

ious solar technologies is lively (Fthenakis et al., 2009; Price-WaterHouseCoopers, 2010). Peters et al. (2011) comparedPV- and CSP-based systems for large-scale solar powerplants (> 50 MWE), and concluded that the cost and e-ciency of storing energy can turn the competitiveness infavour of CSP systems. Another potentially important

Corresponding author. Tel.: +31 152782172.E-mail address: [email protected] (P. Colonna).

Available online at www.sciencedirect.com

Solar Energy 96 (2013Keywords: Thermal energy storage; Organic Rankine cycle (ORC); Complete ashing cycle (CFC)

1. Introduction

It is generally agreed that the clean energy revolution isrequired in order to nally decouple economic growth fromthe adverse environmental impact of the fossil-fuel econ-omy (Galiana and Green, 2009), and its geo-political impli-cations. Variable-output renewable energy technologiessuch as wind and solar PV power systems are negatively

supply of electricity (Inage, 2009). The feasibility of energystorage integrated into a conversion system for renewableenergy sources is a key enabler of such technology. Thedemand for a viable energy storage technology is particu-larly relevant for solar conversion systems, which, accord-ing to recent analysis, could provide up to one-third of theworld nal energy supply after 2060 (IEA, 2011).

The debate over the advantages and disadvantages of var-Communicated by: Associate Editor Robert Pitz-Paal

Abstract

The feasibility of energy storage is of paramount importance for solar power systems, to the point that it can be the technology enabler.Regarding concentrated solar power (CSP) systems, the implementation of thermal energy storage (TES) is arguably a key advantage oversystems based on photovoltaic (PV) technologies. The interest for highly ecient and modular CSP plants of small to medium capacity(5 kWE5 MWE) is growing: organic Rankine cycle (ORC) power systems stand out in terms of eciency, reliability and cost-eectivenessin such power-range. In this paper a thorough investigation on thermal storage systems tailored to high-temperature (300 C)ORCpowerplants is addressed rst, stemming from the observation that the direct storage of the ORCworking uid is eective thanks to its favourablethermodynamic properties. The concept of complete ashing cycle (CFC) is then introduced as a mean of achieving an unmatched systemlayout simplication, while preserving conversion eciency. This is a new variant of the Rankine cycle, whereby the vapour is produced bythrottling the organic working uid from liquid to saturated vapour conditions. The presentation and discussion of a case study follows: a100 kWECFC systemwith direct thermal energy storage, coupledwith state-of-the-art parabolic trough collectors. The proposed turbogen-erator achieves an estimated 25% eciency, which corresponds to a value of 18% in design conditions for the complete system. Consideringsiloxanes as working uids, the estimated values of storage density are around 10 kW eh=m

3storage.Thermal energy storage for socycle

E. Casati a,b, A. GaProcess and Energy Department, Delft University of Tech

bDipartimento di Energia, Politecnico di M0038-092X/$ - see front matter 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.solener.2013.07.013r-powered organic Rankinegines

i b, P. Colonna a,

ogy, Leghwaterstraat 44, 2628 CA Delft, The Netherlands

o, via Lambruschini 4, 20156 Milano, Italy

www.elsevier.com/locate/solener

) 205219

nerbenet ofCSP systems integratingTES, alongwith dispatch-ability, is their ability to provide grid exibility: this featuremight enable higher overall penetration of other variable-generation technologies such as those based on PV cells

Nomenclature

Acronyms

TES thermal energy storageORC organic Rankine cycleHTF heat transfer uidO&M operations and maintenanceSF solar eldHCE heat collecting elementDNI direct normal irradiation (W/m2)CSP concentrated solar powerPV photovoltaicDSG direct steam generationVLE vapour liquid equilibriumSM solar multipleSCA solar collector assemblyEEED Equiv. Elec. En. Density kW hE=m3st

Subscripts

E electricM mechanicalR reduced (w/r to critical value)amb ambient conditions

206 E. Casati et al. / Solar Eand wind turbines (Denholm and Mehos, 2011).Recent studies have underlined the techno- and socio-

economic opportunity of shifting toward a global energysystem which is more integrated and complex than pres-ently, and which heavily relies on distributed generation(IEA, 2012). Within the same context, also small-sizeCSP power plants in the 100 kWE5 MWE power rangehave been investigated (Price and Hassani, 2002; Prabhu,2006). It has therefore been argued that the new develop-ment paradigm of getting bigger by going smaller couldprovide a path to viability for CSP technologies in general,through modularity and economy of production, thusovercoming the bankability issue which is negatively aect-ing the sector (Skumanich, 2011).

Another notable advantage of thermal CSP plants fordistributed generation is the possibility of co-generatingelectricity and useful thermal output for maximum energyutilization and exibility. The thermal energy dischargefrom the primary mover can either be used for industrialor domestic purposes on-site, or/and drive an absorptionchiller for air-conditioning or process cooling (Balaraset al., 2007; Al-Sulaiman et al., 2011).

Among the technologies suitable for high-eciencyconversion of thermal power into electricity and heat inthis range of capacity, ORC turbogenerators stand outin terms of reliability and cost-eectiveness, see, e.g.,Angelino et al. (1984). ORC power plants are steadilyadopted for the increasing exploitation of geothermal res-ervoirs, while the growth of the number of ORC powersystems for the thermal conversion of biomass fuel(Obernberger and Gaia, 2005) and industrial waste heat

sv saturated vapourT thermalCR criticaldes design conditionsst storagesl saturated liquid

Symbols

p pressure (bar)h spec. enthalpy (kJ/kg)v spec. volume (m3/kg)q vapour quality (kgsv/kgtot)m mass (kg)nturn TES turnaround eciencygis isoentropic eciencyT temperature (C)u spec. int. energy (kJ/kg)q density (kg/m3)V volume (m3)hincid incidence angle ()gopt,p peak opt. eciency

gy 96 (2013) 205219is remarkable. ORC-based CSP plants have been widelystudied, and prototypes were put into operation alreadyseveral years ago (DAmelio, 1935; Verneau, 1978); com-mercial power plants went recently on-line (Canada et al.,2006), and new ones are planned or are currently underconstruction.

To the knowledge of the authors, however, no researchhas been published on TES systems specically conceivedto be integrated into ORC power plants. The study docu-mented here stems from the need of a thorough investiga-tion on thermal storage systems tailored to ORC powerplants, and from the observation that the direct storageof high-temperature ORC working uids is eective thanksto its high heat capacity and other favourable thermody-namic properties.

The paper is structured as follows: in Section 2 theorganic uids of the class of siloxanes are briey intro-duced; these compounds are widely used as working mediafor high temperature ORC power systems (turbine inlettemperature 300 C). In Section 3 a brief overview ofTES-systems integration into power stations is reported.Section 4 deals in detail with the proposed technical solu-tions for the direct thermal storage of working uid, dis-cussing their applicability to ORC systems. A case studyillustrating the application of one of the proposed storagesystems is treated in Section 5. Section 6 summarizes theconclusions and the foreseen developments.

2. Siloxanes as high-temperature ORC working uids

The ORC working uids considered in this study belongto the family of siloxanes, see Table 1. These light siliconoils are already employed in commercial high-temperatureORC applications since they are non-toxic, environmen-tally friendly, low-ammable, bulk-produced and highlythermally stable; mixtures of siloxanes are widely employedas heat transfer uids (HTF) in multiple elds, comprisingthe CSP industry (Angelino and Invernizzi, 1993; DOW,2012). Multiparameter equations of state, employing theSpan-Wagner functional form (Span and Wagner, 2003),

of thermodynamic cycles is presented here, whereby thecycle minimum and maximum temperature are xed, anddierent working uids are evaluated in terms of obtain-able conversion eciency and other technological aspects.Note that xing the minimum and maximum cycle temper-ature results also in the specication of the minimum andmaximum cycle (saturation) pressure for the considereduid. The expansion ratio available for work extraction isthus also xed.

From these preliminary considerations, the comparisonof the thermodynamic features of molecularly complex u-ids with those of water along expansions yields interesting

Table 1Main properties of the uids considered in this work. MW: molecular weight, Tboil: normal boiling temperature, pvap@80C: vapour pressure at 80 C,cp@T st;R0:9 : specic heat capacity at constant pressure for saturated liquid at reduced temperature TR = T/TCR = 0.95.

Fluid MW (g/mol) TCR (C) pCR (bar) qCR (kg/m3) Tboil (C) pvap@80C (bar) cp@TR0:95 (kj/kg/K)

Water 18.0 373.9 220.64 322.4 100.0 0.474 12.0D6 444.9 372.7 9.61 246.8 245.0 0.002 2.6D4 296.6 313.3 13.32 301.3 175.3 0.035 2.6MDM 236.5 290.9 14.15 302.9 152.5 0.091 3.1

E. Casati et al. / Solar Energy 96 (2013) 205219 207have been recently developed for these uids (Colonnaet al., 2006, 2008). These thermodynamics models, imple-mented in a software library, are adopted throughout thiswork (Colonna et al., 2012; Lemmon et al., 2010). The spe-cic thermophysical properties of the working uids heav-ily aect the design of the most critical components,namely the turbine and the heat exchangers. Other uidsmay be preferred for the same application, based uponmultiple considerations, see e.g. Tabor and Bronicki(1964) and Angelino et al. (1984). Fig. 1 illustrates the ther-modynamic features of interest in this case for siloxane D4and water using the T s state diagram. A typical analysisFig. 1. Comparison between the T s thermodynamic diagram of D4 and thaTmin = 80 C, Tmax = 250 C, and Tash = 220 C (dash-dotted black lines). CRVLE region, solid-grey: isobaric lines, dashed-black: iso-enthalpy lines. Note tdiagram (b).conclusions:

I A consequence of the large complexity of the uidmolecular structure, thus of the high value of the specicheat capacity, is the so-called retrograde shape (positiveslope) of the bubble line in the T s diagram of theuid, which helps visualizing how the expansion of thesaturated vapour is inherently dry, see, e.g., Colonnaand Guardone (2006). This thermodynamic featureimplies, contrary to what can be observed for a sim-ple-molecule uid like water, that for a complex organicuidt of water: the highlighted temperature levels, common to both uids, are: liquidvapour critical point, solid-black line: saturation line enclosing thehat the scale of the specic entropy in diagram (a) is dierent from that in

(a) Starting from saturated liquid conditions, see statec in Fig. 1(a), an isenthalpic pressure-reductionrepresentative of a ashing process can result inthe uid being in a saturatedvapour state (processc! d, whereby qd = 1). If the pressure is furtherreduced, also superheated-vapour states areattainable.

(b) The previous observation, applies also to an isen-tropic pressure-reduction process (c! e0).

(c) An isentropic expansion starting from saturatedvapour conditions always evolves towards super-heatedvapour states (dvs ! e).

(d) Furthermore, the temperature of the superheatedvapour at the end of the expansion may be so highthat internal heat-regeneration is mandatory ifhigh cycle eciency is required (Invernizzi et al.,2007).

3. Concepts of TES systems for power plants

The basic principle of storage system integration intopower plants is the so-called ow-storage, whereby theTES gets its charge according to several main concepts,corresponding to the plant congurations summarized inFig. 2(a). In solar power plants, storage systems deal withsecondary energy since, as opposed to fuelled thermalplants, storage on the primary energy side (fuel storage)is not possible. Fuel control would be feasible (defocusingof heliostats or collectors), but it is avoided because of theenergy loss it entails. Energy storage integrated into the pri-mary heat transfer loop, see case a of Fig. 2(a), is the mostadopted concept in commercial CSP plants. In this case,sensible heat is accumulated into a liquid, which can bethermal oil and/or molten salt (Medrano et al., 2010); inso-called direct systems the heat transfer uid serves alsoas storage medium, while in indirect systems a separatedsystem is used to store thermal energy, see Fig. 2(b).

208 E. Casati et al. / Solar Energy 96 (2013) 205219II Mainly as a consequence of the higher molecularweight, a lower specic enthalpy drop is associatedwith the given expansion. Thus, the working uidmass-ow through the expander must be larger forthe same power output. In combination, the low spe-cic enthalpy drop and the higher mass-ow rateallow for the realization of comparatively simpleand ecient turbines even for low or very lowpower outputs (Tabor and Bronicki, 1964; Verneau,1978).

III The saturation pressure corresponding to the maxi-mum cycle temperature is lower (e.g., 4.9 bar for D4versus 39.8 bar for water): this is a major advantageif TES is of interest. Conversely, very low condensa-tion pressure (e.g., 0.035 bar for D4 versus 0.47 barfor water) entails technological challenges for othercomponents (e.g., turbines and condenser), see Ange-lino et al. (1984).Fig. 2. Thermal energy storage inThe Spanish Andasol solar power plants, which are inoperation since 2009, are representative of the state-of-the-art for technology based on parabolic troughs (SolarMillenium, 2009). They adopt an indirect thermal storagesystem whereby thermal oil transfers the energy collectedfrom the solar eld to molten salt contained in two tanks,see Fig. 2(b); such layout, involving multiple subsystemswith dierent working uids, is arguably unfeasible forsmall-scale solar power plants due to complexity and cost.

Other concepts, see cases b, c, and d in Fig. 2(a), arebased on the storage of energy in the working uid itself,and have been implemented in steam power plants. Ther-mal storage can be used to make pre-heated feed wateravailable to the steam generator (b), mainly for peakingpurposes, in power plants with regenerative feed-waterheating (Gilli and Beckmann, 1980). The storage vesselcan also supply the turbine with steam in saturated orRankine-cycle power plants.

highly integrated into the power plant, both the storage

et al., 2005). Silicon oils have been already adopted as

the bottom of the vessel, causing the vapour volume to

nersuperheated state, at both live-steam (c) or medium/lowpressure (d) conditions. Also in these cases both indirectand direct system concepts can be implemented.

Indirect systems for working uid storage are exten-sively investigated as TES for direct steam generation(DSG) power plants (Feldho et al., 2012). Direct systemshave been successfully used for decades, and also recentlybuilt CSP plants adopt this TES conguration. A directsystem for the accumulation of working uid is often calledsteam accumulator (Beckam and Gilli, 1984; Medranoet al., 2010). The main advantage of steam accumulatorsis that they are simpler than indirect systems, in that nointermediate uid loop and the related heat exchangersare needed; Section 4 treats in detail these concepts.

The evaluation of the protability of energy systems is acomplicated task, involving a number of considerationsfrom dierent domains. Among the main elements of theevaluation one needs to consider the projected investmentcost and eciency of the complete system, possible envi-ronmental hazards, operation strategy, O&M cost, andthe local regulatory framework, i.e., taris and/or incen-tives (Pilkington Solar, 2000). In particular, when newlyproposed concepts are considered, the uncertainty relatedto equipment costs has a large impact on the reliability ofsuch evaluation (Henchoz et al., 2012). Also for this rea-son, an exhaustive economic evaluation is beyond thescope of the present work, which in turn is aimed at thethermodynamic and technical assessment of a new conceptfor thermal energy storage suitable for small solar-poweredORC plants. The analysis identies and discusses the fac-tors aecting the performance and the projected costs ofthe considered systems, such as their eciency, the temper-ature and pressure levels in the storage system, the volu-metric expansion ratio across the turbine, and thepressure level in the condenser.

The typical thermodynamic performance parameters fora TES system integrated into a thermal power plant are.

1. The storage density qex kW hM=m3st

, which is useful toevaluate the size of the storage unit and, thus, to give arst estimate of its cost. Since thermal energy is storedfor subsequent conversion into work, density of avail-able energy (exergy) has to be considered (Bejan,1982). The parameter EEED (Equivalent ElectricalEnergy Density kW hE=m

3st

), accounting for the subse-

quent conversion into electricity, is also introduced here.This value quanties the equivalent electrical energystored as thermal energy into one cubic meter of liquidat the storage conditions. For a given size of the storagein terms of equivalent hours of storage heq,st, the value ofthe EEED of a certain TES concept allows for a preli-minary estimation of the storage volume and of therequired mass of uid.

2. The turnaround eciency nturn, which accounts for exer-gy losses along the entire charge-standstill-discharge

E. Casati et al. / Solar Ecycle, and depends both on design and on operationalparameters. nturn is typically chosen as the objective var-increase. Additional vapour is produced by evapourationof a small part of the liquid volume, thus causing the pres-sure to decrease slightly. The drains coming back from theworking uid loop have to be collected and stored in a sep-arate cold-storage vessel, which, being at lower tempera-ture and pressure, is also relatively inexpensive. Withrespect to the displacement storage solution (A), the com-the storage uid in these systems (Gil et al., 2010). If thedirect-storage conguration is adopted, pressurization isneeded in order to prevent boiling, and an external pressur-izer may be needed in this case. Hot pressurized uid canthus be extracted from the storage vessel at constant pres-sure, and this is the main advantage in power generationapplications.

4.1.2. Expansion storage at almost constant pressure

In this case liquid and vapour working uid are stored inthermodynamic equilibrium at the saturation temperature.A vapour cushion is present at all times in the upper partof the storage tank. Hot saturated liquid is extracted from4.1.1. Storage at constant pressure

It entails the storage of sensible heat in liquids, usuallyat atmospheric pressure. Two-tank arrangements, as wellas single-tank systems exploiting the thermocline eect (dis-placement storage) are feasible (Kandari, 1990; Brosseauconcept and its discharge mode have to be considered inorder to properly characterize the system. Three main stor-age concepts and three discharge methods have been intro-duced in the past (Beckam and Gilli, 1984): these aredescribed in Sections 4.1 and 4.2 respectively. Section 4.3treats their combination to form several possible storagesystems, refer to Fig. 3.

4.1. Storage methods

The working uid is typically stored in the liquid phase,in order to exploit its greater storage density.iable for the thermodynamic optimization of a storagesystem (Krane, 1987). Systems implementing the directstorage of the working uid attain the highest levels ofnturn, namely up to 95%, mainly as a consequence ofthe absence of any heat exchange process external tothe storage vessel (Beckam and Gilli, 1984).

4. Direct storage of working uid in Rankine power stations

The concepts originally proposed and adopted in steampower plants are introduced here, and their extension toORC systems is discussed. Being direct storage systems

gy 96 (2013) 205219 209plications related to pressurization and thermocline pro-motion can be avoided and the vessel does not have to

planor t

nerwithstand severe thermal gradients during the chargedis-charge phases.

4.1.3. Sliding pressure storage (Ruths accumulator)

Fig. 3. Main congurations of direct storage systems for steam powercorrespond to those adopted for the description in Sections 4.1 and 4.2. Fwithout internal regeneration. Adapted from Beckam and Gilli (1984).

210 E. Casati et al. / Solar EIn analogy with method B, liquidvapour equilibrium ismaintained in the storage vessel. In this case, however, notthe liquid but the vapour forming the cushion is extractedduring the discharge phase. The wide pressure swing duringthe discharge phase, a characteristic of this method, is amajor drawback as far as power production is concerned(Goldstern, 1970). The main advantage of this storagemethod is the fast reaction time, allowing for high dis-charge-rates of saturated steam.

4.2. Discharge methods

4.2.1. Vapour generation by ash evaporation

Internal ashing in the vessel pertains to the sliding pres-sure method of storage. In external ash processes theliquid is extracted from the storage tank, and thus throt-tled; processes featuring multiple ashing steps are alsoconceivable. The obtained vapour stream can be sentdirectly to a turbine.

In case water is the working uid, ashing systemsrequire the adoption of so-called wet-turbines, which implywell-known technical challenges and rather low eciency(Dipippo, 2008). However, in case the working uid is anorganic compound, the ashed saturated vapour can bedirectly fed to a high eciency dry turbine (see Section 2,point Ic). The so-called retrograde characteristic of theworking uid allows also for the complete evaporation ofthe liquid stream by throttling (see Section 2, Ia). Thephase-separator and the relative liquid-drain circuit aretherefore, in principle, unnecessary.

ts, as a combination of storage methods and discharge modes. Labelshe sake of simplicity, only single-stage ash ORC systems are considered,

gy 96 (2013) 2052194.2.2. Feed water storageIn case of conventional thermal power stations, thermal

storage upstream of the steam generator is a proven solu-tion for peak-load generation. With such a method, peak-load can be sustained to an extent limited by the amountof power to be gained by cutting-o all the regenerativebleeds, and by the overload capacity of the main turbinegenerator set (Beckam and Gilli, 1984). In the case ofORC power systems, extractive regeneration is neveremployed, therefore the peaking potential would be dueexclusively to the overload capacity.

This discharge method is not applicable to solar steampower plants as the only storage system; in periods withlow solar radiation, feeding the turbine only with steamcan become impossible (Gilli and Beckam, 1977).

4.2.3. Cascading storageA combination of method 1 and method 2 provides

more exibility for the complete system.In case the working uid is formed by complex mole-

cules, a fourth discharge method can be identied, namely,

4.2.4. Direct liquid expansion

The liquid extracted from the pressurized storage vesselcan be directly fed to an expander. If the working uid is acomplex organic molecule, the so-called wet-to-dryexpansion process becomes possible (see Section 2, Ib).

Wet-to-dry expanders have been proposed and tested withpromising results (Elliott, 1982; Dipippo, 2008). However,since none of them has reached technological maturity,wet-to-dry expansion has not been considered in this study,despite its notable potential.

4.3. Storage systems

Fig. 3 shows the main possible system congurationsobtained by combining the storage methods described inSection 4.1 (A, B, and C), and the discharge methods trea-ted in Section 4.2 (1, 2, and 3), see (Beckam and Gilli,1984). The congurations C2 and C3 are not realizable,while conguration B and C may be combined: vapourmay be taken from an expansion storage vessel in addition

(see Section 3) which are lower than those of direct water-steam systems, and also of state-of-the-art indirect systems(see Section 5). In case the working uid is a siloxane, for agiven power output, the same thermal storage capacityrequires a larger vessel, if compared to a steam powerplant.

It is worth noting that the problems related to uid con-tainment at concurrently high-pressure and high-tempera-ture levels, which have ultimately hindered the diusionof water-steam storage systems, are reduced in case work-ing uid is an organic compound (Section 2, III). Besidethe vessel volume and the pressurization level, also the costof the uid largely contributes to the total investment costof the storage system. At present, siloxanes are approxi-mately two times more expensive than synthetic oils (typi-cally mixtures of diphenyl-diphenyl ether), in terms of cost

conguration allows for a substantial simplication of theoverall layout of the plant, with a benecial eect on its ini-

t

M

rdin

E. Casati et al. / Solar Energy 96 (2013) 205219 211b b1

(a)Fig. 4. (a) simplied plant layout of a CSP ORC power plant working accoto liquid (shown by the lines for congurations B2/C1 andB3/C1).

The A1 scheme has been proposed for nuclear powerplants (Beckam and Gilli, 1984), while the B2/C1 schemefor CSP plants (Gilli and Beckam, 1977). The A3 schemegained acceptance in the late 1920s: the displacement stor-age plant of the coal-red power station in Mannheim,Germany, is well known (Goldstern, 1970). The C1 scheme(pure sliding pressure) found wider application, mainly as asolution for buer-storage. Within the island grid of Berlin,the 50 MWE Charlottenburg plant built in 1929 - hasbeen operated with steam accumulators of 67 MWhE stor-age-capacity for more than 60 years. Sliding-pressure sys-tems have recently been realized (EU, 2006), and thisconcept has been proposed as a solution to supply DSGplants with buer-storage capabilities (Steinmann andEck, 2006). Two-phase refrigerant accumulators workingaccording to this principle are key components in automo-tive air conditioning systems (Li et al., 2011).

It can thus be concluded that all the above mentionedconcepts are applicable in principle to ORC power systems.However, applying the very same concepts for thermalenergy storage to ORC power plants leads to EEED levels

c c

d(q d=1)

ae

f2

Solar field (SF) TES system ORC planthe displacement-storage type, from Casati et al., 2012. (b) cycle state pointslines). CR: liquidvapour critical point, solid-grey line: contour of the vapourtial cost.

5. Case study

In order to evaluate the proposed integrated TES systemfor small-scale solar ORC power plant, the system ofFig. 4(a) _W net 100 kWE has been studied.

The ORC working uid is circulated and heated in theSF, which is composed of parabolic trough collectors withevacuated absorber tubes: the feasibility of such concepthas been preliminarily assessed in a recent study (Casati

(b)s [kJKg-1 oC-1]

T[o C

]

-0.4 -0.2 0 0.2 0.4 0.6 0.8

100

150

200

250

300

350CR

c

d(qd=1)

f

h=h(c)

a

e

b

g to the single-stage ash process, integrating a direct TES system based onper unit of thermal energy delivered (Pilkington Solar,2000). However, contrary to synthetic oils, siloxanes areclassied as non-hazardous materials. Such classicationis expected to play an important role if the proposed tech-nology will be applied, particularly if the distributed energyscenario is considered.

The cost of the storage system is however only a fractionof the nal investment for a power plant. The case studypresented in section Section 5 shows that the direct-storagein the T s thermodynamic diagram of D4 (black points and solid-blackliquid equilibrium region, dashed: iso-enthalpy line.

et al., 2011). The main novelty is the adoption of one of theTES systems introduced in Section 4: the working uidserves also as the storage medium, making the congura-tion completely of the direct type. The selected TES systemis based on a displacement-type storage, with vapour gen-eration through external ashing (type A1 in Fig. 3).

The concept is aimed at maximizing the simplicity of theplant layout, since lowering of initial cost and maintenancerequirements, as well as ease and safety of remotely con-trolled operation, are considered as key aspects for distrib-uted power applications.

The design data adopted here are reported in Table 2:the general specications are common to all the ORCplants modelled in this study (see also Appendices A andB). A relatively high value of the condensing temperatureTcond is chosen (80 C), since avoiding excessively low vac-uum levels in the condenser is mandatory in high tempera-ture applications, because the presence of air due to inwardleaking accelerates the thermal degradation of the workinguid.

The data specic to the proposed exemplary system interms of uid and operating conditions, have been deter-mined based on the treatment described in Appendices Aand B, where the main trade-o existing between system

Table 2Design data for steady-state modelling, common to all the simulatedsystems (see also Appendices A and B). For a detailed description of theadopted HCE and SCA technologies, see Fernandez-Garca et al. (2010)and Burkholder and Kutscher (2009). Dpb0c0 : pressure drop in the SF,

212 E. Casati et al. / Solar EnerDpp,cond: pinch point temperature dierence in the condenser, reg:regenerator eectiveness, gME: electro-mechanical eciency of thegenerator and of all the electrical motors. Dpef, Dpfa, Dpa0b, and DpFan:pressure drops in the regenerator (vapour side), in the condenser (processside), in the regenerator (liquid side), and in the condenser (static, air side)respectively.

Solar eld design data

HCE Schott PTR-70 SCA ET-150gopt,p 0.75 DNIdes 850 (W/m

2)Hincid 0 () Tamb 25 ()Vwind 0 (m/s) Dpb0c0 1 (bar)SM 1

Main design data for the ORC plants_W net 100 (kWE) heq,st 4 (h)Tcond 80 (C) Dpp,cond 15 (C)reg 0.85 gis,turbine 0.85gis,pumpsfans 0.75 gME 0.97Dpef 50 (%pcond) Dpfa 10 (%pcond)Dpa0b 0 (bar) DpFan 50 (Pa)

Design data for the exemplary system

Fluid D4 pcond 0.035 (bar)Tc,R = Tst,R 0.998 Tc = Tst 312.6 (C)pc = pst 14.2 (bar)

Calculated design performance of the exemplary system

_mfluid 1.81 (kg/s) _mair 8.61 (kg/s)gORC 0.251 gglob,SF 0.71gglob,sys 0.18 Amirrors 704 (m

2)VR 246 EEED 6.2 kW h =m3 turb E storageVst 65 (m

3) muid,st 3E4 (kg)eciency, plant simplication, and components designare discussed in detail. Realistic assumptions regardingboth the design of the dry air-cooled condenser and ofthe plates-regenerator are considered. This informationhas been obtained from the preliminary design of thesecomponents, performed with a commercial package forheat exchanger design (Aspen, 2007).

5.1. Working principle

In nominal conditions, the temperatures at the outlet ofthe solar eld Tout,SF, and in the hot region of the storagevessel TST,hot (which, in turn, equals that of the uid fed tothe ORC system), are considered to be both equal to Tc.For modelling purposes this is chosen as the main operat-ing variable (see Appendices A and B), while pc is supposedto be maintained at a level higher than the correspondingvapour pressure, by an external pressurizer (1 bar in designconditions).

Cold uid is extracted from the vessel (b) and pumpedthrough the SF: under normal operating conditions themass ow is controlled by acting on the pump in order tomaintain a set outlet temperature Tc. Also in this case thetemperature at the outlet of the regenerator Tout,ORC, andthat of the stored cold uid TST,cold, are assumed to beequal to Tb. The hot uid extracted from the storage vessel(c) is externally ashed to saturated vapour conditions,before being fed to the ORC turbogenerator (d, withqd = 1, see Section Ia). The superheated vapour leavingthe turbine enters the regenerator (e), and then the con-denser (f). The uid, in saturated liquid conditions (a), isthen pumped back, through the regenerator, to the bottompart of the storage vessel (b).

The mass ow circulating in the SF is determined by theavailable irradiation together with the area of the collectorswhich, in turn, is related to the chosen solar multiple (SM).Optimal combinations of SM and storage-vessel size can bedetermined only through detailed techno-economic optimi-zation (Gilli and Beckam, 1977; Cabello et al., 2011). Thevalues adopted here have therefore to be considered asindicative.

5.2. Flashing Rankine cycles with organic uids

The main implication of the working principle presentedin Section 5.1 is that the plant always operates according toa thermodynamic cycle which includes a ashing evapora-tion process while, usually, the ash process is adoptedonly when the storage is being discharged (Beckam andGilli, 1984). Referring to Fig. 1, the ashing cycle (FC)of the working uid in the temperature-entropy diagramis identied by the state points a, b, c, d (with qd = 1), e,f (1, 2, 3, 3vs, 4 for water). When evaluated for the exploi-tation of thermal energy sources whose thermal capacitycan be assumed as innite, such power cycles feature an

gy 96 (2013) 205219inherently lower eciency compared to the correspondingevaporative cycle operating between the same maximum

els r

)

291.8 0.56 291.8 0.57 1

f 87.6 0.04 2.551

nerdecouples the SF and the ORC power block by means ofa suitable direct thermal storage system, see Fig. 4(a). Aminor eciency reduction can thus be accepted, in viewof the substantial simplication it allows for, both in termsof plant layout and operation.

5.3. Flashing the organic vapour down to saturated

conditions

A further simplication of the plant congurationderives from the possibility of reaching complete vapouri-zation of the working uid by ashing (see Section 2, pointIa). In this way several components become redundant,namely the ashing vessel and the liquid drain circuit.More details are provided in Appendix B. To the authorsknowledge, ORC power systems working according to theashing cycle principle, whereby the working uid is throt-tled down to saturated vapour conditions before enteringand minimum temperature (state points a, b, c, g, h, f, and1, 2, 5, 6 for water) (Dipippo, 2008). However, if the work-ing uids is an organic compound, it can be shown that theeciency penalty aecting the ashing cycle may be com-paratively low. A detailed treatment is reported in Appen-dix A. Flashing ORC power systems for waste-heatrecovery applications have been recently investigated byHo et al. (2012).

The ashing cycle boasts notable benets in case of asolar ORC power system with thermal storage: (i) it avoidsphase transition in the SF, with major advantages (Odehet al., 2000; Eck et al., 2004; Casati et al., 2011); (ii) it

1 30.0 1.01 0.8802 67.0 1.01 1.052Table 3Thermodynamic properties of the state points of the ORC system. State lab1 and 2 refer to the cooling air stream.

state T (C) p (bar) v (m3/kg

a 80.0 0.04 0.001b 205.6 14.20 0.001c 312.7 14.20 0.002d 283.2 8.49 0.011e 234.9 0.06 2.489

E. Casati et al. / Solar Ethe turbine, refer to Fig. 4(b), have not been consideredbefore, thus this concept is named here complete ashingcycle (CFC).

5.4. Results

The steady state modelling of the system is performedwith an in-house code implemented in a well known com-puter language for technical computing (MathWorks,2011), coupled with an in-house library for the accurateestimation of the thermophysical properties of the uids(Colonna et al., 2012). The calculated performance isreported in Table 2, while Table 3 shows the thermody-namic properties of the state points of the thermodynamiccycle. Notwithstanding the selected high design value forTcond, the calculated eciency of the ORC power systemexceeds 25% which, combined with the eciency of theSF, yields a global eciency in design conditions close to18%. This value can be compared to the measured valuesof recently-built state-of-the-art CSP plants. These steampower plants are much larger, and adopt an indirect stor-age system with synthetic oil as HTF, and their eciencyis of the order of 22% (Giostri et al., 2012).

Even if no index of annual performance has been esti-mated yet, ORC power systems are characterized by excel-lent o-design performance. This characteristic canpartially overcome the lower design eciency in a highlydynamic application such as CSP (Casati et al., 2011).

The calculated values of equivalent electrical energydensity (EEED) storage are lower than those characterizingtraditional TES solutions, and this holds for all the consid-ered working uids (see Fig. A.7). The proposed systemapproaches, for the EEED, the limiting value of 6.2kW hE=m3ST, see Table 2, assuming that the storage vesseldelivers its full energy content without any variation in thedischarged uid properties (conditions corresponding tostate c). Thermal losses, as well as exergy losses due to dete-rioration of the stratication (Haller et al., 2010) are thusneglected: such simplications are typically justied fordaily charge-discharge cycles (relatively short standstilltimes). A recently designed displacement storage systemusing synthetic oil as HTF, and proposed as an add-onto the APS Saguaro ORC-based CSP plant (Canadaet al., 2006), reaches the value of approximately 15

233.6 0.59 4.5 0.07 0.0 37.2 efers to the layout of Fig. 4(a) and to the T s diagram of Fig. 1(a). States

h (kJ/kg) s (kJ/kg/K) q (kgsv/kgtot)

172.6 0.43 059.1 0.12

gy 96 (2013) 205219 213kW hE=m3ST (Kolb and Hassani, 2006). The lower valuecalculated for the proposed system is mainly due to thelow specic work extracted from the turbine, which causesthe uid to be injected back in the storage vessel at hightemperature (Tb Tout,ORC).

6. Conclusions

This paper documents a study about extending directow-storage methods applicable to steam power plants toORC power systems. So-called direct thermal storage sys-tems are feasible, whereby the same uid is circulated inthe heat source, serves as thermal storage medium, and isalso the working uid of the ORC turbogenerator. A case

study regarding a 100 kWE solar plant implementing suchconcept is presented: the proposed system features a con-stant-pressure thermocline storage system, with vapourgeneration through external ashing of the liquid extractedfrom the storage vessel.

The thermal storage system can be integrated into theplant, thus decoupling the thermal energy source fromthe ORC power block: the system can be classied as con-stant-parameters storage, whereby the uid enters andleaves the vessel (in principle) in the same thermodynamiccondition, see states c and b in Fig. 4(a). Apart from a sub-stantial simplications in terms of both plant layout andoperational strategy, this conguration ensures high exer-getic performance of the thermal charge and discharge pro-cesses (Beckam and Gilli, 1984).

investigating the binary cycle conguration, whereby directthermal storage is implemented in the topping cycle.

Acknowledgements

This research is supported by the Dutch TechnologyFoundation STW, Applied Science Division of NWO andthe Technology Program of the Ministry of Economic Af-fairs, grant # 11143. The authors thankfully acknowledgethe fruitful discussions with their friends and colleaguesRyan Nannan, Francesco Casella, and Adriano Desideri.

PSF b

214 E. Casati et al. / Solar Energy 96 (2013) 205219The power cycle operates according to a variant of theRankine cycle, whereby a ashing evaporation process pre-cedes the power-generating expansion of the working uid.The properties of the adopted complex-molecule workinguids are such that ashing can lead to saturated or super-heated vapour conditions. This characteristic implies fur-ther simplications of the system if compared toconventional steam power plant systems with thermal stor-age. The eciency penalty of an ORC power plant workingaccording to the newly introduced complete ashing cycle(CFC) may be small, under the described assumptions, ifcompared to the eciency of a conventional evaporativeORC power system.

A design value of the solar-to-electric eciency of 18%is calculated for the exemplary 100 kWE solar ORC powersystem with direct thermal storage and the ashing cycleconguration, while no index of annual performance hasbeen evaluated yet. The advantages in terms of simplica-tion of the plant layout could overcome the relatively lowvalues of storage densities, the need of pressurization,and the specic cost of the uids. A detailed techno-eco-nomic analysis of the proposed system aimed at clarifyingthese open questions will be developed as the next step ofthe project. In order to improve system performance, par-ticularly in terms of storage density, it might be worth

PSFb

g

b

m

Solar field (SF) ORC plant

(a) EC

Fig. A.5. (a) simplied plant layouts of ORC power systems working accordState points correspond to those in the T s chart of Fig. 1(a).b

(b) FCcdvs

d ls

dls

d

m liq

mvap

Solar field (SF) Flash system ORC plant

PflashAppendix A. Thermodynamic comparison between ashing

and evaporative organic Rankine cycles

This section presents the thermodynamic evaluation ofthe steady-state evaporative (EC) and ashing (FC) cyclecongurations, as introduced in Section 5. The eect of dif-ferent working uids is also addressed. The general designdata, common to all the modeled ORC systems, are thosereported in Table 2, and Fig. A.5 shows the conceptualplant layouts of the considered systems. The working prin-ciple is the same described in Section 5, whereby the ORCpower block shown in Fig. 4(a) has been lumped here intoa single component, and no storage system is considered.The thermodynamic evaluation is carried on by varyingthe maximum temperature of the cycle, which is kept thesame, i.e., Tmax Tc = Tg. It is further assumed that:

1. In the EC the working uid exits from the thermalenergy source (the solar eld) as saturated vapour atTmax, i.e., state point g in Fig. A.5(a);

2. In the FC the working uid exits from the thermalsource (c) in the state of saturated liquid at Tmax, andthen undergoes the ashing process. It is assumed herethat the ash evaporation leads to saturated vapour con-ditions at the outlet of the ashing subsystem (processc! d, where qd = 1): a critical assessment of thisassumption is presented in Appendix B. As a conse-quence, no liquid drains from the ash vessel have tobe recirculated _mliq 0.ing to the conventional evaporative cycle, and (b) single-stage ash cycle.

From these assumptions follows that, for each value ofTmax, all the state points dening the two thermodynamiccycles can be determined for a given working uid. Notethat the condensation temperature Tcond is xed and spec-ied. The results of the steady state simulations, performedwith an in-house code implemented in a well-known lan-guage for technical computing (MathWorks, 2011), arepresented in Figs. A.6 and A.7.

The main term of comparison is the global system e-ciency gglob,SYS, dened as

gglob;SYS gORC gglob;SF; A:1where

gORC _W net= _Qin;ORC A:2is the thermal eciency of the ORC system._W net _W turbine _W aux is the electrical power output ofthe plant, decreased of the power consumption for auxilia-

_

collectors. Having assumed a null incidence angle for de-

Tc,R [-]

VR

turb

ine

[-]

0.7 0.75 0.8 0.85 0.9 0.95 1100

101

102

103

104

D6-ECD6-FCD4-ECD4-FCMDM-ECMDM-FC

as a function of Tmax,R.(b) Turbine volume tric expansion ratio VR turb

Fig. A.7. Comparison between ashing organic Rankine cycles fordierent working uids. Throttling down to saturated vapour conditionsis assumed. Equivalent electric energy density EEED as a function ofmaximum cycle reduced temperature Tmax,R.

E. Casati et al. / Solar Energy 96 (2013) 205219 215ries; W net is constrained to be the same for all the simulatedcases. _W aux is obtained by summing the power consump-tion of all the pumps (subscripts P in Eq. (A.3)) and thefans in the system and it is therefore evaluated as

_W aux _W PORC _W PSF _W Pflash _W Fan: A:3_W Pflash in Eq. (A.3) is zero for both the EC and the FC sys-tems (in this last case by virtue of Assumption 2), since noliquid drains from the ash are present. _Qin;ORC is the ther-mal power supplied to the ORC system and, for the FC, itreads

_Qin;ORCFC _mvap hc h0b _mliq hc hd0ls: A:4Here the 2nd term in the right-hand side vanishes because ofAssumption 2. In the EC case, Eq. (A.4) becomes

_Qin;ORCEC _m hg hb0 : A:5

Tc,R [-]0.7 0.75 0.8 0.85 0.9 0.95 1

0.05

0.1

0.15

0.2

D6-ECD6-FCD4-ECD4-FCMDM-ECMDM-FC

glob

,SY

S[-]

(a) Global system effciency glob,SYSas a function of Tmax,R.Fig. A.6. Elements for comparison between corresponding evaporative and throttling down to saturated vapour conditions is assumed.The global eciency of the solar eld is

gglob;SF _Qin;ORC= _Qav; A:6and it accounts for the optical and thermal eciency. Thethermal power made available by the direct radiation of thesun at the given design point is given by

_Qav DNIdes ASF: A:7The area of the solar eld ASF can be evaluated as

ASF _W net=gORC _qabs _qhl _qpiping A:8All the terms of the denominator in Eq. (A.8) representthermal power specic to the m2 of SF aperture area._qabs DNIdes gopt is the thermal power absorbed by theashing ORC systems for dierent working uids. In case of ash cycles,

Table A.4Main information needed to compare solar ORC power systems with thermal s

216 E. Casati et al. / Solar Enersign calculations, the optical eciency gopt is equal to thepeak value gopt,p (Due and Beckam, 2006). _qpiping ac-counts for thermal losses in the piping subsystem of theSF, and a value of 10 W/m2 is assumed here (Forristall,2003). _qhl FT in;SF; T out;SF;Hincid; V wind accounts for thethermal eciency of the solar absorbers, and is evaluatedaccording to the detailed procedure presented in Burkhold-er and Kutscher (2009).1

Fig. A.6(a) shows gglob,SYS (Eq. (A.1)) as a function ofTmax. In order to better compare dierent working uids,reduced temperatures are used (TR,max). As a consequenceof the critical temperature increase with molecular weight(Table 1), and being the condensing temperature the samefor all the simulated cycles, more complex uids attainhigher eciencies for a given Tmax,R. As expected, beingthrottling a purely dissipative process, the FC eciencyfor a given Tmax,R is always lower than that of the corre-sponding EC for the same working uid. For all the uidsthis penalty decreases for increasingly higher Tmax,R, andtends to vanish with larger molecular complexity of theuid.

It can thus be concluded that, if siloxanes are adopted ashigh temperature working uids, and if the maximum cycletemperature is close to the uids critical temperature, theash cycle conguration does not imply severe eciencylosses with respect to the traditional evaporative cyclesolution.

gglob,SYS can be considered as the key merit parameter inthe comparison, since it is directly related with the area ofthe SF and, thus, to the main cost-driver of any CSP instal-lation (Pitz-Paal et al., 2007). However, also considerationsabout other critical components, such as the turboexpand-er and the storage system, should be accounted for in orderto better dene a suitable working uid and the operatingconditions for the given application. In particular the spe-cic cost of the turbine, for small-scale ORC systems,strongly inuences the cost of the power block.Fig. A.6(b) shows the turbine volumetric expansion ratioVRturb _V in= _V outturbine as a function of maximum cyclereduced temperature TR,max. The volumetric expansion

are considered as the working uids.

Fluid gglob,SYS Tc,R Tc pc

D6 0.178 0.895 333.6 6.3D4 0.178 0.998 312.7 14.2ratio strongly inuences the design/complexity of theexpander and therefore its cost (Macchi and Perdichizzi,1981). For a given uid and TR,max, the expansion due tothe throttling process causes the enthalpy drop across theexpander and VRturb to be signicantly lower in the FC

1 The coecients adopted in the correlation proposed in the referencehave been slightly modied, as a consequence of the dierent uids andow regimes, as discussed in Casati et al. (2011).than in the EC case. Smaller expansion specic work andsmaller volumetric expansion ratio allow for the design ofa more ecient turbine in the FC case than in the EC case,if the level of technology (therefore cost) is to be the same.Note that if higher turbine eciency for the FC case isaccounted for, the dierences in gglob,SYS shown inFig. A.6(a) between the FC and EC congurations wouldbe further reduced.

In case ashing is considered as the discharge method ofan hypothetical storage system (see Section 4.2), state c canbe regarded as the state of the uid extracted from the stor-age vessel, such that Tmax Tc = Tst. This holds for thecase-study presented in Section 5, whose storage densityEEED (see Section 3) can be evaluated as

EEED _W net

_mvap _mliq qls3600

kW hE=m3storage

h iA:9

In this case, _mliq becomes zero because of Assumption 2.This simplied approach assumes that the storage, initiallyfully charged with uid in conditions corresponding tostate c, delivers its full energy content without any varia-tion in uid properties. Thermal losses, as well as exergylosses due to deterioration of the stratication (Halleret al., 2010) are thus neglected. Such simplications aretypically justied for daily charge-discharge cycles, that isfor relatively short standstill times. The EEED reaches amaximum value for all the working uids considered here.This maximum value does not correspond to the maximumstorage temperature; furthermore, the EEED line is quiteat in the region where the maximum is reached. The dif-ferences among uids are comparatively large, as well asthe conditions of the stored uid in terms of pressure andtemperature.

In order to summarize these results, Table A.4 reportsthe main values obtained with the simulations that areneeded to select the working uid, if D6 and D4 are consid-ered. Only these uids are evaluated here since they allowfor higher gglob,SYS values. The comparison is then carriedon, aiming at the same value of gglob,SYS, which is taken

torage operating according to the ashing cycle, in case siloxane D6 and D4

pcond VRturb EEED muid (kg/kW hE)

0.002 954 8.2 660.035 246 6.2 74

gy 96 (2013) 205219equal to the maximum value reached in case D4 is theworking uid. Storing D6 at higher temperature is not con-sidered here given the corresponding extremely high valuesof VRturb, though it allows for the higher values of gglob,SYS(up to 0.185).

As expected, the two working uids allow attaining thesame eciency at almost the same value of Tc, which how-ever corresponds to a storage pressure pc 2.2 times larger incase D4 is the working uid. On the other hand, the con-densing pressure in case D6 is the working uid is 16 times

lower. Its very low value constitutes a design criticality forthe condenser and the turbine. The volumetric expansionratio of the turbine, for instance, is almost 4 times largerin case D6 is the working uid, whereby the inlet volumet-ric ows are similar.

The value of EEED is nonetheless 25% lower if D4 is theworking uid, and the corresponding specic mass of uidis 12% larger: these eects, combined with the higher pres-sure needed, would make D6 the preferred working uid ifonly the benets for the thermal storage are considered.

Appendix B. Complete ash evaporation as a working

condition for ORC power systems

The analysis of the performance of a ashing ORC, seeFig. A.5(b), as a function of the ashing conditions is pre-sented in this section. Only the results for working uid D4are reported, since they are representative of all the otherinvestigated systems featuring siloxanes as working uid.The system performance is evaluated according to the pro-cedure and the parameters dened in Appendix A; in this

which, as discussed in Appendix A, can also be seen asthe storage temperature Tmax Tc = Tst; the storage pres-sure is assigned a value of 1 bar higher than the corre-sponding vapour pressure (pmax pc = pst). For eachvalue of Tmax, the value of Td whereby complete ashingevaporation is reached (Td for which qd = 1) is also plotted(ash evaporation is considered as an isenthalpic process).

Fig. B.8(a) shows how, for each maximum temperatureTmax, the system eciency gglob,SYS initially grows fordecreasing Td until it reaches a relative maximum(gglob,SYSMAX): this is a consequence of the total mass owwhich need to be circulated, see Fig. B.8(b), and the corre-sponding power consumption of the auxiliary components.

As it is characteristic of CSP power systems, the thermaleciency of the solar eld gglob,SF is a decreasing functionof the temperature of the uid owing in the collector.Since gglob,SYS includes this eect, lower storage tempera-ture (and pressure) levels lead to comparatively higher val-ues of gglob,SF; such an eect, however, does notcounterbalance the concurrent decrease of gORC.

For increasingly higher Tmax, the corresponding gglob,SYSline tends to atten in the region where g

SF

tem,

E. Casati et al. / Solar Energy 96 (2013) 205219 217case, however, no simplifying assumption based on theabsence of liquid drains can be applied (see Eqs. (A.3)and (A.4)): the liquid drains from the ashing vessel haveto be compressed and circulated back to the heat source,see Fig. A.5(b). This stream is supposed to merge withthe main one in the solar eld, such that temperature equal-ity between the ows is ensured, while the vapour is deliv-ered to the ORC turbogenerator.

Fig. B.8(a) and (b) show the trends of the quantities ofinterest as a function of the ashing temperature Tash =Td, and the corresponding vapour pressure pash = pd.Each curve corresponds to a given maximum temperature

Tflash=Td [oC]

pflash= pd [bar]

A[m

2 ]

100 150 200 250 300

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

500

600

700

800

900

1.00 1.83 3.09 4.92 7.470.500.09 0.22

Nline Tmax,R Tmax pmax Td@qd=1

1 0.808 253.1 6.19 192.12 0.888 278.2 8.86 228.03 0.949 297.3 11.50 257.54 0.998 312.7 14.20 288.2

1 432

10.9

glo

b,SY

S[-]

(a) Solid black lines : global system effciency.Dashed black lines: area of the solar field.

pCR

Fig. B.8. Detailed analysis of the performance of a ashing-cycle ORC sys

(Tmax Tc = Tst), the quantities of interest are plotted as a function of the asindicated. For each value of Tmax, the value of Td whereby complete ashing eglob,SYSMAXoccurs. This implies that, for higher values of Tmax, extend-ing the throttling down to saturated vapour conditions leadsto a comparatively low eciency decrease with respect togglob,SYSMAX. Fig. B.8(a) shows the variation of the SF areaASF, which, having imposed the system net power output, isdirectly related with gglob,SYS.

If EEED (Eq. (A.9)) is considered, Fig. B.8(b) showshow strongly this quantity is dependent upon Td. The g-ure displays that a relative maximum of the EEED corre-sponds to the situation whereby the working uid isashed down to saturated vapour conditions. The locus

Tflash=Td [oC]

pflash= pd [bar]

EEED

[kW

h E/m

3 st]

100 150 200 250 300

0

5

10

15

20

0

1

2

3

4

pCR

29.4 74.705.0 00.1 38.1 90.30.09 10.90.22

NlineTmax,R Tmax pmax Td@qd=1

1 0.808 253.1 6.19 192.12 0.888 278.2 8.86 228.03 0.949 297.3 11.50 257.54 0.998 312.7 14.20 288.2 1 432 o

ut-

st[k

g/s]

m

(b) Solid black lines : equivalent electrical energy density.Dashed black lines : total mass flow entering the flashing

subsytem (i.e., extracted from the storage vessel)

in case D4 is the working uid. For dierent maximum temperature levels

hing temperature (Tash = Td); The corresponding vapour pressure is alsovaporation is reached (Td for which qd = 1) is plotted as a solid grey line.

nergy 96 (2013) 205219of such maxima, for varying Tmax, corresponds to the linefor D4 of Fig. A.7. Since this operating condition allows forimportant advantages (such as maximum storage densityfor the given Tmax, and simplication of the ashing systemlayout), without implying noteworthy eciency penalties,it can be considered as a reasonable working conditionfor a system implementing the FC conguration, especiallywhen storage temperatures close to the critical temperatureof the working uid are considered.

References

Al-Sulaiman, F., Hamdullahpur, F., Dincer, I., 2011. Trigeneration: acomprehensive review based on prime movers. International Journal ofEnergy Research 35, 233258.

Angelino, G., Gaia, M., Macchi, E., 1984. A review of Italian activity inthe eld of organic Rankine cycles. In: VDI Berichte Proceedings ofthe International VDI Seminar, vol. 539. VDI Verlag, Dusseldorf, pp.465482.

Angelino, G., Invernizzi, C., 1993. Cyclic methylsiloxanes as workinguids for space power cycles. Journal of Solar Energy Engineering,Transactions of the ASME 115, 130137.

Aspen Technology Inc., 2007. Exchanger Design and Rating, ver. 7.2..

Balaras, C., Grossman, G., Henning, H.-M., Infante Ferreira, C.,Podesser, E., Wang, L., Wiemken, E., 2007. Solar air conditioning inEuropean overview. Renewable and Sustainable Energy Reviews 11,299314.

Beckam, G., Gilli, P., 1984. Thermal Energy Storage. SpringerVerlag.Bejan, A., 1982. Entropy Generation Trough Heat and Fluid Flow. John

Wiley.Brosseau, D., Kelton, J., Ray, D., Edgar, M., Chrisman, K., Emms, B.,

2005. Testing of thermocline ller materials and molten-salt heattransfer uids for thermal energy storage systems in parabolic troughpower plants. Journal Solar Energy Engineering 127, 109116.

Burkholder, F., Kutscher, C., 2009. Heat Loss Testing of Schotts 2008PTR70 Parabolic Trough Receiver. Technical Report NREL/TP-550-45633 National Renewable Energy Laboratory.

Cabello, J., Cejudo, J., Luque, M., Ruiz, F., Deb, K., Tewari, R., 2011.Optimization of the size of a solar thermal electricity plant by means ofgenetic algorithms. Renewable Energy 36, 31463153.

Canada, S., Brosseau, D., Price, H., 2006. Design and construction of theAPS 1-MWE parabolic trough power plant. In: Proceedings of theASME International Solar Energy Conference 2006, vol. 2006.Denver, CO.

Casati, E., Colonna, P., Nannan, N.R., 2011. Supercritical ORCturbogenerators coupled with linear solar collectors. In: Proceedingsof the 30th ISES Biennial Solar World Congress, vol. 5. Kassel, pp.40564067.

Casati, E., Desideri, A., Casella, F., Colonna, P., 2012. Preliminaryassessment of a novel small CSP plant based on linear collectors, ORCand direct thermal storage. In: Proceedings of the 18th SolarPACESConference.

Colonna, P., Guardone, A., 2006. Molecular interpretation of nonclassicalgas dynamics of dense vapors under the van der Waals model. Physicsof Fluids 18, 056101114.

Colonna, P., Nannan, N.R., Guardone, A., 2008. Multiparameterequations of state for siloxanes: [(CH3)3SiO1/2]2[OSi(CH3)2]-

i=1,. . .,3 and [OSi(CH3)2]6. Fluid Phase Equilibria 263, 115130.Colonna, P., Nannan, N.R., Guardone, A., Lemmon, E.W., 2006.

Multiparameter equations of state for selected siloxanes. Fluid PhaseEquilibria 244, 193211.

Colonna, P., van der Stelt, T.P., Guardone, A., 2012. FluidProp (Version3.0): A Program for the Estimation of Thermophysical Properties of

218 E. Casati et al. / Solar EFluids. A Program Since 2004.DAmelio, L., 1935. Impiego di vapori ad alto peso molecolare in piccoleturbine e utilizzazione del calore solare per energia motrice [On the useof high molecular weight vapors in small turbines and solar energyconversion into mechanical work]. Industria Napoletana ArtiGrache.

Denholm, P., Mehos, M., 2011. Enabling Greater Penetration of Solar Powervia theUseofCSPwithThermalEnergyStorage.TechnicalReportNREL/TP-6A20-52978 National Renewable Energy Laboratory.

Dipippo, R., 2008. Geothermal Power Plants: Principles, Applications, CaseStudies and Environmental Impact, 2nd ed. Butterworth Heinemann.

DOW, 2012. SYLTHERM 800 Heat Transfer Fluid Product TechnicalData. Technical Report Dow Corning Corporation.

Due, J., Beckam, W., 2006. Solar Engineering of Thermal Processes, 3rded. John Wiley & Sons, Inc..

Eck, M., Steinmann, W.-D., Rheinlander, J., 2004. Maximum temperaturedierence in horizontal and tilted absorber pipes with direct steamgeneration. Energy 29, 665676.

Elliott, D., 1982. Theory and Tests of Two-Phase Turbines. TechnicalReport Jet Propulsion Laboratory, California Institute of Technology,Pasadena, CA.

EU (2006). Final technical progress report of the European project NNE5-1999-356: PS10 10 MW Solar Thermal Power Plant for SouthernSpain. Technical Report European Union.

Feldho, J., Schmitz, K., Eck, M., Schnatbaum-Laumann, L., Laing, D.,Ortiz-Vives, F., Schulte-Fischedick, J., 2012. Comparative systemanalysis of direct steam generation and synthetic oil parabolic troughpower plants with integrated thermal storage. Solar Energy 86, 520530.

Fernandez-Garca, A., Zarza, E., Valenzuela, L., Perez, M., 2010.Parabolic-trough solar collectors and their applications. Renewableand Sustainable Energy Reviews 14, 16951721.

Forristall, R., 2003. Heat Transfer Analysis and Modeling of a ParabolicTrough Solar Receiver Implemented in Engineering Equation Solver.Technical Report NREL/TP-550-34169 National Renewable EnergyLaboratory.

Fthenakis, V., Mason, J., Zweibel, K., 2009. The technical, geographical,and economic feasibility for solar energy to supply the energy needs ofthe us. Energy Policy 37, 387399.

Galiana, I., Green, C., 2009. Let the global technology race begin. Nature462, 570571.

Gil, A., Medrano, M., Martorell, I., Lzaro, A., Dolado, P., Zalba, B.,Cabeza, L., 2010. State of the art on high temperature thermal energystorage for power generation. part 1-concepts, materials and model-lization. Renewable and Sustainable Energy Reviews 14, 3155.

Gilli, P., Beckam, G., 1977. Design and economy of solar plants withintegrated thermal energy storage. In: Organization, W.M., (Ed.),Proceedings of the Unesco/WMO Symposium on Solar Energy.

Gilli, P., Beckmann, G., 1980. Thermal energy storage for non-fossilpower plants. In: Silverman, U.J. (Ed.), Energy Storage: A VitalElement in Mankinds Quest for Survival and Progress. PergamonPress, University of Maryland.

Giostri, A., Binotti, M., Astol, M., Silva, P., Macchi, E., Manzolini, G.,2012. Comparison of dierent solar plants based on parabolic troughtechnology. Solar Energy 86, 12081221.

Goldstern, W., 1970. Steam Storage Installation. Pergamon Press, Oxford.Haller, M., Yazdanshenas, E., Andersen, E., Bales, C., Streicher, W.,

Furbo, S., 2010. A method to determine stratication eciency ofthermal energy storage processes independently from storage heatlosses. Solar Energy 84, 9971007.

Henchoz, S., Buchter, F., Favrat, D., Morandin, M., Mercangoez, M.,2012. Thermoeconomic analysis of a solar enhanced energy storageconcept based on thermodynamic cycles. Energy 45, 358365.

Ho, T., Mao, S., Greif, R., 2012. Comparison of the organic ash cycle(OFC) to other advanced vapor cycles for intermediate and hightemperature waste heat reclamation and solar thermal energy. Energy42, 213223.

IEA, 2011. Solar Energy Perspectives. Technical Report International

Energy Agency.

IEA, 2012. Energy Technology Perspectives 2012. Technical ReportInternational Energy Agency.

Inage, S.-I., 2009. Prospects for Large-Scale Energy Storage in Decar-bonised Power Grids. Working Paper International Energy Agency.

Invernizzi, C., Iora, P., Silva, P., 2007. Bottoming micro-Rankine cyclesfor micro-gas turbines. Applied Thermal Engineering 27, 100110.

Kandari, A.M., 1990. Thermal stratication in hot storage-tanks. AppliedEnergy 35, 299315.

Kolb, G., Hassani, V., 2006. Performance analysis of thermocline energystorage proposed for the 1 MW Saguaro solar trough plant. In:Proceedings of the ASME International Solar Energy Conference, vol.2006. Denver, CO.

Krane, R., 1987. A second law analysis of the optimum design andoperation of thermal energy storage systems. International Journal ofHeat and Mass Transfer 30, 4357.

Lemmon, E., Huber, M., McLinden, M., 2010. NIST standard referencedatabase 23: Reference uid thermodynamic and transport properties-REFPROP, version 9.0. National Institute of Standards and Tech-nology, Standard Reference Data Program, Gaithersburg.

Li, B., Peuker, S., Hrnjak, P., Alleyne, A., 2011. Refrigerant massmigration modeling and simulation for air conditioning systems.Applied Thermal Engineering 31, 17701779.

Macchi, E., Perdichizzi, A., 1981. Eciency prediction for axial-owturbines operating with nonconventional uids. Journal of Engineeringfor Power Transactions of the ASME 103, 718724.

Medrano, M., Gil, A., Martorell, I., Potau, X., Cabeza, L., 2010. State of theart on high-temperature thermal energy storage for power generation. Part2-case studies. Renewable and Sustainable. Energy Reviews 14, 5672.

Pilkington Solar, I.G., 2000. Survey of Thermal Storage for ParabolicTrough Power Plants. Technical Report NREL/SR-550-27925National Renewable Energy Laboratory.

Pitz-Paal, R., Dersch, J., Milow, B., Tellez, F., Fernere, A., Langnickel,U., Steinfeld, A., Karni, J., Zarza, E., Popel, O., 2007. Developmentsteps for parabolic trough solar power technologies with maximumimpact on cost reduction. Journal of Solar Energy Engineering,Transactions of the ASME 129, 371377.

Prabhu, E., 2006. Solar Trough Organic Rankine Electricity System(STORES) Stage 1: Power Plant Optimization and Economics.Technical Report NREL/SR-550-39433 National Renewable EnergyLaboratory.

Price, H., Hassani, V., 2002. Modular Trough Power Plant Cycle andSystem Analysis. Technical Report NREL/TP-550-31240 NationalRenewable Energy Laboratory.

PriceWaterHouseCoopers, 2010. 100% renewable electricity a roadmapto 2050 for Europe and North Africa. Technical Report.

Skumanich, A., 2011. CSP at a crossroads: The rst solar electric powerplants are still proving their worth after three decades, so why arentwe seeing more csp reach the development stage? Renewable EnergyFocus 12, 5255.

Solar Millenium, 2009. The parabolic trough power plants Andasol 1to 3: The largest solar power plants in the world. TechnicalReport.

Span, R., Wagner, W., 2003. Equations of state for technical applications.I. Simultaneously optimized functional forms for nonpolar and polaruids. International Journal of Thermophysics 24, 139.

E. Casati et al. / Solar Energy 96 (2013) 205219 219Obernberger, I., Gaia, M., 2005. Biomass-Power-Heat Coupling Based onthe ORC Process State-of-the-art and Possibilities for ProcessOptimization. VDI Berichte, pp. 131148.

Odeh, S., Behnia, M., Morrison, G., 2000. Hydrodynamic analysis ofdirect steam generation solar collectors. Journal of Solar EnergyEngineering, Transactions of the ASME 122, 1422.

Peters, M., Schmidt, T., Wiederkehr, D., Schneider, M., 2011. Sheddinglight on solar technologies-a techno-economic assessment and itspolicy implications. Energy Policy 39, 64226439.Steinmann, W.-D., Eck, M., 2006. Buer storage for direct steamgeneration. Solar Energy 80, 12771282.

Tabor, H., Bronicki, L., 1964. Establishing criteria for uids for smallvapor turbines. In: SAE National Transportation, Powerplant, andFuels and Lubricants Meeting.

The MathWorks Inc., 2011. Matlab, ver. 7.13.0.564. .

Verneau, A., 1978. Application of organic uids in solar turbines[lemploi des uides organiques dans les turbines solaires]. Entropie14, 918.

Thermal energy storage for solar-powered organic Rankine cycle engines1 Introduction2 Siloxanes as high-temperature ORC working fluids3 Concepts of TES systems for power plants4 Direct storage of working fluid in Rankine power stations4.1 Storage methods4.1.1 Storage at constant pressure4.1.2 Expansion storage at almost constant pressure4.1.3 Sliding pressure storage (Ruths accumulator)

4.2 Discharge methods4.2.1 Vapour generation by flash evaporation4.2.2 Feed water storage4.2.3 Cascading storage4.2.4 Direct liquid expansion

4.3 Storage systems

5 Case study5.1 Working principle5.2 Flashing Rankine cycles with organic fluids5.3 Flashing the organic vapour down to saturated conditions5.4 Results

6 ConclusionsAcknowledgementsAppendix A Thermodynamic comparison between flashing and evaporative organic Rankine cyclesAppendix B Complete flash evaporation as a working condition for ORC power systemsReferences