at SciVerse ScienceDirect

Applied Thermal Engineering 50 (2013) 582e592

Contents lists available

Applied Thermal Engineering

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

Performance analysis of conventional and sloped solar chimneypower plants in China

Fei Cao a, Liang Zhao a,*, Huashan Li b,c, Liejin Guo a

a State Key Laboratory of Multiphase Flow in Power Engineering, Xi’an Jiaotong University, No. 28, West Xianning Rd, Xi’an 710049, PR ChinabKey Laboratory of Renewable Energy and Gas Hydrate, Guangzhou Institute of Energy Conversion, Chinese Academy of Sciences, Guangzhou, PR ChinacGraduate University of Chinese Academy of Sciences, Beijing 100049, PR China

h i g h l i g h t s

< The optimum collector angle for maximum power generation is 60� in Lanzhou.< Main parameters influencing performances are the system height and air property.< Ground loss, reflected loss and outlet kinetic loss are the main energy losses.< The sloped styles are suitable for Northwest China.< The conventional styles are suitable for Southeast and East China.

a r t i c l e i n f o

Article history:Received 26 April 2012Accepted 21 June 2012Available online 28 June 2012

Keywords:Conventional solar chimney power plantSloped solar chimney power plantSolar chimneySolar energyNorthwest China

* Corresponding author. Tel.: þ86 029 82668287; fE-mail addresses: xjtu.lzhao@gmail.com, lzhao@m

1359-4311/$ e see front matter � 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.applthermaleng.2012.06.03

a b s t r a c t

The solar chimney power plant (SCPP) has been accepted as one of the most promising approaches forfuture large-scale solar energy applications. This paper reports on a heat transfer model that is used tocompare the performance of a conventional solar chimney power plant (CSCPP) and two sloped solarchimney power plants (SSCPPs) with the collector oriented at 30� and 60�, respectively. The powergeneration from SCPPs at different latitudes in China is also analyzed. Results indicate that the largersolar collector angle leads to improved performance in winter but results in lower performance insummer. It is found that the optimal collector angle to achieve the maximum power in Lanzhou, China, isaround 60�. Main factors that influence the performance of SCPPs also include the system height and theair thermophysical characteristics. The ground energy loss, reflected solar radiation, and kinetic loss atthe chimney outlet are the main energy losses in SCPPs. The studies also show SSCPPs are more suitablefor high latitude regions in Northwest China, but CSCPPs are suggested to be built in southeastern andeastern parts of China with the combination to the local agriculture.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

The solar chimney power plant (SCPP) is composed of threecomponents: a solar collector, a chimney situated in the center of thecollector andapowerconversionunit (PCU). A schematic of the SCPPis shown in Fig. 1. Sunlight transmits through the transparent solarcollector cover and heats the ground below. Ambient cold air entersthe collectors from the periphery of the collectors and is heated as itflows along the collector toward the center. Due to the pressurecreated by the density difference between the warm airflow andambient cold air, the airflow enters the chimney, and with the PCU,the kinetic energy of the airflow is converted into the electric power.

ax: þ86 029 82669033.ail.xjtu.edu.cn (L. Zhao).

All rights reserved.8

1.1. Theoretical study of SCPPs

The concept of SCPP was originally proposed in 1903 by IsidoroCabanyes [1] and then presented in a publication by Günther [2]. Asystemic research on the SCPP was first performed by Schlaich. In1981, with the financial support from the German Ministry ofResearch and Technology, Schlaich began the construction of a pilotSCPP with the peak power about 50 kW in Manzanares, Spain [3,4].This is the most systematic study of the solar chimney powertechnology in practice until now and successfully verifies thefeasibility of SCPPs.

Since then, extensive research has been carried out on the huge-potential of the SCPP over the world. Haaf et al. carried out a basicresearch about the energy balance on the ground surface, energyloss in the chimney and turbine of theManzanares pilot power plant

Fig. 1. Schematic of a solar chimney power plant.

F. Cao et al. / Applied Thermal Engineering 50 (2013) 582e592 583

[3]. von Backstrom et al. analyzed the pressure drop in solar chim-neys [5], the turbine characteristics [6], andperformances of thePCU[7]. Bernardes et al. analyzed some available heat transfer coeffi-cients applicable to SCPP collectors [8], and evaluated some opera-tional control strategies [9]. Pretorius et al. made a sensitivityanalysis of the operating and technical specifications of SCPPs [10],andmade a critical evaluation on the solar collector andheat storagelayer [11]. Zhou et al. analyzed the influence of chimney height onSCPPs [12], and made preliminary research on the special climatearound a commercial SCPP [13]. Schlaich et al. reported the designresults of commercial SCPPs [14]. Yanet al. investigated the influenceof theworking fluid and chimney temperature [15]. Pasumarthi andSherif built a complete mathematical model and analyzed theperformances of SCPPs against the experimental research [16].

In 2005, Bilgen and Rheault proposed a sloped solar chimneypower plant (SSCPP), whose solar collector is laid along the hillside(see Fig. 2) [17]. They concluded that SSCPP has higher thermal effi-ciency at high latitudes. This kind of SCPP is suitable for high latitudeand mountain areas. Serag-Eldin explored the feasibility of an SCPPwith the chimney built over the steep side of themountain [18]. ZhouandYang reportedanovel SSCPPwithfloating chimneyandpredicteditspotential inChina’sDesert [19].Weietal. andCaoetal. analyzed theslope angle effect on receiving insolation and investigated on the

Fig. 2. Schematic of a sloped so

optimal collector slopeof the SSCPP [20,21]. The SSCPP is basedon theSCPP principle and takes advantage of the local geographical features.So, we could recognize the SSCPP in Fig. 2 as the development of thehorizontal solar chimney power plant in Fig. 1, which in this study iscalled the conventional solar chimney power plant (CSCPP).

1.2. Case study of SCPPs

There are some case studies of CSCPPs in literature. Mullet madedetailed analysis about the efficiency of SCPP, and concluded thefeasibility of building SCPPs in developing countries [22]. Zhou et al.investigated the performances of a 100 MW CSCPP in Qinghai-TibetPlateau [23]. Nizetic et al. analyzed the feasibility of implementingSCPPs in the Mediterranean region [24]. Larbi et al. made a perfor-mance analysis of an SCPP in the southwestern region of Algeria [25].Ketlogetswe et al. described a systematic experimental study onamini-solar chimney system in Botswana [26]. Dai et al. explored thefeasibilityofSCPPs in threeregionsofNorthwestChina [27].However,as for the SSCPP, to our knowledge the only case is that a simulationofSSCPP in Lanzhou, Chinawas carried out by Cao et al. [28].

Because of significant meteorological and geographical differ-ences and local economic differences, case studies of SCPPs fordifferent countries or regions are of high value. The authors’previous study suggested that the SSCPP had better performances inspring and autumn days, whereas the CSCPP developed superiorityin summer days [28]. For Northwest China, where local geographicalresources are rich (over 30 mountain chains) and annually solarradiation is also strong (over 5852 MJ/(m2 year)) [29], it is thus ofhigh significance to analyze and compare the performances ofCSCPPs and SSCPPs in such areas throughout the year.

By using a mathematical model based on the heat transfer,thermodynamics and fluid dynamics theories, a comparative studyof the performances of a CSCPP and two SSCPPs in Lanzhou, China isperformed. Main tasks of this study include:

1) To build a simplified mathematical model for SCPPs.2) To compare the performances of the CSCPP and SSCPPs in

Lanzhou.3) To analyze the power generation of CSCPPs and SSCPPs at

different latitudes in China.

lar chimney power plant.

Table 1Configuration sizes and coefficients of the CSCPP and the SSCPP.

Style CSCPP SSCPP

Solar collector Collector radius 550 m e

Collector area 950,000 m2 950,000 m2

Collector heighta 0 0e1378.4 mCollector cover emittance 0.87 0.87Glass extinction coefficient 32 m�1 32 m�1

Glass thickness 5 mm 5 mmRefractive index 1.526 1.526

Chimney Height 547 m 60 mDiameter 54 m 54 m

Ground Material Granite GraniteDensity 2640 kg/m3 2640 kg/m3

Specific heat capacity 820 J/(kg K) 820 J/(kg K)Thermal conductivity 1.73 W/(m K) 1.73 W/(m K)Normal emittance 0.92 0.92Reflectance 0.25 0.25

Turbine Efficiency 0.8 0.8Inlet loss coefficient 0.056 0.056

a Collector heights of SSCPPs depend on the solar collector angles.

F. Cao et al. / Applied Thermal Engineering 50 (2013) 582e592584

2. Mathematical model

2.1. Solar radiation into the collector

Total solar radiation on a horizontal surface Hhor is:

Hhor ¼ Hb þ Hd (1)

where Hb and Hd are, respectively, the beam and diffuse solarradiation on the horizontal surface. The data of the annuallyaverage monthly total and diffuse solar radiation on the horizontal,viz., Hhor and Hd, are obtained from the National MeteorologicalInformation Centre (NMIC) of China.

The received solar radiation Ht on a surface with tilted angleb could be calculated as:

Ht ¼ Ht;b þ Ht;d þ Ht;r (2)

where Ht,b, Ht,d and Ht,r are the beam, diffuse and reflected solarradiation on the sloped surface, respectively. They could be calcu-lated by the next three equations:

Ht;b ¼ HbRb (3)

Ht;d ¼ HdRd (4)

Ht;r ¼ ref $HhorRr (5)

where ref is the ground reflectance [30]. Rb, Rd and Rr are the coef-ficients. They could be calculated in the following equations [31]:

Rb ¼cosð4� bÞcos d sin uts þ

� p

180

�uts sinð4� bÞsin d

cos 4 cos d sin us þ� p

180

�us sin 4 sin d

(6)

Rd ¼ 1þ cos b2

(7)

Rr ¼ 1� cos b2

(8)

where 4 is the latitude, us is the sunset angle, uts is the sunset angletoward the sloped surface, and d is the solar declination angle.

The total solar radiation on a sloped surface finally could beexpressed as:

Ht ¼ ðHhor�HdÞ Rbþ12Hdð1þcosbÞþ1

2ref $Hhorð1� cosbÞ (9)

The monthly average solar radiation absorbed by the south-facing sloped solar collector cover S1 is calculated as:

S1 ¼ ðHhor � HdÞ Rbab þ Hdad1þ cos b

2þ Hhor$ref $ar

1� cos b2(10)

The monthly average solar radiation transmitted through theglass cover and absorbed by the heat storage layer S2 is:

S2 ¼ ðHhor � HdÞ RbðsaÞb þ HdðsaÞd1þ cos b

2

þ Hhor$ref $ðsaÞr1� cos b

2(11)

where a is the absorptance of the collector cover and (sa) is thetransmittanceeabsorptance product of the solar collector. Thesubscript b, d and r refer to the beam, diffuse and reflected radia-tion, respectively. Detailed methods for calculating a and (sa) canbe found in Ref. [31].

2.2. Solar collector

The solar collector is made up of three parts: the collector cover,the heat storage layer and the working fluid between them.Ambient air, the working fluid of SCPPs, enters the channelbetween the collector cover and heat storage layer, and flows intothe chimney connected to the outlet of the solar collector.According to Pretorius [11], glass is chosen as the collector covermaterial because of its high optical and intensity performances.The solar energy received by the collector transmits through theglass cover and is absorbed by the ground. The ground acts as theheat storage layer. Pretorius analyzed various ground types(including sandstone, granite, limestone, sand, wet soil and water),and concluded that stones had better performance than the sandfor heat storage [32]. As for Lanzhou, granite is chosen as theground type in this study. Detailed parameters of solar collectormaterials are shown in Table 1. There are six basic assumptionsused in simulation and they are: (1) steady state conditions, (2) theairflow speed at the collector inlet is ignored, (3) the solar collectoris oriented northesouth and facing toward the equator, (4) noevaporation takes place under the collector, (5) the verticaltemperature profile of the collector air is constant, and (6) energyloss from the airflow to the ambient air at the collector inlet isignored.

The hot ground transfers heat in the method of convection andradiation to the air above it. The schematic of thermal balance in thecollector is shown in Fig. 3. The cold air, with the temperature Ta,enters the collector and is heated by the hot ground. Energybalances in the solar collector are defined as:

Continuity equation:

rfvfAf ¼ rovoAchi (12)

Energy equations for the glass cover

S1 þ UtðTa � TcÞ þ hr�Tp � Tc

�þ h1�Tf � Tc

�¼ 0 (13)

for the ground

S2 þ Ub�Ta � Tp

�þ h2�Tf � Tp

�þ hr

�Tc � Tp

� ¼ 0 (14)and for the airflow

Fig. 3. Thermal balance in the solar collector.

F. Cao et al. / Applied Thermal Engineering 50 (2013) 582e592 585

h1Acoll

�Tc � Tf

�þ h2Acoll

�Tp � Tf

�¼ Q (15)

where

Q ¼ rovoAchicpðTo � TaÞ (16)

Tf ¼ To þ Ta2

(17)

In the above equations, Af is the section area of the solarcollector where the air density and airflow speed are rf and vfrespectively, Achi is the chimney sectional area, Tc is the collectorcover temperature, Tp is the heat storage layer temperature, Tf is theairflow temperature, vo is the airflow speed at the collector outlet,ro is the air density at the collector outlet, To is the airflowtemperature at the outlet of the collector, Acoll is the collector area,Q is the energy absorbed by the airflow, ra is the ambient airdensity, and Ut, hr, h1, Ub and h2 are the heat transfer coefficients.Detailed equations to calculate these heat transfer coefficientscould be found in Ref. [16], and the working fluid densities ro and rfare, respectively, determined at To and Tf by the following equationderived from the air properties between 0 and 70 �C in [33]:

rðTÞ ¼ 0:000012T2 � 0:011167T þ 3:445689 (18)

Energy loss in the solar collector could be divided into threeparts, viz. the radiation that is reflected by the collector cover andthe ground Sref, the energy loss by the collector cover Qloss,t and theenergy loss by the heat storage layer Qloss,b. They could be calcu-lated as:

Sref ¼ rrH�t þ ref H

�ts1 (19)

Qloss;t ¼ UtðTc � TaÞ (20)

Qloss;b ¼ Ub�Tp � Ta

�(21)

where H�t is the daily average total solar radiation on tilted surfaces,

and rr and s1 are the reflectance and transmittance of the glasscover. Note that the ground reflection from the second time has notbeen included in the calculation due to its small and negligibleinfluence on the final results.

2.3. Chimney

The model described below is based on the following assump-tions: (1) Boussinesq approximation is assumed to be valid, (2)energy loss through the chimneywall is ignored, and (3) turbulenceafter the turbine is ignored.

The continuity equation in the solar chimney is:

rovo ¼ routvout (22)

where rout and vout are the air density and airflow speed at thechimney outlet. Noted that rout is assumed to be the density ofambient air (ra) according to the conventional method of SCPPs[3,17].

For a vertical adiabatic chimney, the pressure difference createdin the chimney is:

DPchi ¼ ðrout � roÞgHchi ¼assumedðra � roÞgHchi (23)

The pressure difference between the inlet and outlet of the solarcollector is calculated as:

F. Cao et al. / Applied Thermal Engineering 50 (2013) 582e592586

Zoutlet

20 30 40 50 60 70600

800

1000

1200

1400

1600

Sola

r rad

iatio

n/ G

Wh/

m2

Annually solar radiationAn

nual

ly p

ower

out

put /

GW

h

20 25 30 35 40 45 50 55 60 65 70 7540

80

120

160

200

Solar collector angle

Annually power outputMaximum solar radiation

Maximum power output

Fig. 4. Solar radiation on different slopes and power generation by SSCPPs withdifferent collector angles.

DP ¼inlet

gðra � rðzÞÞdz (24)

where z denotes the height.According to Haaf et al. [3], Bernardes et al. [8] and Bilgen and

Rheault [17], we assume that the air density variation is linearbetween entrance and exit of the sloped collector. Thus, the airdensity variation can be calculated as:

rðzÞ ¼ ra þro � raHcol

z (25)

By integrating Eq. (24) between the inlet and the outlet of thesloped solar collector with r(z) from Eq. (25) and assuming z ¼ 0 atthe collector inlet, the pressure difference generated in thecollector is:

DPcol ¼ ra � ro2

gHcol (26)

The total pressure difference yielded due to buoyancy can beexpressed as:

DPtot ¼ ðra � roÞg�Hchi þ

Hcol2

�(27)

where Hchi and Hcol are the chimney and the collector height,respectively.

The pressure difference is consumed by four parts, i.e., thefriction losses in the collector and the chimney DPf, the kineticenergy losses at the turbine inlet DPin, the kinetic energy losses atthe chimney outlet DPout, and the rest is the effective pressurewhich is used by the turbine to generate electricity DPt.

DPtot ¼ DPf þ DPin þ DPout þ DPt (28)

where DPf, DPin and DPout could be calculated as:

DPf ¼ fLthD

12rfv

2f (29)

DPin ¼ g12rov

2o (30)

DPout ¼ 12routn

2out (31)

where f is the friction loss coefficient, Lth is the length of thechannel, D is the hydraulic diameter and g is the turbine inlet losscoefficient.

2.4. PCU and efficiencies

The power generated by the turbine, Pele, is:

Pele ¼ htDPtvoAchi (32)

where ht is the turbine efficiency.The collector efficiency can be expressed as:

hcol ¼ Q=ðHt � AcollÞ (33)

The system efficiency is:

hsys ¼ Pele=ðHt � AcollÞ (34)

2.5. Theoretical solution procedures

Parameters that indicate performances of the SCPP could beconfirmed through simultaneous Eqs. (12)e(17), (22), (27), (28) and(32). There are 10 unknown parameters and 10 equations aresummarized. So, the equations could be solved through iterativecalculations. During the process, an initial guess of the collector covertemperature, airflow temperature, heat storage layer temperature,airflow speed and airflow density is made firstly. Then, an iterativeprocess is initiated, and all the required heat transfer and friction losscoefficients are calculated based on the initially guessed values. Eachnewassumedresult calculated in theSCPPs is thencomparedwith theold corresponding value. Only the difference between any corre-sponding new and old values is less than the maximal acceptabledifference, would the iteration be finally stopped. By this repetitiveand iterative process, the temperatures and airflow speed in thecollector, the airflow densities in the collector and chimney, themassflow rate, generated power in the turbine, etc. can be obtained.

3. Results

3.1. Configuration sizes of the CSCPP and SSCPPs in Lanzhou

Lanzhou (103.50�E, 36.03�N) is a zonal basin city 1520 m abovethe sea level, with an area of 13,085.6 km2. It is the capital of GansuProvince and locates in the geographical central of NorthwestChina. Its annual global solar radiation is more than 5020 MJ/m2,and sunshine duration is over 2600 h per year. Its annually meantemperature is 9.8 �C.

To carry out the analysis of SCPP performances, we considerthree reference 5 MW SCPPs for examples, whose baselineparameters are also given in Table 1 based on Schlaich et al. [14] andBilgen and Rheault [17]. The main reason for choosing 5 MW SCPPsfor this case study is that though many reports suggest that SCPPswith 100 MW power generation or even higher could have higherefficiency [14,23], huge costs of such SCPPs are overbearing toNorthwest China, which is the economically under-developedregion, and Lorente et al. argued that the principle of “few largeand many small” for SCPPs is land-efficient [34]. Besides, only5 MW typical dimensions of the SSCPP could be found in literature.

3.2. SSCPPs with different solar collector angles

Fig. 4 shows the solar radiation on tilted surfaces and poweroutput of SSCPPs with different collector angles. It is found that the

Density

C1C2C3

35

40

45

50

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec1.15

1.20

1.25

1.30

1.35

Flowrate

Mas

s flo

wra

te/ x

103 kg

s-1

Month in the year

Den

sity

/ kgm

-3

Fig. 6. Air densities and air mass flow rates of C1, C2 and C3 throughout the year.

F. Cao et al. / Applied Thermal Engineering 50 (2013) 582e592 587

tilted surface of 30� receives the largest solar radiation. However,the maximum power generation occurs with the collector angle at60�. This finding agrees well with the result of Li [35] using thecriterion of maximum airflow rate to optimize the tilted angle ofSSCPPs in Lanzhou. Similarly, a report from Sakonidou et al. showedthat to obtain the maximum airflow rate in a solar chimney forventilation, the collector slope is 20e25� larger than the locallatitude [36]. On the other hand, Bilgen and Rheault suggested thatto maximize solar radiation received by the collector, the optimumcollector slopes are smaller than the latitude by 5e7� [17]. It shouldbe noted that for turbines of the SCPPs, maximum airflow rate ismuch more essential than the maximum solar radiation.

Therefore, in the next part of discussion, we take three SCPPsinto comparison: the conventional solar chimney power plant (C1),and the sloped solar chimney power plants with the collector tiltedangle of 30� (C2) (following Bilgen and Rheault’s study [17]) and60� (C3) (following Li’s [35] and Sakonidou et al.’s [36] studies).

3.3. Performances analysis

Fig. 5 shows the ambient air temperature and the airflowtemperatures of C1, C2 and C3. It is found that airflow temperaturesof C1, C2 and C3 have similar tendency: first rising from January toJuly then decreasing from July to December. The peak values occurin July and the temperatures are roughly symmetric with thevertical axis in July. Also, it can be seen that the airflow temperatureincreases of C2 are higher than the other two styles in everymonth,and temperature increases of C1 are higher than those of C3 onlybetween April and August. The airflow temperature increase isa compound result of the solar irradiation and the system pressuredifference: the higher solar radiation would increase the airflowtemperature, and the higher system pressure difference wouldincrease airflow speed but decrease the airflow temperature rise.

Fig. 6 shows that the air densities in summer are smaller, andtheir densities match well with the temperatures in Fig. 5,following the principle that the higher temperature results in lowerdensity of the airflow. However, the variations of air mass flow ratesof C1, C2 and C3 are different, such as:

1) For C1, the highest air mass flow rate occurs in June (45.87 t/s)and the lowest air mass flow rate locates in November(39.39 t/s); the air mass flow rates in summer are higher than inwinter; and the curve is roughly symmetrical with June.

C1 Temperature C2 Temperature C3 Temperature

Tem

pera

ture

/ o C

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec-10

-5

0

5

10

15

20

25

30

35

Airflow temperature increaseAirflow temperature increase

Month in the year

Ambient Temperature

Airflow temperature increase

Fig. 5. The ambient temperature, the airflow temperature of C1, C2 and C3 throughoutthe year.

2) For C2, the highest air mass flow rate locates in March(42.41 t/s) and the lowest air mass flow rate locates in July(38.37 t/s); a small drop first appears in February, and thenfollowed by a long-time gradual drop from March to July; andthe tendencies in the rest months are roughly symmetricalwith those in the first half year.

3) For C3, the highest air mass flow rate locates in January(50.37 t/s) and the lowest air mass flow rates locates in July(40.78 t/s); the tendencies of C3 are similar to that of C2, but thedrop from March to July of C3 is much larger than that of C2.

In addition, it is also observed that the highest total air massflow rate in a year is produced by C3 (396.29 Mt), followed by C1(382.74 Mt) and then C2 (352.90 Mt). For the turbine generator, airmass flow rate is an important parameter to determine its poweroutput.

Fig. 7 shows the input energy and output power of C1, C2 and C3.The input energy is the monthly total solar radiation reaching thesolar collector. It is found that the solar radiation on a horizontalsurface is roughly symmetrical with July, and the highest solarradiation also appears in July. Tilting the collector would decreasethe solar radiation in summer days but increase it in winter days.

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec0

100

200

300

400

500

600

700

800 Power

Sola

r rad

iatio

n/ M

W

Month in the year

Pow

er g

ener

atio

n/ M

WRadiation

C1C2C3

0

2

4

6

Fig. 7. Monthly total solar radiation received by the solar collector and the powergeneration of C1, C2 and C3 throughout the year.

F. Cao et al. / Applied Thermal Engineering 50 (2013) 582e592588

The larger the collector angle is, the more obvious the describedincrease and decrease tendencies would be created. And theannually total solar radiation of C2 is the largest, up to 5066.82 TJ,then followed by C1 of 4634.87 TJ, and finally C3 of 4529.19 TJ.

As for the output power of C1, C2 and C3 throughout the year, itis found that:

1) For C1, the power generation tendency is the same as its inputenergy; and the highest power is produced in June (5.33 MW)and the lowest in December (2.85 MW).

2) For C2, the power generation is much stable comparingwith C1and C3, with two small peaks in March and September; and thehighest power generation occurs in March (3.65 MW) and thelowest in July (3.02 MW).

3) For C3, the power generation is much higher in winter daysthan that of C1 and C2; the power generation decrease of C3from March to July is larger than that of C2; and the highestpower generation is in January (5.90 MW) and the lowest inJuly (3.76 MW).

The average power generation of C1, C2 and C3 throughout theyear is 4.36 MW, 3.23 MW and 4.78 MW, respectively. The powergeneration of C1, C2 and C3 in Lanzhou is smaller than the referenceSCPP of 5 MW. The differences are caused by the different solarradiation and solar duration between Winnipeg [17], Manzanares[14] and Lanzhou. Note that though the power generation of C2 issmaller than C1, the chimney height of C2 is much smaller than thatof C1, and even though chimney height of C3 is smaller than that ofC1, power generation of C3 is larger than C1. From the abovecomparison, it is concluded that:

1) The suggested configuration sizes in literature may not besuitable for other regions. For different regions, specific designwork would be required.

2) Inclining the collector could sharply decrease the solarchimney height (from 547 m of C1 to 60 m of C3, as shown inTable 1.

The system efficiencies and solar collector efficiencies of C1, C2and C3 throughout the year are shown in two vertical groups inFig. 8. The solar collector efficiency describes the ability of solarcollectors converting the solar energy into the airflow’s thermalenergy. It is found that the solar collector efficiency of C1 is smallerthan that of C2 and C3 in each month. The findings agree with

1.0 0.8 0.6

Power efficiency/ %

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

40

Fig. 8. Power efficiencies and solar collector effici

Bilgen and Rheault [17]. The solar collector efficiencies of C1 aresymmetrical with July. The solar collector efficiency of C2 is higherthan that of C3 in summer days. In addition, C2 has the highestaverage solar collector efficiency of 56.43%, followed by C3 of56.11%, and the lowest is C1 of 48.52%. In aword, the tilted collectorwould enlarge the solar collector efficiency.

Besides, the power efficiency in Fig. 8 describes the proportionof generated power to the total solar energy input. Power genera-tion is the ultimate purpose of the SCPPs, thus the power efficiencyis of high importance for the SCPP. It is found that throughout theyear, the rank of power efficiency is C3, C1 and C2. The powerefficiencies of C2 and C3 have the similar tendencies: higher inwinter but smaller in summer; the highest power efficiencies bothoccur in January (0.595% and 0.998% respectively), and the lowestpower efficiencies are both in July (0.534% and 0.862% respectively).However, the power efficiency tendency of C1 is opposite to that ofC2 and C3. The highest power efficiency of C1 is in June (0.814%),and the lowest in December (0.760%).

4. Discussion

4.1. Energy consumption of SCPPs

Energy consumed by the components of SCPPs is shown inFig. 9, with the total solar radiation utilized by C1, C2 and C3 on itstop left. Statistical results indicate that tilting the collector angle to30� would increase the total solar energy input by 6.14%, but risingthe collector angle to 60�, the total solar energy input woulddecrease by 5.11%. It is also found that:

1) The ground energy loss, the reflected energy, the groundenergy storage and the glass cover energy storage are themajorenergy consumptions for the SCPPs. According to Eq. (19), thereflected energy consists of two parts, viz. the glass cover re-flected energy and the ground reflected energy, in which theground reflected energy holds the major proportion. It is thusfound that the ground plays an important role in the energyconsumption. Pretorius et al. compared the power outputs ofsix different ground types: sandstone, granite, limestone, sand,wet soil and water. They found that the SCPPs employing thewet soil and the sand have the lowest and highest poweroutputs respectively [11], and different materials lead tovarying power outputs during the daytime and at night. Pre-torius also concluded that increased ground absorptivity holds

C3 C2 C1

45 50 55 60

Solar collector efficiency/ %

encies of C1, C2 and C3 throughout the year.

0 20 40 60 80 100

0

1000

2000

3000

4000

5000 4529.185066.82

C3C2

Tota

l sol

ar ra

diat

ion/

TJ

C1

4773.91

1.67%

0.70%

17.94%

12.63%

1.58%

0.46%

23.89%

28.01%

2.73%

3.49%

19.61%

20.78%

16.98%

28.25%

16.11%

15.92%

1.00%

10.81%0.86%23.59%

2.66%

16.37%

18.48%

C3

C2

Proportion of the total solar radiation /%

Syst

em s

tyle

s

Energy used by the turbine Energy loss at the chimney exit Friction loss Ground energy loss Glass cover energy loss Reflected Energy Energy storage in the ground Energy storage in the glass

C115.96%

Fig. 9. Energy consumption by components of C1, C2 and C3.

F. Cao et al. / Applied Thermal Engineering 50 (2013) 582e592 589

positive effects on annual solar chimney power output [32].Bernardes et al. evaluated the influence of varying the groundheat penetration coefficient from 1000 Ws1/2/(K m2) to2000 Ws1/2/(K m2) on the performance of SCPPs, and foundnegligible influence on the power output [37]. Accordingly, itcan be deduced that ground types could partly influence thepower generation. Thoughmany different ground types exist atlocations around the world suitable for the construction ofSCPPs, lower heat penetration coefficient materials are sug-gested to be chosen as the heat storage layer. In addition, betterinsulation methods could reduce the energy loss to thesurroundings [32]. With respect to the glass cover reflectedenergy and the glass cover energy storage, the quality of glass isof high significance. In this study, the poor glass with theextinct coefficient of 32 m�1 is used as the collector cover.Pretorius et al. analyzed the performances of the good glasswith extinct coefficient of 4 m�1, and indicated that during thesummer months the better quality glass could cause the poweroutput a little increase throughout the day, and the overalloutput during the colder months was also slightly higher [32].So, the glass with good quality, called the “waterwhite glass”, isa better choice for the solar collector cover. Consequently, theinvestment of SCPP will increase.

2) The energy loss at the chimney exit is the fifth largest energyloss for C1 and C2 but the third loss for C3, which also holdslarge proportion to the total energy consumption. Methods forreducing the energy loss at the chimney exit are rarely reportedin literature. The airflow at the chimney outlet has high kineticand gravitational energy. However, the SCPP using air as theworking fluid could hardly reuse the energy. Kashiwa et al.proposed a novel “solar cyclone” for extracting freshwater fromthe atmosphere [38]. Wang et al. proposed the combined solarchimney system for both power generation and seawaterdesalination [39]. Using water vapor together with air as theworking fluid would partially recover the energy loss at thechimney outlet.

3) The glass cover energy loss and friction loss are smallercomparingwith others.With the glass cover temperature beingthe higher, the cover energy loss of C2 is correspondinglyhigher. The friction loss is the smallest energy loss and ofa great relationship with the airflow speed.

4.2. Major factors influencing SSCPP performances

In Fig. 10, we compare the performances of nine SSCPPs. Two ofthem with the solar collector angle of 30� and 60� respectively arechosen as the reference systems. Then we calculate and diagramthe relative proportions of other systems to the reference systemin Fig. 10(a) and (b) separately. With Fig. 10(a) and (b), it is foundthat:

1) When the solar collector angle is rising from 15� to 80�, thevariations of air densities and airflow temperatures are slight.The maximum air density difference and solar radiation occurat 30�, and themaximum air speed and effective pressure occurat 60�. The height varies from 15� to 60� is significant, whereasa slower height increase is found after 60�.

2) When the solar collector angle decreases from 30� (see Fig. 10(a)), all parameters decrease except the air density. Thedecrease of the airflow speed and effective pressure cause thedecline of the power output.

3) When the solar collector angle increases from 30� to 60� (seeFig. 10 (a)), the solar radiation and the density differences bothdecrease, but the air speed, the height and the effective pres-sure increase. Power generation in this range also increases.

4) When the solar collector angle is larger than 60� (see Fig. 10(b)), all parameters decrease except the air density andheight. However, the power generation decreases in this range.As the inclination of the collector enlarges in this range, twothings occur that work in opposite directions with respect tothe power generation. A higher inclination results in a lowerexposure of the collector surface to solar irradiation and henceyield smaller heat utilization and smaller airflow speed. On theother hand, tilting the collector increases the height as well aspressure difference of the collector and so enhances the airflow.As the influence of collector’s increasing pressure difference isweaker than the decreasing airflow speed, power generationafter 60� finally declines. A similar description is also reportedby Sakonidou et al. on optimizing tilted angle of a solarchimney for maximum airflow rate [36].

Overall, the main factors that influence SSCPP performancesare the air density difference and the height. When the collector

Solar radiation [Ht]

Density [ρo] Airflow speed [vo] Effective pressure [ΔP

t]

-60

-40

-20

0

20

Airflow temperature [(Ta+T

o)/2]

Density difference [ρa-ρo]

Height [Hchi+H

col/2]

Power [Pele

]

Rel

ativ

e pr

opor

tions

to th

e re

fere

nce

plan

t/ %

80o70o60o50o40o36o30o25o15oRel

ativ

e pr

opor

tions

to th

e re

fere

nce

plan

t/ %

-40

-20

0

20

40

60

80

100

b

Collector angle

80o70o60o50o40o36o30o25o15oa

Fig. 10. Relative proportions of SSCPPs with different collector angles to the reference system:Performanceangle � Performancereference

Performancereference� 100%, (a) when the reference SSCPP’s

collector angle is 30� , (b) when the reference SSCPP’s collector angle is 60� .

F. Cao et al. / Applied Thermal Engineering 50 (2013) 582e592590

angle is rising from the horizontal, the effect of the heightincrease plays the leading role. However, when the collector angleis too large (larger than 60� in this study), the main influencewould be created by the air density differences. The air densityvariation with temperature changing is the thermal characteristicof air. In order to enlarge the power generation of the SCPP,a proper working fluid whose density decreases sharply with thetemperature rise is suggested to be the working fluid. However,air is the most cost-effective, clean and abundant working fluidfor SCPPs. Other working fluids would also increase thecomplexity of the system [38,39]. Besides, the mass flow rates ofSCPPs’ working fluid are huge according to Fig. 6. Utilizing otherfluids would be impossible from the techno-economic point ofview. The system height has beneficial effects on the powergeneration. However, the mountain resources, whose surfaceslopes are 60� for example, are limited. Building SSCPPs alonga high slope mountainside would increase the transport and

construction investments. Besides, Zhou et al. reported that thesystem power generation would be decreased by 13%, if consid-ering the irregularity of the surface [19].

4.3. Power generation of SCPPs at different latitudes in China

The performances of CSCPPs and SSCPPs at different latitudes ofChina are summarized in Table 2. It can be found that:

1) For every region, there are always a maximum power genera-tion (MPG) angle and a maximum solar radiation (MSR) anglefor SSCPPs. For low latitude regions, CSCPPs have higher powergeneration than SSCPPs with MSR angles; but for high lati-tudes, SSCPPs with MSR angles have higher power generation.

2) The power generation of SCPPs is not only influenced by thetotal solar radiation, but also has somehow relationship withthe solar duration and the ambient temperature.

Table 2Power generations of CSCPPs and SSCPPs at different latitudes of China.

City Local geographical information CSCPP SSCPP

Maximum solar radiation (MSR) Maximum power generation (MPG)

Latitude/�

Solarduration/h

Ambienttemperature/K

Radiation onhorizontal surfaces/W m�2

Powergeneration/MW

Collectorangle/�

Radiation onsloped surfaces/W m�2

Powergeneration/MW

Collectorangle/�

Radiation onsloped surfaces/W m�2

Powergeneration/MW

Guangzhou 23.10 1628.0 295.2 679.48 4.98 21 689.81 2.72 61 598.03 5.00Wuhan 30.37 1835.1 289.8 600.04 4.31 22 644.54 2.68 61 560.76 4.77Shanghai 31.10 1894.5 289.3 657.61 4.75 26 686.58 3.29 62 606.41 5.25Beijing 39.48 2067.8 285.4 514.25 3.61 38 592.84 4.80 68 540.11 5.40Shenyang 41.44 2468.0 281.5 546.64 3.85 40 637.59 4.38 66 593.10 5.44Urumchi 43.47 2523.3 280.1 553.46 3.91 36 623.06 3.97 64 570.82 5.17Harbin 45.45 2571.2 277.4 538.77 3.77 43 652.70 4.81 67 612.40 5.75

F. Cao et al. / Applied Thermal Engineering 50 (2013) 582e592 591

3) MSR angles are nearly 2e8� smaller than the local latitudes.However, the MPG angles are 25e38� higher than the locallatitudes.

SSCPPs can enlarge the power production and increase theenergy conversion efficiency. However, as is shown in Fig. 2, theperformance of SSCPPs is closely related to the geographicresources. Fig. 11 shows that most long mountain chains and highinsolation areas are distributed in regions with the latitude higher

Fig. 11. Solar radiation and mountain chains distribution in China: the radiation is presentpretation of the references to color in this figure legend, the reader is referred to the web

than 30� N and longitude smaller than 105� E [40,41]. SSCPPs couldbe an adaptation to the local conditions in this region. For southeastand eastern regions, solar radiation is relatively weak, and moun-tain chains are discontinuous and short. CSCPP is a more economicand suitable solution, especially in combination with vegetationunder the collector roof for possible agricultural purposes. Pretor-ius reported that though the performances of the SCPP decreasenearly 50% in the vegetable style of SCPP, the advantage of landsaving and greenhouse effect for grain and power generation are of

ed by different colors, and the mountain chains are identified by numbers. (For inter-version of this article.)

F. Cao et al. / Applied Thermal Engineering 50 (2013) 582e592592

high values [32]. There are only several mountain chains located inNortheast China. Besides, for the Sichuan Basin, the solar radiationis too weak and annually average temperature is high. So, SCPPsmay not be suitable to be built in these regions.

5. Conclusions

SCPPs are promising for large-scale utilization of solar energy,and extensive research has been carried out to investigate its huge-potential over the world. In this study, we compare the perfor-mances, such as the airflow temperature, the mass flow rate, thepower generation and the efficiencies, of a CSCPP and two SSCPPs.The energy consumption, the main factors that influence the powergeneration and the power generation at different latitudes of Chinaare also analyzed. Results show that:

1) CSCPPs, of which the highest power outputs are generated insummer, have symmetrical performances throughout the year.Tilting the solar collector angles from the horizontal wouldimprove SSCPPs’ performances in winter but weaken them insummer.

2) For different regions, configuration sizes of SCPPs are differentbecause of various meteorological and geographic conditions.The SCPP has low power generation efficiency but relativelyhigh thermal efficiency. The ground energy loss, reflected solarradiation and kinetic loss at the chimney outlet are the mainenergy losses in SCPPs.

3) There are two special collector angles for SSCPPs: themaximum solar radiation (MSR) angle and the maximumpower generation (MPG) angle. For Lanzhou (36.03�N), theMSR angle is 6� smaller than the local latitude, and the MPGangle is 24� larger than the latitude. The main reasons causingthe angle difference are the system height and the air ther-mophysical characteristics. For different regions, factors thatinfluence the power generation of SCPPs are the total solarradiation, the solar duration and the ambient air temperature.

4) SSCPPs are suggested to be built in Norwest China, where solarenergy and mountain resources are abundant; whereas CSCPPsare more suitable to Southeast and East China, where it couldbe combined with the local agriculture.

Acknowledgements

This research was funded by the National Natural ScienceFoundation of China (Nos.: 50506025, 51121092) and Programmefor New Century Excellent Talents in University (No.: NCET-08-0440). We would also like to thank the National MeteorologicalInformation Centre, China Meteorological Administration for itsdata support.

References

[1] J. Lorenzo, Las Chimneas solares: De una propuesta espanola en 1903 a deManzanares, De Los Archivos Históricos De La Energia Solar, http://www.fotovoltaica.com/chimenea.pdf (accessed 11.11.11).

[2] H. Günther, In Hundred Years-future Energy Supply of theWorld, FranckhscheVerlagshandlung, Stuttgart, 1931.

[3] W. Haaf, G. Friedrich, G. Mayr, J. Schlaich, Solar chimneys, part I: principle andconstruction of the pilot plant in Manzanares, International Journal of SolarEnergy 2 (1983) 3e20.

[4] W. Haaf, Solar chimneys, part II: preliminary test results from the Manzanarespilot plant, International Journal of Solar Energy 2 (1984) 141e161.

[5] T.W. von Backstrom, A. Bernhardt, A.J. Gannon, Pressure drop in solar powerplant chimneys, Journal of Solar Energy Engineering 125 (2003) 165e169.

[6] T.W. von Backstrom, A.J. Gannon, Solar chimney turbine characteristics, SolarEnergy 76 (2004) 235e241.

[7] T.P. Fluri, T.W. von Backstrom, Performance analysis of the power conversionunit of a solar chimney power plant, Solar Energy 82 (2008) 999e1008.

[8] M.A. dos S. Bernardes, T.W. von Backstrom, D.G. Kroger, Analysis of someavailable heat transfer coefficients applicable to solar chimney power plantcollectors, Solar Energy 83 (2009) 264e275.

[9] M.A. dos S. Bernardes, T.W. von Backstrom, Evaluation of operational controlstrategies applicable to solar chimney power plants, Solar Energy 84 (2010)277e288.

[10] J.P. Pretorius, D.G. Kroger, Sensitivity analysis of the operating and technicalspecifications of a solar chimney power plant, Journal of Solar Energy Engi-neering 129 (2007) 171e178.

[11] J.P. Pretorius, D.G. Kroger, Critical evaluation of solar chimney power plantperformance, Solar Energy 80 (2006) 535e544.

[12] X.P. Zhou, J.K. Yang, B. Xiao, G.X. Hou, F. Xing, Analysis of chimney height forsolar chimney power plant, Applied Thermal Engineering 29 (2009) 178e185.

[13] X.P. Zhou, J.K. Yang, B. Xiao, X.Y. Shi, Special climate around a commercialsolar chimney power plant, Journal of Energy Engineering 134 (2008) 6e14.

[14] J. Schlaich, R. Bergermann, W. Schiel, G. Weinrebe, Design of commercial solarupdraft tower systems e utilization of solar induced convective flows forpower generation, Journal of Solar Energy Engineering 127 (2005) 117e124.

[15] M.Q. Yan, S.A. Sherif, G.T. Kridli, S.S. Lee, M.M. Padki, Thermo-fluids analysis ofsolar chimneys, Industrial Applications of Fluid Dynamics ASME FED 132(1991) 125e130.

[16] N. Pasumarthi, S.A. Sherif, Experimental and theoretical performance ofa demonstration solar chimney model-part I: mathematical model develop-ment, International Journal of Energy Research 22 (1998) 277e288.

[17] E. Bilgen, J. Rheault, Solar chimney power plants for high latitudes, SolarEnergy 79 (2005) 449e458.

[18] M.A. Serag-Eldin, Analysis of a newsolar chimney plant design formountainousregions, WIT Transactions on Engineering Sciences 53 (2006) 437e446.

[19] X.P. Zhou, J.K. Yang, A novel solar thermal power plant with floating chimneystiffened onto a mountainside and potential of the power generation inChina’s deserts, Heat Transfer Engineering 30 (2009) 400e407.

[20] J. Wei, F.H. Wang, L. Zhao, Research on the best slope gradient of slope solarinduced convective flows power generation system, in: D.Y. Goswami,Y.W. Zhao (Eds.), Proceedings of ISES 2007 Solar World Congress Conference,Tsinghua University Press, Beijing, 2007, pp. 1795e1799.

[21] F. Cao, L. Zhao, H.S. Li, L.L. Fan, Y. Han, Angle optimization of sloped solar chimneysystem, in: Proceedings of the ASME-ATI-UIT Conference on Thermal and Envi-ronmental Issues in Energy Systems, Edizioni ETS, Pisa, 2010, pp. 695e700.

[22] L.B. Mullett, Solar chimney-overall efficiency, design and performance,International Journal of Ambient Energy 8 (1987) 35e40.

[23] X.P. Zhou, F. Wang, J. Fan, R.M. Ochieng, Performance of solar chimney powerplant in Qinghai-Tibet Plateau, Renewable and Sustainable Energy Reviews 14(2010) 2249e2255.

[24] S. Nizetic, N. Ninic, B. Klarin, Analysis and feasibility of implementing solarchimney power plants in the Mediterranean region, Energy 33 (2008)1680e1690.

[25] S. Larbi, A. Bouhdjar, T. Chergui, Performance analysis of a solar chimneypower plant in the southwestern region of Algeria, Renewable and Sustain-able Energy Reviews 14 (2010) 470e477.

[26] C. Ketlogetswe, J.K. Fiszdon, O.O. Seabe, Solar chimney power generationproject e the case for Botswana, Renewable and Sustainable Energy Reviews12 (2008) 2005e2012.

[27] Y.J. Dai, H.B. Huang, R.Z. Wang, Case study of solar chimney power plants inNorthwestern regions of China, Renewable Energy 28 (2003) 1295e1304.

[28] F. Cao, L. Zhao, L.J. Guo, Simulation of a sloped solar chimney power plant inLanzhou, Energy Conversion and Management 52 (2011) 2360e2366.

[29] B.Z. Wang, H.S. Zou, Z.Q. Yin, Solar energy distribution in China, Solar Energy 4(1998) 19 (in Chinese).

[30] T. Muneer, Solar Radiation and Daylight Models, Elsevier, Oxford, 2004.[31] J.A. Duffie, W.A. Beckman, Solar Engineering of Thermal Process, third ed.,

John Wiley & Sons Inc, Hoboken, 2006.[32] J.P. Pretorius, Optimization and control of a large-scale solar chimney power

plant, Ph.D. thesis, University of Stellenbosch, South Africa, 2007.[33] S.M. Yang, W.Q. Tao, Heat Transfer, fourth ed., Higher Education Press, Beijing,

2006 (in Chinese).[34] S. Lorente, A. Koonsrisuk, A. Bejan, Constructal distribution of solar chimney

power plants: few large and many small, International Journal of GreenEnergy 7 (2010) 577e592.

[35] H.S. Li, Optimum inclination of collector for slope solar chimney powergeneration systems, Master thesis, Xi’an Jiaotong University, P.R. China, 2008(in Chinese).

[36] E.P. Sakonidou, T.D. Karapantsios, A.I. Balouktsis, D. Chassapis, Modeling of theoptimum tilt of a solar chimney for maximum air flow, Solar Energy 82 (2008)80e94.

[37] M.A. dos S. Bernardes, G. Weinrebe, Thermal and technical analysis of solarchimneys, Solar Energy 75 (2003) 511e524.

[38] B.A. Kashiwa, C.B. Kashiwa, The solar cyclone: a solar chimney for harvestingatmospheric water, Energy 33 (2008) 331e339.

[39] Y.P. Wang, J.H. Wang, L. Zhu, Z.Y. Yang, Z.L. Fang, The study of sea desalinationand hot wind electric power integrated system by solar chimney, ActaEnergiae Solaris Sinica 7 (2006) 731e736 (in Chinese).

[40] China Global Horizontal Solar Radiation, http://www.nrel.gov/gis/pdfs/swera/china/china40kmglo.pdf (accessed 11.11.11).

[41] X.P. Wang, Z.H. Wang, J.Y. Fang, Mountain ranges and peaks in China, Biodi-versity Science 12 (2004) 206e212 (in Chinese).