A non-tracking concentrating collector for solar thermal applications

A non-tracking concentrating collector for solar thermal applications

Applied Energy 200 (2017) 39–46 Contents lists available at ScienceDirect Applied Energy journal homepage: www.elsevier.com/locate/apenergy A non-t...

948KB Sizes 38 Downloads 200 Views

Applied Energy 200 (2017) 39–46

Contents lists available at ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

A non-tracking concentrating collector for solar thermal applications Wattana Ratismith a, Yann Favre b, Maxime Canaff c, John Briggs d,⇑ a

Energy Systems Research Institute, Prince of Songkla University, Songkla, Thailand Joseph Fourier University of Grenoble, Grenoble, France c University of Technology of Belfort-Montbliard, Belfort, France d Theoretical Quantum Dynamics, Institute of Physics, University of Freiburg, Germany b

h i g h l i g h t s  A non-tracking solar collector of novel design is tested outdoor in SE Asia.  A novel performance parameter, the power concentration factor, is introduced.  The header pipe has a new ‘‘wet” connection giving enhanced efficiency.  The collector is superior to a commercially-available non-concentrating collector.

a r t i c l e

i n f o

Article history: Received 19 July 2016 Received in revised form 28 March 2017 Accepted 4 May 2017

Keywords: Solar thermal energy Compound parabolic trough Solar collector efficiency

a b s t r a c t We report the development of a solar thermal collector module based on our proposed design (Ratismith et.al., 2014) of a large acceptance-angle multiple-parabolic trough which surrounds a standard evacuated cylindrical tube containing a planar absorber plate. The concentrator accepts diffuse solar radiation with an intercept factor of near 100% and so is suitable particularly for tropical climates. The module incorporates a novel direct metal-to-water contact resulting in an improved efficiency of heat transfer to the working liquid. Comparison with the performance characteristics, principally power output and temperatures attained, of a commercially-available non-concentrating assembly of evacuated absorber tubes is made. The experimental results, obtained by testing under typical conditions of solar irradiation throughout the day in Bangkok, indicate that the improvement over a non-concentrating collector, suggested theoretically on the basis of ray-tracing studies, is attained in practice. The superior performance of the concentrating collector indicates its suitability both for residential and industrial applications. Ó 2017 Published by Elsevier Ltd.

1. Introduction In a recent paper in Applied Energy [1] on ‘‘Assessment of renewables for energy security and carbon mitigation in Southeast Asia” it was concluded that ‘‘ expanding the share of renewables in the energy mix can bring extensive socio-economic benefits to the Southeast Asian countries”. The study concentrated mostly on power generation and hence the harnessing of solar power in the form of photovoltaics. Here we report on the alternative development of small-scale solar thermal collectors for heating and cooling. Although small-scale, since solar irradiation is considerable throughout the Southeast Asia region, were such installations widely implemented they could lead to a significant reduction in primary electricity demand. The solar collector proposed here is based on a new design of CPC trough proposed in our previous ⇑ Corresponding author. E-mail address: [email protected] (J. Briggs). http://dx.doi.org/10.1016/j.apenergy.2017.05.044 0306-2619/Ó 2017 Published by Elsevier Ltd.

publication [2]. The collector is installed in central Bangkok and has the important characteristics of being non-tracking, therefore cheap to manufacture and maintain. Furthermore, it operates efficiently in both clear and cloudy conditions as pertain in most of Southeast Asia. The collector is completely scaleable in size and therefore suitable for a variety of residential and industrial applications. Since the seventies of the last century there have been many suggestions for the use of compound parabolic concentrators (CPC) for concentrating solar radiation. As early as 1975, in a seminal paper, Winston and Hinterberger [3] elucidated the principles for optimisation of the concentration efficiency of such concentrating collectors for solar thermal and photovoltaic applications. Note that the designation ‘‘CPC” is often used also for collectors not always consisting of parabolas. Some earlier important examples, for both photovoltaics and for solar heating, which is the subject of this paper, we have discussed in detail in Ref. [2]. Interest in CPC continues and more recently, there have been several papers

40

W. Ratismith et al. / Applied Energy 200 (2017) 39–46

on design improvements [4–6]. Other papers concentrate on a wide variety of applications, indicating the versatility of CPC solar concentrators. For example, a non-tracking external CPC (XCPC) with a cylindrical absorber has been implemented [7] to drive successfully a chiller for air conditioning, which is one of the applications described in this paper. In Refs. [8–10] the straightforward heating of water is discussed whilst in Ref. [11] the collector is used to heat air. The industrial application to generate process heat up to 300° is outlined in Ref. [12] and as example, the design of a solar thermal plant for methanol reforming is given in Ref. [13]. In contrast, there are also many proposals to use the CPC solar concentrator for cooling, either refrigeration [14] or to drive a chiller for air conditioning [15] However, almost all of these proposed systems are based on the conventional CPC design with a doubleminimum in cross-section or with a horizontal flat absorber at the base. In some cases only a small acceptance angle is achieved by the CPC. The challenge for any non-tracking collector is to optimise the capture of solar energy throughout the day. This implies having a large acceptance angle but at the same time achieving an intercept factor close to 100%. This intercept factor, which we denote by I ðhÞ, where h is the angle of incidence of radiation with respect to the vertical, is defined as that fraction of the radiation entering the aperture which is captured by the absorber plate. Our unconventional design, which we called the flat-base trough (FBT), is a significant step towards achieving this aim. It has a large acceptance angle and the desirable property that throughout the day almost 100% of solar radiation is captured by a verticallyplaced flat absorber plate. In Ref. [2] we demonstrated the superiority of the FBT with vertical absorber to the conventional CPC having a double parabola with a central peak in the cross-section and a horizontal flatplate absorber. The main aim of this paper is to demonstrate the advantage, in terms of power harvested per tube, of embedding a flat-plate collector tube in our FBT trough rather than simply exposing the tube itself to solar radiation, as is standard in commercially-available flat-plate solar installations. Hence in an experiment we demonstrate the superior performance of a collector composed of fifteen connected FBT’s to that of a commerciallyavailable non-concentrating assembly of sixteen of the same evacuated tubes. (The slightly different number is due to the fact that our collector is composed of modules of three tubes, the commercial collector has modules of eight tubes). It is shown that the expected power concentration factor is achieved by the FBT concentrating collector so that it requires roughly only half the number of tubes as the non-concentrator to harvest the same output power. A further novel feature of the collector design compared to commercial flat-plate collectors is the immersion of the tube header directly into the working fluid (wet connection) rather than the simple conductive dry connection usually employed. This leads to efficient direct transfer of heat to the working fluid as we demonstrate by measuring the efficiency of identical collector modules with both dry and wet connections. As a first application of the FBT concentrator, a modular system installed on the roof of a 12-storey building at Chulalongkorn University in the centre of Bangkok, Thailand is connected to a chiller to drive an air conditioning unit, as described in detail in Section 5. In this way we show that the collector is capable of achieving output temperatures and power sufficient for residential and low-scale industrial applications. The concentrator has been designed to operate in both direct and diffuse irradiation conditions. This is important particularly for use in tropical climates where, for example in Thailand 60% of solar irradiation is diffuse in humid and cloudy conditions. This capture of radiation from many directions is achieved without

tracking by the trough shape which is designed specifically to give a large acceptance angle. A further feature of the collector contributing to wide acceptance is the arrangement of the troughs in modules of three, the two outer troughs being tilted away from the vertical. This arrangement ensures also that a smooth variation in the captured solar power throughout the daylight hours, from roughly 7 a.m. to 5 p.m., is achieved. We describe test results showing that the FBT concentrator achieves higher temperatures and increased power output compared to the non-concentrator under the typical weather conditions of, (1) bright clear skies, (2) intermittent cloud/sunshine and (3) cloudy for most of the day. The potential for medium to high- temperature (70–150 °C) application of solar thermal collectors in industrial applications and in particular, its role in air conditioning and the consequent fuel savings, reduction in greenhouse gases, etc. is described in some detail in Refs. [16,17]. Here we demonstrate that our collector is suitable both for domestic use, where a considerable saving on the number of evacuated collector tubes necessary can be achieved, and for air conditioning of industrial buildings. The structure of the paper is as follows. In Section 2 we describe briefly the salient features of the capture and concentration characteristics of the flat-base trough proposed in [2]. Furthermore we introduce a new dimensionless parameter, the ‘‘power concentration factor”, with which to quantify the performance of a concentrating trough. Then in Section 3 we discuss the structure of the modules of three troughs. In particular we describe the comparison of the efficiencies of concentrating modules with wet and dry connection of the header pipe to the tube in which the water to be heated flows and the improvement of the transfer of heat from the absorber pipe to the working fluid, which in our case is water. The heated water is used, for example, to drive an absorption chiller for cooling purposes. The manifold header pipe from the absorber tube is now inserted directly into the flowing water. This we call a ‘‘wet” connection and is to be contrasted with the conventional ‘‘dry” design in which the header pipe is exterior to the container of the working liquid and makes a simple conductive contact with it. By arranging that the working liquid flows around the manifold header pipe the heat transfer is effected more directly. The testing of an assembly of five connected concentrator modules (15 absorber tubes) in outside conditions is the subject of Section 4 of the paper. We compare directly with the performance of a commercially-available standard configuration of collector tubes without solar concentration and demonstrate, under the weather conditions listed above, that our design of concentrator leads to the expected increase in output power per collector tube and a significant rise in the inflow-outflow temperature difference. In Section 5 we describe the application of the collector to airconditioning and give data on the achievement of the power output necessary to drive the chiller. We discuss further realisable applications of the collector and then present our conclusions in Section 6. A note as to nomenclature; the assembly of three troughs into a single unit will be called ‘‘a module”. The connected set of five such concentrating modules will be called the ‘‘concentrating collector”. The commercially-available assembly of sixteen vacuum tubes will be called the ‘‘non-concentrating collector”.

2. The concentrator trough Here we present the pertinent capture and concentration characteristics of the flat-base trough (FBT) proposed and described in detail in [2]. This trough is composed of two overlapping parabolic segments which are truncated at their intersection and then rotated such that the gradient of the curve at the centre of the cross-section of the trough is zero, i.e. the base is flat. This is shown

W. Ratismith et al. / Applied Energy 200 (2017) 39–46

in Fig. 1. Then the whole trough is truncated in height at the point at which the gradient of the sides of the trough becomes infinite i.e. vertical. The intersection point and the curvature of the parabola decide the precise shape of the trough. The parameters chosen for the FBT of Ref. [2] (length 2.2 m, width 0.22 m and height 0.156 m) means that the trough aperture width is a factor approximately 1.4 times larger than its height. The trough is specifically designed such that, unusually, a vertically-placed absorber plate will capture almost 100% of the light incident on the trough. This is achieved by the shape of the trough bottom giving rise to a ‘‘tea-cup” cusp in reflection, as shown in Fig. 2. Then at zero degree incidence 100% of light incident is captured. When light is incident from a finite angle, this cusp breaks up into two symmetric parts. Since each side of the trough is a parabola whose base is rotated, light incident at an angle will almost pass through the corresponding focus and give a fan of rays intercepting one side of the absorber. This is seen clearly in Fig. 2 for 30° incidence. The examples shown in Fig. 2 are for the period roughly 12 noon to 5 p.m. The corresponding period from 7 a.m. to 12 noon will have the light incident from the opposite angles and the reflection pattern is completely mirror symmetric. The aim of using a concentrator is to enhance the amount of solar radiation entering the absorber plate located in the trough compared to the amount which enters in the absence of the concentrator trough. For a given trough this enhancement is often expressed in terms of a geometric quantity known as the concentration ratio, depending upon the dimensions of the trough and the precise shape and size of the absorber plate. However, our sole aim in this study is to compare the characteristics of our trough with a commercial absorber tube which is placed horizontally (e.g. on a roof) without the concentrating trough. That is, we compare the light-gathering properties of two identical absorber tubes, one within the trough and one without. Then the relevant quantity is the gain in harvested power of our trough, containing a single vertically-placed tube, compared to that achieved by a single ‘‘naked” horizontally-placed tube. Accordingly, to avoid confusion, we will call this factor the ‘‘concentration gain” and denote it by C. In our case the flat absorber plate and trough have the same length L so that the maximum gain from concentration is simply



AL A ¼ aL a

ð1Þ

where A is the width of the trough and a the width of the absorber plate. The intercept factor I ðhÞ of a trough, as a function of the angle of incidence h, is defined as the fraction of the radiation entering the trough that is captured by the absorber, the remainder being reemitted from the trough. This factor must be calculated from ray-

41

tracing simulations for each angle h of incidence of the solar irradiation. The values for the FBT are shown in Fig. 2. In assessing the theoretical trough performance, losses due to transmission and reflection must be taken into account. Following the manufacturers’ specifications and ignoring some incidentangle dependence these can be estimated as: the trough glass cover plate transmittivity is 90%, the trough aluminium surface reflectivity 95%, the glass cover of the absorber plate has transmittivity 90% and the absorbance of the absorber plate 95%. The product of these trough loss factors is 0:9  0:95  0:9  0:95 ¼ 0:73. For the non-concentrating collector the last two losses also apply, hence 0:9  0:95 ¼ 0:86. Therefore, in comparing the concentrating collector to the non-concentrating one must include an additional loss factor defined as L ¼ 0:73=0:86 ¼ 0:85. Although concentration gain and intercept factor are usually considered separately, the meaningful dimensionless factor, the power enhancement of our absorber trough compared to the commercial tube alone i.e. due to concentration, is given by what we will call the power concentration factor (PCF). This factor we define as PðhÞ  Q T =Q a where Q T is the power entering the absorber contained in the trough and Q a the power absorbed by a horizontal absorber without trough. For an insolation of GðhÞ in Watts/m2, the power in Watts entering the horizontal absorber plate without trough at angle h is aL cos hGðhÞ. Similarly, by definition of I , the power Q T entering the absorber plate in the trough is given by Q T ¼ AL cos hGðhÞI ðhÞ. Hence

PðhÞ ¼

AL cos hGðhÞI ðhÞ A L ¼ I ðhÞL ¼ CI ðhÞL aL cos hGðhÞ a

ð2Þ

where L is the factor defined above representing additional losses of the concentrating trough. This shows that the power concentration factor PðhÞ is a trough property independent of the solar power. The details of the method used to calculate I ðhÞ from the ray tracing diagram are given in Ref. [2]. As can be calculated from Fig. 2, for incident angles 60° or lower, the average intercept factor is I ¼ 0:99. The FBT trough has C ¼ 2:2 and so from Eq. (2), with L ¼ 0:85 one has the expected average power concentration factor of P ¼ 2:2  0:99  0:85 ¼ 1:85. This theoretical value will be compared to the experimental value in Section 4. After testing various possibilities, it was concluded that the optimal alignment of the absorber tube in the trough is such that its lower end sits on the trough base and its plane coincides with the vertical axis of the trough cross-section as shown in Fig. 2. Although this diminishes the amount of direct sunlight incident on the absorber plate it leads to interception of radiation through reflection from the trough sides for virtually all angles of incidence as can be seen from Fig. 2. Then an intercept factor of close to unity is obtained throughout the main part of the sunlight hours of the

Fig. 1. The construction of the flat-base trough (FBT). The two parabolic segments of (a) are rotated in opposite senses to give curve (b) which has zero gradient at the origin.

42

W. Ratismith et al. / Applied Energy 200 (2017) 39–46

0

o

o

15

o

30

100%

94%

100%

o

o

75

60

45

100%

100%

o

0%

Fig. 2. Ray-tracing diagrams for different angles of incidence. Blue, orange and green lines are incident, 1st reflected and 2nd reflected rays respectively. The circle is the absorber tube and the vertical thick line represents the flat-plate absorber. The number under each tough is the intercept factor expressed as a percentage. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

increased to 44 cm. Such a trough, Fig. 3c, would then concentrate twice the sunlight of our trough of Fig. 3a. An alternative, shown in Fig. 3d, (see also Ref. [5]) would be to set two tubes of 10 cm diameter one above the other. This gives the same area as a single absorber of 20 cm diameter but, since at different times of day predominantly one or other of the tubes absorbs most of the incident light, this would lead to an effectively higher concentration ratio and output temperature. In principle, this pattern could be continued and f tubes stacked one above the other in one large trough of aperture 22f cm. This would also have a near 100% intercept factor. The necessity of scaling the size of the collector to satisfy residential or commercial requirements is discussed in Section 5.

day, i.e. this intercept factor is maintained out to an incident angle of approximately 60°. Since the absorber is vertical, all radiation is intercepted after reflection at the trough surface. From Fig. 2b one sees that for vertical incidence the reflected rays show a typical cusp shape and 100% of radiation is intercepted. As the angle of incidence increases, the cusp breaks up and for incident angles greater than 30° the reflection pattern is dominated essentially by the nearfocus of the left hand half-parabola of the trough. This also leads to 100% interception. Interestingly, at 15° incidence the break-up of the cusp leads to around 6% of the radiation exiting the trough. Further ray-tracing analysis of the trough properties have revealed its versatility and, above all, scalability. Any absorber shape which has the same projection onto the vertical axis as the planar vertical absorber, will give the same intercept factor for incident radiation. For example, a cylindrical absorber with circular cross-section of diameter equal to the size of the planar absorber (10 cm in our case) will give the same, almost 100%, absorption of incident rays as shown in Fig. 2b. The trough is linearly scalable with no loss of intercept factor or reduction of acceptance angle. This is illustrated in Fig. 3 where the present trough and absorber are shown in Fig. 3a. The trough, with a 10 cm wide absorber plate, has an aperture width of 22 cm and a height of 15.6 cm. An absorber increased or reduced in size by a factor f can be embedded in a trough of the same shape but whose aperture is also changed by the factor f. For example, were one to use a planar absorber plate of 5 cm diameter, as in Fig. 3b, the trough aperture can be reduced to 11 cm and the height to only 7.8 cm. By contrast, one can increase the size of the absorber, e.g. by doubling to 20 cm diameter, when the trough aperture would be

a

b

3. The three-trough module and the metal-to-water contact As depicted in Ref. [2] and shown in Fig. 4, a single collector module is composed of three concentrator troughs with the two outer troughs tilted at 15 to the vertical. Although, due to the large acceptance angle of the troughs, this does not lead to an increase in the total absorbed power integrated over the day, it does lead to a smoother power output as a function of time. Of course in the extreme limit of randomly diffuse irradiation such tilting is unnecessary. The testing of a single module alone is reported in Ref. [2]. The purpose of this paper is to compare the performance of a complete collector assembly of these modules with a non-concentrating collector composed of roughly the same number of tubes under typical outside conditions in the semitropical climate of Thailand.

c

d

Fig. 3. (a) The FBT used in the test with an absorber plate of 10 cm height and a trough aperture of 22 cm. (b) The FBT scaled down by one half. (c) The FBT scaled up by a factor of two. (d) The same as (c) but with a stack of two absorber tubes.

W. Ratismith et al. / Applied Energy 200 (2017) 39–46

Fig. 4. The collector of five concentrating trough modules containing a total of fifteen evacuated tubes.

Many of the evacuated tube solar collectors commercially available consist of modules where the absorber tube header pipe is not directly in contact with the heat transfer fluid. That is they have an exterior metal-to-metal contact and the ‘‘dry” connection to the tube in which the fluid flows consists of a header pipe in thermal contact with the water pipe. The contact area between the two pipes may not be optimal giving a corresponding non-optimal heat transfer. Clearly such a simplified construction leads to losses in the transfer of heat to the working fluid. For these reasons the collector prototype reported here has been equipped with a special manifold heat pipe. The manifold is put directly inside the water flow to ensure efficient heat transfer and minimise radiative losses. This we call the ‘‘wet” connection. In one type of commercial dry connection the header pipe simply rests in a semi-circular groove on the top of the tube in which water flows. In a preliminary comparison, whose results are reported in Ref. [18], we tested the efficiency of a single FBT concentrating module employing the wet connection against the efficiency of a commercial non-concentrating module using this simplest dry connection. The test results of Ref. [18] show that the wet-connected concentrator module provides an increase in temperature gain and an efficiency roughly 25% larger than those of the non-concentrating module. Here we report a more direct comparison in that, as shown in the diagram Fig. 5 of the test arrangement, we compare the efficiency of two identicalFBT concentrating modules, one using the dry connection and the other the wet connection. However, to emphasise that the wet connection is not better simply because the dry connection, as in the above case of contact with a groove in the flow pipe, has a bad thermal connection, we have improved the thermal contact area of the dry connection as illustrated in Fig. 6a. The dry thermal connection has been enhanced by fitting the cylindrical header pipe into a cylindrical tube around which the water to be heated flows, giving all-round contact. Heat transfer proceeds by metal-to-metal conduction. By contrast, the wet connection shown in Fig. 6b has the header pipe immersed in the water flow to give a direct metal-to-water connection. The efficiency of the module is defined as g ¼ Q e =Q i where the input power Q i is equal to the insolation G multiplied by the total collector area and the output power Q e is calculated as

_ C ðT e  T i Þ Q e ðt Þ ¼ m

ð3Þ

_ is the water flow rate in Kg/s and C is the water specific where m heat in J/(Kg °C). According to the ISO standard Ref. [19], the efficiency is plotted against the variable x ¼ ðT m  T a Þ=G, where T a is the ambient temperature of the experiment and the mean temperature is given, for initial water temperature T i and exit water

43

temperature T e by T m ¼ ðT e þ T i Þ=2. To measure the efficiency, experiments were conducted over a short period of two to three hours to ensure relative constancy of solar power and ambient temperature. The inlet temperature is varied between 40 °C and 70 °C and is carefully controlled by four heaters delivering 9 kW each in conjunction with a cooling tower. This temperature range was chosen for testing, since it is precisely the range suitable for domestic water heating [19], which is the main application of the commercial non-concentrating collector with which we are comparing our concentrating collector. The test began with all heaters operating to give a maximum inlet temperature of 70 °C. Then the heaters are switched off successively to reduce the inlet temperature in steps of 10 °C down to 40 °C. In this way the efficiency is measured for varying values of the parameter x. The results are shown in Fig. 7 with linear fits to the efficiency as a function of x. They demonstrate that the wet connection leads to an x ¼ 0 efficiency of 49% compared to 41% for the improved dry connection i.e. a 20% gain in efficiency.

4. Comparison of concentrating and non-concentrating solar collectors The concentrator configuration is a linked assembly of five modules (see Fig. 4) each containing three tubes to give fifteen tubes in all. Each module has a total width of approximately 0.7 m. Its performance was compared to that of a commercial non-concentrating module composed of sixteen tubes, identical to those used in the concentrator, separated by gaps of 0.02 m to give a total width of 1.90 m. Unfortunately this unit is supplied with sixteen fixed tubes so that the number of tubes in the two assemblies is not exactly equal but significant is that the concentrating collector has one tube fewer. Note also that the commercial collector uses a dry heat pipe manifold heat transfer connection whereas the concentrator assembly is fitted with the new wet connection. The assemblies and test circuit arrangement are shown in Fig. 8. All results and efficiency values are from tests conducted according to the ISO 9806 standard [19]. The parameters to be varied are the water flow rate and the inlet temperature. The water flow is controlled by a 0.75 kW pump giving a constant a constant flow rate of 0.5 m3/h for both collector assemblies. A control valve is located at the concentrating collector inlet in order to balance the flow rate in the two assemblies. A bypass valve has been installed to control the outlet temperature of the whole system allowing the inlet temperature to the cooling tower to be controlled. This cooling tower is used as a thermal load for the system. The heaters are used to maintain the inlet temperature at a constant value. After the water passes through the collectors, the outlet temperature T e is measured. The water is propelled by the pump to the cooling tower where it cools below the inlet temperature T i . Afterwards the water is heated again to the fixed inlet temperature and the cycle repeated. The complete procedure is executed for inlet temperatures varying from 40 °C to 70 °C. As there may be some loss of water (because of vapourisation and leakage), the water level has to be maintained constant in the cooling tower in order to avoid damaging the pump and overheating the system. Exemplary test results are shown in Fig. 9, specifically to illustrate the performance on typical days in Bangkok. Here is shown the variation of the global solar irradiance G, in W/m2, on one day over the hours from 11 a.m. to 4 p.m. In Fig. 9a, upper frame, the solar irradiance is typical of a clear, sunny day when there is a slow, smooth variation from around 900 W/m2 at noon falling gradually to around 400 W/m2 by 4 p.m. In great contrast, see the upper frame of Fig. 9b, on a day of intermittent sunshine and cloud the irradiation is erratic with periods of sunshine giving

44

W. Ratismith et al. / Applied Energy 200 (2017) 39–46

Fig. 5. The test assembly to compare wet and dry connections of two concentrating modules connected in parallel.

a

b Fig. 6. Diagram of the dry connection (a) and the wet connection (b) of the manifold header pipe to the water-flow pipe.

Fig. 7. The efficiency of the collector module with; continuous upper line, wet connection; dot-dash lower line, dry connection.

around 800 W/m2 punctuated by cloudy periods where the power drops to 200 W/m2. This strong variation can occur over periods as short as a few minutes. In Fig. 9c is shown the insolation of a completely cloudy day where it varies smoothly but does not exceed around 500 W/m2 and is entirely diffuse. The features shown on the three upper frames, completely clear, partly clear/partly cloudy

and completely cloudy encompass the conditions typical of a semitropical climate. Shown on the lower frames of each of Fig. 9 is the power output of the two collectors as a function of time during the same period. The power output Q e for constant flow is directly proportional to the difference of outlet and inlet water temperatures for the two collectors as given in Eq. (3). The temperature gain T e  T i is marked on the right scale of the figures and shows, for Fig. 9a, a maximum of around 4 °C for the concentrating collector but only around 2.4 °C for the nonconcentrator. Not surprisingly, the power output and temperature difference follow closely the variations in solar power. On the clear and fully cloudy day (Fig. 9a and c) the power output exhibits the same smooth variation of the solar power. Noteworthy is that the output power, and correspondingly the temperature gain, from the concentrator are significantly and consistently higher than for the non-concentrator. For example, for Fig. 9a, at the peak of the insolation the concentrator achieves a power of 2.4 kW but the nonconcentrator only 1.4 kW. For Fig. 9c, the gain of output power compared to the non-concentrator is even greater, indicating the improved performance of the concentrator in capturing diffuse sunlight. On the intermittent cloudy day typified by Fig. 9b again the power output mirrors the solar power. However, the output power smooths to some extent the strong fluctuations in solar power, due to the finite heat capacity of the collectors. Again the concentrator shows consistently higher output than the non-concentrator and, as may be expected, the gain factor is much higher during periods of more intense solar irradiation. In Section 2, the average power concentration factor is calculated to be P ¼ 1:85. This is in satisfactory agreement with the test results shown in Fig. 9a, where for smooth irradiation a consistent increase in output power for the concentrator of around 1.7 compared to the non-concentrator is achieved. The discrepancy can probably be attributed to additional heat losses from the trough construction compared to the ‘‘naked” vacuum tubes. Considering that the concentrator has one tube fewer, this performance implies that the concentrating collector requires only around half of the number of tubes to achieve the same power output as the nonconcentrating collector. 5. Applications Since the collector is suited to semi-tropical conditions, one obvious and environmentally important application would be use

45

W. Ratismith et al. / Applied Energy 200 (2017) 39–46

COOLING TOWER

Te

Ti

Te

T

T

T

CONSTANT LEVEL L

BY-PASS

Ti

Make up Water

T

F

F

FLOW METER

FLOW METER ELECTRIC HEATER

TEMP CONTROL

20 Kw

0.75 Kw

PUMP

Fig. 8. The piping and instrumentation diagram of the collector installation for testing. The non-concentrating collector is on the left and the five concentrating modules are on the right.

Fig. 9. (a) Upper frame; the solar irradiance as a function of time on a clear, sunny day. Lower frame; the output power (left scale) and temperature gain (right scale) of the concentrating (continuous curve) and non-concentrating (dashed curve) collectors. (b) The same as (a) but data taken on a partly cloudy day. (c) the same as (a) but on a fully cloudy day.

by poor rural communities throughout Asia and Africa. Once installed, the collector is relatively free of maintenance costs and can well provide for domestic heat e.g. for cooking, to minimise consumption of wood and fossil fuels. For example, two modules composed of six vacuum tubes can give an output of approximately 1 kW which is sufficient for low-technology households. The concentrator module of three tubes corresponds in terms of output power, to a non-concentrator of six tubes and, even when the cost of trough manufacture is taken into account, represents a capital saving, the precise amount of course depending largely upon the market price of the vacuum tubes. In rural communities the size and shape of the collector is not of major concern. However, in developed countries domestic application of the collector may be subject to planning regulations. Here the scalability of the collector could come into play. As shown in Fig. 3b a trough with a 5 cm absorber width would be only 11 cm broad and 7.8 cm high. Hence, this collector would be relatively unobtrusive and suitable for roofs of residential buildings. Such an assembly of small collectors would give roughly the same power output per unit area of collector as the present collector but would be half the height. Since the number of tubes per unit area is half that of a conventional non-concentrating collector, the weight and capital cost of the installation would be reduced correspondingly. For larger-scale industrial application, for example for air conditioning of factories or for drying processes, as shown in Fig. 3c and

d, the trough size could be scaled up. Particularly efficacious would be the double-tube design of Fig. 3d, since at certain hours only one of the tubes contributes substantially to absorption, this would lead to an effective doubling of the concentration ratio during that time. In turn this would lead to a higher absorber temperature and output temperature of the flowing water. As mentioned in the Introduction, one of the prime applications of this medium-temperature collector is in the air-conditioning of buildings. To demonstrate that this application is feasible our concentrating collector is presently installed on the roof of the Energy Research Institute, Chulalongkorn University, Bangkok. The concentrating collector installation occupies a gross area of around 40 m2 corresponding to 30 collector modules of three tubes each. This supplies approximately 15 kW for maximum solar irradiation. In addition since the non-concentrating collectors are also installed for comparison purposes, they can be employed to augment the power output to the chiller. Three such units of sixteen tubes each produce only up to 5 kW. Then the total heat source of up to 20 kW drives a lithium-bromide single-effect solar absorption chiller, supplying hot water to the chiller at a temperature over 80 C. In these conditions, the chiller can achieve a 12 kW cooling capacity, supplying cooled water at a temperature of 18 °C. This corresponds to a co-efficient of performance of 0.6 which is close to the value of 0.7 expected under maximum output conditions of 20 kW. Hence, although this cooling system is not run at its full capacity, this application shows that the concentrating collector can be used

46

W. Ratismith et al. / Applied Energy 200 (2017) 39–46

in an already existing industrial-scale process that requires medium to high temperature. At the moment this is employed to aircondition an office of 45 m2. Moreover, it runs efficiently, as the coefficient of performance in our case is nearly the same as the theoretical one. When one considers that Bangkok is surrounded by huge areas of low-rise warehouse and factories offering literally millions of square metres of open roof space, the potential for fuel savings by utilisation of our collector for air-conditioning is apparent. A further commercial realisation of the FBT collector, presently in production, is the use of three modules as a simple water heater coupled to a storage tank and pump. The maximum temperature achievable is 150 °C but the unit can also deliver 300 l of water at 60 °C per day which is adequate for domestic purposes. In any possible commercial application the feasibility and economics of employment of any solar collector are governed by the effective number of hours per day of useful operation. This is quantified by the concept of Utilizability [20]. However, as this quantity depends crucially upon parameters determined by the application and size of collector envisaged, it must be evaluated separately for each particular collector design. 6. Conclusions We have described the implementation of a novel flat-base concentrator trough (FBT) with high acceptance angle and intercept factor in a solar collector configuration particularly suited to operation in tropical or semi-tropical climates. A parameter novel to the solar engineering community, the power concentration factor, is introduced as a dimensionless property of a concentrating trough. The collector is non-tracking and so does not require any moving parts but still can concentrate solar irradiation efficiently over the whole of the daylight hours of useable insolation. A new feature designed to improve heat transfer to the working liquid is the immersion of the manifold header pipes of the collector tubes directly in the flowing liquid. This new design feature has been shown to lead to an approximately 20% increase in efficiency. A test comparison against a non-concentrating collector composed of identical tubes demonstrates that the concentrator can achieve the gains in output temperature increase (on average throughout the day a factor of two) and output solar power measured to be a factor 1.7, in accordance with the improvement expected on the basis of ray-tracing analysis of the FBT properties predicted in Ref. [2]. Our collector design has the novelty of great versatility, in that it can be used as a water heater for low-technology domestic applications, particularly in rural areas, but also is being implemented currently for air conditioning of office space and refrigeration on an industrial scale in a commercial fish farm. Acknowledgements Travel support for two of the authors (WR and JSB) was provided by the Thai-German Researcher Mobility Scheme in the Project OESC administered by the Fraunhofer Institute for Solar Energy Systems (ISE) in Freiburg. Special thanks go to Prof. J. Luther and

Dr. P. Nitz of ISE for helpful advice and comments on the manuscript and to Dr. W. Platzer and the Director, Prof. E. Weber, for the hospitality of their Institute. We thank the Ratchadaphiseksomphot Endowment Fund under the Outstanding Research Performance Program of Chulalongkorn University for financial support. We also thank Mr. Pairoj Anantasethakul for his help in developing the manifold heat pipe and Mr. Narong Amornpitakpunt, AMP Metalworks [Thailand] Co. Ltd for his help in designing the solar collector. References [1] Kumar S. Assessment of renewables for energy security and carbon mitigation in Southeast Asia: the case of Indonesia and Thailand. Appl Energy 2016;163:63–70. [2] Ratismith W, Inthongkhum A, Briggs JS. Two non-tracking solar collectors: design criteria and performance analysis. Appl Energy 2014;131:201–10. [3] Winston R, Hinterberger H. Principles of cylindrical concentrators for solar energy. Sol Energy 1975;17:255–8. [4] Borah A, Khayer SM, Sethi LN. Development of a compound parabolic solar concentrator to increase solar intensity and duration of effective temperature. Int J Agric Food Sci Technol 2013;4:161–8. [5] Abdullahi B, AL-Dadah RK, Mahmoud S, Hood R. Optical and thermal performance of double receiver compound parabolic concentrator. Appl Energy 2015;159:1–10. [6] Karwa N, Jiang L, Winston R, Rosengarten G. Receiver shape optimization for maximizing medium temperature CPC collector efficiency. Sol Energy 2015;122:529–46. [7] Widyolar Bennett, Winston Roland, Jiang Lun, Poiry Heather. Performance of the Merced Demonstration XCPC collector and double effect chiller. J Sol Energy Eng 2014;136. 041009-1-13. [8] Helal O, Chaouachi B, Gabsi S. Design and thermal performance of an ICS solar water heater based on three parabolic sections. Sol Energy 2011;85:2421–32. [9] Souliotis M, Quinlan P, Smyth M, Tripanagnostopoulos Y, Zacharopoulos A, Ramirez M, et al. Heat retaining integrated collector storage solar water heater with asymmetric CPC reflector. Sol Energy 2011;85:2474–87. [10] Zou B, Dong J, Yao Y, Jiang Y. An experimental investigation on a small-sized parabolic trough solar collector for water heating in cold areas. Appl Energy 2016;163:396–407. [11] Wang P, Guan H, Liu Z, Wang G, Zhao F, Xiao H. High temperature collecting performance of a new all-glass evacuated tubular solar air heater with Ushaped tube heat exchanger. Energy Convers Manage 2014;77:315–23. [12] Grass C, Schoelkopf W, Staudacher L, Hacker Z. Comparison of the optics of non-tracking and novel types of tracking solar thermal collectors for process heat applications up to 300 C. Sol Energy 2004;76:207–15. [13] Gu Xiaoguang, Taylor Robert A, Morrison Graham, Rosengarten Gary. Theoretical analysis of a novel, portable, CPC-based solar thermal collector for methanol reforming. Appl Energy 2014;119:467–75. [14] Umair M, Akisawa A, Ueda Y. Performance evaluation of a solar adsorption refrigeration system with a wing type compound parabolic concentrator. Energies 2014;7(3):1448–66. [15] Lu ZS, Wang RZ. Experimental performance investigation of small solar airconditioning systems with different kinds of collectors and chillers. Sol Energy 2014;110:7–14. [16] Nkwetta Dan Nchelatebe, Smyth Mervyn. Performance analysis and comparison of concentrated evacuated tube heat pipe solar collectors. Appl Energy 2012;98:22–32. [17] Nkwetta Dan Nchelatebe, Smyth Mervyn. Comparative field performance study of concentrator augmented array with two system configurations. Appl Energy 2012;92:800–8. [18] Ratismith W, Favre Y. Novel non-tracking solar collector with metal-to-water contact. In: Proceedings of the 15th IEEE international conference on environment and electrical engineering, June 10–13, 2015. Rome, Italy; 2015. [19] Fraunhofer Institute for Solar Energy Systems ISE. Test report: KTB Nr. 200624-en [20] Klein SA, Beckmann WA. Review of solar radiation utilizability. J Sol Energy Eng 1984;106:393–402.