Scholarly article on topic 'A Tubed, Volumetric Cavity Receiver Concept for High Efficiency, Low-cost Modular Molten Salt Solar Towers'

A Tubed, Volumetric Cavity Receiver Concept for High Efficiency, Low-cost Modular Molten Salt Solar Towers Academic research paper on "Materials engineering"

CC BY-NC-ND
0
0
Share paper
Academic journal
Energy Procedia
OECD Field of science
Keywords
{"Solar Thermal" / "Solar Tower" / Receiver / Cavity / Volumetric / "Molten Salt / "}

Abstract of research paper on Materials engineering, author of scientific article — P.J. Turner, C. Sansom

Abstract Small molten salt modular towers linked together to feed into a large power block, including storage, offer the potential to significantly reduce the cost of solar thermal energy. This is primarily through the significant increase in solar field and receiver efficiency that are achieved while still retaining the benefit of scale in the power block. Such towers would use cavity type receivers that are inherently more efficient than an external receiver. This paper examines the potential for a new concept for a cavity receiver, suitable for molten salt,which can increase efficiency and reduce metal hot-spot temperatures. By distributing the tubes within the volume of the cavity and arranging for the cooler inlet tubes to take the highest flux, the metal temperatures can be reduced close to the outlet salt temperature. The proposed design concept has the potential to solve a number of design issues that increase the cost of receiver systems. The paper provides a first-stage, simplified, theoretical analysis to show how receiverefficiency (from a radiative perspective) and hot-spot temperatureare affected by the number of heat transfer layers and the degree to which each layer blocks the radiation. The work shows promising results that needs to be taken forward in a number of areas.

Academic research paper on topic "A Tubed, Volumetric Cavity Receiver Concept for High Efficiency, Low-cost Modular Molten Salt Solar Towers"

CrossMark

Available online at www.sciencedirect.com

ScienceDirect

Energy Procedia 69 (2015) 553 - 562

International Conference on Concentrating Solar Power and Chemical Energy Systems,

SolarPACES 2014

A tubed, volumetric cavity receiver concept for high efficiency, low-cost modular molten salt solar towers

P.J. Turner* and C Sansom

Cranfield University, College Road, Cranfield, Bedfordshire, MK43 0AL, UK

Abstract

Small molten salt modular towers linked together to feed into a large power block, including storage, offer the potential to significantly reduce the cost of solar thermal energy. This is primarily through the significant increase in solar field and receiver efficiency that are achieved while still retaining the benefit of scale in the power block. Such towers would use cavity type receivers that are inherently more efficient than an external receiver. This paper examines the potential for a new concept for a cavity receiver, suitable for molten salt, which can increase efficiency and reduce metal hot-spot temperatures. By distributing the tubes within the volume of the cavity and arranging for the cooler inlet tubes to take the highest flux, the metal temperatures can be reduced close to the outlet salt temperature. The proposed design concept has the potential to solve a number of design issues that increase the cost of receiver systems. The paper provides a first-stage, simplified, theoretical analysis to show how receiver efficiency (from a radiative perspective) and hot-spot temperature are affected by the number of heat transfer layers and the degree to which each layer blocks the radiation. The work shows promising results that needs to be taken forward in a number of areas.

© 2015TheAuthors. Publishedby Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/4.0/).

Peer review by the scientific conference committee of SolarPACES 2014 under responsibility of PSE AG Keywords: Solar Thermal, Solar Tower, Receiver, Cavity, Volumetric, Molten Salt,

* Corresponding author. Tel.: +44 1234 752955; fax: +44 1234 2946 . E-mail address: peter.turner@cranfield.ac.uk

1876-6102 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer review by the scientific conference committee of SolarPACES 2014 under responsibility of PSE AG doi: 10.1016/j.egypro.2015.03.064

Nomenclature

Ay number of absorbed rays emitted from surface i and absorbed by surface j

B; Blockage factor for the ith surface layer

cp specific heat capacity of molten salt

DH hydraulic diameter of heat-transfer tube

G irradiation, the incident radiation energy on a surface

h heat transfer coefficient

k thermal conductivity of molten salt

Nray total number of rays emitted from each surface in the ray tracing calculationns

Nu Nusselt number

P probability

Pr Prandtl number, cp^/k

Q emitted (radiated) power from a surface i or of a ray

Re Reynolds number, pvDH/^

T temperature

v fluid velocity

a absorptivity

8 emissivity

^ dynamic viscosity

dynamic viscosity at the tube wall temperature

p density of molten salt

0 Stefan-Boltzmann constant, 5.67x10-8

1. Introduction

Cavity receivers, such as those used in PS10 and PS20 [1], have tube-bank, heat absorbing surfaces that make up some part of the inside of the cavity and an aperture through which the radiation is beamed. These tube-bank panels are also used in molten salt external receivers such as Solar Two [2] and is shown in Figure 1a.. Tubes containing a heat transfer fluid are mounted adjacent to one another to provide a continuous heat-transfer surface. A volumetric receiver, on the other hand, can be described as one where heat is transferred within a volume of a material or a material structure. The most obvious benefit is to increase the heat transfer surface area, reduce hot-spot temperatures and to arrange for the hottest parts of the receiver to be furthest from the external surface so that radiation losses are minimized. An obvious example of such a system would be one in which cool air is drawn into and through a porous, irradiated surface; the air gets hotter as it flows further into the receiver leaving the cooler parts of the receiver close to the surface to limit the amount of radiation lost to help increase receiver efficiency. A review of volumetric receivers is given in [3]

This paper introduces a combination of both approaches that offers the potential to avoid a number of technical challenges that can add to the cost of external receivers and, at the same time, improve the efficiency of a cavity receiver. It distributes a tubed heat transfer surface within the cavity with the result that the receiver is suitable for liquid heat transfer fluids, such as molten salt, and creates a design that may be cheaper to manufacture and have a higher efficiency. It is therefore described as a "tubed volumetric cavity receiver". Such receivers are envisaged for CSP plants that use multiple, small, molten salt solar towers, feeding a large central power block and thermal store.

Previous work [4], comparing the economics of this and other CSP concepts, has shown that a multiple, modular molten salt tower approach has, potentially, the lowest levelised cost of energy of any of the currently commercial CSP technologies. It combines the significant benefit of higher solar field and receiver efficiencies, but still retains the economic and thermodynamic efficiency benefit of a large-scale, central power block, with thermal storage enabling greater utilisation of the power block as well as providing more reliable power. Work on this concept started to develop a design of a small solar test tower of less than 10MWth. Although this was not progressed, budget quotations for conventional tubed panel receivers were sought and found higher than expected. Part of this

will have been due to the small-size and one-off nature of the potential contract and the need for the price to include all the design costs. However, these high costs stimulated the development of a "wish list" of potential attributes and features that would make the development of such receivers simpler, more cost effective and easier to integrate with the solar field and its control system. One of the issues that were encountered was the difficulty in contractually separating the receiver manufacture from the solar field and its control, due to the risk that problems in solar field control could lead to premature failure of the receiver.

This paper examines an alternative cavity receiver design by identifying a number of ideal design attributes. It shows how a tubular, volumetric, cavity receiver design may be developed to meet many, if not all of the design challenges identified and uses a simplified modelling methodology to provide a first-stage proof-of-principle. Further work, both simulation modelling and detailed design is required but it is the aim of this paper to provide a theoretical justification to show that the concept has potential to be used in an array of small molten salt, towers and may have the potential to be simpler, cheaper and more efficient than the normal single tube-bank approach currently used in cavity or large external receivers. A simplified form of the design concept is shown in Figure 1b.

(a) (b)

Fig. 1 (a) External molten salt receiver design from Solar Two, (b) Simplified section through tubed volumetric cavity receiver

2. Design challenges

Conventional molten salt solar receivers with flat panel tube-bank receivers, such as those used in Solar Two and Gemasolar [5] have a number of design challenges that adds to their complexity and cost. The design for the Solar Two receiver, reproduced from [6], shown in Figure 1a, shows the multiple parallel paths in each tube-bank panel and the series connections of each panel. A conventional molten-salt cavity receiver will have a similar arrangement of tube-banks and panels, albeit arranged within a cavity. Figure 1b shows a simplified cross-section through the tubed volumetric receiver concept where a number of layers of tubes are positioned within the volume of the cavity, each layer only partially blocking radiation. Complexity and cost factors in Figure 1a include:

• Multiple series and parallel tube paths are used to ensure equalisation of the outlet temperatures from each parallel path. This usually requires header manifolds at the bottom and top of each tubed panel and with a number of panels joined in series and linked with other parallel paths [6,7]. These parallel paths are arranged to try and ensure that changes in the direction of solar radiation from the solar field do not yield significantly different outlet fluid temperatures

• High numbers of welded connections between the thin-walled tubes and each header manifold and between manifolds all need to be carefully made and pressure tested, adding to the manufacturing costs.

• Thermal stresses between the tubes in contact to one another are caused by tubes being restrained. This can lead to potentially high levels of thermal stress if temperature distributions are uneven between tubes and between the front and back.

• Peak metal temperatures are dependent on the heat transfer coefficient between the metal surface and the bulk fluid, the temperature of the fluid and the solar flux. Although tubed panel receivers may be designed

with sections of panel nearest the hot outlet to be in regions of lower solar flux, it is often unavoidable to have metal surface temperatures significantly higher than the bulk fluid outlet temperature.

• Expensive nickel-based super alloys may be unavoidable for corrosion resistance to molten salt above 600deg.C but operating with metal temperatures significantly above this has often meant that Incoloy 800-H or other expensive alloys are proposed [8,9,10,11]

• Multiple high temperature drain valves are necessary in each section to enable the molten salt to be drained at night and for plant shut-down. Such valves and the piping to them are expensive.

• Integration of receiver to solar field control is necessary since receivers are generally sensitive to the distribution of solar flux across the heat transfer surface. The design of the receiver and of the solar field and its control are linked and must often be procured within a single contract to avoid commercial problems arising from risks to the receiver from poor solar field flux control. Commercial competitiveness would be greater if the receiver and solar field could be uncoupled as separate contracts but this would require a receiver design sufficiently robust to accept any reasonable solar flux distribution that a solar field could conceivably create.

3. Receiver comparison

3.1. External Single Surface Receiver

This paper represents the first stage in an assessment of the potential benefit of distributing heat transfer tube surfaces within the volume of a cavity receiver rather than the conventional approach of covering only some part of the internal surface of the cavity with a single heat transfer surface using a conventional tube-bank. As such, a simple one-dimensional comparison is used to compare the difference in radiation losses between a single flat surface and one that has a number of radiation porous layers. To make equivalent comparisons it is first necessary to model the heat transfer through the wall of a heat transfer tube using reasonable assumptions for key parameters. A small 10MWth receiver has been considered using tubes 20mm internal diameter and key parameters are shown in Table 1. The values for tube dimensions are similar to those of Solar 2 [2]; other values are chosen to be indicative of a small solar tower and are for illustrative purposes only. A more detailed analysis of heat transfer and stresses is given in [7]. Here the heat transfer coefficient, h, between the molten salt and the internal tube surface has been calculated using a standard correlation formula for the Nusselt number for turbulent flow in terms of Reynolds and Prandtl numbers [12], i.e.

Table 1 shows the assumed values used to determine a typical temperature differential across the tube wall and between tube internal surface and the average salt temperature.

A key result for the assumed design parameters chosen is the magnitude of the AT temperature differentials across the tube wall (63°C) and between the wall and the salt (81°C) for an assumed 1MW/m2 solar (absorbed flux). This indicates that for fluxes at this level and with an outlet salt temperature of 600°C, tube metal temperatures in contact with the salt will operate at around 681°C and surface metal temperatures of nearly 750°C will occur.

Calculation of the radiation emitted from the surface assumes a temperature distribution over a surface that follows the linear profile between inlet and outlet temperatures. Integrating the standard radiation equation, eoT4, between two absolute surface temperatures, T1 an T2 gives an effective emitted radiation per unit area, Q,xtt as

Nu = 0.027 Re08 Pr1/3(p/pJ h = Nu k / DH

(1) (2)

Qext = e a (T25 - T15 ) / ( 5 ( T2 - T1))

For a flux of 1MW/m2, an absorptivity and emissivity of 0.9, salt inlet/outlet temperatures of 250/600°C and a metal to salt AT of (typically) 150°C at lMW/m2, then the theoretical efficiency of a flat plate receiver is 87.32% (without considering convection) with a peak metal temperature of 735°C. Note that equation (3) represents surface radiation

emission only. For simplicity, additional absorbed and exchanged radiation from distant surfaces at ambient temperatures and the sky has been omitted from the analysis since, at the elevated operating temperatures considered, the effect is negligible.

Table 1. Example of 10MWth Receiver Thermal Analysis.

Parameter Value Unit Comment

Receiver Thermal Rating 10 MWth Potentially typical of a small modular tower system

Tower Height 70 m Used as part of the pumping loss calculation

Peak Solar flux absorbed 1000 kW/m2 Representing a typical maximum for a tubed receiver

Salt Inlet Temperature 250 °C Value above the Solar Salt solidification temperature

Salt Outlet Temperature 600 °C Temperature where corrosion affects increase

Tube internal diameter 20 mm Similar to Solar Two

Tube wall thickness 1.25 mm Similar to Solar Two

Parallel flow paths 7 This affects the path length and flow velocity. It changes the pumping power and wall to bulk salt AT

Mass flow 25.9 kg/s A function of the rating and inlet/outlet salt temps

Minimum Active Path length 63 m Assumes the peak solar flux is maintained over the whole surface area i.e. 10m2

Peak tube-wall AT 63 °C A function of the tube thickness and thermal cond.

Tube to average salt AT 84 °C

Total surface to av. salt AT 147 °C

Pumping power 0.8 % A max. value of 1% was chosen as a design limit

In comparing a single surface with a distributed surface, one further theoretical possibility is to extend the receiver area and to control the solar flux in an exponential decay so that the radiation flux and temperature differential reduces to restrict the metal surface temperature to some fixed, constant value. While this limits peak surface temperatures and radiation fluxes, it increases the area of the receiver and the overall radiation loss. It is also virtually impossible to control solar fluxes to this kind of precise profile. However, if metal surface temperatures are limited to 610°C, the theoretical radiation efficiency reduces to 85.68% and the receiver area becomes x1.65 larger than the theoretical minimum. This enlarged single surface is denoted 1-Lextended in section 5. Note that convective heat losses have been intentionally ignored in this comparison so that the effects on receiver efficiency can be more clearly investigated. However such losses are not zero [13]. Convective losses and heat transfers between tubes will occur and will need to be taken into account and examined in later design phases.

3.2. Tubed volumetric receiver

Although volumetric receiver designs are usually associated with porous, high temperature air receivers [3], the benefits of distributing and extracting heat flux throughout a volume can be equally applied to a receiver comprising of tubes containing molten salt. In this volumetric tubular cavity receiver, molten salt is pumped through long tubes formed to create multiple heat transfer layers, each tube being separate from one another to create layers semipermeable to radiation. Allowing the radiation to penetrate through the layers allows both front and back surfaces of the tube to receive reflected radiation and radiation emitted from all the other layers and also to tolerate a greater degree of thermal distortion without upsetting its function so potentially reducing thermally induced stresses.

The tube path is arranged so that the hottest parts are furthest from the cavity aperture and receive the lowest radiation flux densities to keep metal temperatures close to the molten salt outlet temperature. The higher radiation losses from the hottest parts are also furthest into the cavity and are partially intercepted and absorbed by the tube layers in front. This provides a receiver (radiative) efficiency advantage. Another efficiency advantage comes from reducing peak metal temperatures, since radiation power is proportional to absolute temperature to the forth power. These benefits are in addition to the fundamental benefit of utilising a cavity rather than an external receiver so that the radiating aperture area can be minimized and be significantly less than the internal heat transfer area, although this advantage is not accounted for in this simplified analysis.

Sensitivity to daily changes in solar flux distribution across the cavity aperture is minimized by using a minimum number of parallel paths, each designed to pick up a similar solar flux irrespective of the dominant angle of the most efficient heliostats in the solar field. It may also be possible to arrange the tube paths to maintain a steady incline

that will allow a single valve at the top of the receiver to be capable of draining the receiver prior to shut-down, but practical design details that may enable this have yet to be investigated. As a result many of the design challenges outlined in section 2 may be potentially addressed.

4. Analysis

As a first-stage, proof-of-concept, the analysis examines the comparison between a single heat transfer surface and a receiver comprising of a number of radiation porous heat transfer layers ending with a non-porous back surface. This is shown in Figure 2a showing the way the porous heat transfer surface is represented as a series of flat plates with gaps between and Figure 2b, modelled as a series of layers in which each ray/surface interaction causes any ray to be either absorbed, reflected or transmitted to the next layer.

(a) (b)

Fig. 2 Simple 1-D layer radiation model. a) Physical model, b) Ray interaction model

A simple Monte Carlo ray-tracing algorithm determines whether a ray is absorbed, reflected or transmitted. By summating the rays absorbed by the different surfaces it is possible to calculate the distribution of heat fluxes on the front and back surfaces of successive layers. (Note: Heat transfer layers comprise front and back surfaces, so layer 1 is represented by surfaces 2 and 3.) For the purpose of this analysis absorptivity and emissivity of the heat transfer surfaces has been assumed to be the same and equal to 0.9 for both solar and infra-red surface radiation. The last surface is non-porous and represents a ceramic lining.

Each ray interacting with a surface has a probability of being absorbed, Pab, reflected, Pr, or transmitted, Pt given

Pab =EB, (4)

Pr = (1-e)B, (5)

P, = 1-Bt (6)

Surface 1 represents the source of incoming solar radiation. The last surface is a non-porous insulated surface that does not extract heat but is in thermal radiation equilibrium. Rays emitted by each surface, including surface 1, are traced to determine which surface eventually absorbs them. Once all rays from all surfaces are traced, a matrix of absorbed rays is available, [A] with coefficients Aij. The power of a ray from surface 1 represents the dominant solar source and is given by

Qray} = Qi / Nray (7)

where the incoming receiver solar flux, Qi is assumed to be 1MW/m . For this analysis Nray was chosen to be 10 which was found sufficient to remove most of the random ripple in the results.

For an initial estimate of the net radiation flux and the power absorbed by each surface, only Qray1 is used. Molten salt temperatures are then calculated based on the energy absorbed by successive layers and the metal surface temperatures determined by adding on the temperature differential, AT, between the molten salt and the metal surface (150°C at 1MW/m2) scaled in proportion to the actual surface heat flux that is absorbed. This provides a first estimate of the starting and finishing metal surface temperature for each surface and this can then be used to determine the power of rays emitted from each surface, proportional to eoT4. Assuming a linear temperature distribution in each surface, a weighted average surface temperature, Tmw, is determined that is representative of the overall radiation from the surface.

Tmw = [ (Tend — Tstart ) / {5 (Tend — Tstart) } ] (8)

Where Tstart and Tend are the start and end surface metal temperatures of a particular surface. This weighed average temperature of the ith surface, Tmwi, can then be used to calculate the power of each ray emitted by each surface. For the ith surface, Qrayt is

Qrayi = Bi Si a Tmwi " / Nray (9)

The total net power absorbed by each surface can then be determined as the summation of the rays from all surfaces. For surface j, the power it absorbs, less the power of the rays it emits, is given by

Qsurfacej = £= to n Qrayi A,d - Qrayj Nray (10)

This calculation modifies the net power absorbed by each layer and a simple iterative scheme is used (5 to 10 iterations) to converge to a stable condition where salt and metal surface temperatures are self-consistent with the power of each surface ray calculated from the resulting metal temperatures.

5. Results

The obvious variables to examine are the effect of blockage factor and number of heat transfer layers on receiver radiation losses and efficiency. Figure 3a shows the way in which receiver radiation losses change as a function of the blockage factor. As the blockage factor tends towards 1, the losses, irrespective of the number of heat-transfer layers modelled, tend towards that of a single layer.

(a) (b)

Fig. 3. Receiver efficiency per m2. (a) Loss per m2, (b) Theoretical efficiency (radiative effects only)

If the blockage factor tends to zero, losses increase rapidly as there is less heat transfer surface to intercept the incoming radiation or the radiation emitted and reflected from the rear surface. The minimum loss occurs at an intermediate blockage factor that reduces as more layers are added. For four layers the optimum blockage factor is about 0.5 and for 5 layers 0.42. This may indicate that the optimum is tending towards one where the blockage factor multiplied by the number of layers is tending to a value of about 2 which is also a measure of the additional tube surface used above that of a single, fully blocked panel. Reductions in radiation loss occur as the number of layers increases but the improvements reduce and 4 or 5 layers appear to obtain most of the possible benefit.

Figure 3b shows the receiver efficiency (based on reflected and radiated energy) which peaks with the 5-layer model with a blockage factor of 0.42 at 90.9%. (Note that the results are very sensitive to the value of absorptivity/emissivity used. Increasing the absorptivity/emissivity value to 0.95 leads to the efficiency increasing to 94% for the same blockage factor and the improvement over a single layer becomes less pronounced. As the absorptivity/emissivity tends to 1, the benefit of the volumetric design reduces since there is significantly less solar radiation being reflected from each surface; distributing the layers only helps to intercept some of the radiation emitted from the layers behind.

Peak Metal Temperature

o oj tta os os i

Blwl^jjo F actor

Fig. 4. Metal surface hot-spot temperature

Peak metal hot-spot temperatures in each layer occur at the end of each layer and are a function of the salt temperature and the temperature differential across the tube wall, AT, that is proportional to the absorbed heat flux. Figure 4 shows how the overall hot-spot temperature of all layers is affected by the blockage factor and the number of heat transfer layers. The peak metal temperature does not necessarily occur in the last (salt-outlet) layer and small ripples in the results are caused by this overall peak moving from one layer to another in successive studies.

What is particularly encouraging is the way in which the peak metal temperature can be brought down to values close to the exit salt temperature of 600°C which will reduce corrosion problems that are temperature sensitive [10,11]. Also added on Figure 4 is the theoretical hot-spot temperature for a single layer (600°C + AT) and for a single layer with a theoretically optimally designed extended surface that assumes the incoming flux can be controlled and reduced to limit the peak metal surface temperature. It has been shown here for a value of 610°C corresponding to an area increase of 1.65 over the theoretical minimum and is denoted as 1-Lextended in Figures 3 and 4.

6. Cavity receiver design

The results given in Section 5 are for a succession of layers with molten salt flowing from the cool inlet in layer 1 to the hot outlet from the last layer furthest from the solar source. Given that a typical cavity receiver will have an

aperture angled down towards the centre of the solar field, it may be possible to arrange the flow to run from a bottom inlet to a top outlet; jumping between layers to create a continuous upward gradient so that the receiver can drain easily although the practicality of this needs to be explored further. The analysis can be used to give an indication of the likely metal hot-spot temperatures for such a situation and whether such a design is still able to keep peak metal temperatures close to the salt outlet temperature. Figure 5a shows the distribution of temperatures along the molten salt flow path and Figure 5b the distribution of peak metal temperatures for each tube. The molten salt temperature steadily increases with the gradient dependent on the net power absorbed from the different layers. Peak metal temperatures fluctuate; reductions in metal temperature are due to the lower radiation fluxes as the flow moves to layers further from the solar radiation source while the sudden increases in temperature are due to the flow moving to the front layer where radiation fluxes are highest. It is encouraging to see that such a flow path is still able to control metal temperatures effectively.

(a) (b)

Fig. 5. Typical distribution of peak metal temperatures within the receiver.

7. Further work

While this proof-of- principle is encouraging, a significant amount of further work is necessary to determine if a practical design is possible. The following work is necessary:-

• Development of a 2- and 3-D ray tracing model is necessary to explore the potential benefit of modelling the tubes and the increased likelihood that they will block radiation reflected and radiated from internal layers. It should also be used to explore whether there is any potential advantage in optimising the profiling of the blockage factors for different layers.

• The way in which the tubes can be mechanically supported needs to be examined. Any support will be exposed to radiation and although any supports may be optimally oriented, they are unlikely to benefit from any significant heat conduction to the tubes they touch and will therefore take up a temperature in radiation equilibrium. Specific engineering options need to be examined, including the possibility of self-supporting tubes, and the thermal and material implications of this need to be examined in more detail.

• The possibility of a simple self-draining receiver depends on maintaining a steady gradient in the flow path. Detailed receiver design and flow draining calculations (and perhaps experiments) will be necessary to understand what gradient can be achieved, depending on the size and rating of the receiver, and whether self-draining is indeed a practical possibility.

• The effect of convection heat losses needs to be examined and ways to minimise this need to be investigated.

8. Conclusions

This paper compares the conventional single-layer tube-bank approach to this volumetric cavity receiver concept and shows that surface radiation losses are reduced by two separate mechanisms; the volumetric approach and by reducing peak radiating metal temperatures. As well as the potential to create a simple and cheap design to construct, the reduction in tube to header junctions should also reduce parasitic pumping losses while the flexible support of individual tubes may reduce thermally induced stress loading.

One obvious simplification is the use of flat "tubes" in the simple 1-D model. Modelling the actual tubular surfaces will affect the results since radiation, particularly from the rear, will be at more random angles than the incoming solar flux. Tubes will create a greater effective blockage factor stopping radiation from escaping out of the front of the receiver and this should increase the benefits observed in this simplified analysis. The analysis also does not take any advantage of a cavity effect in limiting radiation losses so actual receiver efficiency should increase once this is included.

This concept is in an early stage of development, initial results are encouraging and indicate that it is possible to design a tubed, volumetric receiver in which peak metal temperatures can be reduced close to the molten salt outlet temperature. This simplified proof-of-principle study shows that by distributing a tubed heat transfer surface as a series of radiation porous layers, the receiver efficiency of the receiver can be increased and temperatures can be more easily controlled. By reducing the radiation flux levels on successive layers, peak metal temperatures can be reduced close to the outlet salt temperature and that this may reduce material costs. It shows that a tubed volumetric cavity receiver may have a number of potential advantages that should make it cheaper to manufacture, be more efficient and more easily managed and controlled. Such a receiver is envisaged to be applicable to a central receiver tower system comprising many small modular towers. A significant plan of future work is envisaged. This includes detailed design work, both thermal and mechanical, and extension of radiation and thermal modelling in 2-D and 3D to optimally design and assess the efficiencies and internal temperatures.

References

[1] Osuna R et al. PS10, Construction of a 11MW solar thermal tower plant in Seville, Spain. SolarPACES Conference, Seville, Spain June 2006.

[2] Pacheco JE. Final Test and evaluation results from the Solar Two project. Sandia Report SAND2002-120, January 2002

[3] Avila-Marin A.. Volumetric receivers in solar thermal power plants with central receiver system technology: a review. Solar Energy 85 (2011) 891-910

[4] Allen N, Guest P, Riedl K and Turner P. Development of a levelised cost model for different CSP technologies. Solar PACES conference, Las Vegas, USA (2013)

[5] Burgaleta, JI, Santiago A, and Diego R. Gemasolar, the first tower thermosolar commercial plant with molten salt storage. SolarPACES, Granada, Spain (2011): 20-23.

[6] Reilly HE and Kolb GJ. An evaluation of molten-salt power towers including results of the Solar Two project. Sandia Report SAND2001-3674, November 2001

[7] Flores O, Marugan-Cruz C, Santana D and Garcia-Villalba M. Thermal stresses analysis of a circular tube in a central receiver. SolarPACES 2013, Energy Procedia 49 ( 2014 ) 354 - 362

[8] Kolb GJ. An evaluation of possible next generation high temperature molten salt power towers. Sandia Report SAND2011-9320, December 2011

[9] Reynolds, DJ., Jance, MJ, Behnia, M and Morrison GL. An experimental and computational study of heat loss characteristics of a trapezoidal cavity absorber. Solar Energy, Volume 76, Issues 1-3, January-March 2004 p229-234.

[10] Goods SH, Bradshaw RW, Prairie, MR and Chavez JM. Corrosion of Stainless and carbon steels in molten mixtures of industrial nitrates. Sandia report SAND94-8211, March 1994.

[11] Diaz BG and Gray J et al. Corrosion in very high-temperature molten salt for next generation CSP systems. Savannah River National Laboratory

[12] Sieder EN and Tate GE. Heat transfer and pressure drop of liquids in tubes. Ind. Eng. Chem., Vol.28, 1429 (1936)

[13] Fang JB, Wei JJ, Dong XW and Wang YS. Thermal performance simulation of a solar cavity receiver under windy conditions. Solar Energy 85 (2011) p 126-138.