900-724 AN INITIAL STUDY OF SOLAR POWER PLANTS USING A DISTRIBUTED NETWORK OF POINT FOCUSING COLLECTORS EM 342-308 July 1975 JET PROPULSION LABORATORY CALIFORNIA INSTITUTE OF TECHNOLOGY PASADENA, C A L I F O R M A
900-724 AN INITIAL STUDY OF SOLAR POWER PLANTS USING A DISTRIBUTED NETWORK OF POINT FOCUSING COLLECTORS EM 342-308 July 1975 by Richard S. Caputo Thermal Energy Conversion Group Vincent C. Truscello, Supervisor Thermal Energy Conversion Group JET PROPULSION LABORATORY CALIFORNIA INSTITUTE OF TECHNOLOGY PASADENA, CALIFORNIA
900-724 DISTRIBUTION Internal M. E. Alper (3) W. Bachman H. Bank K. Bartos J. Becker R. Bourke R. Caputo (3) 0. Citron V. C. Clarke, Jr. E. Costogue E. S. Davis K. Dawson R. Dickinson J. Doane L. Dumas (3) N. Eddy T. English T. Fujita J. Goldsmith J. Kelley R. Manvi R. Miles W. Owen R. Phen J. Plamondon N. Riise D. Ross L. D. Runkle P. Scher D. Schneiderman H. Schurmeier M. K. Selcuk D. Shirley R. Stephenson M. Swerdling V. Truscello R. Turner A. Uchiyama W. Victor R. Wallace P. Wiener T. Williams
DISTRIBUTION External R. Andrews (NASA Johnson) H. Bloomfield (NASA Lewis) (3 copies) A. Bond (NASA Johnson) S. Copps (NASA HQ) J. Craig (NASA Johnson) R. Currie (NASA Marshall) H. P. Davis (NASA Johnson) W. R. Downs (NASA Johnson) A. Edwards, Jr. (Raytheon) S. W. Fordyce (NASA HQ) D. Ginter (NASA HQ) P. Glaser (A. D. Little) E. Greenblat (Econ) M. Gutstein (ERDA HQ) (2 copies) J. Holdren (U.C. Berkeley) G. Kaplan (ERDA) (5 copies) R. LaRock (NASA HQ) L. Lees (Caltech EQL) W. B. Lenoir (NASA Johnson) J. Leonard (Sandia Albuquerque) (5 copies) P. A. Lowe (ERDA) S. V. Manson (NASA HQ) (5 copies) 0. Maynard (Raytheon) C. A. Nathan (Grumman) B. D. Newsom (NASA Ames) J. Nicol (A. D. Little) R. Noll (Caltech EQL) R. Ragsdale (NASA Lewis) L. Shure (NASA Lewis) A. Skinrood (Sandia Livermore) (5 copies) L. W. Slifer, Jr. (NASA Goddard) G. Stevens (NASA Lewis) C. J. Swet (ERDA HQ) (2 copies) G. F. von Tiesenhausen (NASA Marshall) J. Ward (NASA Lewis) W. Whitacre (NASA Marshall) (5 copies)
ACKNOWLEDGMENTS The author wishes to acknowledge those who contributed to this study. Dr. R. Turner provided the energy transport material while Tad Macie developed the electric power collection data. H. N. Riise provided the parabolic dish performance estimates. The author expresses sincere thanks to Mrs. P. Panda for editing, typing and general preparation of this document. Without their support this report would not have been possible.
FOREWORD The NASA Office of Energy Programs is presently conducting a study of the potential utility of large orbital central power stations as a possible means to help meet our country's demands for electricity. As part of this study, JPL has been directed to perform a survey of potential terrestrial energy conversion systems for comparison with orbital central power stations. The candidate terrestrial options being reviewed include conventional power plants and both solar thermal and photovoltaic conversion. Among the solar thermal options, the following are being studied: low temperature flat plate collectors, medium temperature single-axis tracking linear concentrators, and high temperature two-axis tracking point concentrators, namely the dispersed parabolic dish and the central receiver concepts. This report presents a preliminary survey of solar thermal power plants using a distributed network of point focusing dish collectors. Such important characteristics as size, performance, operating temperatures, cost and state-of-the-art of the technology are given. Maximum use of existing literature data was made. The work is being performed under the technical direction and guidance of Mr. Simon Manson of the Energy Technology Applications Division.
Distribution Lists.............................................................................................................. ii Acknowledgements................................................................................................................... iv Foreword................................................................................................................................ v Summary ................................................................................................................................ 1 I. Introduction....................................................................................................................... 5 II. System Description......................................................................................................... 5 2.1 Distributed Generation ................................................................................... 8 2.1.1 Collector Subsystem............................................................................................. 11 2.1.2 Brayton Engine Subsystem ........................................................................... 14 2.1.3 Electric Power Collection................................................................................ 17 2.2 Central Generation ............................................................................................ 19 2.2.1 Collectors...............................................................................................................20 2.2.2 Energy Transport ............................................................................................ 21 2.2.3 Rankine Engine...................................................................................................... 28 III. Subsystem Performance Characteristics .............................................................. 29 3.1 Parabolic Dish Collector.....................................................................................29 3.1.1 Performance...............................................................................................................29 3.1.2 Collector Cost......................................................................................................33 3.2 Energy Transport ................................................................................................ 41 3.2.1 Fluid Transport......................................................................................................41 3.2.1.1 Analytical Approach.........................................................................................41 3.2.1.2 Pipe Network and Cost Data.................. .....................................................45 3.2.1.3 Economic Performance ............................................................................... 51 3.2.1.4 Initial Energy Investment............................................................................56
3.2.2 Electric Power Transport ............................................................................ 60 3.3 Heat Engine....................................................................................................................64 3.3.1 Brayton Heat Engine.............................................................................................. 64 3.3.2 Rankine Heat Engine.............................................................................................. 68 3.3.2.1 Rankine Plant Performance.............................................................................68 3.3.2.2 Rankine Plant Costs.......................................................................................... 72 IV. System Performance............................................................................................................... 75 4.1 Brayton Distributed Generation System............................................................. 75 4.1.1 Sizing for Optimum Performance and Cost..................................................... 75 4.1.2 Preliminary Impact Characteristics ...................................................... 82 4.2 Central Generation ............................................................................................. 85 4.2.1 System Performance and Economics .......................................................... 86 4.2.1.1 Water Transport..................................................................................................88 4.2.1.2 Steam Transport.................................................................................................. 91 4.2.1.3 NaK and Helium Transport ....................................................................... 95 4.2.2 Impact Characteristics ............................................................................... 97 V. References................................................................................................................................ 100 VI. Appendix.................................................................................................................................... 102
SUMMARY Parabolic dish collectors can be coupled to heat engines in a number of ways. Both distributed generation using a Brayton engine at each dish and central generation with a single large Rankine plant is possible. The central generation plants using steam or chemical energy transport have about a 20% cost advantage over the distributed generation concept with electric energy transport. Table 1 shows the capital cost breakdown for major subsystems for each of these three approaches and represent direct overnight construction costs. The total costs range from 1140 to 1435$/kWe. This difference may vanish when construction costs are considered such as interest during construction, escalation, spares, contingencies, and startup. The reason is the partial generation of power by the dish-Brayton combination during construction. Later economic analysis will evaluate this factor. The central generation plants using NaK and helium for energy transport proved to be too expensive, costing over 400$/kWe more than the steam or chemical approaches. The cost of pressurized water as the transport fluid was only 100$/kWe more than the cost of steam. However, there were additional penalties of a longer morning startup time (0.4 hour compared to 0.15 hour for steam and nil for chemical), but easier control with variable solar input. Table 1 also indicates the energy cost based on direct costs (no 0 & M), land area, and several other parameters. All three plants have a nearly neutral effect (using a first order of magnitude calculation) on the area heat balance compared to having no plant at all. The central generation plants use wet cooling towers and require 1100 acre ft/yr of cooling water. This may be a problem in the arid Southwest region. Using a dry cooling tower in the central generation plants, the direct capital costs increase by 150$/kWe, and the overall efficiency is reduced to ~0.9 of the efficiency obtained with a wet cooling tower.
Table 1. Summary of System Characteristics
An additional consideration for the distributed generation approach is that it is modular at the 20 kWe size level. It can fit irregular sized land in remote or in urban areas. Also, it has a low visual profile (~ 36 ft diameter dish) with no large buildings or cooling towers. It is air cooled and can be used for total energy systems to supply electric power and heat to community, commercial and industrial sites. It has a quick startup (several minutes). The capacity at a particular site can be built up gradually. The dish collector requires development as does the closed cycle Brayton engine. The electric power collection equipment is available. The central generation approach (steam or chemical) should be used in a size greater than 100 MWe, and can use a single, large,relatively circular plot of land. It has large buildings and cooling towers in the center of the solar field. The energy transport subsystem and Rankine plant technologies are available. The chemical thermal dissociation and recombination (catalyst) components must be developed for this application. Both systems may start up in less than 1 hour if the Rankine plant is preheated in some manner. Several approaches are possible for the thermal link between the dish cavity receiver and local heat engine (distributed generation), or to the fixed piping network for the central generation approach. Candidate approaches are a flexible fluid line, or a fixed cavity receiver on a probe through a slot in the dish surface, or a Cassegrain reflection of energy to a small heat engine hung at the counterweight position with a flexible electric connection to the ground, or a double mirror Cassegrain approach where the cavity is fixed on the ground. A detailed design-costing study should be carried out to choose the best approach in this area. 2 Major uncertainties exist in the area of dish collector cost. Nearly $13/ft was estimated (no cavity), but estimates varied widely. Also the small Brayton
engine cost is somewhat uncertain with the estimated cost being about 2.5 times the Rankine plant cost. The chemical energy transport has promise because of projected low cost, low losses and potential for storage of CO + gases. The potential of using a combination of distributed and central generation with a Brayton-Rankine topping cycle should be investigated. The waste heat from a Brayton engine on each dish can be carried to a centra! Rankine engine via water transport at 315°C. Storage was not addressed in this review of sun following plant operation, but will be considered in the next phase of this program.
SECTION I INTRODUCTION The paraboloid of revolution (parabolic dish) is one type of solar energy concentrator that uses two-axis tracking. The other major two-axis tracking system uses a field of heliostats focusing on a central receiver (Ref. 14). The parabolic dish solar collector can achieve relatively high concentration ratios (500 to 2000), and as a result rather high potential temperatures; values in the range of 500°C to 1400°C are achievable. The thermal energy collection efficiency is also quite high; for example, the efficiency is estimated to be 70% at 800°C for a reasonable cavity receiver design. These characteristics indicate both the strength and weakness of the parabolic dish solar collector. The high concentration ratio and high heat fluxes at the focal zone make it difficult to design a receiver with low temperature differences between tube wall and fluid. Heat pipes, spiral fluid tubes in insulated cavities, and porous ceramics are examples of design approaches for a suitable receiver. A three-dimensional surface (dish) with sufficient surface accuracy and structural rigidity is difficult to achieve at low cost. Also the structural anchoring needed to withstand winds due to the large area cross section is a major design concern along with accurate two-axis tracking. The high potential temperatures and efficiencies are extremely attractive for driving a heat engine, but the high temperature makes it difficult to collect heat from a large field of dish collectors and to bring the heat to a central power plant. Two approaches have been considered in the application of the parabolic dish for electric power production. The first is central generation where heat is collected from a field of dishes and transported to a central site either through thermal transport or chemical transport for eventual energy conversion. Large central steam Rankine engines are efficient even at moderate temperatures (500°C) and
are relatively inexpensive (~$150/kWe). A substantial heat or chemical transport subsystem is required for central electric generation. The extra costs and losses of this subsystem must be traded off against the lower costs of the central Rankine plant. The second approach is distributed generation where the heat engine is located near the dish. Electric power is generated at each dish and collected at a central site for external transmission. This second approach can use much higher temperatures since there is a very short (~10's of feet) heat transport line to the heat engine. Energy conversion devices which are suitable at high temperatures (> 500°C) are gas Brayton engines, Stirling engines, advanced liquid metal Rankine systems, and thermionic conversion devices. In this study only the closed cycle Brayton engine is considered. The approach evaluated for distributed generation in this study considers only a small amount of power produced at each dish (—20 kW). The small heat engines tend to be expensive and less efficient than large central heat engines. This disadvantage may be partially alleviated by grouping a small number of dish collectors to provide energy to a larger heat engine. The modular approach of the distributed generation system has the advantage of partial power production early in construction. This tends to minimize the costs of tying up capital for a long construction period with interest, escalation and startup difficulties. The other advantage of a modular distributed generation is that small, irregular parcels of land can be used effectively. This characteristic is most useful in near urban or urban areas where large blocks of land are difficult to assemble, or in newly developing urban areas where land can be put aside for power generation as the area develops. The visual impact is much less for this low profile system compared to centra! tower designs. Finally, the use of a closed cycle Brayton engine has the advantage of high (95-150°C) heat rejection temperatures
where air cooling can be used to eliminate dependence on water cooling.* This is an important characteristic in many arid regions of the Southwest where both land and solar insolation are available in greater than average quantities. The alternate concept in which a relatively small number of dish collectors (#50) can be used to drive a larger dosed cycle Brayton engine which would have lower cost and higher efficiencies may also be attractive, but is not evaluated in this report. The relatively short distances involved (~200 ft) could accommodate even an 800°C heat transport loop. A topping cycle is possible with a Brayton engine at each dish; water cooling could carry the waste heat at approximately 300°C to a central Rankine plant. Two basic concepts have been evaluated in this study in an attempt to bracket the cost and performance characteristics of solar power plants using parabolic dish collectors. They are distributed generation with a small Brayton engine at each collector, and central generation in which heat is collected from the field and delivered to a central Rankine power plant. The suitability of each approach only partially depends upon direct "overnight" construction costs. Questions of plant site (remote versus near urban), water availability (greater ease of using air cooling), plant reliability (many generators versus a single generator), total energy systems, and partial generation during construction all have a bearing on the attractiveness of these types of solar power plants. Only sun following plants are considered (i.e., no storage) in this report. These plants will be integrated with various storage systems in further studies. * A dry cooling tower can be used with the Rankine steam plant, but overall system performance will decrease about 10%, and the cost of the heat rejection subsystem will approximately double (see Section 4.2.2).
SECTION II SYSTEM DESCRIPTION Two basic solar power plant systems are described which use parabolic dish solar collectors. The first is distributed generation of electric power in which small closed cycle Brayton engines are coupled to each collector. The second is central generation of electric power in which heat is collected from the field of dish collectors and is used to drive a central heat engine (steam Rankine). The only operational mode considered in the present study is the sun following mode. That is, the solar plant is used to generate electric power during the daytime hours only and no storage is provided. Storage subsystems are presently being characterized, and they would be combined with the sun following solar plant for peaking, intermediate and baseload operation at a future time. 2.1 DISTRIBUTED GENERATION Small closed cycle Brayton engines are placed at each parabolic dish to minimize the distance required for high temperature heat transport. Preliminary studies indicate that the Brayton engine should not be placed at the focal point of the collector because although these engines are small and relatively light, the extra costs in the collector structure and counterweight would offset gains. A short gas duct can connect the heat absorption cavity to the engine. A critical area is the minimization of the pressure drop and heat losses in this short duct. The Brayton heat engine could also be located at the base of the dish with a flexible line, or actually mounted on the dish as the counterweight. Another approach is to locate a Cassegrain second surface mirror at the parabolic dish focal point. The optical energy would reflect through the center of the dish to a cavity receiver located behind the dish surface. No detailed design study has been carried out at this time to compare these various approaches. For this preliminary evaluation,
an insulated cavity is assumed to be at the focal point with helium gas passing through spirally wound tubes inside the cavity. A short flexible gas line connects the cavity to the closed cycle Brayton engine located at the counterweight position. A schematic of one module is shown in Figure 1. Electric energy would be collected from each Brayton engine and brought to the transmission terminal for power conditioning. The Brayton cycle operating at the current state-of-the-art temperature of 815°C is a relatively high efficiency (30 to 36%) device. An air heat exchanger, which is equivalent to a dry tower, it used to reject heat to the environment. Normally, ambient air will be heated from 20°C to about 120°C to cool this engine. The amount of power generated by economically attractive designs of a dish coupled to a Brayton engine is small by central plant standards. A dish which is 36 ft (11 m) in diameter can collect, over the day, about 52 kW of thermal power out of 74 kW of incident direct solar power. This figure is based on an annual average during the day for a good location in the Southwest. At a peak insolation 2 of 1 kW/m , the power collected would be 65 kWt. The use of a small closed cycle Brayton engine at each collector to avoid long-distance transport of high temperature heat limits the heat engine conversion efficiency to about 33%. Thus, an average of about 17 kWe is generated at each dish collector during the day. This combination of equipment has certain advantages compared to most central generation systems. These are the very short startup time and very low initial energy investment. Engine startup is accomplished in several minutes, rather than 10's of minutes or even an hour, after achieving full temperature in the working fluid. A solar plant which can generate electric power in the 10's of kWe at each dish can be used in a variety of ways. By using a large number of these modules (10,000's), large central station plants can be formed in remote areas to take advantage of good solar insolation. Substation power can be generated near an
Figure 1. Parabolic Dish Solar Power Plant - Distributed Generation
urban area by using a smaller number of modules, thus eliminating transmission costs and inefficiencies. Community size power can be generated by integrating this type of plant info newly expanding suburban areas with the possibility of using waste heat for commercial and residential space conditioning. Finally, it is possible to integrate these devices into low rise industrial sites for electric and heat generation. For the purposes of this evaluation, only the central power station application will be considered. 2.1.1 Collector Subsystem In general, a parabolic dish collector with two-axis tracking intercepts the most solar energy per m2 of aperture area compared to all other types of solar collectors. Figure 2 illustrates the relative differences for various types of collectors in their ability to intercept sunlight in June (based on Ref. 1). The parabolic dish with two-axis tracking always makes the maximum use of its collector surface since the dish aperture is perpendicular to the sun over the entire day. Varying atmospheric absorption of the solar input accounts for the variation shown over the day. Each of the other collectors has an additional geometric effect either on a daily or seasonal basis which further reduces the effectiveness of the collector. An artist's sketch of the parabolic dish collector is shown in Figure 3(a). Short insulated ducts carry the heated fluid from the cavity receiver along the strut to the closed cycle Brayton engine. The engine is on the ground adjacent to the base of the collector pedestal (not shown). Figure 3(b) shows the cavity receiver at the focal point where the inert gas is heated to temperatures in excess of 800°C. Although several types of reflection surfaces could be used for the mirror, backsilvered glass has been assumed. This type of reflection surface has both a high reflectivity and long life. The substrate which provides the structural support for the glass can either be steel with a thermal coefficient similar to glass,
Figure 2. Energy Intercepted Compared to Noon Peak - June (from Reference 1)
Figure 3. Parabolic Dish & Cavity Collector System
or foamed glass which is a lightweight structural material made of glass. This latter concept is presently being developed at JPL. The optimum size collector is found by trading off the dish cost versus size with that of the Brayton engine cost versus size. 2.1.2 Brayton Engine Subsystem The Brayton heat engine is a good match with the parabolic dish collector for distributed power generation. The high temperature potential of the dish collector is suited to a gas turbine power plant. The closed-cycle gas turbine is a well established technology and related technologies are currently available for unit sizes from 5 kWe up to 250 MWe. A schematic of a closed cycle Brayton engine is shown in Figure 4. A recuperator is used to increase cycle efficiency. A single shaft BRU (Brayton rotating unit) has been selected in which the generator is enclosed in a hermetically sealed housing. An external generator could be used for near term application, or if gearing were necessary for coupling to the generator. The hermetically sealed generator with gas bearings has the potential advantage of long life without any maintenance (5 to 15 years). . Small closed cycle Brayton engines were developed for the aerospace program (Ref. 2) in the 5 to 10 kWe range. An even smaller engine (500 W to 2200 W) is currently being developed at AiResearch for NASA (Ref. 3). Devices similar to this are now used commercially for auxiliary power on DC-lOs, although they are open cycle, use jet engine combustion products and are unrecuperated. Two current commercial programs are directly relatable to this application. A 19 kWe generator for a refrigerated railroad car and a 7.5 kWe gas fired refrigeration power unit are currently being developed (Ref. 4). Both units operate at 815°C to avoid hot spot problems and ensure long life (>10 years). The expected efficiency is 28% based on its having low capital cost. The thermal
Figure 4. Closed Cycle Brayton Engine
efficiency is dependent upon the degree of complexity provided in the system such as compressor intercooling, regeneration and reheat. Large units (> 400 kWe) with high pressure levels, low pressure drops, maximum complexity and axial compressors could reach an efficiency of nearly 50% at 815°C (Ref. 5). For machines in the 10 to 50 kW size, radial compressors are used; the overall efficiency is about 36% rather than 50% without reheat and intercooling. These machines have large heat exchangers with a low compressor inlet temperature. The mass produced commercial cost would be close to $400/kWe for these small machines with oversized heat exchangers (Ref. 5). Attempts to reduce cost result in an estimate of nearly $300/kW, but with a loss of efficiency. The efficiency is reduced to the 28% to 31% range with costs at $300/kW. Engines of this size would have to be developed, and an estimate of the R&D cost is approximately $25M (Ref. 5). Larger closed cycle machines (~30 MW) may be as low as $130/kW with about 46% efficiency at 800°C. Such a machine would require a development program of approximately $100M (Ref. 5). Improved performance is possible mainly by increasing turbine inlet temperatures. The current designs are limited to an 815°C turbine mixed average inlet to keep the tube hot spot temperature due to combustion products in the combustion to less than 900°C. This temperature constraint is based on using Inconel 713 as the cavity heat exchanger and turbine material. Stainless steel is used for the recuperator. Turbine inlet temperatures of 1230°C may be possible with ceramic heat exchangers, and refractory alloy (moly) turbines with advanced cooling. Ceramic materials such as SiC or Si^N^ would be used for the recuperator. Efficiency could reach 46% for small machines (< 50 kW) with costs similar to those of the 815°C machines. The hot spot would be limited to 1370°C for this advanced design (Ref. 5). In summary, relatively near-term development programs could produce 30% to 36% efficient Brayton machines costing $300/kW to $400/kW. Advanced high
temperature versions might reach 46% even in these small sizes (10 kW to 50 kW) at similar costs. Nearer term (1 to 3 year) Brayton engines would have reduced performance. Performance-cost data are discussed in Section 3.3. 2.1.3 Electric Power Collection The distributed generation power plant produces electric power throughout the field of solar collectors. The collection of this electric power to a central site for external distribution is a problem of detailed design tradeoffs using commercially available equipment without need of advanced R&D. The major assumptions used in the study of this subsystem are 1) no energy storage (sun following operation), 2) automated controls, 3) fail-safe characteristics for the system, 4) extremely high reliability and 5) local repairs possible while the rest of the system is operating. The major questions to be resolved concern the type of generator such as ac synchronous, ac induction or de shunt wound machine; the use of ac or de in the collection network; and the techniques used for power conditioning and control. The use of a de shunt wound generator is ideally suited to parallel operation. Startup is relatively easy, and constant voltage operation is possible with variable speed on the shaft. The output voltage depends upon the shaft speed (n) and the level of excitation (0), i.e., E = kn0. Thus, variations in speed due to varying solar input can be compensated by controlling the excitation to maintain constant voltage output. Controls can easily be automated. Synchronization is not a problem at all, but the generator is about 150% as costly as equivalent ac synchronous machines. Maintenance may be high unless brushless versions are used at even higher initial cost. Inversion from de to ac is required even if de external transmission is used. The generator currently used almost universally is the synchronous ac machine. It is more efficient than other types of generation. Startup is difficult since
the generator must be brought up to synchronous speed by the prime mover and then the circuit breaker switches must be closed when the generator is in phase. This generator may fall out of synchronization easily if perturbed, and hunting between generators working in parallel may be a problem. The reactive power generated can be adjusted by the de field excitation. An ac induction generator (squirrel cage type) is a third choice although it is not commonly used for generation. It is inexpensive (about one third that of a synchronous generator) and easy to start. Line power can run the machine as a motor to near synchronous speeds with a slight frequency lag. This will spin the Brayton engine, and as power is provided to the shaft, it will generate power as the speed of the machine is increased to slightly greater than the synchronous speed. It is easy to run and inexpensive to maintain. It is about 2% less efficient (on an absolute basis) than the synchronous generator, and reactive power is obtained by additional cross-the-1ine capacitors. Since the de generator is both more expensive itself and requires the additional expense and inefficiency of dc-ac inversion, an ac generator is considered more suitable for this application. Even though it is 2% less efficient, the ac induction generator is cheaper and easier to start, operate at synchronous speed in great numbers and maintain than the synchronous ac generator. (A 100 MWe power plant may require approximately 5000 generators.) It has, therefore, been selected for the present study. Underground aluminum cables are considered most suitable for power collection. Rapid burial is available via mechanized methods. Aluminum cables are generally larger in diameter than copper, and heat dissipation to the ground is easier. Aluminum is also lower in cost. The cost for the buried cable is about $10/ft of which 90% is installation and 10% is material cost. Thus, the wire size affects the cost in only a minor way and has been left out of this study. The exception
is the consideration of line losses which do have a direct impact on system costs. Simplified methods have been used to evaluate an electric power collection subsystem; the results are shown in Section 3.2.2. The techniques used to achieve transmission level voltages are transformers with circuit breakers for startup and repair isolation. Capacitor banks are used for power factor correction and underground ac cables rather than de are used for power collection. 2.2 CENTRAL GENERATION The alternative approach to generating electric power with a field of parabolic dish collectors is collection of heat over the entire field and transporting the heat (or chemical energy produced by the heat) to a central energy conversion plant. The currently available steam Rankine plant is the most appropriate energy conversion device based on price and efficiency at moderate temperatures (-= 550°C). There are many energy transport techniques such as pumped water, steam, liquid metal, inert gases or chemical transport. This approach is sensitive to operating temperature since the heat leak from a widely distributed piping network must be carefully minimized. The temperature also defines the type of material required for the pipe wall and combined with the line pressure defines the required wall thickness. The major tradeoff within a particular scheme (specific heat transfer fluid, temperature range and pressure) is the balancing of the pipe diameter to control the fluid pressure drop and required pumping power against the thermal heat leak. The pipe diameter and amount of insulation are varied to find the minimum total cost.
Another difference between central and distributed generation is the optimum ground cover ratio (GCR). The electric power collection system of the distributed generation approach is relatively insensitive to GCR and therefore system performance is weakly related to GCR. This is not the case for the central generation approach where high temperature fluid lines are used for energy transport. Based on earlier studies (Ref. 1), a GCR equal to 0.4 is used as near optimum for the various versions of the central generation system. Other issues of interest are the startup time from a cold start and associated difficulties, as well as overnight cooldown. This energy investment to achieve design operating condition in some cases is not a negligible amount of energy compared to the daily energy collected. The Rankine steam energy conversion plant can operate over a range of temperatures (250°C to 550°) and operating conditions (saturated steam to superheated steam). When neither saturated nor superheated steam is produced by the field of collectors, a steam generator (heat exchanger) is used between the energy transport fluid and the steam. Figure 5 illustrates the general layout of a central generation plant. The three major subsystems of this plant are the collectors, energy transport network and the Rankine energy conversion plant; these subsystems are discussed in the following sections. 2.2.1 Collectors The parabolic dish "point" focusing solar collector was described in a general manner in Section 2.1.1. All of that material is applicable to central generation systems. The differences are in two areas; these are the cavity receiver design and the optimum size of the dish. Different transport fluids can be considered such as water, steam NaK, He and gases such as CO and H2 (chemical transport). The design of the receiver will
Figure 5. Central Generation Layout
vary with the fluid chosen due to the widely varying heat transfer characteristics and the different tubing materials and wall thickness required. The cavity receiver approach as suggested in Section 2.1.2 is considered an appropriate design. 2.2.2 Energy Transport A solar thermal power plant with a field of collectors which locally heat some transport fluid requires a pipe network system for eventual delivery of energy to the power generation equipment. The pipe network will transport cool fluid from the feedwater heaters of the central power station and distribute it to all of the dish collectors. The heated fluid then transports energy to the power plant for electrical conversion. Since the transport network is only one component of the solar power plant, its design and cost must be compatible with the rest of the system. Nevertheless, useful information can be generated by isolating a proposed fluid-energy transport system from all other components of the power plant and estimating its technical and economic performance. In this section five types of candidate energy transport fluids and their associated transport systems are considered. The installed cost of each subsystem has been calculated, the primary factor for comparison being installed cost per delivered thermal kilowatt ($/kWt). Heat exchanger equipment at the collector and at the Rankine plant are included elsewhere, and fluid pumps and other equipment perpherial to the transport subsystem are not considered in this economic analysis as these components add little to the total cost. The transport and system costs include expenses for pipes, insulation (when required), fluid to fill the pipes, pipe supports and installation labor costs. Correction is also made for pumping power requirements (expressed in kWt) during transport and heat loss through insulation when appropriate. The five systems considered are listed in Table 2. The temperature level considered most appropriate for each heat transfer fluid is specified in this table.
Table 2. Energy Transport Subsystem Options
System 1. Pressurized Hot Water Transport In this heat transport concept pressurized water (liquid) exits from the collector at 315°C with a pressure in excess of 1545 psia. The hot water is collected to a central site, passes through a heat exchanger (steam generator) and vaporizes a working fluid in the Rankine power cycle. Alternatively it is flashed and becomes the working fluid. The return line contains water at a temperature of approximately 200°C. Both hot and cold lines can be fabricated from carbon steel piping; thermal insulation is required. System 2. Steam Transport with Water Return In this concept the collector subsystem produces superheated steam at 510°C. This steam is fed to an accumulator (thermal flywheel to accommodate transitory insolation fluctuations) and then is piped directly to the turbine generators. The cold return loop from the condenser is at a temperature in the range of 40°C to 260°C. The hot loop at 510°C will be fabricated from low alloy steel which is about four times more expensive than carbon steel for equivalent size pipe, and will be insulated. The liquid water return piping will be carbon steel. System 3. Liquid Metal Transport In this concept liquid metal (NaK) flows through a stainless steel heat collection-transmission loop. The stainless steel is necessary to resist corrosion attack by the hot liquid metal. The hot loop is assumed to be at a temperature of 650°C and the cool loop is at a temperature of approximately 370°C. Some severe technical problems (such as system reliability and safety problems) may exist if a liquid metal transport system is used. These have not been considered in the economic analysis. Special consideration must be given to all heat exchanger and collector piping components in contact with the NaK; it is necessary that these components be fabricated from stainless steel. The NaK does not require
pressurization and a line pressure of 100 psia has been assumed. System 4. Hot Inert Gas Transport For this concept helium is heated by the solar collector system to 650°C and returned from the power generation system at a temperature of approximately 370°C. The pressure must be as high as the pipes will allow in order to maintain the highest possible density (which reduces pipe size and required pump work); a pressure of 1500 psia has been considered in this study. The hot line is composed of low alloy steel and the cooler return line uses carbon steel. System 5. Chemical Transport Although there are many chemical systems which could be used,the methane approach is used as an illustrative example. Heat from the solar collector at 700°C causes a mixture of methane (CH^) and steam (H20) to thermally dissociate to hydrogen (H2) and carbon monoxide (CO). Thus,a large quantity of heat is stored chemically and this heat in turn is released after transport to a central site where the H2 and CO are passed through a catalyst bed and reacted chemically with each other. After the hot reaction products (with stored chemical energy) leave the collector they are passed through a heat exchanger and their sensible heat is transferred to the incoming reactants. The incoming reactants were previously at ambient temperature as they moved from the central plant to the collectors in the field. Thus, all transport lines between the dish collectors and the Rankine plant are at ambient temperature and do not require insulation or expansion loops. The energy transport subsystem analyzed is shown in Figure 6. Condensed water is separated from the gas whenever possible and run through a separate line; thus, there is one line containing product gases leading from the collector, and two lines containing CH4 (gas) and water (liquid) returning to the collector. The pressure will be 1500 psia in all lines, and the pipe material will be carbon steel. Since most of the heat from the collector goes into chemical storage with a high reaction enthalpy, the system
Figure 6. Chemical Energy Transport
flowrates will be much smaller than for systems which deliver sensible heat or even for system 2 which delivers sensible heat and latent heat in steam. The costs of counter flow pipe heat exchangers which could simply be two concentric pipes and water separators are not included in the study, but are expected to be minimal. These five approaches to connecting the field of dish collectors to a central Rankine power plant have been evaluated. The capital costs and cost of losses are minimized for each approach and expressed as a function of the total heat transported. The temperature rise in the collector is an important factor, as is the value assumed for the solar insolation, the collector efficiency and the ground cover ratio. The analysis of the energy transport subsystem considers these various effects in an attempt to make reasonable decisions for the entire system. Section 3.2 develops this analysis, and Section 4.2 integrates these data into a complete power plant evaluation. The question is often raised of how difficult it is to maintain and operate an extensive heat transport system that has, as in system 2, a high pressure (1500 psi) and high temperature (~950°F) fluid. Current practice in the utility industry uses higher temperatures (1100°F) and pressures (3500 psi) in the piping system between the boiler and the turbine; the transit distance is typically between 100 and 200 ft. This high temperature piping network is considered a high reliability component by the utility industry. When this piping network was applied to the solar plant of this study, the steam conditions were derated somewhat, but the length of travel was increased an order of magnitude to approximately 2000 ft for a 100 MWe sun following plant. Although it is an area that requires some further study, by using current installation practice, expansion loops, drains and traps, the use of an extensive piping network to transport heat to a central site is considered a reasonable concept. The impact on cost and performance are evaluated in Section 3.2.
2.2.3 Rankine Engine Existing commercial technology for the Rankine central plant is considered for use in the central generation system. The turbine inlet conditions vary from nearly saturated steam at approximately 290°C to superheated steam at 1500 psi and 510°C (220°C superheat). Since the heat source (field of collectors) is distributed over a wide area, interstage reheat by returning the fluid to the collector field is not considered as a realistic possibility. Reheat is possible by introducing a heat exchanger and using the fluid from the collector field as the heat source. Fossil reheating is also possible. However, for this evaluation reheat has not been considered. Feedwater heating is one of the major tradeoffs between the energy transport and Rankine engine subsystems. A useful reference to evaluate Rankine plant performance and cost is the Black and Veatch contribution to Reference 1. This has been used to obtain a consistent basis for Rankine plant evaluation. Section 3.2.2 will develop the cost-performance data for this subsystem.
SECTION III SUBSYSTEM PERFORMANCE CHARACTERISTICS The three major subsystems associated with the parabolic dish solar collector power plant are the collector itself, the energy transport subsystem and the heat engine. Each of these subsystems is discussed below and they are integrated for performance in Section 4. The technical performance of each subsystem and the state-of-the-art are considered along with the projected economics. 3.1 PARABOLIC DISH AND CAVITY COLLECTOR SUBSYSTEM 3.1.1 Performance The performance of this type of collector is shown in Figure 7. The collector thermal efficiency is shown versus collection temperature for several concentration ratios. The inherent assumptions are the following: 1) Glass transmissivity = e , where K (extinction coefficient) = 0.6 and T is the glass thickness (0.05 inches) 2) Silver reflectivity = 0.90 3) Cavity receiver absorptivity = 1.0 (black body) 4) Direct solar flux = 0.080 W/cm2 (254 BTU/hr-ft2 ) 5) 5% conduction and convection heat leak from cavity and the lines to the Brayton engine The reflectivity of a 0.05 inch thick plate of glass which was backsilvered was reported to be 0.94 in tests conducted early 1975 at JPL. The cavity receiver is nearly a black body, and the major heat loss mechanism from such a heat receiver is radiation through the aperture. Convection and conduction are small for an insulated cavity. Even the non-cavity receiver in Reference 1 operating at 815°C with a concentration ratio of 1000 has a heat leak of only 0.75 kW out of the 20 kW which are absorbed. This is about 4%; the
Figure 7. Paraboloid Dish So!ar Collector Performance
cavity design reduces the heat leaks even further. The heat leak along the path from the receiver cavity to the Brayton engine at the counterweight position is considered to be about 5% of the incident energy. This can be achieved by a well insulated gas line between the focal point and the Brayton engine. A flexible link is required between the two-axis tracking dish if the engine is positioned on the ground. An alternate approach would be the use of a secondary reflecting surface at the focal point to the receiver placed behind the dish surface. This Cassegrain approach would place the cavity receiver much closer to the Brayton engine and minimize line heat leaks, but introduce additional reflector losses. A fixed receiver concept was introduced by Honeywell (Ref. 1) using a probe holding the receiver through a slot in the dish surface. These various approaches have to be evaluated in detail to trade off initial capital cost versus heat leak. Many questions of performance and reliability are raised when a flexible fluid line is used to carry the heat from the two-axis tracking dish to a fixed position on the ground. The probe through the slot approach avoids a flexible joint, as does an approach of using a Cassegrain geometry to reflect the energy back through the dish center to a fixed receiver on the ground. For the case of the small heat engine located at each dish, a Cassegrain reflector could be used to deliver the heat to the heat engine hung at the counterweight position of the dish. A flexible electric connection would be used to transport the power from the two-axis tracking dish to the ground. Although several approaches have been identified, a detailed study is required to find the most appropriate concept. The receiver cavity can be open to air at temperatures up to 900°C if nickel alloys such as Inconel are used as the heat exchanger tubing material (Ref. 6). Inconel has a stable oxide coating in air up to 900°C. For a long life, design stresses should be limited to 1000 psi. This is possible with low pressure closed
cycle gas turbines which use inert gases such as He or He-Xenon mixture up to 100 psi. Alternate materials exist such as TD nickel (DuPont) for a tubular heat exchanger in the cavity receiver. Alternate cavity designs exist such as a porous graphite receiver slab in a high temperature glass pipe such a concept would be operated at lower temperatures due to glass strength limits (Ref. 7). Open cycle (air) Brayton cycles are also being considered with various porous ceramic receivers. The dish collector can be used to heat a variety of fluids for the central generation system, or a gas for the distributed generation system. Although the materials, temperature level and specific design will be different, a cavity receiver similar to that for inert gases can be used. Thus, the cavity receiver is attractive for several reasons. It is probably not necessary to use a vacuum to minimize convection heat leaks which both simplifies the design as well as eliminates the transmission decrease of an additional glass surface. Radiation heat leak is limited to only the small amount reradiated through the small aperture. The heat transfer problems which result from such high concentration ratios are minimized since the amount of tubing surface inside the cavity is relatively independent of the concentration ratio and aperture size. This is especially important when a gas heat transfer medium is used. A collection efficiency prediction of 70% is estimated for 815°C at a concentration ratio (CR) of 1000 as shown in Figure 7. The collection efficiency approaches 80% at temperatures of 550°C. This study makes the simplifying assumption that the absorber fluid exit temperature is relatively close (<50°F) to the effective cavity temperature. The collector temperature of Figure 7 is therefore considered to also be the fluid temperature. This assumption applies very well to good heat transfer fluids such as NaK and boiling water. Poorer fluids such as helium and subcooled water and superheated steam do require significant temperature differences between the cavity and
RkJQdWJsaXNoZXIy MTU5NjU0Mg==