全球变化知识资源中心

Integrating solar thermal systems into Rankine-cycle power plants can be done with minimal modification to the existing infrastructure. This presents an opportunity to introduce these technologies into the commercial space incrementally, to allow engineers to build familiarity with the systems before phasing out fossil-fuel energy with solar electricity. This paper shows that there is no thermodynamic barrier to injecting solar thermal heat into Rankine-cycle plants to offset even up to 50% fossil-fuel combustion with existing technology: with better solar-to-electricity efficiencies than conventionally deployed solar-thermal power plants. This strategy is economically preferable to installing carbon-capture and compression equipment for mitigating an equivalent amount of greenhouse-gas emissions. We suggest that such projects be encouraged by extending the same subsidy/incentives to the solar-thermal fraction of a ‘solar-aided’ plant that would be offered to a conventionally deployed solar-thermal power plant of similar capacity. Such a policy would prepare the ground for an incremental solar-thermal takeover of fossil-fuel power plants.

The need to cut greenhouse-gas emissions is well recognized. However, thermal power plants burning fossil fuels, which are the largest single source of global greenhouse-gas emissions, supply most (~70%) of the world’s power. Although some renewable technologies for power production, notably wind and solar photovoltaics, can reasonably claim to have reached ‘grid-parity’^{1, 2}, this is true only intermittently: as long as the sun is shining or the wind is blowing. The cost of electric storage (for example, battery or pumped hydrostorage) and the significant changes to power distribution infrastructure required make intermittent energy technologies still uneconomical for providing base-load or dispatchable power; as the German experience will warrant^{3, 4}. Hence, notwithstanding the forceful arguments in favour of mothballing coal-fired power plants^{5, 6}, this is not a feasible medium-term solution even in the developed world⁴. On top of this, it now seems that the renewables renaissance is under pressure from a new era of cheap hydrocarbons⁷. Although an excellent case can be made for pricing/taxing carbon dioxide emissions^{6, 8} as a means to encourage greater adoption of renewables, the design of such instruments is non-trivial and any step towards enforcement is likely to encounter stiff resistance from a wide swathe of the socio-political spectrum in the developed world and find practically no traction in the developing world⁹. The current thermal power infrastructure is fully depreciated, has decades of operational life left and hence the immediate effect of applying carbon pricing will merely be an increase in electricity/fuel prices that will disproportionately burden the poor; that is, it could easily become a regressive tax that will need complex instruments (‘green cheques’ and so on) to offset. The problem is severe¹⁰ enough that the founders of Google’s ambitious RE < C program have despaired of finding a solution with current technology¹¹.

This paper presents a strategy by which the existing fossil-fuel power plant infrastructure can be co-opted to progressively mitigate greenhouse-gas emissions in an economical and sustainable manner.

Solar-thermal electricity generation, accompanied by cost-effective thermal storage, is widely held to be the renewable resource utilization strategy that promises to deliver base-load and/or dispatchable electricity in the medium term^{12, 13}, pending development of radically cheaper electricity storage devices or perhaps massive deployment of wind power¹⁴. Such systems are conventionally deployed in the manner of the Solar Electricity Generation System (SEGS) systems of the Californian desert and constitute most solar-thermal power producers¹⁵. Notwithstanding the fossil-fuel ‘backup’ (used to tide over periods of lower solar intensity and/or extend operation beyond sunlit hours), these conventionally deployed plants are conceived as primarily solar powered and thought of as replacements for conventional fossil-fuel-fired thermal power plants¹³. However, as things stand, solar-thermal power production is on the back-foot especially in comparison with solar photovoltaics¹⁶, with several solar-plant developers having shifted from the former to the latter in the recent past¹⁷. A major reason for this situation is that, unlike photovoltaics projects, conventionally deployed solar-thermal projects do not allow modularity. They must be installed in blocks of tens of megawatts of solar-thermal capacity (sufficient to power an economical Rankine cycle) before they can even begin to produce electricity with the large solar fractions necessary to attract subsidies.

The alternative is to integrate solar thermal technology into the heat cycle of a conventional fossil-fuel-fired power plant; a straightforward implementation because the power block of both is essentially the same. The idea of such synergetic solar-aided fossil-fuel power plants has been investigated for some time¹⁸ and such plants are shown to be significantly more cost effective than the conventionally deployed solar-thermal plants¹⁹. Notable plants executing this strategy are the Martin Next Generation Solar Energy Centre in South Florida (with installed solar capacity corresponding to 2% of the 3.8 GW plant), Kogan Creek Solar Boost in Australia (6% of 750 MW), ISCC Kuraymat in Egypt (15% of 140 MW) and ISCC Hassi R’Mel in Algeria (17% of 150 MW). The 100 MW Shams-I plant in the United Arab Emirates, with natural-gas-fired steam superheaters following solar-thermal boilers, may also be considered an example, although it would have been more efficient to pass the natural gas through a Brayton cycle first.

However, this strategy has not drawn the attention it deserves from the stakeholders in the field. As a result, these plants constitute less than a tenth of the total installed solar-thermal power capacity around the world¹⁵ and find no mention in comprehensive renewable energy roadmaps by the authors of refs 20, 21 and only a passing mention in the solar thermal roadmap of the International Energy Agency (IEA) in 2010 (ref. 13). A separate report by the National Renewable Energy Laboratory (NREL) in 2011 (ref. 22) suggests that the solar-augment potential of American thermal power plants is less than 5%.

A possible explanation for this neglect of a promising strategy could be a ‘one-plant-one-fuel’ paradigm in the area of power production. We must point out that this neglect is not benign. Solar-thermal subsidies (for example, in Spain and the United States) are designed to reward conventionally deployed solar-thermal power producers with an upper limit set to fossil-fuel co-firing. Such a policy clearly precludes solar-aided fossil-fuel power plants, with minority solar fractions, from its rewards.

This paper presents a comprehensive analysis of the benefits of injecting solar thermal heat, harvested by various technologies, into the Rankine cycle of a coal-fired power plant. We have shown that there is no thermodynamic barrier to reaching 50% fossil-fuel offset using existing line-concentrating technologies and with solar-to-electricity conversion efficiencies several percentage points better than those obtained by conventionally deployed solar-thermal power plants. The fuel-offset values will, of course, be subject to constraints including availability of land, solar resource and thermal storage. We also bring out several other subtle synergies of such integration; not least of which is that it is more profitable than installing and operating carbon-capture and compression equipment for mitigating equivalent greenhouse-gas emissions.

It is therefore the contention of this paper that solar thermal technology should not be treated on par with intermittent power production systems such as photovoltaics or wind turbines but treated as an independent category in itself for policy purposes. An example of the result of such a policy change would be to relax the requirement for large solar fractions in the electricity generated by a solar-thermal power plant for attracting government assistance; but rather the solar-thermal contribution to solar-aided power plants should be monitored for extending support to this technology. Policy recommendations in this direction are already made while keeping in mind the fossil-fuel backup of conventionally deployed solar-thermal power plants¹³ but their scope needs to be extended to the solar fraction of solar-aided thermal power plants.

Figure 1 is a schematic diagram of a typical coal-fired power plant operating a Rankine cycle that raises high-pressure steam in a boiler, superheats it and depressurizes it through several steam turbines (with intermediate reheating); the exhaust steam is condensed in a condenser. The cold condensate is preheated in several stages by steam bled from the turbines before injection into the boiler. This use of steam bleeds significantly improves efficiency of the cycle (as can be seen from Table 1). Heat to the Rankine cycle is supplied by hot flue gases resulting from combusting fuel with ambient air preheated by cross-exchange with exiting flue gases. Flue gases are exhausted at their acid dew point, that is, the temperature below which water vapour would condense on the heat-transfer surfaces, and this dew is acidic enough to corrode them. For this analysis, the ambient air temperature is 30 °C and the acid dew point is 130 °C. Heat from solar radiation is harvested with one (or more) of several types of solar thermal harvester and transferred to a heat-transfer fluid (HTF). Figure 1 also shows how this solar heat can be injected into the Rankine cycle using one of two strategies: standard and feedwater heating (FWH). The former (standard) is so called because it does not affect the steam cycle of the plant at all. It directly supplements the fossil-fuel enthalpy in the economizer and boiler (and perhaps in the superheaters) while feedwater preheating is accomplished with steam bled from the turbines. The latter (FWH) uses solar heat to preheat feedwater, allowing steam bleeds to be suppressed and increasing power production because the steam not bled can do further work in the turbine; thus, it significantly modifies the steam cycle of the plant. The FWH strategy does not preclude use of solar heat in the boilers and superheaters. We have considered coal-fired plants operating on the subcritical and supercritical Rankine cycles and a stand-alone (no fossil-fuel input) solar thermal plant. Table 1 reports the parameters used in the calculations. If thermal-to-electric conversion efficiency were the only criterion, then it would be most efficient to supply solar heat at the highest possible temperature to raise steam while using steam bleed to preheat condensate. However, solar radiation capture efficiency has an inverse relationship with the temperature. This effect can be modelled straightforwardly for solar thermal devices (see Methods). Table 2 lists the various types of collector we have considered in this article (details of each technology can be found in standard reference books²³).

Co, condenser; PP1–PP6, feedwater pumps; BP, boiler pump; E, economizer; B, boiler; AP, air preheater; C, combustion region; S1, first superheater; S2, second superheater (reheater); HP/MP/LP, high-, medium- and low- pressure turbines. S1 and S2 share a common furnace. Also, four points of solar-thermal injection are shown: SLPH, SLBE, SLSP1, SLSP2—solar supplement to preheater, boiler and economizer, first superheater and second superheater respectively. Solid, dotted-and-dashed, dashed and dashed-and-double-dotted lines depict the working fluid, working fluid diverted for solar heating, flue gas and HTF circuit respectively.

Figure 2 is a summary of our results. A detailed discussion appears in Supplementary Discussion and all data points are available in the Supplementary Spreadsheet. The observation most pertinent here is that HTF heated to 400 °C (the limit of state-of-the-art HTF Dowtherm A) can be used to offset 57% and 28% of fossil-fuel combustion in a subcritical and supercritical (respectively) Rankine-cycle plant. This is with solar-to-electric efficiencies in excess of 16% (the value for an SEGS type solar-thermal plant operating with C50 and no fossil-fuel input during sunlit hours). Further, these numbers are for the standard configuration, which would necessitate, at most, a reconfiguration of the plant’s heat exchangers. With the FWH strategy, it is possible to use cheap and easy-to-maintain designs, for example, vacuum tube collectors, to generate power. Whereas feedwater preheating can in principle offset about 20% of fossil-fuel combustion^{24, 25, 26}, it has an upper limit set by the quantum of additional steam that the turbines can take (<20% above rated capacity) that limits fossil-fuel offsets to 6%. This may, however, constitute ‘low-hanging-fruit’ for power plants seeking to meet emission norms. More subtly, feedwater preheating can avoid efficiency penalties accompanying suppressing steam bleeds to meet temporary spikes in demand; thereby making a base-load coal-fired power plant dispatchable to the tune of 6% of its operating capacity. This fraction can be higher if there is a spare turbine (for example, the ‘spinning reserve’) to accept the excess steam. Alternatively, the excess steam could be used to drive the regenerator of an absorption-based carbon-capture unit. This also suggests a strategy for the incremental introduction of a solar fraction into power plants: begin with the low-cost collectors and gradually move up the technology ladder.

a–f, Results of injecting solar thermal energy into subcritical and supercritical Rankine cycles. a,b, Temperature-entropy diagrams of the respective cycles (see Table 1) with temperatures of flue gas (dashed-and-dotted line) and HTF (bold-dashed line) corresponding to each temperature of the working fluid (solid line). c,d, Effect on various parameters (including solar radiation collection efficiencies) of injecting solar heat using HTF heated to the temperature shown on the ordinate. Labels refer to the corresponding solar thermal harvester. Standard strategy is assumed unless the label is qualified with FWH. e,f, Overall solar-to-electricity efficiencies. Some lines are omitted for clarity.

A major thrust of the policy change that we recommend is to break away from the ‘one-plant-one-fuel’ paradigm. This paradigm is implicit in practically every nation’s renewables policy; for example, policies of the United States and Spain actually prescribe an upper limit to the fossil-fuel co-firing permissible in a solar-thermal power plant and then reward the entire output of the plant—a policy that excludes the minority solar-thermal fraction in a solar-aided fossil-fuel power plant. Our argument is that, with the solar-aided power plant paradigm, the conventional power infrastructure can actually prepare the way for development and deployment of renewable power.

Our policy recommendation is therefore quite simple: that policymakers should drop the insistence on majority solar fractions in solar-thermal power plants and instead offer the same incentives to the (initially minority) solar-thermal fraction of solar-aided power plants that would be offered to the corresponding conventionally deployed solar-thermal power plant: including the favourable feed-in tariffs (a point echoed in the IEA Solar Thermal Roadmap 2010 (ref. 13) but not made forcefully enough).

There is therefore no need to call for mothballing of coal-fired power plants^{5, 6}: a strategy that has proved unworkable in Germany³, let alone the developing world. The strategy of carbon pricing is at once politically untenable and stands a real chance of becoming a regressive tax. Instead, we propose incentivizing renewable power irrespective of its source using funds raised by a progressive taxation regime. Not only will this, as a subsidy rather than a penalty, be politically palatable, it will give power plants the breathing room needed to familiarize themselves with solar thermal technologies. A further course of action in this direction would be to develop protocols for estimating the fraction of power output that can be attributed to the solar-thermal section³⁰. This will facilitate the solar takeover of thermal power plants.

Solar-thermal power production is indispensable for a renewable future but has not seen the support from policymakers, entrepreneurs or researchers that solar photovoltaics has enjoyed. The time is ripe for a revolution in the field of solar-thermal power production similar to that which led to Swanson’s law² for photovoltaics. Perhaps this small policy change will help kickstart that revolution.

This exercise is based on a regenerative Rankine cycle with reheat and is applicable to coal-fired power plants. The calculations are simple enough to be done on a macro-enabled spreadsheet (details appear below). For all plants, we have considered six-stage direct-contact preheating of condensate with steam bled from the turbines. The properties of steam at various conditions of temperature and pressure have been taken from IF-97 steam tables using the MS-Excel Macro written by M. Holmgren (http://www.mycheme.com/steam-tables-in-excel). All mechanical devices are deemed to operate with 90% efficiency. Heat from combustion of fossil fuels is supplied by hot flue gases: modelled as air at atmospheric pressure. The gases exiting the furnace are used to raise steam in the boiler, then to preheat boiler feedwater in the economizer, and then to heat incoming fresh air (at T_amb = 30 °C) in the air preheater before being vented at a temperature of 130 °C (the acid dew point). Minimum temperature driving force between flue gas and water/steam in the superheater, boiler/economizer and air preheater are 60 K, 50 K and 100 K respectively; and between the heat-transfer fluid (HTF) and water/steam it is 50 K in the superheater and 20 K elsewhere. The specific heat capacity of the HTF is that of Dowtherm A (http://www.dow.com) and fitted to the linear form C_p^HTF (kJ kg⁻¹ K⁻¹) = 0.00293(T_K) + 0.7 (T_K in K). To keep from distracting from our main point, we have kept as many parameters constant across the various systems studied as we could without compromising the peculiarities of each system. Hence, for the purposes of this article, solar capture efficiency is calculated from the following averaged values: ambient temperature T_amb = 30 °C, radiation absorptivity of the receiver α_abs = 0.85 (includes absorptivity of surface and transmissivity of protective glass layer), reflectivity of mirror (if any) α_ref = 0.9, emissivity of receiver ε = 0.1. Area-cosine losses are taken as α_area^no−track = 0.8 for untracked solar harvesting and α_area^track = 0.85 for solar collectors mounted along the north–south direction and tracked in one dimension and α_area^tower = 0.75 for central-receiver towers. Global insolation is taken as I = 800 W m⁻² and direct normal irradiance is I_DNI = 600 W m⁻², that is, α_DNI = 0.75. The concentration ratio is C_R. At any point on the receiver surface, the incident radiation flux is given by I_inc = α_absα_cosineI(1 + α_refα_DNI(C_R − 1)), and the flux radiated away is R(T) = εσ(T_K⁴ − T_{amb, K}⁴),