The intermittency of renewable power generation systems on the low carbon electric grid can be alleviated by using nuclear systems as quasi-storage systems. Nuclear air-Brayton systems can produce and store hydrogen when electric generation is abundant and then burn the hydrogen by co-firing when generation is limited. The rated output of a nuclear plant can be significantly augmented by co-firing. The incremental efficiency of hydrogen to electricity can far exceed that of hydrogen in a standalone gas turbine. Herein, we simulate and evaluate this idea on a 50 MW small modular liquid metal/molten salt reactor. Considerable power increases are predicted for nuclear air-Brayton systems by co-firing with hydrogen before the power turbine.
Motivation and Background.
As the U.S. moves toward a low carbon electrical grid, the only power sources that will be available to any great extent will be the renewables, solar and wind, and the dispatchable, nuclear. For these three types of sources to provide a stable grid, a strategy must be developed for them to work together. The intermittent characteristics of solar and wind sources require that an energy storage system be developed that can smooth the differences between power production and power demand. One possible solution is a nuclear component that develops its own storage. There will clearly be times when a solar system is incapable of generating power. The same is true of a wind system. At those times it would be very useful to have nuclear systems online. Of course when the sun is shining and the wind is blowing, nuclear systems may not be needed. This poses a dilemma for nuclear systems as they produce very little revenue when shut down. Quite simply the electric grid needs an energy storage system, and it needs to be flexible. Here, we describe several nuclear power systems that can meet this need. The total system has multiple components (Fig. 1). The reactor provides heat to the power cycle—the subject of this paper. The reactor operates at base load. The system can operate at base-load on nuclear heat. To produce peak power, hydrogen is added to the compressed air after nuclear heating in the power cycle to raise temperatures and power levels. At times of low electricity demand, electricity from the grid is fed to a high-temperature electrolysis system to produce hydrogen. Hydrogen is stored and used as fuel for peak electricity production. Hydrogen as a storage system has the advantage of low-cost storage using the same underground storage technologies as used for natural gas. However, the challenge is round-trip efficiency of electricity to hydrogen to electricity. High-temperature electrolysis is the most efficient method to produce hydrogen. A nuclear air-Brayton cycle (described herein) is the most efficient method to convert hydrogen to electricity. This combination creates new options for using hydrogen for energy storage.
A nuclear system is needed to store energy when its production is not needed and recover that energy when it is needed to augment the nuclear system's baseload capability. One possible solution to this challenge is to develop a nuclear air-Brayton combined cycle (NACC) system. The standard gas turbine combined cycle (GTCC) would be modified to use nuclear energy rather than chemical energy. This can be accomplished by replacing the gas turbine combustion chamber with a heat exchanger. A nuclear-driven heat exchanger cannot reach the high temperatures that current gas turbine combustion chambers do as it requires a temperature drop from its walls to the working fluid (air) rather than a temperature drop from the working fluid to the walls of the chamber. However, there are compensating benefits. In particular, a heat exchanger can be built with a smaller pressure drop in the working fluid, and the air can be heated multiple times. Heating the working fluid multiple times is very common in steam systems. Since this is an advanced concept and system performance is always better at peak cycle temperatures, only liquid metal systems and molten salt systems have been considered for this study. High-temperature gas systems could also be considered, but the cycles would look enough different that they need to be analyzed in a different fashion. It is also possible to develop a nuclear air-Brayton recuperated cycle (NARC) system with an intercooler using either water (NARCw) or air (NARCa) to cool the intercooler. Liquid metal and molten salt heat exchangers were developed in the 1960s for the aircraft nuclear propulsion program and performed quite well . Sodium-to-air heat exchangers were used on the Fast Flux Test Facility to dump the heat generated in the nuclear core. Currently, there are designs to use sodium-to-air heat exchangers for safety heat dumps on sodium fast reactors. Apparently such heat exchangers are not as complicated as sodium-to-water heat exchangers and do not have some of the severe safety issues . The system of interest is described in Fig. 2. It is an open cycle air-Brayton topping cycle mated to a steam-Rankine bottoming cycle similar to most gas turbine combined cycles. The Brayton system consists of one or two compressors and three gas turbines followed by a heat recovery steam generator (HRSG). The HRSG drives three steam turbines in the bottoming cycle. To improve the normal operating efficiency of the system, a recuperator can be included after the HRSG. If this is done, an intercooler can also be included between the two compressors.
The recuperator is required to take advantage of the efficiency improvements provided by the intercooler. Three gas turbines with three reheat cycles were chosen as this system performs better than the two gas turbine system and very close to the performance of four gas turbines. The three gas turbines consist of two turbines connected directly to the compressor(s) and a power (or free) turbine connected only to the driven electrical generator. The system has three heat exchangers, one before each turbine. The power turbine approach seems to perform slightly better than a system with all three turbines, the compressor(s), and the generator on the same shaft. The turbines for the steam system can be connected to the same shaft as the power turbine or to a separate shaft driving a second generator. The basic cycle without the recuperator and intercooler (RIC) has been identified as an NACC. When RIC is added, it is identified as an NACC&RIC. Since the RIC seemed to make the combined cycle more efficient, a system was considered that did not include the bottoming cycle with the turbine exhaust going directly to the recuperator. This system has been identified as an NARCw or NARCa. The calculated efficiency of each of the three systems is compared with the supercritical carbon dioxide (SCO2) system  under study by DOE in Fig. 3.
The basic NACC system is significantly less efficient than SCO2 system when both operate at a turbine inlet temperature (TIT) of 500 °C, typical sodium reactor output temperatures , but improves significantly as temperatures approach those expected from molten salt reactors . Adding the recuperator and intercooler brings the performance of the NACC&RIC system up considerably at the lower temperatures. It appears that the NARCw system outperforms all three on an efficiency basis. Typical compressor pressure ratios for the NACC and NACC&RIC systems are given in Fig. 4. Adding the recuperator lowers system operating pressure as expected. This would work as follows: The nuclear reactor system would operate continuously in somewhat of a baseload fashion. When the sun is shining and the wind is blowing, the electricity produced by the nuclear system will be used to produce and store hydrogen by high-temperature electrolysis . Any excess electricity produced by the renewables could also be diverted to produce and store hydrogen. When the sun goes down and the wind stops blowing, the nuclear system will produce the electricity required to meet the grid's demands. In addition to the rated nuclear plant output, the stored hydrogen can also be burned in a combustion chamber and its exhaust fed to the power turbine to “superheat” the airflow. This process is called co-firing. Augmentation of the nuclear system's output in this fashion is the purpose for considering NACC and NARC systems as a type of energy storage mechanism for the grid of the future. Other energy storage mechanisms have been proposed that may be more efficient than producing and storing hydrogen, such as heating firebricks electrically and recovering the energy by flowing the air over the firebricks prior to entering the nozzle for the power turbine . Only the hydrogen production and burning scheme will be considered here. Of course NACC/NARC systems have other advantages than advanced efficiencies and serving as a grid energy storage mechanism. They will take advantage of the broad manufacturing base of gas turbine components and technology. And they will significantly reduce the water requirements for waste heat removal, reducing such requirements by at least a factor of 2, compared to current light water reactors. But the focus here is on the energy storage mechanism.
The goal of this effort has been to develop models that will enable us to identify the characteristics of optimal systems. Once the optimal systems have been identified, then a search can be undertaken for off-the-shelf components that come close to meeting the requirements of an optimal system. If no such components exist, then a development need has been identified, and specific requirements can be set. There exist a number of cycle analysis codes that can address NACC/NARC systems such as thermolib (Aachen, Germany), thermoflex (Southborough, MA), and Aspentech (Houston, TX). Many of our results have been checked against thermolib, but it was desirable to build a code that would allow not only cycle calculations but also sizing estimates and possibly a bottom up economic analysis. The cycle modeling follows the techniques described in Refs. [8–10]. Size estimates for turbomachinery follow the scaling parameters given as in Ref. . Heat exchanger sizes are calculated based on preliminary designs. Air compressor and turbine polytropic efficiencies follow the published scaling rules . Steam turbines were simply set at a 90% isentropic efficiency. All the heat exchangers were designed to meet a 1% pressure drop and a 95% effectiveness. Recuperators and intercoolers are counterflow, and all other heat exchangers are cross flow. A pinch point temperature difference of 10 deg was baselined. Predicted efficiencies have been compared against currently published data for gas turbine combined cycles and agree very closely [12,13].
Nuclear Air-Brayton Combined Cycle Scenarios Analyzed
The objective of this effort was to analyze the impact of co-firing the power turbine of an NACC system with a hydrogen burn to augment its power output. Since the normal turbine inlet temperatures produced by a liquid metal or molten salt heat exchanger are well below typical gas turbine inlet temperatures, two burn temperatures were chosen as being of interest. The first was 927 °C which corresponds to the peak temperature that uncooled turbine blades are advertised to withstand . The second was 1427 deg, the temperature advertised as the peak that GE H series turbines can achieve . Clearly, the nuclear heated only systems can operate with uncooled turbine blades. If they can operate with cooled blades without the cooling air losses at reactor outlet temperatures, then it is worth considering co-firing at the 1427 deg limit. At the 1427 deg limit, it may well be necessary to require ceramic heat exchangers for the HRSG and its superheaters. Estimated HRSG inlet temperatures in this case are about 1154 deg. It should also be pointed out that the power turbine in all the cases will have to be fed by a variable area nozzle to accommodate the higher temperature gas and maintain input pressure and mass flow rate. The operating pressures in the steam bottoming cycles were held constant between the normal operating conditions and the co-firing environment. This was done to eliminate one control variable in the analysis. It may be useful to address this restriction in the future. The near optimum pressure for the NACC system was 9 MPa, and the near optimum pressure for the NACC&RIC system was set at 12 MPa. The NACC efficiency improves slightly at lower pressures, and the NACC&RIC improves slightly at higher pressures due to higher temperature air reaching the recuperator. But the improvements in performance are not significant, and these two pressures are in the range of current GTCC systems.
The power level chosen for this study was 50 MW (e), typical of some small modular reactors currently being considered. For both systems (NACC and NACC&RIC), the Brayton air cycle produced about 70% of the power, and the Rankine steam cycle produced 30% of the power at rated power output. A crucial parameter of interest is the ratio of the mass flow rate of steam to the mass flow rate of air. For the NACC system, this ratio was about 3% at 500 deg and 6% at 700 deg. For the NACC&RIC system, it was approximately 1% lower at all the temperatures. Normal turbine inlet temperatures are used to characterize the systems analyzed as they can easily be related to the technology involved.
Nuclear Air-Brayton Combined Cycle With Only Co-Firing
For both the NACC and NACC&RIC systems, there are several design strategies that can be pursued. Consider first the NACC system. For this system, there are no downstream temperature limitations. So when the gas exiting the power turbine is routed to the HRSG, the entry temperature for the 927 deg power turbine inlet temperature is 763 deg, and the HRSG exit temperature is 277 deg if the normal turbine inlet temperature is 500 deg. The pinch point temperature difference is 348 deg. For this power turbine inlet temperature when the normal turbine inlet temperature is 700 deg, the HRSG inlet temperature is 666 deg and the exit temperature is 405 deg. The pinch point temperature difference is now 163 deg. The increase in power as a result of the hydrogen co-firing is shown in Fig. 5 for both the 927 deg and 1427 deg cases.
At a normal 500 deg TIT, the power goes up about 55% when co-fired to 927 deg and more than doubles at 1427 deg. The hydrogen is burned with an efficiency of between 20% and 32% for both the 927 deg case and the 1427 deg case. This low efficiency occurs because the temperature drop between the air and the steam in the evaporator is quite large. The pinch point temperature difference goes from 10 deg to 153 deg. This is a major penalty and probably not a reasonable method of operation. The efficiency of the hydrogen burn is improved significantly with other designs, as shown.
Nuclear Air-Brayton Combined Cycle With
Co-Firing and Increased Steam Flow
The large pinch point temperature differences with simple co-firing imply a large increase in irreversibility or exergy. The easiest way to bring the pinch point temperature difference back down to a more efficient case (10 deg) is to increase the steam flow in the bottoming cycle. This would imply that during normal operation, the bottoming cycle would operate at flow rates significantly less than maximum. The loss in efficiency for operating in this manner has not been calculated. However, since the steam system only produces 30% of the total power during normal operation, the loss in efficiency will be mitigated. Figure 6 describes the increase in steam flow required to lower the pinch point temperature difference to its value before co-firing.
At a 500 deg nominal TIT, the increase in steam flow is 300% for 927 deg and 500% for 1427 deg. This means the normal bottoming system would be running at 25% of maximum if a 927 deg burn is desired and at 17% of maximum if a 1427 deg burn is desired.
However, the increase in power is significant as shown in Fig. 7. For the normal 500 deg TIT case, the power can be increased by 200% for a burn at 927 deg and 450% at 1427 deg. At a normal 700 deg TIT, the increases are not as great, but still very significant—200% for co-firing at 1427 deg and 100% for 927 deg. Figure 8 presents the hydrogen burn efficiency for this case.
NACC&RIC With Co-Firing—Maximum Recuperator Temperature Limit
When the recuperator and intercooler are added, the normal operation efficiency increases, but it sets a limit on the peak temperature that the recuperator can add to the flow entering the first heat exchanger. No matter how much heat the recuperator adds during co-firing, it will affect the heat transferred in the first heat exchanger. The limiting case occurs when the air temperature entering the first heat exchanger reaches the normal TIT. In this case, the heat exchanger transfers no heat to the air, and the reactor power must be reduced to compensate for this lack of heat transfer. This does not appear to be an impossible condition, but all of the ramifications have not been explored at this time. Therefore, the following analysis is based on limiting the hot air out of the recuperator to the normal TIT for the system. This can be accomplished in three ways. First, it is possible to limit the burn temperature in the co-firing process to less than the maximum. Second, it is possible to dump the hot gas going into the recuperator so as to not exceed its output limit. Third, it is once again possible to increase the steam flow in the bottoming cycle to pull down the fluid temperature entering the recuperator so that it only transfers enough heat to bring the air flow entering the first heat exchanger up to its normal TIT. The second option will not be considered as it is obviously a waste of energy.
Limiting the Co-Firing temperature
The limit on the co-firing temperature as a function of the normal system TIT is plotted in Fig. 9. The limit is less than 927 deg, up to about 560 deg. So at normal system TITs above 560 deg, the full 927 deg co-firing can be accomplished. The peak co-firing temperature for the 700 deg case is about 1317 deg, so the 1427 deg co-firing cannot be reached even in this case.
Figure 10 describes the power increase possible by choosing to limit the maximum co-firing temperature as described earlier. The increase varies from 35% to 70% as a function of the normal system TIT. Figure 11 presents the required reactor heat output as a function of its rated power during co-firing at the maximum allowed temperature.
The reactor power must be reduced because the first heat exchanger is no longer transferring heat to the air. Figure 12 presents the hydrogen burn efficiency for this case. Note that the efficiency for this case is significantly greater than those for the basic NACC system and 10–15% greater than any achieved by current gas turbine combined cycle systems.
NACC&RIC With Co-Firing—Increased Steam Flow to Meet Recuperator Temperature Limit
The steam flow in the bottoming cycle can be increased to limit the temperature out of the recuperator to the normal operating TIT similar to the way it was for the NACC system. In this case, it is not the pinch point temperature difference that matters, but rather the recuperator outlet temperature entering the first heat exchanger. Figure 13 plots the required increase in steam flow to limit the recuperator outlet temperature. Note that for normal system TITs above about 560 deg, the steam flow does not have to be increased for the 927 deg co-firing. The peak increase is about 150% at a system normal TIT of 500 deg. When co-firing a 700 deg system at 1427 deg, the peak increase in steam flow is about 15%. The increase in power achieved by increasing the steam flow in the bottoming circuit is shown in Fig. 14.
At the 500 deg limit, increasing the steam flow by 140% increases the power output by slightly over 100% when co-firing at 927 deg. When co-firing at 1427 deg, increasing the steam flow by 400%, increases the power output 340%. At the 700 deg limit, the steam flow does not need to be increased but a 25% increase in output is achieved by co-firing at 927 deg. Co-firing at 1427 deg increases the power output 90% for a 15% increase in steam flow. The reactor power requirements are plotted in Fig. 15. For the 927 deg co-firing, the power requirements decrease with normal system TIT until an increase in steam flow is no longer required. When the increased steam flow is no longer required, the recuperator outlet temperature to the reactor drops and the reactor power must be increased to maintain the nominal system TITs. This causes a rather dramatic increase in reactor power above 550 deg. When co-firing occurs at 1427 deg, increased steam flow is always required, so the reactor power requirement continuously decreases from about 40% to 30%.
The efficiency of the hydrogen burn is plotted in Fig. 16. For the 927 deg co-firing, the efficiency starts at 45% but rapidly rises to over 70% and peaks for the 700 deg system at almost 90%. This is because the hydrogen is introduced at the normal TIT temperature and must raise the fluid temperature to only 927 deg. For the 1427 deg co-firing, the burn efficiency starts at 47% and eventually increases to over 75% for the 700 deg system.
Nuclear Air-Brayton Recuperated Cycle Scenarios Analyzed
By dropping the steam bottoming cycle, it is possible to build a nuclear air-Brayton recuperated (only) cycle. With a water cooled intercooler (NARCw), this cycle appears to be the most efficient as plotted in Fig. 3. The relative amount of heat that must be dumped to the environment for an NACC&RIC system and an NARCw system is plotted in Fig. 17. The NARCw system is significantly better than an NACC&RIC system. However, it is also possible to build an NARC with an air-cooled intercooler (NARCa). These systems will not require any environmental water for a heat dump.
Since the same problem with the exit temperature from the Recuperator not exceeding the system nominal TIT occurs for NARC systems, it is better to set the burn temperature for hydrogen augmentation based on this limit. Figure 18 plots the peak burn temperature for this condition. Note that the peak burn temperature in all the cases does not significantly exceed the 927 deg temperature for uncooled turbine blades.
Similar to the case for an NACC system, the hydrogen is burned at a very high efficiency. This data is presented in Fig. 19. In the case of recuperated systems, this efficiency is always over 70%. Since instead of producing the hydrogen on site and storing it, natural gas could be burned with a slightly different turbine strategy and achieve a similar efficiency. This means that the most efficient use of natural gas would be to augment an NARC system . Of course this means a larger carbon footprint, but not as large as burning it in a gas turbine of any variety.
The power augmentation due to co-firing for NARC systems is not as significant as it is for NACC systems. The power increases are presented in Fig. 20. The augmentation varies from about 16% to 26% from low to high nominal TIT. Since the NARC systems do not have the power enhancement capability of the steam bottoming cycle, the gains due to co-firing are not as significant. However, there is also less concern about efficiencies over the operating range of the steam cycle. The NARCa is a little less efficient and produces a smaller increase in power during co-firing.
Finally, it is worth talking about overall system efficiencies during co-firing and the sizes of these nuclear air-Brayton systems. The overall efficiency for burning both nuclear and hydrogen fuel for the NACC and both NARC systems is presented in Fig. 21. For the two systems requiring water as a heat dump, the overall efficiency starts at about 50% for 500 deg and climbs to almost 60% for a 700 deg system.
Estimates of overall system sizes are presented in Fig. 22. These are for the nominal 50 MW(e) system. The NACC and NARCw systems are very comparable in size to the proposed NuScale system, a small modular light water reactor. NARCa is significantly larger due to the size required for an air to air intercooler. However, it is worth pointing out that this system requires no cooling water source and can be built or installed anywhere in the world.
Nuclear power plants can serve the role of energy storage systems for the national electric grids by producing hydrogen when there is an excess of generation due to strong renewable sources and burning the hydrogen to augment their rated capability when renewables are not available. The analysis has been performed for a 50 MW small modular reactor, using sodium/molten salt as a primary coolant. Scale up for a power plant can be accomplished by adding more SMRS or expanding this to a large reactor. Significant power increases can be obtained for nuclear air-Brayton systems by co-firing with hydrogen before the power turbine. Power increases in the range of 300–400% can be achieved for NACC that systems normally operate at 500 deg and increases in the range of 100–200% for systems normally operating at 700 deg. The most effective strategy appears to be increasing the steam flow in the bottoming cycle when co-firing. This works for both recuperated and unrecuperated systems. Nuclear air-Brayton recuperated (only) systems can achieve higher normal operation efficiencies and comparable overall efficiencies to NACC systems. The power enhancement provided by NARC systems is significantly less, in the range of 20%, but easier to implement. The NARCa system is significantly larger and has a slightly lower efficiency, but it requires no cooling water to operate. The air-Brayton power conversion systems provide a very flexible storage capability for a low carbon grid.
Idaho National Laboratory (Grant No. DE-AC07-05ID14517).
- p =
- T =
- a =
- w =
- GTCC =
gas turbine combined cycle
- HRSG =
heat recovery steam generator
- NACC =
nuclear air-Brayton combined cycle
- NARC =
nuclear air-Brayton recuperated cycle
- RIC =
recuperator and intercooler
- SCO2 =
supercritical carbon dioxide
- TIT =
turbine inlet temperature