Comparison of Energy Storage Methods for Solar Electric Production

Abstract

Energy storage is key to expanding the capacity factor for electric power from solar energy. To accommodate variable weather patterns and electric demand, storage may be needed not just for diurnal cycles, but for variations as long as seasonal. Five solar electric systems with energy storage were simulated and compared, including an ammonia thermochemical energy storage cycle, compressed air energy storage (CAES), pumped hydroelectric energy storage (PHES), vanadium flow battery, and thermal energy storage (TES). To isolate the influence of the storage system, all systems used the same parabolic concentrator and Stirling engine. For CAES, PHES and battery, the engine directly produced electricity, which was then converted and stored. For TES, heat transfer fluid was heated by the dish and stored, and later used to drive the engine to produce electricity. For ammonia, the dish heated an ammonia dissociation reactor to produce nitrogen and hydrogen, which was stored. Heat was recovered to drive the engine by reforming ammonia from the stored gases. Each system was simulated in TRNSYS with weather data for Louisville, KY and Phoenix, AZ with subsystem efficiencies and storage losses estimated from previous experimental results. All systems including the ammonia cycle involved time dependent storage losses. Losses from the receiver included convection and emitted radiation, both of which depend on receiver temperature. Overall (solar-storage-electric) efficiency of the ammonia cycle depended strongly on synthesis reactor temperature, ranging from less than 1% to ∼18% for both Louisville, KY and Phoenix, AZ, at 500 K to 800 K, respectively. In contrast, the effect of dissociation reactor temperature was less. Overall (solar-electric-storage-electric) efficiencies of the CAES, systems in the limit of zero storage time ranged from ∼10% to ∼18% for solar receiver temperature of 500 K to 800 K. The vanadium flow battery and PHES efficiencies ranged from ∼9% to ∼17% for the same conditions. TES initially provided 12 to 23% efficiency over the same range of temperature. When time-dependent storage losses were included, however, efficiencies for all systems declined rapidly except the ammonia cycle in both locations and PHES in Louisville. The ammonia system had the highest efficiency after one month of storage, an advantage that increased with time of storage. The simulations showed that TES was most efficient for diurnal-scale storage and the ammonia cycle for longer storage. Full capacity factor for solar electric production may be most efficiently accomplished with a combination of direct solar-electric production and systems with both diurnal and long-term storage, the proportions of which depending on weather conditions and electric demand profiles.