Rock bed thermal storage for concentrating solar power plants

Allen, Kenneth Guy (2014-04)

Thesis (PhD)--Stellenbosch University, 2014.

Thesis

ENGLISH ABSTRACT: Concentrating solar power plants are a promising means of generating electricity. However, they are dependent on the sun as a source of energy, and require thermal storage to supply power on demand. At present thermal storage – usually molten salt – although functional, is expensive, and a cheaper solution is desired. It is proposed that sensible heat storage in a packed bed of rock, with air as heat transfer medium, is suitable at temperatures of 500 – 600 °C. To determine if this concept is technically feasible and economically competitive with existing storage, rock properties, packed bed pressure drop and thermal characteristics must be understood. This work addresses these topics. No previously published data is available on thermal cycling resistance of South African rock, and there is limited data from other countries in the proposed temperature range for long-term thermal cycling, so samples were thermally cycled. There is rock which is suitable for thermal storage applications at temperatures of 500 – 600 °C. New maps of South Africa showing where potentially suitable rock is available were produced. Dolerite, found extensively in the Karoo, is particularly suitable. Friction factors were measured for beds of different particles to determine the importance of roughness, shape, and packing arrangement. Five sets of rock were also tested, giving a combined dataset broader than published in any previous study. Limitations of existing correlations are shown. The friction factor is highly dependent on particle shape and, in the case of asymmetric particles, packing method. The friction factor varied by up to 70 % for crushed rock depending on the direction in which it was poured into the test section, probably caused by the orientation of the asymmetric rock relative to the air flow direction. This has not been reported before for rock beds. New isothermal correlations using the volume equivalent particle diameter are given: they are within 15 % of the measurements. This work will allow a techno-economic evaluation of crushed rock beds using more accurate predictions of pumping power than could previously be made. Thermal tests below 80 °C show that bed heat transfer is insensitive to particle shape or type. A heat transfer correlation for air in terms of the volume equivalent diameter was formulated and combined with the E-NTU method. The predicted bed outlet temperatures are within 5 °C of the measurements for tests at 530 °C, showing that the influence of thermal conduction and radiation can be reasonably negligible for a single charge/discharge cycle at mass fluxes around 0.2 kg/m2s. A novel method for finding the optimum particle size and bed length is given: The Biot number is fixed, and the net income (income less bed cost) from a steam cycle supplied by heat from the bed is calculated. A simplified calculation using the method shows that the optimum particle size is approximately 20 mm for bed lengths of 6 – 7 m. Depending on the containment design and cost, the capital cost could be an order of magnitude lower than a nitrate salt system.

AFRIKAANSE OPSOMMING: Gekonsentreerde son-energie kragstasies is n belowende manier om elektrisiteit op te wek, maar hulle is afhanklik van die son as n bron van energie. Om drywing op aanvraag te voorsien moet hulle energie stoor. Tans is termiese stoor – gewoonlik gesmelte sout – hoewel funksioneel, duur, en n goedkoper oplossing word gesoek. Daar word voorgestel dat stoor van voelbare warmte-energie in n gepakte rotsbed met lug as warmteoordrag medium geskik is by temperature van 500 – 600 °C. Om te bepaal of dié konsep tegnies gangbaar en ekonomies mededingend met bestaande stoorstelsels is, moet rotseienskappe, gepakte bed drukval en hitteoordrag verstaan word. Hierdie werk spreek hierdie aspekte aan. Geen voorheen gepubliseerde data is beskikbaar oor die termiese siklus weerstand van Suid-Afrikaanse rots nie, en daar is beperkte data van ander lande in die voorgestelde temperatuurbereik, dus is monsters onderwerp aan termiese siklusse. Daar bestaan rots wat geskik is vir termiese stoor toepassings by temperature van 500 – 600 °C. Nuwe kaarte van Suid-Afrika is opgestel om te wys waar potensieel geskikte rots beskikbaar is. Doleriet, wat wyd in die Karoo voor kom, blyk om veral geskik te wees. Wrywingsfaktore is gemeet vir beddens van verskillende partikels om die belangrikheid van grofheid, vorm en pak-rangskikking te bepaal. Vyf rotsstelle is ook getoets, wat n saamgestelde datastel gee wyer as in enige gepubliseerde studie. Beperkings van bestaande korrelasies word aangetoon. Die wrywingsfaktor is hoogs sensitief vir partikelvorm en, in die geval van asimmetriese partikels, pakkings metode. Die wrywingsfaktor het met tot 70 % gevarieer vir gebreekte rots, afhanklik van die rigting waarin dit in die toetsseksie neergelê is. Dit is waarskynlik veroorsaak deur die oriëntasie van die asimmetriese rots relatief tot die lugvloei rigting, en is nie voorheen vir rotsbeddens gerapporteer nie. Nuwe isotermiese korrelasies wat gebruik maak van die volume-ekwivalente partikel deursnee word gegee: hulle voorspel binne 15 % van die gemete waardes. Hierdie werk sal n tegno-ekonomiese studie van rotsbeddens toelaat wat meer akkurate voorspellings van pompdrywing gebruik as voorheen moontlik was. Termiese toetse onder 80 °C wys dat die warmteoordrag nie baie sensitief is vir partikelvorm en -tipe nie. n Warmte-oordragskorrelasie vir lug in terme van die volume-ekwivalente deursnee is ontwikkel en met die E-NTU-metode gekombineer. Die voorspelde lug uitlaat temperatuur is binne 5 °C van die meting vir toetse by 530 °C. Dit wys dat termiese geleiding en straling redelikerwys buite rekening gelaat kan word vir n enkele laai/ontlaai siklus by massa vloeitempos van omtrent 0.2 kg/m2s. n Oorspronklike metode vir die bepaling van die optimum partikelgrootte en bedlengte word gegee: Die Biot-getal is vas, en die netto inkomste (die inkomste minus die bed omkoste) van n stoomsiklus voorsien met warmte van die bed word bereken. n Vereenvoudigde berekening wat die metode gebruik wys dat die optimum grootte en lengte ongeveer 20 mm en 6-7 m is. Afhangende van die behoueringsontwerp en koste, kan die kapitale koste n orde kleiner wees as dié van n gesmelte nitraatsout stelsel

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/86521
This item appears in the following collections: