Tom Konrad

Thermal energy storage (TES) will allow CSP to provide not only clean, renewable electricity but also dispatchable power to ease the integration of variable generation methods such as wind and PV

Heat Transfer Fluids

CSP companies are pursuing several different paths towards integrating TES. The ability and efficiency of a technology to accommodate thermal storage is a function of the heat transfer fluid and working temperature. According to Greg Kolb of the US National Solar Thermal Test Facility (www.sandia.gov/Renewable_Energy/solarthermal/nsttf.html), the ideal solution is direct storage, whereby one fluid serves for both heat transfer and thermal storage. Indirect storage is comparatively complex and costly.

Three heat transfer fluids have been demonstrated to date: steam (in power towers and troughs), mineral oil (in most parabolic trough plants), and molten nitrate salts (in power towers.) The working temperature for steam is limited by the potential for corrosion. Molten salts and oil break down at high temperatures, with molten salt and steam capable of achieving the highest temperatures (about 565° C for nitrate salts.) ⁱ

Steam – not so hot

Steam is the least suitable for TES. Babul Patel of Nextant (www.nextant.com), speaking at the second annual CSP Summit US, said steam was only suitable for short-term buffer storage, used primarily for addressing cloud transients on the solar field.

Abengoa (www.abengoa.com) has been operating a commercial power tower, the PS10, using low-temperature (257°C) steam since 2007. According to Kolb, Abengoa probably chose steam to avoid the technical risks of molten salt while they familiarised themselves with other aspects of tower technology, such as the heliostat field. The Abengoa website says the PS10’s “thermal storage… allows full production for 30 minutes, even after the sun goes down.”

This may be an exaggeration. Kolb says PS10 operators do not bother to charge the steam buffer unless they see a cloud approaching: the lower temperature of steam from the buffer tank leads to low (approximately 70%) round-trip efficiency. When it is in use, the turbine will not be operating at full production.

Lower temperature steam is also the working fluid for Ausra (www.ausra.com,) a company working to commercialise the Compact Linear Fresnel Reflector (CLFR) geometry. CLFR breaks up a trough into a series of narrow, nearly flat reflectors, saving on the high cost of carefully focused troughs. Ausra has talked about developing phase-change steam thermal storage in the past, an approach Kolb calls “very tricky.”

Oil’s sticking point

Mineral oil is commonly used as the heat transfer fluid in parabolic trough systems because it does not freeze at night (nitrate salts freeze at 221°C) and operates at lower pressure than steam. According to Bill Gould, chief technical officer of Solar Reserve (www.solar-reserve.com), such systems have peak operating temperatures of 375°C.

Unfortunately, mineral oil is very expensive for thermal storage, costing approximately three times as much as nitrate salt. For this reason, Acciona (www.acciona-na.com) says its Nevada Solar One parabolic trough plant does not incorporate storage beyond “30 minutes used to minimise the effects of transients”.

Solar Millennium’s (www.solarmillenium.com) Andesol 1 parabolic trough plant in Spain does include eight hours of indirect molten salt storage. Solar Millennium’s president, Rainer Aringhoff, says this was driven more by Spain’s regulatory 50MW capacity limit rather than the underlying economics of a trough plant.

Economics alone would not justify adding so much indirect storage. Kolb says the average cost of energy does not change when indirect storage is added to a trough plant, although it drops with direct storage added to a tower. According to Gould, the Andesol storage system has 93% round-trip energy efficiency. This is confirmed by Greg Glatzmaier, a senior engineer on the National Renewable Energy Laboratory’s (NREL) CSP research team (www.nrel.gov/csp), who estimates the round-trip efficiency of such systems at “around 90%”, and Kolb, who says such systems can be designed for practically any round-trip efficiency, with higher efficiency leading to higher complexity and expense.

A pinch of molten salt

Solar Reserve is working to commercialise the nitrate salt/power tower combination which was demonstrated at DOE’s Solar Two in the late 1990s (for which Gould was project manager), to capture the more favourable economics of higher temperatures and direct storage.

Gould calculates that a trough plant will require three times as much molten salt (along with larger tanks to store it) as a power tower to store an equivalent amount of energy. To store the equivalent of 1kWh of electricity at a trough plant requires approximately $90 of capital cost, compared with about $30 at a tower. These numbers are confirmed by Glatzmaier and Kolb, who put the relative cost of indirect and direct storage between 2.5 and three to one.

Dr Arnold Leitner, president of Skyfuel Inc. (www.skyfuel.com) estimates it will cost between $500 million and $700 million to commercialise power towers. Solar Reserve plans to overcome this barrier with a performance gaurantee from United Technologies (www.utc.com) up to the value of the contract, or $200 million, but in the current financial climate financing remains difficult.

Leitner hopes to make the transition incrementally by evolving parabolic troughs to accommodate molten salt and direct storage. His idea is to use a hybrid parabolic trough/CLFR configuration called a linear power tower (LPT), enabled by the reflective film ReflecTech (www.reflectechsolar.com). By increasing the diameter of the receiver and focusing sunlight 85:1, he hopes to allow the salt to stay molten in the solar field, enabling the use of direct molten salt storage with a variant of parabolic troughs.

FACTS: Nitrate Salt

Sources: Fosberg, Peterson and Zhao¹; Interviews with Greg Glatzmaier, NREL and Bill Gould, Solar Reserve.