Renewable Energy Solar Energy Wind Energy Geothermal Energy Bioenergy Hydropower

The Cost of Utility-scale Solar: PV vs. CST

By Mike Taylor, Solar Electric Power Association
May 13, 2008 | 14 Comments

Do you like this opinion & commentary?

Bookmark

Feed

On a utility scale, which of these two technologies is cheaper: photovoltaics (PV) or concentrating solar thermal (CST)? -- Lisa B, Denver, Colorado

The short answer would be that on an apples-to-apples comparison in the southwestern United States, concentrating solar thermal parabolic troughs (CST) have a lower levelized cost of energy (LCOE) per kilowatt-hour (kWh) than photovoltaics (PV) today. Outside of the broad southwest (which includes 7-8 states), PV is really the only game in town (at least for now, a CST project in Florida not-withstanding).

Looking forward, PV probably has a faster cost reduction curve than CST and their LCOE will cross paths within the next five years, i.e. PV will become cheaper. Concentrating photovoltaics (CPV) is the wild card that could also reach parity with CST over the next ten years.

The next obvious question is that given cost parity between the three technologies, which one is preferred by utilities?

The cost of energy is not the only criteria for utilities seeking renewable energy. The ability to generate during periods of high demand, including into the evening, makes any renewable technology much more valuable. Additionally, minute-to-minute operating characteristics on the grid, such as the potential for rapid changes in output, are also important, especially at the utility scale.

CST projects can offer both better peak capacity characteristics, with 6-8 hour thermal storage, as well as a smoother short-term fluctuations. Intermittent clouds going over a PV system will cause output to spike widely. But with a CST plant, which has thermal inertia in the heat transfer fluids, an operator can have some warning of the clouds, slow the fluid flow rate, which increases the fluid temperatures, and ride through short-term cloud events. (For reference, a wind turbine probably falls somewhere between the two technologies-there is kinetic inertia in the wind and the turbine blades, which provides a somewhat better short-term profile than PV.)

Of course, utility scale storage technologies for PV are being developed, and in some cases demonstrated. Initially these will provide similar ride-through capabilities during short-term cloud events, but as the storage technologies scale-up, they could potentially offer hourly storage into the evening. But right now in the southwest, CST has the cost and storage advantage, but PV essentially has a geographic monopoly everywhere else.

Solar Energy

The information and views expressed in this article are those of the author and not necessarily those of RenewableEnergyWorld.com or the companies that advertise on its Web site and other publications.

14 Reader Comments

Comment
1 of 14

paul-passarelli-76582

May 13, 2008

We've actually designed a DG scale CSP panel and a matched engine that should deliver energy in the $90/MWh range, for systems from 50kW to the low MW peak capacity.

The reasons are bases on at least 5 major improvements: less expensive reflectors & collectors, single fluid loop, direct in tube boiling, optimized installation rates, and an ORC engine that captures the latent heat of vaporization of the working fluid.

http://www.solarandthermal.com/ for more information.

Comment
2 of 14

jeff-kelly-146257

May 13, 2008

There are at least a couple of large projects in Calif now signed by SCE, PGE and others that plan to use solar thermal. They might be troughs or towers systems, steam or molten salt storage. The ability to ride through low sun periods and generate into the night is not a small detail, it is enormously important to the utilities. A project that doesn't have this ability to some extent is almost useless to the utilities. Large PV will have to have huge banks of batteries to be useful to the utlities. Also, solar thermal uses well understood generators for AC current where large scale PV will have to use large scale electronic inverters.

You are correct about the issue of water consumption of cooling towers for the thermal cycle; that is probably the biggst rap on solar thermal. With good design and management, it is possible to significantly reduce the water consumption below what is the large amount that is wasted now. This is a state of the art in industry that needs to be pushed. With proper design and investment, the huge amount of water that is wasted in "blow down" of cooling towers can be partically recycled and reused. It's more money, so the economic incentive has to be there for it.

Comment
3 of 14

michael-grunow-44746

May 13, 2008

I agree on the storage potential for solar thermal but none of the US projects plan to use it, right? My best math suggests that Nevada Solar One (solar thermal, built last year, no storage) had an unsubsidized COE of 22-24 cents per kwh and the large utility scale solar projects currently being developed (in Germany using Thin Films at ~ $4.25 to $5.00 per watt) end up with an unsubsidized COE range from 15-18 cents if sited in the US Southwest. Although the recent solar thermal PPAs have been commited by utilities at PPA rates in the ~15 cent range, I am not sure if these projects will break ground without the 30% ITC. Also, the "no moving parts" is a strong argument in favor if thin films in a remote setting. Conversely, solar thermal works fantasticly in deserts that have a lot of water (???). If my comments are accurate, then thin film looks best for the US SW once the Europeans stop dishing out the fantastic FITs.

Comment
4 of 14

ronald-wagner-75806

May 14, 2008

Please explain the need for water for cooling. The object is to create heat, and use it in Stirling heat engines, or steam engines , isn't it? It seems simple to keep any needed coolant fluid enclosed in a large enclosed reservoir. The pressure could be enclosed there, until it cools, and the coolant is reused.

Isn't cooling just a waste of the energy that the system is designed to maintain and use.

Comment
5 of 14

eugene-lucas-60935

May 14, 2008

You are all missing the point! Storage is only a problem in baseline electricity. Peak electric use occurs between 10 A.M. and sunset, when solar thermal efficiency is at its peak. Here in California, we have no baseline electricity problem, as a matter of fact we pump water back up into reservoirs at night (with the surplus baseline power), so that it can be released in the daytime to augment peak power needs.

Slightly off topic is the fact that CST used directly and locally to heat and cool has twice the efficiency of CST used to run steam generators. Its not applicable to residences, but it is very cost-effective for businesses and industries.

Comment
6 of 14

jeff-kelly-146257

May 15, 2008

To Ronald Wagner:

The need for water cooler for large solar thermal solar power plants (and of power plants in general) comes from the fact that most power plants use heat to generate steam, which drives a turbine and generator. This is the basic Rankine Cycle, and it requires that the steam be cooled and condensed back to liquid in order to run. This generally done with evaporative cooling towers, which use large amounts of water.

There ARE alternatives to these evaporative cooling towers, which the previous post mentioned

http://www.nrel.gov/csp/troughnet/pdfs/2007/dersch_dry_cooling.pdf

There are variations of evaporative cooling that use far less water than the standard cheap/dumb cooling tower with huge amounts of blowdown wasted just to keep the water clean. From an operations standpoint, however, those hybrid/Heller types of cooling towers are more expensive, less reliable, and maintenance intensive. Dry cooling can be an option, but in hot desert air it is less efficent that using water evaporation to cool. The previous slide suggested this lower efficiency adds 5% to the cost of power generation. Without claiming any particular expertise other than some familiarity with power plant operation, that may be optimistic. Perhaps 10% would be more accurate. However, basic dry coolers are mechanically simple, which means lower maintenance and downtime, which utilities like. They probably will be used in these plants eventually, if for no other reason that the owners can't defend the water consumption of standard cooling towers from a green PR standpoint.

Unfortunate to learn that Stirling is having so much trouble.

So, have any of these big solar projects actually broken ground in California yet? Are they awaiting the PTC resolution in Washington? Is it all just political hype to meet the California RPS on paper? Say it isn't so...

Comment
7 of 14

david-ahlport-66657

May 15, 2008

Waterless solar thermal? Yes.
http://www.nrel.gov/csp/troughnet/pdfs/2007/dersch_dry_cooling.pdf

Especially when we are talking about this type.
http://greyfalcon.net/csp

Comment
8 of 14

hussain-alrobaei-50048

May 15, 2008

Current drastically increasing fuel prices and power shortages for summer daytime peaking power in southern Europe suggest that Concentrated Solar Power (CSP) systems will find their prime market segment in summer season on peaks.Any renewable energy supply strategy aiming to take over the major part of electricity supply in the decades to come has to consider CSP as this technology option is capable of contributing with reliable, dispatchable power, specifically for daytime-demand peaks.In addition to that Integrated Solar Combined Cycle power Plants (ISCCP) with their capability of thermal energy storage and of solar/fossil hybrid operation can provide firm capacity and thus are a key element for grid stabilization and power security in such a well-balanced electricity mix.The desired effect of integrating a Gas Turbine Unit (GTU) with a Solar parabolic trough Power Plant (SPP) is not just to add the power produced by the GTU to that produced by the SPP but indeed to augment the latter.Beside that, the increase in the output of solar collector arrays and, subsequently, in solar power generation, will also useful to offset the normal reduction in performance experienced by GTU during the summer season [1].
It is also important to note that, according to the World Bank, the expected evolution of total electricity costs is that they will drop to 8 to 7 € cents/kW.h in the medium term (100 MWe Rankine-cycle plant or 100 Mwe ISCCP, both with storage) and to 5 € cents/kW.h in the long term (200 MWe Rankine-cycle plant with storage) for high insolation sites with an annual direct normal solar radiation of more than 2800 kW.h/m2.
[1] Hussain Alrobaei,2006, Integrated Gas Turbine Solar Power Plant/ The
Energy Central Network/ nergycentral.com/centers/knowledge/whitepapers.

Comment
9 of 14

jay-rosenberg-53116

May 16, 2008

My company is intent on commercializing $.05/kWh Concentrated Solar Thermal, Its doable. An analysis of CST should include CHP (Combined Heat/Power, add light, A/C etc). Take a cold East Coast state, CST delivering BTU's is quite cheap, no conversion to electricity then to heat. The tragedy is those that understand the subtleties, are too busy making money and growing 30%/ year. Both Unlike, and like the PC paradigm shift, renewable energy produces nameless shapless commodities called electricity, BTU's etc. The utilities have a legacy investment momentum, a little like legacy software keeping afloat the sailing ship computers with clever solutions like 1's complement arithmetic having 2 binary representations for zeroes! Osborne computers went belly towards the north star when they announced their next, better version of their pc (portable no less).was also cheaper. Thus, no sales that quarter before the announcement! Faster, better, cheaper means cannibalizing existing product lines. A tricky technical managerial move, which so far only done successfully by Intel.
The existing high rolling renewable firms understand this, and are not interested in more than 1-2 quarters ahead. And, BTW, its still rather hard technically, once you have geothermal to estimate how long the rocks will stay hot. Some regions its easier. However, the real power of Geothermal is not the flashy hot spots, but what is now called Heat Mining. Per recent study, under the US, and not going too deep, is an energy equivalent of 30,000 times our electrical production,
Jay R- CEO, Sannerprojects, JRIAM1945@aol.com. Sun's down, time for tools down.

Comment
10 of 14

john-broughton-40988

May 17, 2008

Some utility scale CSP's have been completed and are operational. other CSP's are contracted and in the works for the near future. Even still though this technology requires utility scale transmission and lots of its, due to the locations of maximum sun.

On the other hand PV's are readily available and can be installed across the board at multiple locations, they may not meet what utility comapnies used to view as feasible projects to own, however they are begiining to factor into their mix.

So in theory CSP's have a cheaper levelized cost than PV if we exclude transmission and availability. But in reality PV can be built now at sites with easy access to transmission and therefore more feasible short term.

Comment
11 of 14

bonkim2003

May 28, 2008

Re- discussion on solar thermal - the condensing part of the thermal cycle in desert region can be done using air cooled condensers which does not involve loss of water - however involves air blast fans and of course energy.

Where some water is available, cooling towers can be used which of course will incur some evaporative loss of the recirculating water.

It is all a question of optimising the different methods at a given location and climatic conditions.

Comment
12 of 14

paul-passarelli-76582

June 21, 2008

On June 20 someone posted to my website anonymously:
---
Nice Shamless[sic] self-promotion on the Renewable Energy website.
Please keep your comments on-topic and non self-centered.
---

The topic is the cost and scale; I posted a cost figure & scale. I've got nothing to be ashamed of. I'm only guessing that this was the post referred to.

The author presented a conclusion that PV is or will be superior. I cannot agree with that, and had to dispute it. The $/W & $/kWh numbers for PV are inferior to CSP by ~3:1. All this demonstrated is that they have more room to improve (when you're at the bottom, up is the only direction)

The geographic monopoly, is political, not technical. Since most legislators do not understand CSP the word "solar" and "PV" are synonymous. Ignorance is easily curable; stupidity can only be mitigated with great effort.

No one can argue that PV is sexy when compared to more traditional CSP, those shiny purple crystals and whatnot. But the appeal fades when examined more closely. Utilities care about one thing: *COST* writ large.

If a CSP plan costs $4/W installed and a pile of PV panels of equal rating costs $5/W the PV proponents will neglect to mention the cost of BOS or installation when comparing the two systems. They will also discuss the incentives available to the buyers of PV technology, making it appear more cost effective.

We cannot fix the problem of the rising cost of energy on a subsidized industry. CSP can do today what PV has been looking to the 20-year horizon, for 4 decades. In the meantime CSP plants have been operating in the desert SW for nearly 30 years.

http://www.solarandthermal.com/ for more information.

The Light is Green!

Comment
13 of 14

cyrus-tashakkori-171973

February 27, 2009

Great discussion. I would argue with a few of the points that have been made for and against CSP. On the one hand, the storage value to the utilities is significant, and not to meet off-peak load. In markets like Las Vegas and Phoenix, the peak load extends into the evening, meaning that solar without storage would have to be augmented by conventional peaker plants, which defeats part of the purpose for building renewable peaker plants. PV can't meet the load between 6 and 8 PM in July in Phoneix!

On the other hand, there is no reason to think that the CSP PPA prices will be met. The technology risk with CSP is considerably higher, with many technologies with large PPAs never having been demonstrated on large scale. While these numbers are used to argue that CSP is "cheaper" than PV, that has yet to be demonstrated. Meanwhile, PV costs (notice I didn't say PV prices) continues to decline over time.

In the end, I think that developments in both technologies will continue to drive costs downward, not just in the technologies themselves, but also in their installation, construction, operations, etc. In the current market glut in silicon, I wouldn't be surprised if PV plants at record-low PPA prices (below $150) were possible.

Comment
14 of 14

MarkGabriel

December 17, 2010

There appears to be some confusion when it comes to utility peak which is THE driver of costs on utility systems. System peaks occur between 5-8 p.m. all across the U.S. so while the energy component of solar thermal is useful, it too lacks the capacity that make it a complete replacement for fossil fuel or nuclear plants. This makes storage the critical factor in order to be able to shift the energy to the utility/system peak.

Add Your Comment

Registered users, please make sure to Sign-In. We and others want to know your ideas and opinions. If you are not yet Registered -- it's quick and easy. Just click below.
Thanks!

NEW Smart Phone App

Mike Taylor

About: Mike Taylor is the Director of Research for the Solar Electric Power Association (SEPA) and leads SEPA’s research publications and materials for over 1000 elect... more »