On Thursday, Sunpower, Optisolar and PG&E agreed to a 800MW deal, with undisclosed financial terms. Its a big leap for Photovoltaic solar in terms of scale - previously, utility sized systems (100MW+) were the domain of solar thermal power and companies like Ausra, but with this project, PV systems scale up to utility scale. The question is how much does this really cost...
First, let me note that there are two separate components to this project. One is a 250MW agreement with SunPower, and the other agreement is with OptiSolar and is 550MW. The area for the project is in the San Luis Obispo County, California. The most important thing to note is that both projects are contingent on the extension of the renewable energy tax credits that Congress will take up when they resume in September, under the compromise energy bill that includes both renewable energy incentives as well as selected offshore drilling.
The SunPower agreement is for 250MW in San Luis Obispo County, California. This is located about half way between LA and San Francisco. The project is slated for California Valley, which is a dry, arid, and is agriculturally undevelopable due to high alkali levels in the soil. This diminishes the value of the land greatly. SunPower has their patented Tracker systems that allow for approximately 1MW for every 6.2 acres. Thus, the 250MW system would require about 2.5 square miles, or 1600 acres.
The Optisolar agreement is for 550MW also in San Luis Obispo County, California, near the California Valley as well. This project will use about 9 and a half square miles, or roughly 11 acres per MW of generation. This is due to the cheaper, but far less efficient thin film solar panels. Also, the panels are fixed-mounted and do not track the sun like the SunPower project, which will result in a lower output per installed DC nameplate capacity.
If we analyze the cost from a cents/kWh perspective (or $/MWh), the SunPower system will generate 550,000 MWh/year. At the standard solar installation cost of $7/W, plus maintenance (0.5% of capital costs), plus $10,000/acre buy or equivalent lease costs, the cost for the power is roughly $110-$133/MWh over 30 years. Which isn’t out of line - summertime peak power in California can cost between $125-$175/MWh, winter daytime power is roughly $60, and spring/fall daytime power is around $80. Now granted more of the power will be generated in the summer (specifically May, June, July) than in the winter (November-March) - the five months of April - September will generate about the same as the seven months of October - March - so a simple average doesn’t fairly asses average sold power. It turns out its roughly a 65/35 split from a revenue standpoint - 65% of the revenue from the 5 months of high demand and 35% of the revenue from the low price months. If we apply this to the numbers above we’ll find that the plant will have an ASP (average selling price) of $120/MWh. This figure does not include and tax credits that are available or will be available with the renewable tax credit passed (hopefully) by Congress.
Optisolar has a similar business case, while their panels are cheaper due to the use of thin-film technologies, the panels have lower efficiency ratings (% of light turned into energy). So while PV may cost $4/W, thin-film panels have a stated target of $1/W. If we go with the $1/W figure, plus $4/W installation (due to the fact that there are more panels per MW installed vs standard PV), and all the other same constraints from above, we see that 1.1M MWh will cost between $105-126/MWh.
In summary, unless the price of energy decreases dramatically (HAH!) or the earth stops rotating and Ben Stiller isn’t around to save us, the plants are providing affordable power at a small premium now, and probably cheaper further down the road.
Would solar thermal been cheaper? Probably - in the neighborhood of $4/W installed. Why they didn’t choose that, I don’t know. Perhaps the sun wasn’t intense enough - California Valley is not within the Mojave Desert proper, and that would have negatively impacted capacity, increasing $/W costs. Or maybe SunPower and OptiSolar managed to bring their costs down enough to make it worthwhile.
The best hope is that thin-film efficiencies improve. Panels getting 150W (10-12% efficient) instead of 75W could really turn around the solar industry, making rooftop installations worthwhile. While they might not always be generating power - cloudy days - the south and southwestern US is suited for solar power. Every little bit helps - if we can reduce grid demand and take the dirtiest, most inefficient peaking plants out of operation, the impact is disproportional to the amount of effort put into clean energy.
Edit: I reran the numbers at 7% (per comments below and some info I found from the CEO of Ausra), and the price went up to about $180-$220/MWh. This figure also includes CA taxes and the federal ITC.