Making solar power more predictable

uk.ibtimes.com
September 29, 2010

One thing that complicates solar power projects is the fact that the power supply is not constant: clouds can affect the output of a large-scale plant. If there were a way to predict how much they would affect the performance, then it would be easier to convince utilities to build them. To address that problem, researchers from Sandia National Laboratories decided to measure the effect. They decided on a solar photovoltaic array in Hawaii, on the island of Lana'i, which provides 1.2MWs of power, largely to a local resort and the island's 3,000 residents.

Joshua Stein, one of the lead researchers on the project, said previous studies have looked at small-scale solar panels, but none have directly measured how much power the panels produce under varying amounts of cloud cover, while showing how that changes over time. Joshua Stein, one of the lead researchers on the project, said the solar plant provides a relatively large amount of the island's power (up to 30% of peak demand) and that conventional power sources are much more expensive than on the mainland.

That makes it much more akin to the situation larger facilities in the U.S, are likely to face in the future, and it gives a better idea of what a 'real world' environment would be like if more communities in the U.S, use solar power on a large scale. "There are some concerns from utilities as to how solar facilities will interact with existing generation", Stein said.

Stein's team put sensors on the solar array that measure how much sunlight the panels get. They then measured the power output at one-second intervals. Because the array is large - several hundred square yards - the sensors could show which parts of the array were under cloud cover and which were not, as well as showing cloud direction and speed.

The data has helped answer the question of just how much effect scattered clouds -- common in Hawaii - will have on a solar power array. Stein said thus far, the effect seems to be linear. While individual power cells can lose a lot of power when they are in shadow, that gets "averaged out" over the entire array.

Stein also said that since the price of solar power decreases as you increase the size of the arrays, the power loss from scattered areas of cloud cover should create smaller fluctuations in overall output. That seems to be the case in Lana'i. "The ultimate goal is to understand irradiance patterns and how they will affect plants larger than any yet built", he said. With the new data, the team will eventually be able to build models that will help when designing larger power plants, making solar power a bit more predictable - and palatable - to investors and utilities.