UCSD professor's cloud predicting technology could solve solar power dilemmas

Carlos Coimbra's work includes computer algorithms

SAN DIEGO - A UC San Diego professor believes he has cracked the code of the clouds and it could pay big dividends when it comes to the energy that powers your home.

After six years of research and scouring the skies, engineering professor Carlos Coimbra believes he has found order in the elusive nature of clouds.

It comes in the form of computer algorithms and years of cloud data. His algorithms aim to predict how the clouds move across the sky, change shape and most importantly, if they block the sun.

Right now, weather models can predict percentage of cloud cover by the hour. But in the world of solar energy, the information is not that useful.

"The cloud cover could be anywhere in the sky or could block or not block the sun," said Coimbra.

In contrast, Coimbra's modeling – which tracks the clouds that mask the sun – gives a forecast of the amount of solar energy available and when it will be available.

Coimbra says the modeling is much more accurate the similar attempts.        

"We can be 20 to 40 percent more accurate in most time horizons," said Coimbra.

The accuracy could mean big savings for solar power plants and utilities, which can better time supply to meet demand and cut back solar resources when the sun is not shining.

"Having the ability to know those spikes and drops will help you manage the power grid," said Coimbra.

For utilities, the forecasting can also help predict when customers might be using the air conditioners.    

Eventually, the technology may also help airports predict delays and fire crews factor in cloud cover during a firefight.

The state's independent system operator, which manages the power grid, is about to compare the UCSD forecasts against its own to see if it should adopt the methods and possibly save ratepayers.

Print this article Back to Top