HyRef Helps Put Renewable Energy on the Grid

In order for renewable sources to generate a significant part of our energy base-load, they need to be predictable. At any given location, it’s difficult to know how much sunshine to expect or how much wind will be blowing on a particular day, and without that information we can’t predict how much energy will be produced by solar or wind farms. It’s been estimated that up to 40% of the renewable energy that’s generated goes unused because it’s produced at times when it’s not needed. It’s impossible for base-load coal and nuclear power plants to simply produce less energy on demand. Those plants need hours or days to change their output levels, so they can’t just scale back when a wind or solar farm has an especially productive day. Grid-level storage partially addresses this problem, but that doesn’t make renewables a part of the generating base-load. To remedy this, IBM has developed HyRef, the Hybrid Renewable Energy Forecasting system.

HyRef uses a cadre of sensors, cameras, weather models, and satellite images to track cloud movement and wind patterns. With that data, HyRef is able to predict the local weather up to a month in advance, in 15-minute intervals. (I wonder if an IBM executive launched this project so he could plan his golf schedule, and an engineer decided to apply it to renewable energy. Probably not, but you never know!)

After producing a local forecast with all that data, HyRef then predicts the amount of energy that can be generated at a particular site. That information will help the utilities route energy over the grid and adjust the production from coal and nuclear plants. Remember, it can take a long time for those plants to change their output levels. Knowing several days in advance that a wind or solar farm will produce a lot of power gives utility operators enough time to make adjustments.

Power that’s placed on the grid isn’t always used locally. The grid is sort of an “energy cloud” where, under ideal conditions, the power taken out of the grid exactly matches the power put into the grid. Renewable sources are intermittent on a local level (the wind isn’t always blowing here), but constant on a global level (there’s always wind blowing somewhere.) With forecasting tools like HyRef, utilities will know how much energy renewables can collectively put on the grid, and that knowledge will help make renewables a base-load source rather than just “extra” power that may or may not be needed at that moment.

IBM is currently testing HyRef with China’s Zhangbei 670 MegaWatt demonstration project, the world’s largest renewable energy site, which includes solar, wind, storage, and energy  transmission. IBM thinks that HyRef will initially increase the integration of renewables on the grid by 10%. That’s not a bad starting point, given today’s grid, but with advances in smart grid technology, I expect to see that number increase significantly.