Skip to main content

How IBM uses local weather forecasts to boost clean energy

<p>Wind farm operators create more accurate predictions about clean power generation using IBM&#39;s HyRef technology.</p>

Predicting weather may be an inexact science, but it can help make wind energy more powerful. IBM researchers are testing an approach that uses sensors, advanced imaging technologies and sophisticated analytics to make forecasts of wind conditions far more accurate, ahead of time.

The technology, called Hybrid Renewable Energy Forecasting, or HyRef, is targeted initially at wind farms, where it can be used to monitor wind speed, temperature and direction. In theory, it also could be relevant for solar generating projects. HyRef has its roots in the IBM Deep Thunder research project, begun in 1996 to help provide "hyper-local, short-term forecasting and customized weather modeling for clients."

By coupling the sensor-collected data with information about movements of clouds in the sky, operators can better manage the performance of individual wind turbines, and closely estimate how much renewable energy will be generated in the coming hours. This will allow utilities to more accurately assess how much electricity must be stored or redirected to the grid. All this information is critical as communities seek to integrate more intermittent generating sources into the power infrastructure.

"The heart of the problem is around predictability," said Lloyd Treinish, IBM distinguished engineer and chief scientist of Deep Thunder, which is leading tests of HyRef. "Accuracy within power forecasts is required for these intermittent generators to be usable on the grid."

Usually, sensors are focused mainly on telling operators what already has happened or is happening, but HyRef is meant to heighten the value of that information by offering educated insight into future conditions at intervals that approach real time. At the heart of the solution is powerful supercomputing technology. Deep Thunder projects are also focused on things such as transportation logistics at airports, which are also highly dependent on real-time meteorological data.

The platform is already in use as part of phase one of the massive, 670-megawatt Zhangbei demonstration project in China, which seeks to combine wind and solar power, energy storage and transmission. State Grid Jibei Electricity Power Co., a subsidiary of the State Grid Corp. of China, is using HyRef to support a 10 percent increase in the amount of clean, renewable energy being added to the grid. The idea is to use as much wind and solar electricity as possible, without having to curtail production.

The video below offers more information about the Zhangbei project.

Dennis McGinn, president and CEO of the American Council on Renewable Energy (ACORE), said technology such as HyRef will be instrumental in helping utilities reach a worldwide goal of generating about 25 percent of their power needs from renewable or close sources by 2025.

"The weather modeling and forecasting data generated from HyRef will significantly improve this process and, in turn, put us one step closer to maximizing the full potential of renewable sources," McGinn said in a statement.

This is just the latest example of how the Internet of Things, coupled with analytics and cloud services, will play a role in helping the electrical grid become smarter. "We are in discussion with other clients about deploying this to address similar problems," Treinish said. IBM, for one, is involved with more than 150 smart grid projects around the world, although they won't all necessarily use HyRef.

Other software applications, such as the visualization solutions developed by Space-Time Insight, are being employed by utilities around the world to guide more informed decisions and to predict potential problems before they occur.

Image of wind turbines by majeczka via Shutterstock

More on this topic