Predicting the power of wind

Lei Yang is using information from individual wind turbines to create wind power data models

Lei Yang's team is modifying the Markov Chain in order to capture all of the variants in wind power output.


2/2/2018 | By: Nicholas Ruggieri |

While trying to predict the outcome of future events may not seem like something new to Reno, Nevada, Lei Yang, assistant professor in the computer science and engineering department of the University of Nevada, Reno, is working on creating models to better predict wind and wind energy output in order to make the renewable resource a more reliable source of energy.

"In the power system, most of the resources we use are controlled, so we can decide at what time we need a certain amount of power in order to keep the balance between consumption and demand," Yang said. "With wind, it's controlled by nature, so we can't just decide when we need more, and predicting it isn't an easy task, because it depends on so many factors, from the location of the wind farm and its turbines, to just your local weather conditions."

In order to better our ability to foresee and utilize wind activity, Yang is looking beyond the actual physicality of wind and wind turbines, and turning towards data analytics.

"The data driven model is our best method, because the underlying physical process is too difficult to characterize," Yang said. "At each individual turbine we can measure the energy output, wind speed, direction, temperature, and humidity. Basically we have a series of data measurements which we use as a base to try and build wind models."

The statistical models will ultimately provide more accurate estimates of wind power, as well as better prepare wind farms to deal with wind ramping, which occurs when a sudden increase or decrease in wind speed interrupts a steady force.

"We deal with ramping effects the same way we are working with general wind and power prediction, by developing data models," Yang said. "If you observe the starting point of the increase or decrease in speed, and you know what slope it has, we can predict where the ramping will go and create models to predict the increase or decrease in power production."

Yang's National Science Foundation funding runs through May 2018, and he hopes his work can lead to more consistent use of renewable energy in the power system.

"The idea is to try and help the generators run at a lower cost and run more effectively, and our data models can make that possible with the use of wind energy," Yang said.

Yang's project was featured as a top story on NSF Science 360 News website, and you can read more about Yang's work at Nexus In Nevada.


For more news on the University of Nevada, Reno, follow @unevadareno on Twitter.

Get Nevada Today in your Inbox!