Siemens is “teaching” wind turbines how to automatically optimize their operation in line with weather conditions.
The turbines are learning to use sensor data on parameters such as wind speed to make changes to their settings. These changes ensure the turbines can optimally exploit the prevailing conditions. Wind power facilities can’t always generate their maximum electrical output when wind speeds are moderate or low.
Specialists for learning systems at Siemens Corporate Technology (CT) developed the self-optimization software for wind turbines in cooperation with Technische Universität Berlin and IdaLab GmbH in the ALICE project (Autonomous Learning in Complex Environments), which is funded by Germany’s Ministry of Education and Research.
The researchers are presenting the results of their work at the CeBIT trade show (March 10–14) in Hanover. Their solution enables turbines to produce around one percent more electricity annually under moderate wind conditions, while also reducing wear and tear.
The researchers have a demonstration wind turbine unit that uses its own operating data and gradually increases its electrical output. The scientists’ approach combines reinforcement learning techniques with special neural networks.
A neural network is a software algorithm that operates in a way similar to the human brain. For several years now, Siemens CT has been developing neural networks in order to model and predict the behavior of highly complex systems, such as wind farms, gas turbines, factories, or even stock markets.
The software programs learn from historical data, which also enables them to forecast the future behavior of a system. A model can thus be created that predicts the electrical output of a wind turbine under specific weather conditions.
The researchers examined a large amount of very noisy data to identify relevant attributes that would enable the efficiency of a wind turbine to be improved by changing settings such as rotation speed. Patented neural networks were then used to create a so-called reinforcement learning policy from the analysis results.
The system thus learns to change certain wind turbine settings in a manner that ensures the maximum possible amount of electricity is generated in a given situation. After just a few weeks, the system is able to define and store the optimal settings for common weather occurrences.
After an additional extended period of training, it can even regulate electrical output under rare and exceptional weather conditions. The technology was successfully tested at a Spanish wind farm last year.
Ongoing analyses of relevant oper-ating parameters ensure the system can continually improve itself through repetition. The methods used here can be employed in many other fields, which means additional Siemens products can also be taught to optimize their own operation.
Dr. Norbert Aschenbrenner | Siemens InnovationNews
Ion implanted, co-annealed, screen-printed 21% efficient n-PERT solar cells with a bifaciality >97%
04.09.2015 | Institut für Solarenergieforschung GmbH
Fraunhofer ISE Develops Highly Compact Inverter for Uninterruptible Power Supplies
03.09.2015 | Fraunhofer-Institut für Solare Energiesysteme ISE
In a survey of NASA's Hubble Space Telescope images of 2,753 young, blue star clusters in the neighboring Andromeda galaxy (M31), astronomers have found that M31 and our own galaxy have a similar percentage of newborn stars based on mass.
By nailing down what percentage of stars have a particular mass within a cluster, or the Initial Mass Function (IMF), scientists can better interpret the light...
Researchers at the Fraunhofer Institute for Solar Energy Systems ISE have developed a highly compact and efficient inverter for use in uninterruptible power...
China's Loess Plateau was formed by wind alternately depositing dust or removing dust over the last 2.6 million years, according to a new report from University of Arizona geoscientists. The study is the first to explain how the steep-fronted plateau formed.
China's Loess Plateau was formed by wind alternately depositing dust or removing dust over the last 2.6 million years, according to a new report from...
The leaves of the lotus flower, and other natural surfaces that repel water and dirt, have been the model for many types of engineered liquid-repelling surfaces. As slippery as these surfaces are, however, tiny water droplets still stick to them. Now, Penn State researchers have developed nano/micro-textured, highly slippery surfaces able to outperform these naturally inspired coatings, particularly when the water is a vapor or tiny droplets.
Enhancing the mobility of liquid droplets on rough surfaces could improve condensation heat transfer for power-plant heat exchangers, create more efficient...
Longer, more severe, and hotter droughts and a myriad of other threats, including diseases and more extensive and severe wildfires, are threatening to transform some of the world's temperate forests, a new study published in Science has found. Without informed management, some forests could convert to shrublands or grasslands within the coming decades.
"While we have been trying to manage for resilience of 20th century conditions, we realize now that we must prepare for transformations and attempt to ease...
03.09.2015 | Event News
20.08.2015 | Event News
20.08.2015 | Event News
04.09.2015 | Power and Electrical Engineering
04.09.2015 | Machine Engineering
04.09.2015 | Materials Sciences