Google and DeepMind apply AI to predict the output of wind farms

In a statement released today 27/2, Google said it has found a way to statistically predict the electricity output produced by wind farms, which is through the use of electricity. artificial intelligence software of subsidiary DeepMind (based in London).

By using DeepMind's machine learning algorithm to predict the power output generated by wind power from the farms Google uses for its green energy initiatives, the company says now. They can make specific plans for operating storage, distribution and power supply systems, which bring much higher value than currently used standards.

Google and DeepMind apply AI to predict the output of wind farms Picture 1

1. Google AI application in detecting eye diseases

According to Google, the software has helped improve the 'value' of wind-generated electricity, which is providing 20% ​​more than storage, distribution and supply without using. Real-time predictions. However, the American giant did not explain clearly whether the value was in monetary terms or in terms of power output, as well as the way Google deployed it on wind farms, but most likely will be in the Midwest region of the United States, where some major data centers are present in the country.

Last year, Google said it finally reached a notable milestone in compensating its energy use with 100% renewable energy from renewable energy. This is largely due to Google's energy sales and investment contracts with solar and wind farms, helping to power the company's data centers, as well as with plans to use renewable energy to compensate for the use of standard power grids in other markets.

Google and DeepMind apply AI to predict the output of wind farms Picture 2

1. China continues to launch the virtual female radio announcer run by the world's first AI

However, when it comes to wind power, the use of electricity generated from this type of energy can be more difficult because of how much electricity a wind farm can generate, how What is the best way to store that power and then how to deliver it effectively is not easy. According to Google, the changing nature of the wind makes it a very difficult source to predict, or to say it is impossible, because it has to rely on nature to estimate the electricity needs of the grid. electricity.

'We certainly cannot eliminate the wind change in calculations, but the initial results of the experiment show that being able to use machine learning to predict wind power output at many The time with many different natural variations is possible. In addition, this approach also helps to bring rigor to the analysis and release of data for the specific activities of wind farms, simply because machine learning can help farm managers. The wind carries out complex assessments in a smarter, faster, and more accurate way, especially in terms of estimating how much of the electricity from wind farms can meet the demand. use electricity.

This is not the first time DeepMind's AI expertise is used in this way. Back in 2016, Google announced that it had successfully cut about 15% of its data center's electricity costs by helping the AI ​​lab. In 2018, Google went further and gave these AI systems more control over the calculation of power consumption for the entire system. In addition, there was a report in 2017 indicating that DeepMind is cooperating with the UK national grid agency to help this agency calculate and balance supply and demand for output. electricity throughout the British territory.

Google and DeepMind apply AI to predict the output of wind farms Picture 3

1. AI engineer Facebook talks about deep learning, new programming languages ​​and hardware for artificial intelligence

In addition, these projects also show DeepMind's great prospects in transforming the operating model, from the form of a test center to a profit-taking direction. In 2017, DeepMind 'spent' its parent company Google on a budget of up to $ 368 million for research projects involving artificial intelligence, while the profits were not significant. If DeepMind's software can be used in real-world situations outside the lab, the company has a full potential to become a huge profitable business for Google.

Above all, once again we see the practicality and effectiveness of AI for all areas of life!