How AI Can Solve The Renewable Energy Problem
Dec 12, 2023 6:31 am
How AI Can Solve The Renewable Energy Problem
4 min read
Artificial intelligence (AI) is a set of technologies that enable computers to perform a variety of advanced functions, including the ability to see, understand and translate spoken and written language, analyse data, make recommendations, and more. Machine learning is a subset of artificial intelligence that uses algorithms to train data to obtain results. Energy storage, land use, reliability, intermittency, and high upfront costs are just some of the major problems with renewable energy. Some of these problems can be solved using AI and machine learning.
Machine learning, like a human brain, needs input to gain information and learn from that information. This helps solve complex problems that would take humans months or even years to solve, in minutes. Machine learning has thus been used to advance various industries such as healthcare, manufacturing, transportation, banking & finance, agriculture and energy.
How AI is used in renewable energy
AI can be used to overcome the problem of intermittency by forecasting sunlight for solar and airflow for wind power. It uses weather data gathered from the past to accurately forecast weather in the future. Power generation companies can use AI to plan when to produce energy and when to store it for later use. This can make renewables more reliable, cheaper, and more efficient.
AI can also be used to model optimal layout and pick the ideal geographical location for renewable energy plants. This will improve the energy production rate as well as the total energy output of the plant.
AI can be used to collect data from sensors installed on panels, wind turbines or some components on the electric grid. This data is then analysed to sense abnormalities and even failures. This allows for the automation of monitoring as the system will simply send an alert message when issues are detected. It can also help schedule maintenance on turbines.
Real life applications
Solar Energy Technologies Office of the US joined forces with IBM to produce an AI powered system called Watt-Sun (excuse the pun) – a technology that uses machine learning to sort through weather report data and help reduce the unpredictability of solar energy output and reduce the need for excessive energy storage systems. The result – increased accuracy in weather forecasting relating to energy output by 30%. The full report can be read here.
Renewable energy company, Siemens Gamesa, partnered with AI and machine learning/semiconductor chip company, Nvidia, to develop Nvidia Modulus and Omniverse, an AI powered digital twin platform that performs high speed simulations of wind farms. A digital twin is a virtual model that accurately mimics the original object. This helps Siemens Gamesa’s engineers make quicker more informed decisions through simulations of physics concepts more accurately and up to 4 000 times faster than their previous AI platform. It will take the new system 11 minutes to solve an engineering problem that the average system would take 1 month to solve.
Nvidia Modulus will also help renewable energy companies save costs and reduce risk during product development. The system does this by modelling the new product (e.g. wind turbine) and show how it would behave in real weather conditions. Nvidia Omniverse creates digital twins – virtual worlds which can mimic real life weather conditions. This allows engineers to see how renewable energy plants such as wind farms can produce optimal energy by exposing weaknesses in the layout that can create issues such as turbulence. This increased Siemens Gamesa’s energy output by 20%.
Danish company, Vestas, is using the Microsoft Azure machine learning and Microsoft partner, Minds.ai, to reduce the effect of wind and other weather conditions. They use a controller design platform called DeepSim to make controllers accurately respond to variables in the wind farm environment such as wind direction, wind speed, yaw (the act of twisting or oscillating about a vertical axis), and many others to improve efficiency and output of wind farms. This boosts accuracy and efficiency which can drastically save costs and increase revenue.
Light Me Up has been able to use Nividia’s Jetson Edge AI platform to make smart electricity meters that can take measurements tens of thousands of times per second. This can allow homeowners to control the flow of solar energy in their homes, which can reduce electricity usage. Because the measurement is so accurate and constantly up to date, the smart meters can limit the amount of solar energy flowing into the home’s appliances to just enough electricity and store any excess in the batteries.
Challenges
- AI and machine learning are based on past data. Lack of new data can leave AI wanting during unprecedented events. Lack of training data or unclean data can also lead to inaccurate predictions.
- For newer AI technologies, there may not be enough existing data to train the machine learning system.
- AI is very complex, which makes it easy for one small problem in the process to ruin predictions.
Running AI models is a very slow process, taking a lot of time – something we don’t have in the climate fight. It also takes a lot of computational power, which requires a lot of electricity and water to cool the machines. If the goal is to reduce energy consumption and increase sustainability, then using a solution that is water & power intensive may defeat the purpose.
The ‘wake effect’ is the trail left by each wind turbine on the next wind turbine that is rotating at a higher speed. IEE Spectrum estimates that the potential wind power lost to wake effects is about 10%. With that much renewable power being lost and global warming getting worse, any method of renewable energy improvement should be welcomed. AI is a powerful tool and is becoming evermore useful as time continues. It can really help in solving the world’s climate issue.
So what do you think of AI in renewable energy? Should we quickly embrace it or should we be cautious about it? Share your thoughts by replying to this email.
For more information on AI in renewable energy, or any other renewable energy related articles, view the other blog posts on my website and be sure to open my weekly emails every Tuesday at 9AM CAT.
Interesting Updates
America's first electric road can wirelessly charge EVs as they drive
