The role of AI in the renewable energy sector


Artificial intelligence (AI), if deployed well, could revolutionise India’s energy sector. The capability of AI solutions to predict demand as well as the supply of renewable energy, to automate routine tasks like billing and queries, as well as control power theft, make for a powerful case for their use in the energy sector.

By Ayushee Sharma

The global energy market is undergoing a huge transformation. It is in the process of making energy cleaner and more reliable, to deal with problems like climate change. According to Bloomberg New Energy Finance (BNEF), almost 50 per cent of the world’s electricity will come from renewable energy sources like wind and solar by 2050. This is making it difficult for energy companies that operate using only traditional methods. Utilities as well as other power companies need to find ways to bridge the energy demand and supply gap by increasing the use of renewable energy. Technologies like artificial intelligence (AI) offer new solutions to manage these changes, and be ready for the next generation of the grid.

As renewable energy sources like solar and wind are variable in nature, the supply of power from them may not be in sync with consumer needs. There are times when conventional sources are still needed as backup. Users also act as providers when the energy generated is more than their needs, and they send the remainder back into the grid. There is a massive amount of data available from multiple sources within the utilities market. When the use of smart meters was growing and traditional meters were being replaced in the US, Oracle had reported that this would generate one billion customer data points each day, which was 3000 times more information than that from the old meters. This number has grown even more now. Smart grids, with embedded sensors for sending information to software and monitoring systems, can thus use AI to digitise the energy sector. In tandem with technologies like Big Data and the Internet of Things (IoT), AI will make it easier to balance demand and supply.

AI can be deployed to forecast renewable energy production as well as weather conditions. It can even predict short term, medium term or long term electricity demand. AI algorithms like those developed by Google’s DeepMind can handle zettabytes of structured and unstructured data collected from wind or solar panels to identify patterns, and make predictions and recommendations based on them. DeepMind was acquired by Google in 2014 to improve energy usage. It managed to cool off Google’s data servers, thereby decreasing the expenditure. In another such example, the data mining company Nnergix uses satellite data from weather forecasts along with machine learning (ML) algorithms to make more accurate forecasting in a region. Satellite images are used to generate weather models and ML algorithms use this data.

AI can help in proactively maintaining infrastructure and predicting equipment failure. AI-driven optimisation is vital to maximising efficiency with real-time monitoring. Predicting what energy output a solar or wind farm in a specific location might achieve can attract investments. US based PowerScout’s AI platform collects data from millions of households and predicts whether or not a given household will be investing in solar energy. It can easily filter out potential investors, who can then be offered great returns. This decreases the risks involved significantly, and makes the energy sector more competitive compared to previously used energy sources.

Fraud or even energy theft is more common than one might think. AI can combat energy piracy by enabling utilities to scan individual consumers’ usage patterns and payment history, and alert them about suspicious discrepancies between billing and usage data.

The technology also empowers consumers through a two-way communication model with the utility companies. The customer experience is a top priority for any business. Repetitive customer-side tasks like payments, basic inquiries related to billing, complaints, changes in customer details and the like can be completed through virtual agents and chatbots, saving time and money. In rural areas especially, invoices and other records are still hand written. With technologies like natural language processing, AI can help in digitising these records. Users can also be instructed on how to install or activate products related to energy management via AI and computer vision, if they want to set them up in their homes. DroneDeploy uses computer vision and AI to create maps and 3D models for electric utilities by analysing data.

The result is better integration of large scale renewable energy systems, thereby lowering peak demand, ensuring efficient power transmission, and reducing operational and management costs—all of which are essential to meet economic and environmental needs.
In 2018, the SLAC National Accelerator Laboratory at Stanford University came up with a software platform called VADER (visualisation and analytics of distributed energy resources). The platform can model potential changes in connectivity and the behaviour of various resources on the grid using AI, for resource optimisation.

In India, startups like Avrio Energy, Energly, The Solar Labs and many more are working in this sector. According to the spokesperson from Avrio Energy, “The energy landscape of the world around us is evolving. The transition towards less carbon-intensive consumption is the need of the hour to save our environment. This transition is made possible due to the recent technological advancements, including AI and cloud computing. They are making the industries responsive to real-time data, leading to innovations. Artificial intelligence, combined with the Internet of Things, is disrupting conventional wisdom. It is enabling the continuous collection and synthesis of an enormous quantity of data from millions of sensors to make wise, data backed decisions. At Avrio Energy, we believe this to have the potential of changing the business of the energy sector, as we know it today.”

While the training speed of algorithms has improved due to better computational power, the use of AI comes with threats like cyber attacks on an automated power grid. Applying AI in the right way is crucial to successfully deploy it on a large scale. The poor quality of training data can lead to false predictions, ending up in a lot of time and effort being spent on data categorisation and integration, rather than on the task at hand. Other issues include lack of staff with the necessary expertise, as well as problems linked to data transparency, performance and latency. But given the potential benefits, as technologies become more advanced and pain points like these are solved, there will be enormous adoption of AI worldwide in the energy sector.


Please enter your comment!
Please enter your name here

Are you human? *