The AES Corporation is a large energy company which generates and distributes electricity to 17 countries and has more than 19,000 employees. It produces or distributes more than 36 gigawatts (GW) of power capacity and has revenues of $14 billion. It services the electricity needs of more than 10 million people worldwide. But, the bold move here is that AES is building another 5-GW capacity, along with a rapidly growing energy storage unit that sells battery systems.
In addition, any spike, fluke, or otherwise unique data can be flagged and pinpointed, and then analyzed and sent to the user before these can be noticed by humans. These are unique problems which would not be normally detected by conventional means.
With this big an operation, the company is moving towards artificial intelligence (AI) to explore how it can improve efficiency and maintenance of its electric grid systems. The use of AI promises to improve the electricity system. Though it is still the early days of using AI, it is already considered as one of the most exciting things brewing in the industry.
Various companies are now involved in developing AI systems for electric generation, prediction, as well as for consumer use. General Electric (GE) and International Business Machines Corporation (IBM) are creating AI applications for forecasting models and maintenance. Startups are tackling niche energy applications which can help lower solar panel cost, as well as making office buildings more energy efficient.
In broad terms, AES is interested in AI systems for neural network design, machine intelligence, and natural language processing. The company is creating a suite of data science tools to take advantage of the large amount of data they accumulate on a daily basis. The data comes from the power plants they manage, including solar energy farms, batteries, and gas plants. AI is required to make sense of all the data being acquired, and help the company improve the daily operations of the energy farms, either for more efficiency, produce more power or to lower cost of operations.
AI – Predicts Outcomes
With the use of AI systems, it is possible to predict output depending on the weather forecast. This information can be used to plan ahead for any shortfall which may necessitate buying energy from other sources, or for transferring energy from other plants. Battery use can also be optimized with the help of AI. Electrical transmission and distribution can also benefit from using AI for current location-related consumption data.
AI and machine learning depends on a large volume of data which electrical systems generate on a regular basis. These tools can sift through the data and understand what is normal, and what is not, and can be taught to backtrack for historical reasons of their normality. The processes are iterative and straightforward, and can be modeled after what users would normally do, but cannot do because of the mountain of available data.
The benefit of machine learning is that at some point, the users do not need to come up with questions. The AI itself would be able to come up with questions based on gathered data. In addition, any spike, fluke, or otherwise unique data can be flagged and pinpointed, and then analyzed and sent to the user before these can be noticed by humans. These are unique problems which would not be normally detected by conventional means.
Using AI on the power systems can help the company improve in a short time. It could also introduce fundamental changes, not just evolutionary changes to the way electric systems are operated.