States are increasingly turning to green energy to reduce their dependence on fossil fuels. According to the International Energy Agency, by 2025, such energy systems could displace one of the main sources of energy - coal. AI is being used to accelerate this process.

Indian energy company Husk Power Systems designs, builds, and owns independent hybrid mini-grids. These grids run on a combination of solar energy, energy from rice husks processed into biogas, and diesel generators. Thanks to Husk Power Systems, households, medical clinics, schools, and businesses in remote rural areas of India and Nigeria have access to electricity.

Husk's hybrid installations provide cheap energy during sunny periods and switch to generators at night and on cloudy days. Rice husks for Husk are supplied by agricultural enterprises. Previously, they were buried in the ground, where they rotted, producing methane. Burning 50 kg of rice husks per hour can generate 32 kW of energy, enough to power 500 rural homes.

Husk mini-grids have become more automated with the introduction of artificial intelligence and the Internet of Things (IoT). Husk now uses AI to monitor supply and demand. The company can now not only predict how much energy a household or someone else will need but also produce the required amount. As Husk explained, the most challenging part of the implementation process was data collection. Initially, Husk installed IoT devices at each of its facilities to collect the necessary information: measuring temperature, energy generation, and consumption depending on the time of day. This data was then used to train the AI model.

After the introduction of AI, the accuracy of demand and supply forecasts reached 75%. The implementation of artificial intelligence technologies has reduced the cost of delivering energy to end consumers, enabled real-time monitoring of mini-grids, and allowed for a quick response to system failures.

Within five years, Husk plans to deploy at least 500 mini-grids in Nigeria. According to the company, about 90 million of the country's 210 million people live without electricity. Husk Power Systems is also involved in social programs. Local residents who work for the company are responsible for maintaining the installations and preventing energy theft. In addition, Husk Power Systems sells light bulbs and essential goods through its network and produces incense sticks.

The UK's National Grid ESO and the non-profit Open Climate Fix are using AI to create detailed weather forecasts. With it, experts can accurately predict how much energy can be generated depending on the amount of sunlight.

The Open Climate Fix AI model was trained on satellite data collected at five-minute intervals across Europe, the Middle East, and North Africa. Additional data sources include multi-year hourly weather forecasts, topographic maps, information about the time of day and the position of the sun in the sky, and real-time readings from solar panels across the UK. Interim results of AI implementation showed a 15-20% reduction in forecast errors. Open Climate Fix has also launched a project called Quartz Solar AI Nowcasting, aimed at predicting cloud movement trajectories four hours in advance. Quartz Solar AI Nowcasting analyzes historical satellite images and weather data to identify patterns in cloud movement and transformation.

These projects are crucial for the UK's energy system, as the country doesn't have an abundance of sunny days. The country is also prone to precipitation, and Glasgow, London, and Manchester are among the cloudiest cities in Europe. These megacities have around 1,400 hours of sunshine per year, less than half that of Los Angeles. Therefore, the lack of accurate weather and cloud movement forecasts hinders the efficient use of solar energy. In foggy and rainy conditions, the UK's National Grid operator relies on fossil fuel power plants rather than solar panels.

SmartHelio software from a Swiss startup remotely diagnoses problems with solar power plants, predicts failures, and provides solutions to improve the performance and lifespan of solar panels. 

SmartHelio was developed by a team of experts in climate, physics, and data processing. SmartHelio explained that most fault monitoring solutions on the market focus on data collection and presentation, but do not make predictions. As a result, faults are not detected until the system's performance deteriorates significantly or stops working altogether. This affects their economic efficiency and also leads to additional costs.

To fill this gap, SmartHelio developed AI-based software capable of analyzing the operation of any existing solar power plant. SmartHelio reads real-time data (current, voltage, temperature) from solar power plants, collects data from third-party sources (satellites, weather stations, etc.), and runs algorithms to automatically detect faults. The software can be synchronized with any platform. AI will analyze the data received from the solar power plant. Once the information is collected, the program will create a virtual model of the object (digital twin) within a few minutes and then provide a report on its operation. If vulnerabilities or disruptions are detected, the maintenance team will receive notifications and take action.

American company Hydronalix Inc. is developing an autonomous device called the Sea Remote Controlled Hydrographic Explorer and Recorder (SEARCHER) to find optimal locations for installing wave power plants. Such stations are usually located in water to generate electricity from the kinetic energy of waves.

SEARCHER is designed to study bioindicators. Bioindicators are living organisms whose presence signals the health of an ecosystem. The team uses AI to train the SEARCHER system to find and map the locations of sand dollars. AI processes information received from cameras and sensors on the autonomous SEARCHER device. Algorithms count sand dollars, track their movements, and even distinguish them from other living organisms.

According to research, the distribution of sand dollars is linked to the level of wave energy. Sand dollars choose places with high wave energy at the ocean floor to provide themselves with more opportunities to find food. Scientists believe that installing wave power plants can change the amount of wave energy reaching the seabed, leading to the migration of sand dollars. And analyzing the distribution of sand dollars before and after the installation of the facility will indicate whether there has been a decrease in energy.

The ocean covers 70% of the Earth's surface, but there is little data on this resource, explain scientists at the University of Delaware who are training SEARCHER. Meanwhile, collecting field data is a costly and time-consuming process. A better understanding of the ocean floor will explain everything from how and where storms develop to how climate change will affect different regions.