We're hiring in our ML team! 💻
Learn more
Over the past few years the world has seen an exponential growth in ambition and investment towards renewable energy. No country better exemplifies this than India. Currently almost three quarters of India’s electricity is generated by coal, creating greenhouse gas emissions and leaving India with air quality levels below WHO standards for more than 140 million people, and India is moving at a breakneck pace to address the challenge.
The country has highly ambitious goals including decarbonising energy by 50% and installing 500 GW of renewable capacity by 2030. Infrastructure projects are already underway, with the world’s largest solar park currently being built by Adani Green in Gujarat. When completed, the park will have a capacity of over 30GW. That is almost double the entire solar capacity of Great Britain in a single solar farm, and is enough to power 26 million homes within the UK, whilst reducing carbon emissions by 35 million tonnes per year.
With these ambitious renewable energy initiatives, comes new challenges for electricity grid operators. Integration of renewables means that our electricity is more weather dependent than ever before. To ensure a steady supply of power, grid operators need to be able to forecast a day ahead of time to know exactly how much renewable energy generation there will be each day, known as grid balancing. If forecasts are inaccurate, operators must purchase carbon-powered electricity to make up the shortfall. These errors not only increase carbon emissions, but are costly mistakes. We know from our partners it can cost a single state ₹10 million ($120k) in a day.
Traditional weather forecasting methods rely on Numerical Weather Prediction models (NWPs). Used throughout the 20th century, they struggle to accurately predict the peaks and troughs in variable generation, particularly at short-term intervals. Artificial Intelligence (AI) has been an essential technology to innovate in this space. At Open Climate Fix, we developed a cutting-edge, AI-powered solar PV generation forecast that combines real-time photovoltaic (solar panel) data, NWPs and satellite imagery to produce incredibly accurate generation forecasts. This increased accuracy has already helped the grid operator in Rajasthan reduce their carbon emissions and costs.
The accuracy of AI-powered forecasts aren’t only important for grid operators. In India, solar and wind farm operators must forecast their generation ahead of time for the grid operator. They can modify this schedule up to 2 hours before they generate. Forecasting at short horizons is key, as penalties are applied by the distribution companies for mis-forecasts. Penalties vary across states, but a common penalty level is for any deviations over 15%, and these penalties are being tightened by distribution companies - they are expected to be 5% next year. These penalties can harm the business case for renewable investments, making accurate forecasts more essential than ever before.
Looking towards the future, India is starting an ambitious scheme for peer-to-peer trading across small businesses and domestic customers, allowing one customer to sell their solar power to another household. This has the potential to make India a world-leading innovator in renewables, but will increase the need for quality solar forecasts even further.
Our AI-powered forecast has already had a proven impact in India. Currently we are forecasting both wind and solar power generation for Rajasthan, the state grid operator with the largest renewables portfolio in India. We are also working with Adani Green, forecasting for Khavda solar park, the previously mentioned 30GW solar farm.
We are currently in discussions with more partners and have exciting plans to expand our innovative Cloudcasting solution to India. This will further help the electricity industry optimise the use of renewables, helping India reach its ambitious energy targets and finally pushing coal generation off the grid.
Are you interested in partnering with us to test our cutting-edge forecasting technology? Our CEO Dan Travers will be in India 24-27 March.
Reach out to us to arrange a meeting and see how we can help increase your savings.
Contact us: bhavika@openclimatefix.org
