Estonia’s Enefit uses global hackathon to develop new power forecasting model
Talinn, Estonia. Image courtesy 123rf
Estonian public energy company Enefit has used a global hackathon competition to develop a solution to reduce imbalance costs, enhance grid reliability and streamline the integration of prosumers into the energy system.
The forecasting model aims to enhance the accuracy of energy predictions, coming out of a hackathon competition on Kaggle, a Google-owned global data science platform.
For the competition, Enefit provided participants with comprehensive datasets, including weather conditions, relevant energy prices and details on installed photovoltaic capacities.
The competition focused on the emerging issue of energy imbalance caused by the increasing number of households installing solar panels.
In Estonia alone, the number of small electricity generators has grown from 3,000 to 21,000 in the past five years. These prosumer households, which both consume and generate energy, often have discrepancies between predicted and actual energy usage and/or production.
Inaccurate predictions then lead to substantial fines for the energy company, ultimately impacting consumer prices. Depending on the market, the cost of the imbalance for the balance manager can reach up to several million euros in one month.
Have you read:
Estonia and Latvia sign on fourth transmission interconnector
E-REDES passes 6 million smart meter milestone
“We were overwhelmed by the level of interest in our data challenge. This reflects the growing appeal of the energy sector among technical professionals. Our positive experience with Kaggle certainly inspires us to continue leveraging innovative, on-demand methods to engage with the world’s top talent,” commented Kristjan Kuhi, a board member at Enefit.
The solution developed through the Kaggle competition, states Enefit, will significantly reduce imbalance costs, enhance grid reliability and streamline the integration of prosumers into the energy system.
In the background, Enefit has also developed a smart energy optimiser that enables prosumer households to maximise the profitability of their home energy solutions, such as solar panels with battery storage or other combinations.
“All preparations on our end are complete, and we are ready to integrate the new code into our system,” added Kuhi. As the smart energy optimiser will ensure people that their energy usage is effectively managed, it could encourage more people to become prosumers and thus promote the adoption of renewable energy.”
For the competition, teams had three months to develop their models and compete for a $50,000 prize fund. The competitors used Kaggle’s proprietary time-series API to generate their predictions on fresh-data and were then ranked based on the error of their prediction.
According to Enefit in a release, with participation of 2,715 teams from Japan, US, China and elsewhere, the event marked the largest energy-related hackathon globally.
