More information
Description
Sedam IT is a leading provider of ICT solutions and services on Croatian market, with headquarters in Zagreb. Their main areas of work include:
1. Development, implementation and servicing proprietary software solutions,
2. ICT infrastructure solutions and services.
This project aims to develop a wind power production forecasting solution (based on the static and dynamic inputs), which will be adjusted for the specific wind conditions along the Dinara mountain range and custom tailored to the characteristics of regional wind power plants. The solution will use historic and near real time data from wind power plants, and weather forecast data from Croatian Meteorological and Hydrological Service and/or other forecast data (e.g. weather research and forecast model from US National Center for Atmospheric Research and University Corporation for Atmospheric Research). Prediction will be presented through the visualization module.
Using advanced machine learning tools and the in-depth knowledge of the project partners Sedam IT and Energy Institute Hrvoje Požar, the solution will decrease the root-mean-square deviation (forecast error) of the production forecast, leading to a lower balancing energy cost.
Solution has confirmed its functionality in lab conditions and this project is aiming at finishing its development and bringing it to the market.
Reduction of forecasting errors increases the necessary system capacity for balancing, thereby allowing a larger share of intermittent power plants (wind and solar) in the energy mix. Furthermore, lower costs of balancing for wind farms is also additional motive to invest capital into wind farms. With lower operating costs of wind power plants when compared to fossil fuel power plants, the share of fossil fuel power plants in the energy mix gets reduced, thereby contributing to less CO2 emissions from fossil fuel power plants.
Summary of project results
The project aimed to tackle the challenge of inaccurate wind power production forecasts caused by the Bora wind''s uneven blowing patterns along the Dinara mountain range. These inaccuracies result in high balancing energy costs, making wind farm investments less attractive and limiting the share of renewable energy in Croatia''s energy mix.
The project developed a wind power production forecasting solution that integrates static and dynamic inputs, using historic and near real-time data from wind power plants and weather forecasts. Advanced machine learning tools and the expertise of project partners (Sedam IT and the Energy Institute Hrvoje Požar) were utilized to create a solution tailored to the specific wind conditions of the Dinara mountain range. The output includes a forecasting system with a visualization module that reduces forecast errors and confirms functionality in lab conditions, ready for market deployment.
The project achieved a functional, market-ready forecasting solution that reduces forecasting errors, lowers balancing energy costs, and increases the capacity for intermittent renewable energy sources like wind and solar in the energy mix. This benefits wind farm operators, investors (through lower operating costs and increased investment appeal), and society at large by contributing to a cleaner energy mix, reducing reliance on fossil fuels, and lowering CO2 emissions.