Australian researchers improved predictions of solar generation by 14.6% by integrating advanced AI methods with seasonal knowledge of First Nations, and offers a new potential for more accurate planning of renewable energy.
Researchers from Charles Darwin University (CDU) in the Northern Territory of Australia have developed FNS-Metrics, a solar prediction system that uses seasonal information from First Nations Calendars. The team has entered this data in Conv-Enterble, a new AI prediction model that they have designed.
The outcome of the study was an error percentage of the solar for the foremost less than half of that of the current prediction models, which represent an increase in the accuracy of 14.6% and an error reduction of 26.2% compared to a strong baseline model.
Researchers developed the AI model using the Tiwi, Gulumorgin (Larrakia), Kunwinjku and Ngurrungurrudjba First Nations Calendars, together with a modern calendar known as Red Center. They said the system has the potential to bring about a revolution in prediction technology.
The CONV and strip model uses ConV1D layers to detect large-scale patterns in the data. It contains transformer and long short -term memory (LSTM) networks to refine detailed trends. These components are combined using a machine learning technique called weighted function -a switch to generate the most accurate prediction.
To test the approach, the researchers drew solar energy and weather data from the Desert Knowledge Australia Solar Center (DKASC) in Alice Springs, with results that show that the model can predict the energy generation of solar energy with a lower error percentage.
CDU co-author, PhD student and Bundjalang-Man Luke Hamlin said that the environmental knowledge that is held in the calendars is an invaluable source.
“Intake First Nations Seasonal knowledge in predictions for solar energy generation can significantly improve accuracy by aligning predictions to natural cycles that have been observed and understood for thousands of years, “said Hamlin,” in contrast to conventional calendar systems, these seasonal insights are precisely rooted in local ecological signals. “
Hamlin said that by integrating this knowledge, predictions can be adjusted to display more detailed shifts in environmental conditions, leading to more precise and culturally informed predictions for specific regions throughout Australia.
This content is protected by copyright and may not be reused. If you want to work with us and reuse part of our content, please contact: editors@pv-magazine.com.
Popular content
