DNVGL To Deliver 5-Minute Energy Forecast Pilot For Australia’s Ararat Wind Farm

DNV GL, the world’s largest resource of independent energy experts and certification body, has won funding from the Australian Renewable Energy Agency (ARENA) to trial advanced 5-minute short term forecasting at the Ararat Wind Farm in Victoria.

The project will introduce machine learning approaches and provide valuable new information to the market on the potential accuracy of 5-minute self-forecasts.

At present, the Australian energy market uses half-hourly forecasts provided by the Australian Wind Energy Forecasting System. By moving towards a 5-minute timeframe and allowing market participants to conduct self-forecasts, the trial’s objective is to reduce generation forecast error, thereby improving grid efficiency and reducing settlement costs.

DNV GL will deploy the advanced wind forecasting approach at the 242 MW Ararat Wind Farm with the objective of providing both short and medium-term wind generation forecasts of the highest possible accuracy.

This will be achieved by integrating machine learning with its existing wind Forecaster service, which has been demonstrated to be accurate and reliable for over 50 GW of installed capacity globally.

Nicolas Renon, Executive Vice President APAC at DNV GL – Energy said: “Forecasting is a strategic enabler that will allow greater penetration of renewable energy into the grid, making it a key facilitator of the energy transition.

“We look forward to leveraging our deep domain knowledge and local expertise to help Australian market operators optimize central dispatch and improve the accuracy of market outcomes, while reducing costs to generators.”

For this project, DNV GL has partnered with Ararat Wind Farm Pty Ltd, owned by OPTrust and Partners Group, to provide live data for the project. DNV GL has also partnered with RES Australia Pty Ltd, the original developers of the wind farm, to provide high level analysis and support to the project.

Numerical Weather Prediction (NWP) outputs and live data will be combined to optimize the effectiveness of different forecasting approaches. DNV GL’S Forecaster team has previously demonstrated that applying machine learning to the processing of NWP data can improve medium to long-term forecasts.

During this project, such approaches will be applied to live site data and assessed, the most promising of which will then be implemented to further explore potential improvements to forecast accuracy.