Wind energyProject Aeolius Wind Forecasting Project
This Lessons Learnt Report outlines the key learnings Aeolius Wind Systems has learnt in the six month period of April 2019 – October 2019 in regard to their involvement in the ARENA Short Term Forecasting trial.
- The short range scanning Doppler lidar provided capacity to map windfield in high spatial resolution (100m) at Macarthur windfarm with a consistent range of 3 km, sometimes longer when atmospheric aerosol loadings are heavier. The maximum distance where data of sufficient quality and quantity could be obtained to derive vector wind fields using AWS proprietary data post processing strategies was approximately 2.5 km. The range capability provided opportunities to evaluate the performance of a limited number of turbines which fell within the horizontal scanning plane (6 in the case of this deployment).
- Understanding wake behaviour in large windfarms with multiple turbines is critical to the success of physical based lidar forecasting strategies. This includes quantifying the impacts of “upstream” turbines on leeward turbines to evaluate impacts on power production, and understanding the role of wake transport processes in both the horizontal and vertical plains;
- The width, length and turbulence intensity of the observed wakes at Macarthur windfarm were within the theoretical limits identified in the scientific literature. Wake behaviour (in terms of transport and dispersion) was strongly influenced by local atmospheric boundary layer conditions;
- It was possible to differentiate between the free flow windfield and turbine wake signatures in the lower boundary layer area using an appropriate lidar scanning strategy and signal processing techniques;
- An analysis of the impact of wakes on downwind turbines is required to quantify additional wear on turbines and the consequent increase in maintenance costs. A recent study of almost 3,000 onshore wind turbines in UK shows that wind turbines will continue to generate electricity effectively for just 12 to 15 years, compared to their industry and financial expectation of 20 – 25 years. This work is outside the scopes of the current project;
- The assembled data is suitable for the planning for the dual doppler and neural network forecasting demonstration, and refining the forecast strategy at the site once operational.
- It’s possible to deploy and implement a lidar based monitoring strategy at an Australian windfarm at relatively short notice (several weeks) if the proponent has access to hardware with range measurement capability beyond 6 km, advanced data post processing software, and analytical skills.
- Access to reliable, high speed telecommunication capability is critical to the success of the lidar wind forecaster. Whilst suitable infrastructure may be in place at the windfarm control centre, remote connection may not be available at the site where the forecaster is installed. This challenge needs to be addressed in the early stages of planning a monitoring demonstration; 2018/ARP/169 Aeolius Wind Systems – Lessons Learnt Report Page 3 of 4
- Reliable, timely and competent onsite technical support is required to ensure the forecaster operates within acceptable limits. Whist the AWS technology is designed to run autonomously and is controlled remotely by specialised staff, there are occasions where onsite assistance will be required. This includes restarting of the instrument following (localised) power blackouts or unscheduled shut down. Whilst the task of restarting the equipment is relatively simple, it requires someone to go to the lidar location, inspect the equipment to ensure no evidence of damage, and throw a switch.
- For practical and legal (insurance) reasons this task needs to be done by a suitably trained and motivated windfarm employee. In many situations this will be a technician employed by a site management subcontractor such as turbine supplier;
- Hence a commercial arrangement is required with the selected party to ensure fast and competent response to onsite maintenance challenges.