Optimal control of a wind turbine with a de-icing system using icing forecasts and observations
Wind power
Voir toutes les réalisations Le projet
Icing is a major source of energy loss for wind farms. Two wind farms operated by Capstone Infrastructure Corporation—Saint-Philémon in Quebec and Glen Dhu in Nova Scotia—were particularly vulnerable to this issue. Together with this partner, we developed a system that automatically controls the de-icing of Enercon wind turbines, without human intervention, based on frost forecasts.
By combining weather forecasts from Environment and Climate Change Canada with GPEO, our frost model, the system anticipates frost events at wind farm sites and estimates the associated energy losses. When an event is forecasted, the de-icing system automatically activates at the optimal time, limiting ice buildup before it even reduces production. This allows operators to make the most of their de-icing system proactively rather than reactively. This project received financial support from the Government of Quebec, through InnovÉÉ, and from the Government of Canada, through NSERC.
The Objectives
- automatically alert the operator to upcoming icing events
- Send commands to the wind turbine controller using Python
- optimize energy savings based on each frost event
- Keep this control at a lower priority than the wind turbine's internal control and the wind farm manager
Our Solutions
- Analysis of the results from our GPEO frost model and sending of email alerts as soon as a frost event is forecast
- Development of a solution tailored to the controller (OPC server) for each fleet, based on existing Python libraries
- development of control strategies based on energy losses estimated by GPEO
- Automatic cancellation of the inspection and an alert to the operator when the wind turbine's parameters do not match the expected values
.
Highlights
-
Frost advisories were issued for Saint-Philémon and Glen Dhu, then expanded to include 3 parks in Nova Scotia and 7 in Ontario
-
Automatic de-icing activation before the wind turbine detects any loss of performance
-
Energy savings comparable to proactive human intervention on the wind turbines tested
-
Ice alerts sent automatically every 6 hours.
-
Creation of an automated control system for the wind turbines at the two wind farms.
