The CleverFarm project Working towards an Intelligent Wind Farm

摘要

The CleverFarm system is a new integrated approach to monitoring and control of remote wind farms. It gives authorised persons access to the system from local networks and via internet. It includes information on the current running conditions, short-term prediction of the wind farm electricity production, video camera surveillance of the wind farm, and early detection of developing wind turbine faults. This combination enables the wind farm owner to effectively plan the maintenance of the wind farm, and, furthermore, the short-term predictions add increased value to the electricity produced. A utility can use the power predictions for its grid quality and power management. The rationale for such a system stems from the fact that many wind farms in the future will be located at remote sites, offshore or otherwise. Operators of such wind farms can use the system for maintenance planning, taking into account the weather conditions at the wind farm to allow for accessibility to the farm and safe working conditions for the maintenance crew. The fault predictions can be used to prioritise the maintenance schedule and allows condition dependent repair actions. Furthermore, the maintenance crew will have on-line access to the systems status and history at all times via the internet. This paper will describe the objectives and preliminary results from an EU EESD funded project (ERK6-CT-1999-00006). Two CleverFarm systems have been installed in wind farms in Denmark and Germany. The system is implemented using Java in a plug-in architecture. The plug-ins are CleverBeans, leaning on the JavaBean terminology. These CleverBeans can also come from third parties, thereby enabling eg sensor manufacturers a quick time to market without worrying about data handling and transfer themselves. Manufacturers of turbines or subsystems can easily get access to data from other sensors distributed throughout the turbine, and thereby use more input to their own fault predictions. The autonomous control of the wind farm is another item on the agenda. The advantages of a farm controller (supporting, not replacing the turbines own controller) are plentiful. Single turbines can be regulated to extend their lifetime or just the time until the next maintenance. A broken anemometer on one turbine does not force a shut down, since there are many more anemometers operating in the farm. The wind field information in upwind turbines can be used for downwind turbines in the farm and so on. Actually, the design regulations for Danish offshore wind farms even require a farm controller capable of reducing the electrical output of the farm by 80% in the course of a few seconds.