About Aeolus

Research area: ICT-2007.3.7 Network embedded and control systems

Project Acronym: AEOLUS
Project Reference: 224548
Start Date: 2008-05-01
Duration: 36 months
Project Cost: 3.36 million euro
Contract Type: Collaborative project (generic)
End Date: 2011-04-30
Project Status: Execution
Project Funding: 2.5 million euro

Background

A key socio-economic challenge for Europe is: how to deal with a climate change, while meeting rapidly increasing demand for energy and ensuring security of supply? Wind energy can be a significant part of the answer. The new frontier of the wind industry is large-scale offshore wind farms. While promising, considerable research and development tasks remain to be carried out before it reaches its full potential in terms of the efficient, stable, safe, predictable and controllable supply of energy. Closed loop control of wind power installations has historically been decentralized and a collection of wind turbines in farms is a highly complex system with interdependencies through the shared resource, the wind. Wind turbines are affected by the wind but they also changes the wind field within the farm through the control. To address objectives related to cost, quality of power and mechanical loads, models and control paradigms must be developed that allow wind resource allocation to individual turbines.

Inspired by the industrial case of complex large-scale distributed offshore wind farms, the Aeolus project will research and develop models that allow real-time predictions of flows and incorporate measurements from a set of spatially distributed sensor devices. In Aeolus we will use the flow information as a basis for new control paradigms, centralized and distributed that acknowledges the uncertainty in the modelling and dynamically manages the flow resource in order to optimise specific control objectives. The model and control principles are used for control of a wind power farm to increase energy quality and reduce the fatigue loads. The usefulness of our techniques will be validated on a case study and by physical experiments on a scaled wind power farm.

Objectives

The objective of Aeolus is to research and develop:

  1. models that allow real-time predictions of flows and incorporate data from a network of sensors, and

  2. control paradigms that acknowledges the uncertainty in the modelling and dynamically manages the flow resource in order to optimise specific control objectives.

More specifically the Aeolus project objectives are:

  1. Generic quasi-static flow models relating single turbine production and fatigue load to the map of wind speeds. A time averaged quasi-static flow model derived from fluid dynamics and based on meteorological and wind turbine related measures.

  2. Dynamic flow models describing deviation from a static model due to rapidly changing flow effects. Dynamic flow models that allow predictions of the flow at all measurement locations. A modelling framework that supports online flow measurements and bridges the gap between a quasi-static flow model and individual wind turbine control.

  3. Principles for supervisory farm power/load optimisation. Centralized control principles for farm level optimisation of flow resource allocation in order to meet farm level control objectives. The control principles incorporate knowledge of the wind flow variations, provide robustness by addressing reconfiguration and at the same time minimize the (extreme and fatigue) loads experienced by turbines.

  4. Principles for decentralized control of the wind power and fatigue load relations. The basic approach for decentralization is to split the global design objective for the wind farm into separate utility functions for each turbine, then let the turbines cooperate by buying and selling support from each other in an on-line virtual market system.

  5. Case study, dissemination and exploitation.

Models and simulation software to support coherence between the activities in Aeolus (both at the methodological and at the validation level), their combined application on a case study, their realization as components for farm level and turbine level control, and, finally, the dissemination of the project's results to industry and the study of potential exploitation routes.

Expected Impact

  • Complexity. Aeolus will support a paradigm shift from single turbine control to farm level control, by investigating a centralized and a decentralized control principle.

  • New markets. Aeolus help develop the European leadership in the new market for large scale wind farms.

  • Efficiency and flexibility. Aeolus provides optimal solutions to maximize power production while minimizing structural loads. Flexibility is achieved by a distributed control paradigm.

  • Low-cost monitoring. Aeolus will provide new knowledge on how to generally predict dynamic flows based on monitoring of the speed and direction of the flow through a network of spatially distributed sensors.

Technical Approach

Six work-packages aim at providing coherent and scalable models and control principles for complex large-scale distributed offshore wind farms bridging the gaps between fluid dynamics, wind field modelling, computer and control engineering. The work-packages are:

  1. Management

  2. Generic quasi-static flow models relating single turbine production and fatigue load to the map of wind speeds

  3. Dynamic flow model describing deviation from a static model due to rapidly changing flow effects

  4. Principles for supervisory farm power/load optimisation

  5. Principles for decentralized control of the wind power and fatigue load relations

  6. Case study, dissemination and exploitation