Overview of control technology/algorithm

In this work package an overview will be generated of the required software/algorithms for wind farm
control. The work package will be organized by characterizing the different building blocks which define
the majority of wind farm control algorithms. For each of these building blocks the aim is to get
an overview of the available options, their specifications, their ability to work with uncertainty and their
Technology Readiness Level (TRL). Combining the individual building blocks will result in a full landscape
of the solution space, an identification of barriers and a roadmap.

The following building blocks are defined for wind farm control:

  • Control technologies: The control technology defines how the control algorithm interacts with the
    physical system (e.g. pitch actuation, yaw actuation, lidar measurements). Important aspects here
    are: controllability, observability, loads, bandwidth, costs and availability
  • Objective function: The (multi-)objective for which wind farm is employed (e.g. loads, power, predictability)
    and how this can be mathematically quantified. Important aspects are: convexity,
    constraints, scalability and predictability.
  • Internal models The model used for decision making e.g. engineering model, empirical model. Important
    aspects are: accuracy, computational load, coding language, included physics, disturbance
    modelling and fidelity.
  • Estimators The decision making typically depends on the fusion between models and data collected by
    the sensors. Algorithms have to be developed that bring these two together. Important aspects
    are: accuracy, sampling time, sensors required, type of model knowledge required and convergence
    speed.
  • Optimizers The objective function conditioned by the internal model will be used to make decisions
    on how actuators are employed over time. The important aspects are: reliability, convergence,
    convexity, adaptability, robustness, ability to work with uncertainty, and computational complexity.

 

The overview of the different building blocks will give a full landscape of the required software. Different
combinations can be made of the different building block leading to a full landscape of the solution space.
In this work package an overview will be given of this landscape with a focus on the scalability and the
complexity/performance trade-off.