Back to the programme printer.gif Print

Wednesday, 12 March 2014
11:15 - 12:45 Advanced control concepts
Science & Research  

Room: Llevant
Session description

The application of advanced control can be utilised to improve turbine performance. The topics addressed in this session include wind turbine/farm control to provide frequency support including droop control, the collective control of a number of wind turbines through the use of a common bus bar and converter, and the stabilisation of floating wind turbines. The control design techniques used include model-predictive, nonlinear and robust control design.

Lead Session Chair:
William Leithead, University of Strathclyde, United Kingdom

Marta Barreras, Gamesa, Spain
Adam Stock Strathclyde University, United Kingdom
Adam Stock (1) F P William Leithead (1) Shona Pennock (1)
(1) Strathclyde University, Glasgow, United Kingdom

Printer friendly version: printer.gif Print

Presenter's biography

Biographies are supplied directly by presenters at EWEA 2014 and are published here unedited

Adam Stock is a researcher at Strathclyde University conducting research into the use of controllers to allow more flexible operation of wind turbines and wind farms.


Providing Frequency Droop Control Using Variable Speed Wind Turbines with Augmented Control


A Power Adjusting Controller (PAC) has been developed that allows wind farm operators far more flexible control of their assets. The controller allows a wind farm operator to change the power output of wind turbines in a farm accurately by an amount ΔP, set by the operator.

One possible application of the PAC is droop control. This paper details how the PAC can be used to provide droop control and evaluates the effectiveness, impact on energy capture and impact on the wind turbine blade and tower loads.


A power adjusting controller has been developed in order that allows the power output of a turbine to be altered by an amount ΔP specified by the operator [1]. The controller acts as a jacket around the central controller and so general operation of the turbine is not affected. In addition, this allows the controller to be retrofitted to older machines – a key advantage.

Fast changes to the generated power can be made by directly adjusting the power demand to the converter. Except when an increase in power in below rated conditions is required, the resulting imbalance between the turbine input and output power can subsequently be removed by adjusting the pitch angles of the rotor blades albeit at a slower rate. While power imbalance remains, the turbine rotor speed changes as energy is stored or released by the rotor. Consequently, a hierarchical structure is utilised, with an inner faster layer acting on demanded generator torque and an outer slower layer acting on demanded pitch angle. The former incorporates constraints to prevent rotor speed changes bringing the wind turbine into undesirable operating conditions.

Previous work has shown that the PAC is capable of providing some grid support in the form of synthetic inertia [2]. This paper explores expanding the use of the PAC to include the provision of droop control. In this scenario the turbine is operated with a set power reduction ΔPred during below rated operation. This allows head room for increased power production for prolonged periods of time. The requested change in power ΔP is then linked to the frequency input f by a linear relationship.
Initial development of the controller was performed in Simulink. The controller was then converted to C code and simulations completed using GL Bladed. Throughout the work, models of the Supergen Exemplar 2MW wind turbine are used.
The accuracy of the change in power delivered is assessed as well as the difference in loadings on the wind turbine compared with normal operation and the impact on energy capture.

Main body of abstract

The UK Government is legally obliged to meet the target of 15% of energy from renewable sources by 2020 [3]. Wind energy will most probably make up a large proportion of this target due to the UK’s excellent wind resource, and thus wind will contribute a much greater proportion of the wider energy mix than at present. High penetration of wind energy could mean that grid support services such as frequency support are required to be provided by wind generation in addition to synchronous thermal generation.

Fluctuations in grid frequency are automatically reacted to by synchronous generators in the system, by their contribution to system inertia and their droop characteristics. As a higher proportion of wind energy connects it would be beneficial to the system for wind power sources to provide some of this response to changes in frequency. In the future, it is conceivable that this will become a requirement.

Previous work has shown that using a novel “Power Adjusting Controller” (PAC) it is possible to provide some grid support in the form of synthetic inertia [2]. This paper utilises Bladed simulations of multi-megawatt, variable speed wind turbines to determine the feasibility of using the PAC to also provide droop control.
In order to provide droop control, additional power is required when the grid frequency falls below 50Hz. The turbine must therefore be de-rated in below rated conditions, as additional power cannot be provided indefinitely in these conditions. In above rated conditions it is possible to provide additional power by over rating the turbine. As such, no de-rating of the wind turbine is required.

Grid codes state that synchronous generation must have a droop capability of 3-5% [4], that is to say a change in frequency of 3-5% translates to a change in power output of 100%. It is also stated that the frequency should be kept between 49.8 and 50.2Hz, giving a max droop requirement in normal operation of 4% of rated power. As such, the turbine power output must be curtailed by 4% of rated power when operating below rated wind speed to allow a prolonged increase in power output. Clearly this will lead to reduction in the life time energy capture of the wind turbine. This loss of power is estimated as being approximately 3.5%.

To simulate droop control, an input of 500 seconds of real grid frequency data was used from the Balancing Mechanism Reporting System [5]. This was converted into a demanded change in power via a linear relationship ΔP = kf, where k is the droop constant, f is the deviation in grid frequency and ΔP is the requested change in power output. Shown below is the average change in power output across five wind turbines in above and below rated conditions. Note that the level of noise on the signal is caused by a misalignment between the natural oscillations in the power signal. The noise level reduces due to averaging as more wind turbines are used.

As droop control may be used for prolonged periods of time – possibly at all times, the effect on fatigue loads is significant. The percentage change in damage equivalent loads when providing droop control compared to normal operation is therefore investigated.

For large periods of time, the grid frequency is often more tightly controlled, only varying by +/- 0.1Hz. In these conditions it may be possible to provide droop control utilising half the wind turbines in the farm. This would increase the energy capture compared to using all the wind turbines for droop control. Allocation of the frequency support between turbines would be performed by a farm level controller at a higher hierarchical level. This could be based on all the available information for each turbine e.g. wind conditions, maintenance schedule, position within the farm etc. If only half the turbines were used for the majority of the time then the power loss would be approximately halved to 1.75%. It is expected that well designed farm level controllers will be able to reduce this figure further.

Simulations are also therefore conducted whereby the turbine offers a droop capability of 8%, limited to frequency changes of +/-0.1Hz, the results are shown below.


As wind energy provides a higher proportion of total generation, grid frequency droop control from wind turbines becomes desirable. A novel method of providing grid frequency droop control through augmented control of variable speed wind turbines is presented. The controller is shown to work for a cluster of five two mega Watt (2MW) machines, providing additional power proportionally to a supplied grid frequency error.

The controller is retrofittable as it takes the form of a jacket around the central controller. This capability is highly important if legislation is brought in requiring all wind turbines to provide grid support regardless of their age.

When operated above rated wind speed there is no net loss in energy capture. In below rated operation the requirement for head room for additional power leads to a reduction in energy capture of 4%. The total reduction in energy capture across the operational envelope is approximately 3.5%. Utilisation of wind farm level control may allow the total reduction in energy capture to be reduced to 1.75% or lower. A reduction in energy capture is a necessary cost of this method of control.

Using a Rayleigh wind speed distribution at a mean of 10m/s the tower damage equivalent loads when using the PAC for droop control are reduced by approximately 3%. The effect on the blade damage equivalent loads is negligible (less than 0.5%).

This technique generally reduces the loads on the turbine in below rated operation. Across the full envelope there is no negative effect upon the blade or tower loads.

Learning objectives
This work demonstrates that droop control from wind turbines is feasible using augmented control. The augmented control requires no alteration of the central controller and so this technique is retrofittable.
Whilst the provision of droop control requires some reduction in energy capture, this can be minimised through farm level control.

[1] A. Stock, “Flexibility of operation”, Department of Electronics and Electrical Engineering, University of Strathclyde, SUPERGEN Wind Energy Technologies Consortium Report, 2013
[2] A. Stock and W.E. Leithead, “Providing Grid Frequency Support Using Variable Speed Wind Turbines with Augmented Control”, Scientific Proceedings of the European Wind Energy Conference, pp152-156, 2012
[3] Department of Energy and Climate Change, “UK Renewable Energy Roadmap”, July 2011
[4] National Grid, “Guidance Notes, Synchronous Generating Units2, Issue 12, 2012
[5] Balancing Mechanism Reporting System, [Online]. Available: