Back to the programme printer.gif Print

Delegates are invited to meet and discuss with the poster presenters in this topic directly after the session 'Advanced control concepts' taking place on Wednesday, 12 March 2014 at 11:15-12:45. The meet-the-authors will take place in the poster area.

Sung-ho Hur University of Strathclyde, United Kingdom
Sung-ho Hur (1) F P William Leithead (1)
(1) University of Strathclyde, 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

Dr. Sung-ho Hur has been working in the wind industry for over 3 years. He is currently a Research Associate with the department of Electronics and Electrical Engineering (EEE) at University of Strathclyde. He received the B.Eng. degree in EEE from the University of Glasgow, the M.Sc. degree (with Distinction) in EEE from University of Strathclyde, and the Ph.D. degree in Control from the department of EEE at University of Strathclyde. His research interests are in the areas of control (particularly in wind turbines/farms and sheet-forming processes), condition monitoring and process modelling.


Curtailment of wind farm power output through flexible turbine operation using wind farm control


The standard approach to curtailing the power output from a wind farm requires shutting down some turbines. This paper proposes a novel wind farm controller that provides improved flexibility in regulating the power output. Further benefits over the standard approach include faster response to reach the wind farm power demanded, reduction in the variability in the wind farm power output by avoiding shutting some turbines down and provision of synthetic inertia [1].


Each turbine is equipped with an existing central (full envelope) controller and the Power Adjusting Controller (PAC) [2]. The central controller causes the turbine to track its design operating curve as depicted in Figure 1; that is, a constant generator speed (i.e., 70 rad/s) is maintained in the lowest wind speeds (mode 1); the Cpmax curve is tracked to maximise the aerodynamic efficiency in intermediate wind speeds (mode 2); constant generator speed (i.e., 120 rad/s) is again maintained in higher wind speeds (mode 3); and above rated wind speed, the rated power (i.e., 5MW) is maintained by active pitching (mode 4).

The PAC has been developed to provide fully flexible operation adjusting the power output from each turbine; that is, reducing the power or increasing the power for a limited time if required. However, if an increase or decrease in the power output is sustained, the turbine operating state could move away from the design operating curve.

The wind farm controller regulates the wind farm power output ensuring, at the same time, that each turbine (with the central controller and the PAC) operates within the safe operating region defined by the thresholds in Figure 1. In below rated wind speed, the turbines operating inside the inner thresholds could be allocated greater adjustments in power (delta P) than the turbines operating outside the inner thresholds. The turbines operating outside the outer thresholds will be allocated zero adjustment in power to bring them back onto the design operating curve. The same strategy is applied in modes 3 and 4 except that there are no inner bounds since the deviation from the design operating curve caused by the PAC in above rated wind speed is smaller than the deviation in below rated wind speed.

Main body of abstract

Matlab/Simulink and BLADED models of the Supergen 5MW exemplar turbine are used. The BLADED model provides greater details for the structural loads. The Matlab/Simulink model enables many turbines to be included in a wind farm model. The wind farm model thus consists of 9 Matlab/Simulink models and 1 BLADED model. The two software packages are connected using a commercial software package to fully integrate the simulation. Due to the high computational demand, it is assumed that the wind farm contains only 10 turbines.

As shown in Figure 2, the wind farm controller consists of the Network Wind Farm Controller (NWFC) that determines the adjustment in power, delta P, required from the wind farm power to meet the network requirements. The Turbine Wind Farm Controller (TWFC) allocates the adjustment in power to each turbine based on the current status of the individual turbine; for instance, protecting those turbines that are experiencing greater loads. In this paper, it is assumed that every turbine has the same status except that they operate in different wind speeds. The wind speeds are modelled taking into account the correlation over the layout of the wind farm [3].

Different wind speeds cause the turbines to operate on different parts of design operating curve in Figure 1. The wind farm controller also adjusts the power for each turbine based on its status with reference to the inner and outer thresholds. The allocation and reallocation of the power adjustment should take place in a smooth manner, which avoids the introduction of large transients, discontinuities and steps in the wind farm power output. The switching between the various modes (Figure 1) should also take place in a smooth manner.

For example, assume that the wind farm is required to produce a constant power of 25MW at a mean wind speed of 10m/s. At this mean wind speed, the central controller causes the turbines to switch between the Cpmax tracking (mode 2) and constant speed (mode 3) operations. In Figure 3, constant wind farm power output at 25MW is shown in comparison to an unadjusted wind farm power output. Figure 4 shows how the power adjustment is shared between the individual turbines. The results demonstrate that the allocation and reallocation (Figure 4) takes place smoothly (Figure 3).

The behaviour of the turbines on the speed vs torque plane is depicted in Figure 5. To analyse the simulation result in more detail, attention is drawn to Turbines 1, 2, 6 and 7. Turbines 1 and 2 operate in mode 2 while Turbines 6 and 7 operate in mode 3. The figures show that Turbines 1 and 2 operate within the thresholds and Turbines 6 and 7 cross the inner threshold. Consequently, Turbines 6 and 7 should be reallocated a reduced power adjustment in comparison to Turbines 1 and 2. Figure 6 (just before 400s) demonstrates that the additional power reallocation takes place as expected at the switching region.

The change in wind farm power output is determined as the difference between the unadjusted wind farm power output and the wind farm power demand. This could create a feedback effect and alter the dynamics of the central controllers. However, as the number of turbines in the wind farm increases, the number of turbines that share the adjustments to the wind farm power output increases and the feedback effect decreases. However, since there are only 10 turbines in the wind farm, it is important to ensure that the wind farm controller does not create a significant feedback effect. For the example reported above, the power spectrum of fore-aft tower bending moment (TBM) is depicted in Figure 6. The spectra of TBM for the situations with and without a feedback effect are also shown for comparison purposes. To model the situation with a feedback effect, the wind farm controller is applied to a single turbine model, and to model the situation without a feedback effect, a power adjustment of a constant is applied. The power spectra for the wind farm controller and the situation with no feedback effect are similar. The only significant difference that is the peak at 1.8 rad/s corresponding to 3P is caused by the wind farm controller, which enables the turbines to operate off the design curve. The results demonstrate that the wind farm controller achieves its objective, avoiding creating a significant feedback effect even for a wind farm with only 10 turbines.


The wind farm controller introduced in this work exploits an existing central controller and the PAC, which has been developed to provide fully flexible operation of each turbine. The wind farm power output meets the wind farm power demand as determined by the grid side operation requirements for the wind farm, taking into account the status and the operating state of each turbine.

In the example presented in this abstract, it is assumed that the status of every turbine is the same other than the difference in operating conditions arising from differences in the wind speeds for each turbine. The results demonstrate that the wind farm power demand is achieved while keeping each turbine in a safe operating region. The allocation and reallocation of the power adjustments between the turbines takes place in a smooth manner, which avoids the introduction of large transients, discontinuities and steps in the wind farm power output. Furthermore, the power spectra demonstrate that the wind farm controller does not cause any significant feedback that could reduce the effectiveness of the turbines’ central controllers. The feedback effect is even weaker for a wind farm with a larger number of turbines.

The benefits of the wind farm controller include the fully flexible operation of the wind farm and its power output, faster response to demanded changes in the wind farm power demand, provision of synthetic inertia and smaller variability in the wind farm power output.

The authors wish to acknowledge the support of the EPSRC for the Supergen Wind Energy Technologies Consortium, grant number EP/H018662/1.

Learning objectives
The issues related to wind farm control and its design are described. A novel wind farm controller that addresses these issues is presented and its performance investigated using Matlab/SIMULINK and BLADED models of a 10 turbine wind farm.

[1] A. Stock and W. E. Leithead, “Providing grid frequency support using variable speed wind turbines with augmented control” in proceedings of European Wind Energy Association (EWEA) Conference, Copenhagen, 2012.
[2] A. Stock, “Flexibility of operation”, Department of Electronics and Electrical Engineering, University of Strathclyde, SUPERGEN Wind Energy Technologies Consortium Report, 2013
[3] W. E. Leithead, “Effective wind speed models for simple wind turbine simulations,” in proceedings of 14th British Wind Energy Association (BWEA) Conference, Nottingham, 1992.