Search

The control of large-scale offshore wind power plants (OWPPs) has historically been handled byseparating the control into different hierarchical levels in a wind power plant (WPP). A WPP control is traditionally implemented in a centralized processing unit, which receives and sends information from all the turbines and system operators. This is commonly known as a centralized control scheme, implemented in conventional networks. But for large-scale OWPP clusters with hundreds of wind turbines (WTs), the centralized control scheme may be impractical owing to the huge computation burden in order to process the large volume of information. Even the conventional control schemes such as droop-based or PI with/without optimization might fail to achieve optimal performance for large-scale OWPPs. This talk describes the need to breakdown the control problem into manageable sub-problems, possibly weakly related, such that the overall plant is no longer controlled by a single controller, but by several independent controllers. Thus, instead of having a fast powerful computer centralizing the control of the entire WPP cluster, the control responsibility can be shared among the wind turbines. This gives rise to two additional control architectures besides centralized control scheme: distributed and decentralized control. Modern WPPs can be visualised as smart grids with several intelligent agents (WTs) exchanginginformation with their neighbours to achieve a desired power profile and agree on a global outcome. Sucha control scheme, commonly known as distributed control, consists of a number of local controllers, each ofwhich controls a subset of the system, with a capability of communication between them. However, itpresents challenges such as the appropriate design of a distributed algorithm, reliability of communicationnetwork and coordination among agents with limited information exchange. The local controllers mightalso be designed to operate in an independent manner as in a decentralized control architecture.However, the local decentralized control centres could share the information to solve the larger problem.Strong interactions between the inputs and outputs of different controllers can even prevent one fromachieving stability.

This report presents a collection of models and control strategies developed and implemented in the power system simulation tool PowerFactory DIgSILENT for different wind turbine concepts. It is the second edition of Risø-R-1400(EN) and it gathers and describes a whole wind turbine model database built-op and developed during several national research projects, carried out at Risø DTU National Laboratory for Sustainable Energy and Aalborg University, in the period 2001-2007. The overall objective of these projects was to create a wind turbine model database able to support the analysis of the interaction between the mechanical structure of the wind turbine and the electrical grid during different operational modes. The report provides thus a description of the wind turbines modelling, both at a component level and at a system level. The report contains both the description of DIgSILENT built-in models for the electrical components of a grid connected wind turbine (e.g. induction generators, power converters, transformers) and the models developed by the user, in the dynamic simulation language DSL of DIgSILENT, for the non-electrical components of the wind turbine (wind model, aerodynamic model, mechanical model). The initialisation issues on the wind turbine models into the power system simulation are also presented. The main attention in the report is drawn to the modelling at the system level of the following wind turbine concepts: 1. Fixed speed active stall wind turbine concept2. Variable speed doubly-fed induction generator wind turbine concept3. Variable speed multi-pole permanent magnet synchronous generator wind turbine conceptThese wind turbine concept models can be used and even extended for the study of different aspects, e.g. the assessment of power quality, control strategies, connection of the wind turbine at different types of grid and storage systems. Different control strategies have been developed and implemented for these wind turbine concepts, their performance in normal or fault operation being assessed and discussed by means of simulations. The described control strategies have different goals e.g. fast response over disturbances, optimum power efficiency over a wider range of wind speeds, voltage ride-through capability including grid support. A dynamic model of a DC connection for active stall wind farms to the grid including the control is also implemented and presented.

This paper describes the development of wind farm controllers for different types of wind farms. The overall aim of the wind farm controllers is to enable the wind farms to contribute to the control of voltage and frequency in the power system. Still, the controllers should meet the conventional aims of wind turbine controllers, which are first of all to maximise the production and to minimize the structural loads and lifetime consumption on the wind turbine components. To meet these aims, the idea is that the wind farm controllers use wind speed predictions.

Nowadays, most of the ancillary services such as reserve capacity, inertia and frequency control relies on large conventional power plants. Approaching future power systems with high penetration of renewable energy sources (RES) has resulted in imperative need for the evaluation of ancillary services. This research focuses on the frequency stability which must be ensured in order to maintain the grid stability against imbalances between generation and load. This large conventional power plants that provide ancillary services are called “must-run” units. These facilities are generation power plants necessary during certain operating conditions and they are responsible for providing enough ancillary services to ensure a reliable operation of power systems. Given a high RES penetration in the future, must-run units are expected to be reduced or totally decommissioned reducing the power system inertia. This may result in insecure operation threatening the reliability of the power supply. This project investigates the frequency stability support from renewable energy generation such as wind power plants (WPPs) and solar photovoltaic systems (SPVSs) in future power systems with high penetration of RES and without must-run units. Sensitivity studies for frequency stability are performed on a simulated 2030 scenario for western Denmark (DK1) power system. The objective of this master thesis is to study the DK1 power system to analyse the ability of modern controllable WPPs to provide frequency stability without must-run units in a future scenario dominated by RES generation. This project examines the primary frequency control in DK1 simulating an overfrequency event islanding DK1 from the CE power system with high wind forecast. The main results of this project reveal that the fast deploy of active power by the RES generation counterbalances the reduced inertia in the power system, which can operate without a lack of stability of the power supply for overfrequency events without must-run units. However, there are technical capabilities and limitations that curtail the RES penetration. Recommendations on the parameters of the WPPs frequency control are made according to the droop, the ramp rate and the RES penetration. The virtual inertia is recommended for frequency control of WPPs and increases its importance when the RES penetration is high. The support of HVDC interconnections is an interesting facility to increase the RES penetration allowing the power system to operate with even less inertia online maintaining a stable supply. Although, the measurement and communication delay by the frequency controllers increases its importance when increasing the RES penetration as faster power deploy is needed.