Modelling, Operation and Control of HVDC Connected Oshore Wind Farms

Adelina Agap, Madalina Cristina Dragan, Borja Imanol Iurre, Adrian Constantin January 8, 2009

Chapter 1
Abstract of the semester project report
New and feasible methods of integration of large amount of electricity produced by the large oshore wind parks have to be found in order to maintain the grid stability and to be economicaly convenient. One of these possible methods that is studied in this project is the HVDC-VSC connection of the oshore wind farm. The interaction between large amounts of electricity produced by the wind farm and the grid involves investigation of voltage and frequency stability, active and reactive power control and ride through capability of the wind turbines. Kriegers Flak located in the Baltic Sea, between Denmark, Sweden and Germany is another opportunity for large scale oshore wind power generation. Due to its shallow depth, Kriegers Flak is well-suited for oshore wind turbines. A 330 MW wind farm was already approved by the German part and the same for the 640 MW on the Swedish part of Kriegers Flak. At interest for this project is the Danish part that is estimated to have a total capacity of 450 MW. Up til now all wind farms have used AC connnection. This was a feasible solution mainly because of the low cost for the station and cables at a short distance to the shore, also it is a proven reliable technology and it uses standard components (cables, transformers, reactors, switchgears) [2]. Advances in power electronic technology have led to the development of HVDC systems more reliable and cost eective, a suitable solution for connecting a large oshore wind farm. The Table 1.1 shows the most signicant factors that are between the HVAC and a HVDC solution based on VSC. The DC transmission oers low cable cost, it is more efective for power transmission on long distances and has high power capabilities [2].
Factor AC DC (VSC type)

Cable voltage Cable power capacity (per circuit) Number of cables Cable losses at peak load (100 km length) Substation losses

Up to 245 kV as 3-core Up to 350 MW 3/GW ~5% 0.1% at 50% loading

Up to +/-150 kV Up to 1200 MW+ 2(1 pair)/GW ~1% 1.5% at 50% loading

Table 1.1: Comparison of HVAC and HVDC transmission systems [2],[3]

Therefore a reliable solution for connecting large oshore wind farms as it is the case of Kriegers Flak to the transmission system may be the HVDC - VSC connection. In the Figure 1.1 below the conguration that has been studied within this project is presented. On the oshore park side each wind turbine has a rectier. All the output power of the wind turbines will be linked to a common point oshore
1

followed by a step up DC/DC booster and eventually, the HVDC cable will connect it to the shore side. Another DC/AC converter is required on the onshore side in order to transform the DC voltage into AC. Galvanic isolation will be provided by a typical AC two windings transformer.

Figure 1.1: Wind farm with full scale converter and HVDC transmission system

Nowadays a special concern is given to the grid code compliance of the new generation of wind turbines. The Danish TSO has formulated grid codes regarding wind turbines. In these grid codes special requirements regarding voltage levels in the Point of Common Coupling (PCC), the active and reactive power regulation, as well as the ride-through capabilities during fault events in the grid are specied [4]. This project investigates how to connect oshore WFs and how to integrate this high capacity of power into the main grid. Dierent study cases will be performed, having as purpose the validation of the model and also to see the impact on the main grid.
1.1

Project Goals

The major interest in the project is to study the grid connection of a large oshore wind power plant. Based on the previously mentioned considerations, the goals of this project are:
modelling of the oshore wind farm with all its components (modelling of the

wind, generator, drive train for one wind turbine, pitch controller, aerodynamic model); DC cable modelling and the converter control structure necessary for reliable operation; the studied system;

modelling of the HVDC system, including the two types of converters, the

proper sizing of each of the components in order to meet the rated values of analysis of the HVDC-VSC solution of connecting the oshore park to the grid;

 analysis of wind perturbance characteristics on the wind farm`s output power; analysis of the power ow from the wind farm to the main power grid and if

possible, a verication of the eciency of the studied system;

analysis on active power reserve allocation feature that enhances the exibility

of the wind farm in respect to active power control and grid stability; that also contributes to the main power grid stability.

study on the reactive power control capabilities of the wind farm, enhancement

1.2

Wind farm and HVDC studied system

The project contains contains models implemented in DIgSILENT Power Factory. Matlab software was also used in order to obtain some characteristics of the parameters.

Figure 1.2: Studied wind farm and HVDC transmission system

Figure 1.2 shows the overall structure of the considered oshore wind farm and its connection to the grid. It is based on [1] and developed for the specic purpose of the project. A number of three identical wind turbines connected in parallel are modeled for the wind farm conguration. The generating side of each WT is connected with the AC/DC side of a full scale power converter, through a phase reactor in parallel with an AC lter. The phase reactor regulates the current and the AC lter is used for switching frequency currents rejection. All the three wind turbines are connected to the same connection point on the oshore side, the sending end station. The link with the onshore is made by a VSC-HVDC transmission system. The main function of the VSC-HVDC is to transmit constant DC power from the rectier to the inverter. For the oshore HVDC transmission cable, the units are DC to high voltage DC converters for boosting up the DC voltage from the wind farm side to the transmission side. On the onshore side, the HVDC cable goes to the receiving end station where the DC to AC transformation is performed. Finally on the main grid side the receiving end station is coupled to the PCC through a set of phase reactors and AC lters. A transformer is also used for AC voltage boosting and galvanic isolation.
3

The following models have been designed and modelled: wind turbine system, wind turbine full scale power converter control, grid side converter control. All the other models have been used as built-in models provided by the simulation software available. Dimensioning of each of the components has been performed according to case specic procedures. Typical parameter values have been used for all of the components.

1.3

Wind Turbine Modelling

The wind turbine's rotor extracts the kinetical power from the wind and transforms it into mechanical power. Then the mechanical power is converted into electrical energy by the generator and fed into the grid. The modelling of the components implied in this process is presented in this section. The overall structure of one wind turbine is shown in Figure 1.3. It consists of the aerodynamic, the mechanical and the electrical model. The inputs of the wind turbine model are the wind provided by the wind simulation block and the value of the power reference (Pref ) required by the grid. Mainly, the Pref value is the rated power of the wind turbine's generator. Depending on the demands of electricity, the grid operator can increase or decrease Pref in order to maintain the stability of the grid. The equivalent wind speed used for the wind turbine simulations is the result of the wind model. Its inputs are the average wind speed of the area in which the wind turbine is placed and the rotor position calculated in the mechanical model block. Based on the wind speed, the pitch angle () from the control block and the speed of the rotor (r ) outputed by the mechanical model, the aerodynamic model calculates the rotor torque (Tr ) . Using the rotor torque and the speed of the generator's rotor (r ), the mechanical power is resulting from the mechanical model. Next the electrical model represented by the SCIG transforms the mechanical power into electrical one. The wind turbine's models are presented in the followings subchapters.

Figure 1.3: General structure of the wind turbine model

The control model limits the output power of the wind turbine to the its reference value in case if the wind is higher than the rated value. Based on the measured power from the output of the generator and the reference value for the power, the pitch
4

control has to adjust the angle of the blades in order to maintain a constant reference power.

1.4

Generator Side Converter Control

The main issue regarding the control of the full power scale converter for the wind turbine is to ensure system stability and maximum operating point of the generators by providing suitable voltage and frequency set points on to the grid. Constant voltage-frequency control has been adopted in this model, maximizing power output through a simple control algorithm. Therefore, the voltage stability is in close connection with the demand of reactive power. A small error in the voltage reference can generate very large reactive currents that may produce instability. Additionally to constant voltage-frequency control, in order to obtain the required stability a voltage and frequency droop control has been implemented which does not require any communication between the converters [1]. Figure 1.4 represents the overall control of the full scale power converter driving the wind turbine, where voltage and frequency set points are provided by a voltagefrequency control.

Figure 1.4: Control structure for the sending end converters

The control of the rectier is composed of voltage-frequency control, frequency droop control block, voltage droop control block and PI regulators.
1.5

Grid Side Converter Control

The grid side converter (receiving end station) control is based on [1] and has the main diagram as in Figure 1.5. The receiving end station is connected to the main power grid and must provide means to control the reactive power in the PCC. Another objective is to maintain and control the DC link voltage. The control system of the receiving end station is based mostly on standard control procedures described in [7, 8, 9] and used for example in [1, 5], resulting the
5

overall control structure described in Figure 1.5. As it can be seen the system is controlling DC voltage and reactive power by sending references to a current control loop designed in a synchronous rotating frame. The outputs of the current loop are the real and imaginary parts of the modulation index P m.

Figure 1.5: Main control diagram

Figure 1.5 depicts also the design of the control model with two branches: d and q current component control. Each of the branches implements a fast inner current loop and a more slower outer loop for setting the DC voltage and the reactive power in the PCC. In the diagram it is also sketched the PCC grid voltage Vpcc which is considered in the control system to be a perturbance since it has small variations considering the time interval of the PI loops. Standard PI controllers have been implemented for all the blocks using DSL scripts.
1.6 Study Cases - Project Results

The following presents dierent study cases so as to verify and illustrate the performance of the control system for the entire structure of a small oshore wind farm connected to an HVDC line transmission system. To test that the wind turbines operate correctly, in Study Case 1 , the pitch angle and the power production of the WTs are analyzed. The analysis is made for dierent wind speeds for which the WTs will operate in optimization and limitation modes. Study Case 2 , presents the analysis for grid stability. Production follows demand at all times in order to maintain the power balance and secure power system stability. In order to ensure grid stability, active power reserve margins are provided by the TSO. In this set of simulations, the active and reactive power production are analyzed when the reserve margins are applied. Study Case 3 deals with the dynamic control of reactive power transfer to the main power grid. Control of the reactive power in the PCC provides support to the main power grid. References for the reactive power are provided by the TSO to the wind farm supervisory system, which will further on control the output of the grid side converter. The performance of the DC voltage controller is also analyzed. Losses on the HVDC line are recorded in dierent power output conditions.

1.7

Study case 1 - Pitch Controller

Four dierent simulations are done to prove that the pitch controller implemented in Power Factory works as expected. This study case was done by connecting, rst only one wind turbine and after, connecting the entire WF. For these simulations as input for the WT and WF, was entered the wind provided by the wind model based on the average wind speed. The power regulation in these simulation is done only by using the pitch control.

1.7.1 Power Optimization mode

Figure 1.6: Operation of WT in power optimization

In power optimization mode, the strategy of the pitch control is to keep the pitch angle to zero in order to obtain maximum power output. In this case the wind varies from 4m/s to 11.9 m/s for 350 seconds. Figure 1.6 is composed by two graphics: - The rst one where is presented the variation of average wind speed and the wind outputted by the wind model. As it can be seen, the wind generated by the wind model follows the real wind which was applied. - The second graphic shows the power variation, which is dependent on the wind speed. In Figure 1.7 it can be observed the pitch angle which is equal to zero. Therefore, it can be concluded that the pitch control of the WT works properly. When the wind speeds are below the nominal one the pitch angle is held constant in order to maximize the power output.

Figure 1.7: Pitch angle in power optimization mode

1.7.2 Power Limitation mode

For this simulation the wind is increased, from the rated value 11.9m/s until it reaches 22m/s. This simulation is also run for 350 seconds.

Figure 1.8: Operation of WT in power limitation

In power limitation mode the PI controller is active. The selection of this mode is done function of the wind. As in the previous simulation (power optimization), in Figure 1.8 is depicted the wind speed and the active power variation. When the
8

rated value of the wind is exceeded the blades start to be pitched and the power is held constant to its nominal value, with a small deviation about 1.5 %. For wind speeds above the rated value the PI controller is looking to nd the proper value for the pitch angle in order to limit to power to its rated value. Figure 1.9 shows the modication of the pitch angle for dierent wind speeds. For this study case the only active control of the WTs is the pitch control, therefore is can be seen that the variation of the pitch is small, until 60 .

Figure 1.9: Pitch angle in power limitation mode

1.7.3 Transition mode

To test the functionality of the controller, during the transition period between power optimization and power limitation, the wind speed is changed from 9 m/s to 15 m/s. The switch is done when the nominal wind speed (11.9 m/s) is exceeded. Figure 1.10 shows the wind prole that is applied for this simulation. When simulation starts, for the rst 10 seconds, the wind is at its rated value. In Figure 1.10, it can be seen that for this time interval the power is at the nominal value and pitch angle is zero (see Figure 1.11). In second 10, when the wind speed increases and it is above 11.9m/s, switch is made and the PI starts working. In this precise moment it can be seen that the power has an overshot but it soon arrives to its rated value (see Figure 1.10). As Figure 1.10 shows, from the second 100 the wind speed starts to decrease til it reaches 4m/s and after raises to 15 m/s. During the same simulation time interval, Figure 1.11 illustrates how the pitch angle is modied in order to have the desired power, depending if power limitation or power optimization mode is selected.

Figure 1.10: Operation of WT in transition mode between power limitation and power optimization

Figure 1.11: Pitch angle in transition mode

1.7.4 Power control of the Wind Farm

For this simulation all three wind turbines are connected. Figure 1.12 illustrates the wind speeds applied to each of the wind turbines. For this part of the study case only the wind provided by the wind model was used.
10

The purpose of this simulation is to prove that power regulation is accomplished using only pitch control. In the rst time interval the wind speed is varied between the nominal wind speed and the cut-in vcutin = 4m/s. After that the wind is raised from 4 m/s until 22 m/s. During 350 seconds the pitch control block is tested. As it can be observed Figure 1.13 shows how the active and reactive power change together with the wind.

Figure 1.12: Wind speed applied to each of the WTs

In the time interval [50,150] the wind drops from 11.9 m/s to 4 m/s. The pitch controller is set on power optimization mode and the pitch is kept to zero. Figure 1.13 shows how the active and reactive power of the three WTs starts to be regulated. From the second 150 until 350 the wind speed increase from the cut-in speed to 22 m/s. During this period, Figure 1.13 shows that the PI controllers of the three WTs are activated and start limiting the active power generated by the WTs. The active power for the WF is kept constant to the nominal value, with a small deviation about 1.5 %. The same goes with the reactive power (see Figure 1.13), when the wind speed is greater then the rated wind speed the reactive power is limited to the nominal value. The aim of this study case was to demonstrate that the pitch controlled modelled in Power Factory was correctly implemented. Four simulation were made, three made only for one WT and one for all the WT connected together, and as it can be seen in the resulted graphs the power regulation is accomplished.

11

Figure 1.13: Active power generated by the WTs

1.8

Study Case 2- Active Power Reserve Margin Allocation

Because of the variations of the wind energy, an important issue for the stability of the grid is the allocation of the operating reserve. The supervisory wind farm control (SWFC) has the objective to assure that the active and reactive power outputted by the wind farm comply with the TSO requests. Based on the consumptions from the grid, TSO sends commands to SWFC about the needed active or reactive power from the wind farm. Having the information from TSO, SWFC has to be able to calculate the reserve margin for each wind turbine. The reserve margin for every wind turbine is dierent, function of the wind speed aerent to each of them.[51] This study case analyzes the allocation of the active power reserve margin for one wind turbine at rated wind speed and also for all the wind farm in typically wind speeds conditions.

1.8.1 One Wind Turbine at Rated Wind Speed

It is considered that the wind speed has the rated average value 12 m/s. Figure 1.14shows the variations of the wind speed within the simulation time of 350 s for the rated wind speed. This is the input of the wind turbine system for performing the reserve margin allocation.

12

Figure 1.14: Wind speed variation

The wind turbine outputs active and reactive power function of an operation set point imposed by the SWFC. If no reserve margin is needed, the reference set point is 1p.u. This means that the wind turbine has to produce the rated active and reactive power. For the variation of the wind speed from the Figure 1.14, if the set point for the active power imposed by the SWFC is 1 p.u.(the reserve margin is zero), the wind turbine will produce the rated active power 1 p.u. It is supposed that SWFC imposes dierent reserve margins within the simulation time of 350s. The reserve margin is varying from 0 to 50%. In Figure 1.15 Pref is the active power reference set point calculated by SWFC in order to obtain the desired reserve margin. Pdelivered is the output of the wind turbine when Pref is applied in the rated wind speed condition.

Figure 1.15: Variation of the active power imposed by SWFC

Figure 1.16 shows the variation of the reactive power absolute values function of
13

the active power reference value. If no reserve margin would be applied, the reactive power would have the rated value (1p.u.). When the reference point for the active power is applied as in the Figure 1.15, the reactive power fallows the same variation as the active power.

Figure 1.16: Variation of the reactive power when the active power reserve margin is applied

In Figure 1.17 it can be seen how the power coecient varies in order to obtain the requested active power. The wind speed has small variations around the rated value 12 m/s. The air density and the blade radius are constant parameters. Therefore CP has to be modied in formula ??to obtain the desired value of active power.

Figure 1.17: Power coecient when the active power reserve margin is applied

In direct relation with the power coecient is the pitch angle, CP (, ). The tip speed ratio varies in a small range because of the small variations in wind speed. Thus, the pitch angle, has to be modied to obtain a variable power coecient. As it can be seen in Figure 1.18, the pitch angle is zero in the 0-35 and 40-45 when the wind speed has small variation above the rated value and the reference set point for the active power is 0.98 rst 20 s and it's decreasing in the next 25. The maximum absolute value for the pitch angle is 9o when the reserve margin is largest, 50%.
14

From the simulation time 300s when the reserve margin is 0%, the pitch angle is not going to 0o because the wind speed has small variations above the rated value.

Figure 1.18: Pitch angle when the active power reserve margin is applied

Figure 1.19 shows the how the torque applied at the rotor shaft side of the wind turbine is varying. Based on formula ??, the rotor torque is not reaching the rated value 1 p.u. because of the aerodynamic power and the rotor speed parameters. The maximum average rotor torque value is 0.35 p.u. when the set point for active power is kept to 1 p.u.

Figure 1.19: Rotor torque when the active power reserve margin is applied

The speed of the rotor and the speed of the generator are depicted in Figure 1.20. Both are following the same variations. Between the values in p.u. system of these parameters it is a dierence of 0.007% that occured because of the calculation approximations.

15

Figure 1.20: Speed of the rotor and generator when the active power reserve margin is applied

1.8.2 Wind Farm in Typical Wind Speed Conditions

In this subsection of the reserve margin study case, the behavior of the wind farm is presented when variable reserve margin is imposed and the wind speed is also a variable parameter. The same average wind speed is applied to all three wind turbines as it can be seen in Figure 1.21. The equivalent wind speed for each WT is dierent, therefore the resulted active and reactive power outputted by each of them will not be the same. Two simulation cases were performed:
The reserve margin is zero, meaning that the reference set point is 1 p.u. con-

stant;

The reserve margin is varying in the simulation time due to the reserve margin

commanded by the SWFC.

For both cases the same average wind speed is applied.

Figure 1.21: Wind speed applied to each WT 16

It is supposed the variable reserve margin is imposed by the SWFC. The resulted reference active power from Figure 1.22 is applied to all three wind turbines.

Figure 1.22: Imposed reference active power for all three WT

Figure 1.23 presents the active power outputted by each wind turbine when the reference set point is modied due to the reserve margin modication and also when no reserve margin is applied. At the beginning of the simulation the wind speed is around the rated value and also the rated power is requested. For the same wind speed the reference active power is decreased to 0.8 p.u. (2% reserve margin). Between the simulation time 100-150s, when the average wind speed is lowest (4m/s), the reference point for the active power is increasing from 0.8 p.u. to 1 p.u. and it is kept constant. From the simulation time 250s the average wind speed has the rated value, but the reference set point starts at 300s to decrease from 1 p.u. to 0.6 p.u.

Figure 1.23: Active power outputted by WTs with and without the reserve margin

In Figure 1.24 is depicted the resulted active power of the wind farm for both of the cases, when the reserve margin is requested and also when it is not. Due to the fact that the same reference set point is requested for each wind turbine, the
17

variations of the total active power are function of the wind speeds applied to the wind turbines.

Figure 1.24: Total active power of the WF with and without reserve margin

The modications in active power and by keeping the power factor constant leads to the variation of the demand of reactive power for each wind turbine. The resulted reactive power absolute values for each wind turbine when the reference set point is modied, can be seen in Figure 1.25. In p.u. system, the reactive power has the same variations as the active power when the reserve margin is zero and also when it is modied as in Figure 1.22.

Figure 1.25: Reactive power with and without the reserve margin

From the simulation results it can be observed that the wind velocity should be high enough in order for the wind farm to provide reserve margins. At low speeds the wind turbines are not able to produce rated active power, therefore if the reserve margin is requested, the uctuation of power is not signicant.

18

1.9

Study Case 3 - Reactive power control in PCC

The goal of the following simulation is to verify the reactive power control algorithm implemented in the HVDC-VSC transmission system. The reactive power operating setpoint for the main power grid in the wind farm's PCC is usually calculated by the TSO and sent in real time to the wind farm controller. Afterwards, it is the plant's controller duty to change the operation point of the grid side converter in order to achieve the desired value for the reactive power in the point of common coupling.

1.9.1 Rated Wind Speed

A simulation has been performed in order to verify that the reactive power output from the receiving end station converter is being maintained into the normal operation limits. In Figure 1.26 the reactive power control is described when the wind farm operates at rated active power. A unity positive and negative control signal is being fed to the receiving end station controller, while the reactive power is regulated accordingly. Maximum positive and negative values for reactive power are continuously veried in order to keep the converter at normal operation. The reactive power is measured at the converter's terminals in order to be able to obtain the maximum variation from feeding to consuming reactive power. The only purpose of the forementioned simulation has been to show the limits in which the converter is operating. The following simulations will control the reactive power in the PCC, therefore, the measurement must be performed in this node.

Figure 1.26: Reactive power control at the receiving end station's terminals

As explained in Chapter 2, the reactive power depends on the voltage levels in the PCC and the PWM Converter output. Consequently, changing the amplitude dierence between the PCC voltage UP CC and the converter's voltage UP W M will control the reactive power ow:
if UP CC > UP W M then QP CC < 0, meaning the converter is in reactive power

consumption mode.

if UP CC < UP W M , then QP CC > 0, equivalent with reactive power generation

into the PCC.

The reactive power setpoint required from the TSO has been dened to cover mostly the reactive power generation side, where Q > 0 since these are the typical requirements in real situations. As it is expected, the maximum allowed reactive power
19

value that can be supplied by the grid side converter in nominal operation conditions is also the value at which the control limits its reactive power, therefore, providing stability to the entire system. Figure 1.27 best describes the results obtained in this set of simulations and the conrmation of the reactive power control objectives.

Figure 1.27: Reactive power regulation in PCC at rated power

Following the voltage levels in Figure 1.28 conrms the relation between the PCC and converter voltages and the reactive power production. The voltage on the converter side (with blue) is varying according to the desired reactive power setpoint, while the PCC AC voltage (with red) was constant during the entire time of simulation, thus denoting that this node is modelling a sti grid.

Figure 1.28: Voltage levels in the PCC and AC side of grid converter

The active power sent into the main power grid when operating the wind farm at constant rated wind speed should have a constant value. Also, the DC voltage on the transmission cable is kept almost constant at rated, only with small uctuations due to dierent reactive power setpoints. The power losses on the DC transmission cable have also been monitored. Since neither the DC voltage nor the DC current is modifying its values during the simulation, a conrmation of constant power losses on the DC cables has been obtained with a typical value of 2.2%.
20

1.9.2 Typical Wind Speed Conditions

The analysis of the reactive power output of the grid connected converter when typical wind speed characteristics are applied to the wind turbines provides more insight on how the wind farm is able to supply the energy demand from the TSO. It has also been found a correlation between the average wind speed and the limits in which the reactive power can be set maintaining the normal operation conditions of the system. A typical mean wind speed characteristic has been used in the following set of simulations as shown in Figure 1.29. The reference wind speed (blue) is fed into the aerodynamic model which will output the wind speed (red) that will further on determine the power output of one wind turbine. The analysis was made using three wind turbine systems as it has been done in Study Case 1.

Figure 1.29: Study case 3 - Wind speed and wind turbine power for one simulated wind turbine system

Three wind turbine systems along with their full scale power converters have been used in the model of the wind farm. The wind turbines are fed with constant rated wind speed input values, providing therefore constant rated active power in the system. The resulting wind turbine mechanical power and the total wind farm power output is detailed in Figure 1.30.

Figure 1.30: Wind farm power output versus individual mechanical power output

Figure 1.31 presents the simulation results describing the operation of the grid side converter when various wind speeds are applied to the wind turbines. It can be observed that if low active power is generated to the grid, the converter has the capability of modifying the reactive power generation or consumption to high
21

values, see the 3rd setpoint (75 - 100 seconds of simulation) in Figure 1.31, when the reactive power output is almost the rated apparent power. In the time interval from 200 to 250 seconds, the wind speed is increasing, therefore, the active power generated by the wind farm is also increasing. In this case, the reactive power is continuously adjusted negatively, thus maintaining the balance in the system.

Figure 1.31: Grid side converter operation with dierent reactive power setpoints

More important the AC voltages on the PWM converter side and the PCC side are presented in Figure 1.32. Once more, the close relationship between the reactive power output and the dierence in these two voltages can be observed. In the time interval 200 until 250 seconds, the AC voltage at the PWM converter's terminals is also varying proportionally with the reactive power output.

Figure 1.32: AC Voltages at the converter's terminal and in the PCC

Figure 1.33 shows the AC current imaginary and real components measured on the AC side of the receiving end station. It can be seen that in Figure 1.33, the real part of the current is proportional to the active power ow in the PWM converter and the imaginary part is proportional to the reactive power transfer between the converter and the main grid.

22

Figure 1.33: Real and imaginary current components of the grid side PWM converter

DC voltage levels are presented in Figure 1.34. The DC voltage is being kept within the varying limits dened in the DC voltage controller. The preset variation of 5% can be seen from full power output to low power output, a value that is typically being used in common practice.

Figure 1.34: DC Voltage levels on the HVDC cable

Power losses on the transmission cables have a characteristic similar to the active power output. For the rated power output a typical value of 2.2 percent losses in the DC cable were obtained. In this study case, the control of the receiving end station has been tested. The main purpose was the stable operation of the converter and the possibility of modifying in real-time the reactive power output to preset values given by the grid operator. The most important simulation results have been outlined. It is important to mention that the reactive power control has been tested successfully along with the DC voltage controller that regulates the DC link voltage. Whether the wind farm is operating at low power or full power, the reactive power demand can or cannot be totally fullled. The receiving end station controller is not able to control the active power generated by the wind turbines.

1.10

Conclusions & Future Work

The goal of the project was to implement and simulate a dynamic model of an oshore wind farm connected to the main power grid using an HVDC transmission system.
23

Stability simulations presented in the results section provide a knowledge base for further simulation studies and validate the developed system. The wind farm model represents a starting point for future studies dealing with the grid integration of oshore wind farms. The future work of this project may include: - modelling, implementation and simulation of an aggregated model of an oshore wind farm, consisting of an aggregated model for each of the WT components (wind model, pitch controller, drive train model, generator model); - an in depth research regarding the study of power stability aspects and the impact which large oshore power generating units will have on the main grid; - a comparison between using HVAC and HVDC to connect an oshore WF, from the economical point of view and also taking in consideration the power eciency;

24

Bibliography
[1] Florin Iov. VSC based DC Transmission System for connecting Active Stall Wind Farms to Grid . Ris. Aalborg University 2004 [2] Lars Hojbjerg Nielsen. Grid Development and Integration. EcoGrid Kick O. May15.2007. Energinet.dk. [cited 2008 October 8]. Available from: http://www.energinet.dk/NR/rdonlyres/9F295A1F-5621-4714-B8709EA2B2E1324D/0/05GriddevelopmentandintegrationbyLAN.pdf [3] Econnect Project No: 1312. Study on the Development of the Oshore Grid for

Connection of the Round Two Wind Farms

[4] Florin Iov, Remus Teodorescu, Frede Blaabjerg. Grid Code Compliance of GridSide Converter in Wind Turbine Systems . Institute of Energy Technology. AAU, Denmark [5] ABB AB, Grid Systems - HVDC. It's time to connect - Technical description of HVDC Light technology. Available from www.abb.com/hvdc [6] Morten H. Hansen, Anca Hansen, Torben J. Larsen, Stig ye, Poul Srensen and Peter Fuglsang. Control design for a pitch-regulated, variable speed wind turbine .Ris-R-1500(EN). Ris National Laboratory. Roskilde Denmark. January 2005. ISBN 87-550-3409-8 [7] S. Douangsyla, P. Indarack, A. Kanthee, M. Kando, S. Kittiratsatcha and V. Kinnares - Modeling for PWM Voltage Source Converter Controlled Power Transfer, 2004. International Symposium on Communications and Information Technologies, Sapporo, Japan [8] Cuiqing Du , The control of VSC-HVDC and its use for large industrial power systems, 2003. Department of Electric Power Engineering CHALMERS UNIVERSITY OF TECHNOLOGY - Goteborg, Sweden. ISSN 1651-4998 [9] DIgSILENT GmbH, Markus A. Poller, Doubly-Fed Induction Machine Models

for Stability Assessment of Wind Farms

25