A Detector for Modifying the Aggregated Wind Farm Model According to the Running or Tripping Wind Turbines in a Wind Farm
Most studies and researches [5, 9, 10, 13, 15, 16] have focused on compiling all the wind turbines in a wind farm considering that all turbines are in service. For instance, if the WF consists of 100 turbines, they aggregate the 100 turbines without taking in mind that some of these turbines may be out of service due to high incoming wind speed, three phase short circuit, one or more units in maintenance and etc. No researches have been presented in case of any number of turbines was shut down during the operation, which means that off wind turbines should be taken off from the aggregated model, where every turbine took off from the whole system should be removed from the AWF model. An aggregated model can’t determine some wind turbines trip and others do not in a large wind farm. This means that they have to stop the operation (simulation) and take off the separated wind turbine from the AWF model. Suppose also that some wind turbines became out of service at a certain time and after time, some of these turbines come back in service. This leads to stop the running simulation every period to update the status (on/off duty) of wind turbines in the AWF model.
The aggregation technique (MFAM_EWS) has concluded from chapter two that has the highest accuracy in approximation to the whole WF comprising different wind turbines (different rated power). This chapter is working on this type of aggregation technique (MFAM_EWS). It works on linking the complete (detailed) wind farm and the aggregated model (MFAM_EWS) by a detector as illustrated in the following section.
3.2 The Proposed Detector
As shown in Fig.(3-1) the wind farm is laid out in a matrix with i rows and j columns. The detector detects the running wind turbines in WFs and sends this information to the aggregated model. For high incoming wind speed (exceed cut off wind speed), the wind turbine in a WF is separated, so the detector detects the position of this separated turbine in which rows and columns and separated the speed and power curve of the same wind turbines in the AWF model by sending a signal to the AWF model. In this way, we are saving time and effort to stop the execution (simulation) and modulate the wind turbines status in the AWF model by ejecting the turbines, which was off duty and taken out during operation. With this detector, we can simulate the whole wind farm without the off units or with the units that returned back to service.
Fig.(3-1): The function of the detector
If uf,i,j for one of the farms changed there is a variation in uf,i,j to u’f,i,j. The output power Pf,i,j changes to P’f,i,j where P’f,i,j may be zero for u’f,i,j < uf cut in or u’f,i,j > uf cut off, where f refers to number of farms, i refers to the rows and j refers to the column in the farm. The detector scans wind turbine matrix if the scanned output is zero, then nothing changes. During the scan, if one wind unit or more have changed its status, then the detector should enter the changed speed to the aggregated model to find a new output of the aggregated system by ejecting the power curve and speed of the tripping wind turbines. Fig.(3-2) presents a flowchart of the proposed detector.
Fig.(3-2): Flowchart of the proposed detector
This chapter presents two wind farms composed of PMSG and SCIG wind turbines. The system composed of ten of wind turbines of the type PMSG and ten of wind turbines of the type SCIG. Fig.(3-3) shows the connected schematic of this wind farm. Chapter 2 studied the detailed model of PMSG wind turbine. In this chapter, we first study the detailed model of SCIG wind turbine then show the simulation studies of the proposed detector.
Fig.(3-3): Wind farm structure
3.3 SCIG Wind Turbine Model
The first type of wind generators is fixed speed where squirrel cage induction generators (SCIG) are often used. That kind of generators is characterized as robust, easy to use, and cheap; there are no electrical connections between stator and rotor system.
SCIG wind turbine is the simplest type of wind turbine technology. It has a turbine that obtains mechanical energy from the kinetic energy of the wind, then turns it into electrical energy and delivers the energy directly to the grid. The windings of stator directly are connected to the grid while the windings of the rotor are short-circuited.
Due to its structure, an induction generator consumes a significant amount of reactive power, which increases along with the active power output. So, a capacitor bank must be provided in the generator terminal in order to compensate reactive power consumption so the generator doesn’t load the grid.
The SCIG wind turbine mechanical power is directly turned to a three-phase electrical system, so no complex controller in the electrical part of it for stall regulated wind turbine. However, for pitch regulated wind turbine, a pitch controller is used to regulate the pitch angle of the turbine . The schematic structure of SCIG wind turbine is described in Fig.(3-4).
Fig.(3-4): Configuration of SCIG wind turbine
The model of generator (SCIG)
With neglected stator transients, The SCIG is represented by the third order transient state model of the induction generator (IG), as a usual for representing the IG on studies of power systems transient stability . For per unit model of generator, the differential equations of are expressed as:
(〖de〗_d^’)/dt=-1/(T_o^’ ) (e_d^’-(X_s-X_s^’ ) i_qs )+sω_s e_q^’ (3-1)
(〖de〗_q^’)/dt=-1/(T_o^’ ) (e_q^’+(X_s-X_s^’ ) i_ds )-sω_s e_d^’ (3-2)
T_e=〖1/ω_s (e〗_d^’ i_ds+e_q^’ i_qs) (3-3)
SCIG active and reactive powers can be obtained by the following equation:
P_e=e_d^’ i_ds+e_q^’ i_qs (3-4)
Q_e=e_d^’ i_qs-e_d^’ i_ds (3-5)
T_o^’=(X_σr+X_m)/〖ω_s R〗_r , X_s^’=〖(X_σs+X〗_m)-(X_m^2)/(X_σr+X_m ) (3-6)
Taking into account that the wind turbine of SCIG employs local capacitors for compensating, so the total current transmitted from the SCIG wind turbine at the generation node is given by:
i_dg=i_ds+i_dc=i_ds+1/X_c u_qg (3-7)
i_dq=i_qs+i_qc=i_qs-1/X_c u_dg (3-8)
Fig.(3-5): SCIG wind turbine equivalent circuit
At the PCC of a wind farm, the active and reactive powers generated from the wind turbine can be expressed as:
P_g=u_dg i_dg+u_qg i_qg (3-9)
Q_g=u_qg i_dg-u_dg i_qg (3-10)
The equivalent circuit of the generation system presented in Fig.(3-5) including compensating capacitors and the induction generator.
Pitch angle controller
With regard to the pitch regulated SCIG wind turbines, it contains a controller for the pitch angle for limiting the power to the rated which generated from the wind turbine when exceeding the rated wind speeds. As depicted in Fig.(3-6), to set the reference of pitch angle, the control scheme employs a PI controller .
Fig.(3-6): Pitch angle control
3.4 Simulation Studies
As shown in Fig.(3-3) the wind farm is composed of smaller wind farms with different types of wind turbines. These wind turbines in the wind farm are modeling by MATLAB/SIMULINK. The capacity of this farm is 21.6 MW consisting of 10*1.5 MW PMSG and 10*660 KW SCIG wind turbines. The rated speed of each PMSG and SCIG wind turbine with ωr =1 are 11m/s and 15m/s respectively. These smaller wind farms are coupled to a 66 kV distribution system then exporting power to a 220kV AC grid through a 30km transmission line. We compare between the whole wind farm and MFAM_EWS, which is used due to the farm consisting of different wind turbines as concluded from the previous chapter. The reactive and active power exchange between the WF and power system are the variables taken into consideration in the comparison. SCIG wind turbine main parameters are provided in Appendix F while Appendix G depicted the speeds of the wind received by the PMSG and SCIG wind turbines. This chapter is working in four cases.
3.4.1 Case 1 : Normal operation
If all wind turbines in the complete wind farm are running. Comparison between the complete wind farm and the MFAM_EWS (the type of AWF model) during operation (running all wind turbines) are described in Fig.(3-7).
Fig.(3-7): Comparison during all running wind turbines at the PCC
Fig.(3-7) compares between the complete WF and MFAM_EWS in active and reactive power at the PCC. The figure depicted that MFAM_EWS has a better approximation to the detailed WF model. This aggregated model (MFAM_EWS) can represent the whole system with a reasonable accuracy and reduce computation time. Where the whole WF executes with 540 sec and aggregated model executes with 26 sec.
3.4.2 Case 2: tripping some wind turbine without a detector
Now we suppose that the wind speeds incident on each wind turbine without fluctuations and during the simulation three of SCIG wind turbines become out of service due to high incoming wind speed (exceed the cut-off wind speed) at t=5 sec and two of PMSG wind turbines become also out of service at t=10 sec, then three of PMSG wind turbines becomes off duty at t=15 sec. Fig (3-8) shows a comparison between the complete wind farm and AWF model.
Fig.(3-8): Comparison at PCC without a detector
From the previous figure, it is clear that the AWF model is closer to the whole wind farm, but only during the period of operation of all wind turbines, while when some of the wind turbines are shut down, the difference became very clear between the AWF model and the output of the wind farm. This difference is growing with increasing of separated wind turbines.
When the wind speed exceeded its limit (cut-off speed) in some turbines, the complete WF model could separate these turbines from the farm, as the output power of the complete model decreased as shown in Fig.(3-8). However, in the AWF model, it could not separate these turbines, as the output power of the AWF model increased with the excess of the speed of these turbines; the energy did not decrease like what happened in the complete model.
3.4.3 Case 3: tripping some wind turbine with a detector
The main goal is to change the wind turbines position of in the AWF model from time to time according to the separated turbines without stopping the running simulation. It consumes time to take off the separated turbine from the AWF model. The main objective of this study is that the operation will continue in the case of all turbines in the operating state, or some of them have been separated. So this paper uses a detector to link between the whole wind farm and the AWF model. This detector sends numbers and data (wind speed and power-speed curve) of all wind turbines in the complete WF, data of separated turbines and data of running turbines. Without stopping the run simulation, it can modulate the wind turbines position of in the AWF model. The following figure shows the function of this detector.
Fig.(3-9): Comparison at PCC with a detector
Fig.(3-9) clarifies that the detector may possess the ability to separate the turbines in AWF model that separated during the operation of the whole wind farm, without the need to stop the simulation.
3.4.5 Case 4: retrieve some wind turbine with a detector
From Fig.(3-8) the output power of the wind farm is about 10.24 MW with all wind turbines running and about 6.4 MW after some of the wind turbines are separated as shown in Fig.(3-9). Now, I suppose that some of these separated wind turbines are coming back to the operating condition again. I want to see the reaction of the detector for these returned wind turbines. Suppose that two of PMSG turbines are coming back to service after eight of wind turbines are tripping (five of PMSG and three of SCIG wind turbines). Fig.(3-10) shows the ability of the detector in this case.
Fig.(3-10): Comparison at PCC in case 4
Fig.(3-10) shows the response of a detector about the returned turbines to the operation again. It has the ability to retrieve the separated wind turbines in the AWF model according to the data received from the complete wind farm. So the detector will modify the status of turbines, whether within the operating range or not, as well as be able to retrieve the turbine into operation again without any need to stop the execution to modify the position of wind turbines in the AWF model.
Performance Analysis of SMES Integrated with Offshore Wind Farms to Power Systems through MT-HVDC
There are many benefits of offshore wind farms (OWFs) compared to conventional onshore wind farms. Such as, higher wind speeds, allowing the larger size of wind turbines and the unlimited available locations in the sea to install new wind farms. So, in the recent years, the number of offshore wind farms has increased. The output power from OWFs connected to the main AC network using transmission systems based on AC or DC technology. The selection of these technologies depends on the cost of the installation, the transmission distance, and power output.
Long distance cable connections HVDC technology is the best solution for OWFs. Regarding DC transmission systems, There are two possible technologies: VSC-HVDC or LCC-HVDC (HVDC classical) technology . The main purpose of this chapter is to study the connection of an OWF with the main grid through VSC-HVDC. For large scale offshore wind generation, multi-terminal DC based on voltage source converter (VSC-MTDC) becomes an attractive solution. By VSC-MTDC, the wind farm can be connected to more than one onshore grid that may or may not be synchronized to each other . It is capable of supplying both active and passive networks where the control of the converter station that directly connects the active system usually is in constant DC voltage control mode and the control of the converter that directly connects the passive system is in constant AC voltage control mode .
Recent years energy storage systems are becoming popular for transmission and grid applications. In the market, there is a variety of storage technologies , one of them is superconducting magnetic energy storage (SMES) as it has faster response time than any other storage systems. The important section of SMES system is the superconducting coil, which is placed in a cryostat or dewar consisting of a vacuum vessel and a liquid vessel. Liquid vessel protects the system temperature by saving proper cooling setup cryogenic system, also keeps the temperature below the critical temperature .
SMES is a device like a grid that stores or discharges large quantities of power almost instantaneously. The system is capable of compensating high levels of power during sudden loss or dip in line power. The capacity of the SMES unit is based on the application and charging/discharging duration.
Various applications of the SMES are power quality improvement, custom power, stabilization, voltage/VAR control, load leveling, dynamic response, minimization of power and voltage fluctuations, and frequency control application . It has also the ability to retain the grid active power stable in the face of any kind of disturbance that may occur in the power system since this might extend to the grid and affect or even damage other power devices.
This chapter describes offshore wind farms consisting of permanent magnet synchronous generator (PMSG) wind turbines and doubly fed induction generator (DFIG) connected to Active network (AC grid) and Passive network (loads) through Multi Terminal High voltage direct current(MT-HVDC) transmission system. This chapter discusses also the effect of using a Superconducting Magnetic Energy Storage (SMES) unit in a hybrid power system contains OWF. In this chapter, we have aggregated 150 wind turbines of 1.5 MW PMSG and 150 wind turbines of 1.5 MW DFIG using an aggregation technique (MFAM_EWS). We have used a detector to detect any tripping of any wind turbine and substitute the shortage of power due to this loss of wind turbines immediately through SMES. The Active network in this chapter should have a minimum of 150 MW power to be supplied by controlling the SMES unit (absorbing or providing power according to the system requirement). MATLAB/SIMULINK program is used to carry out this system to test the effectiveness of the SMES unit during tripping some of the turbines, fluctuation in wind speeds, load change and voltage dips.
This chapter showed the system description firstly then studies each part separately then shows the simulation study of the system.
4.2 System Description
As shown in Fig.(4-1), the system consisting of 450 MW OWFs merging 150 * 1.5MW DFIG wind turbines and 150 * 1.5MW PMSG wind turbines, which are aggregated at the common bus of 145KV through 30 km offshore cable and 690V/145 KV step up transformer. They are aggregated by using MFAM_EWS as concluded in chapter 2. The output generation power is transferred through MTDC transmission system to the 380KV AC grid (active system) with 50 Hz frequency and three phase pure resistive load (passive system) by using submarine DC cables where the parameters of transmission line for AC and DC cable are given in Appendix H. The rating of the VSC-Based HVDC Transmission Link is 500 MVA (450 MW/0.9), +/- 200kV.The SMES unit is coupled to the DC side system of the WF via DC-DC chopper. It is located between the OWFs and the AC grid and a load as an interface device.
Fig.(4-1): Offshore wind farm with HVDC system and SMES
Voltage Source Converter Model
Fig.1 shows transmitted power of wind farm to the grid through three-level VSC based HVDC system. VSC-HVDC transmission system basically consists of converter transformer, 3 level voltage source converter, shunt AC filters on the AC side, DC line capacitor and phase reactor as shown in Fig.(4-2). The high-frequency filters are used for blocking higher frequency harmonics. The converter reactor and transformer leakage reactance can control the converter’s output terminal voltage and the output power. On the DC side, the DC capacitor acts as the DC voltage source to maintain the power balance between the AC and DC power. The VSC is modeled as an average value model as presented in .
Fig.(4-2): Components of converter station
In HVDC transmission system, there are two voltage source converters one acts as a rectifier connecting to offshore wind farm and the other acts as an inverter connecting the AC grid. The control strategies of each one are depicted in part 4.3.2 and 4.3.3.
4.3.1 Mathematical model
Conversion of abc to rotating dq reference frame, the converter can be mathematically modeled in terms of decoupled direct and quadrature converter current components id and iq respectively :
L (di_d)/dt=-Ri_d+ωLi_q-u_dconv+u_d (4-1)
L (di_q)/dt=-Ri_q+ωLi_d-u_qconv+u_q (4-2)
where udconv and uqconv are used in order to control the converter currents id and iq.
The relationship of the power balance between the dc output and ac input is given as:
p=3/2 (u_d i_d 〖+u〗_q i_q )=V_DC i_DC (4-3)
The vector of the grid voltage is known to be along with the direction of d-axis, so a virtual grid flux vector can be supposed to be acting along with the direction of q-axis. With this alignment, uq = 0 and the active and reactive power absorbed from or injected into the AC system are given by :
p=3/2 u_d i_d (4-4)
q=-3/2 u_d i_q (4-5)
Hence, the current are split into two portions according to the conversion to rotating d-q coordinate system oriented with respect to the vector of the grid voltage. The first portion determines the contribution which gives desired power flow into the DC side and the second portion defines the condition of reactive power. The possibility to control the components of two current independently is shown in the previous equations.
4.3.2 VSC connected to OWF and its controller
The aim of the offshore VSC is to transmit the active power generated by the OWFs and to set a voltage reference for the wind farms. This is carried out by using an AC voltage controller consisting of a PI controller as shown in Fig.(4-3). A fixed nominal frequency (f) is supplied to the offshore VSC output voltage .
Fig.(4-3): Control scheme of Offshore VSC
4.3.3 VSC connected to AC grid and its controller
The controller of onshore VSC has the objectives to regulate the DC voltage and reactive power exchanged with the AC grid. In order to obtain a decoupled control of reactive and active power, the vector control is applied . As shown in Fig.(4-4), the control scheme of the Onshore VSC consists of a phase-locked loop (PLL), an inner controller for the current, a limiter for this current, and two outer controllers. The objective of the inner current controller is to track the current reference values generated by the outer controllers, to obtain the reference of converter voltage values (ud* and uq*).
Fig.(4-4): Control scheme of Onshore VSC
SMES System and Controlling
The SMES unit is one of the significant storage system solutions. SMES is a device that stores energy in the magnetic field which is formed by a DC current flowing through a large superconducting coil. The ability of this coil is to retain the magnetic energy for a long time with almost no losses this due to the coil has cryogenically become cool to a below temperature its superconducting critical temperature. This means that during operation ohmic losses will be very low, close to zero .
The stored energy (ESMES) and rated power (PSMES) in the coil are:
E_SMES=1/2 L_SMES I_SMES^2 (4-6)
P_SMES=(dE_SMES)/dt=V_SMES I_SMES (4-7)
As shown in Fig.(4-5) the SMES unit consists of two main parts, DC-DC chopper (Type – D chopper)  and the superconductor coil, which has an extremely low resistance.
Fig.(4-5): SMES unit components
The configuration of the type D chopper with SMES coil is shown in Fig.(4-5). When the two choppers (IGBT 1, IGBT 2) are ON the SMES is in charging mode and the relationship between the DC link capacitor and the SMES coil voltage is expressed by:
When the two choppers are OFF and diodes D1 and D2 start conducting, the SMES is in discharging mode and the relationship between the voltages are:
The SMES coil is in standby mode (freewheeling mode) when one of the two choppers is ON and the other is OFF. During this operation mode, the DC current is continuously circulating in a closed loop through the SMES coil with no significant loss due to low resistance.
As shown in Fig.(4-6), the control scheme of the DC-DC chopper. It is designed to make the grid stable without providing any power to the load during any disturbances. The coil of SMES is discharged or charged by adjusting the average (i.e., DC) voltage across the coil to be negative or positive values by means of the duty cycle of DC-DC chopper. It compares the required reference power of AC grid and the measuring power on the grid bus to PI controller. According to parameters of PI controller, it reduces the wind generator output power fluctuations as an effect of the wind speed variability. The power magnitude and direction exchanged between the AC grid and the coil of SMES are determined by the duty cycle (D). When D is larger than 0.5 the coil is charging and when the duty cycle is less than 0.5 the coil is discharging .
Fig.(4-6): Control of the chopper
This chapter works in connecting PMSG and DFIG OWFs to active and passive networks through MTDC systems based on VSC. Normally, the passive network will have first priority over the other terminal (active network) for the power generated by the OWFs. If the power generated of OWFs is not enough for the load demand (passive network), active power will supply the residuum power to the load demand. This paper aims to make the grid (active system) stable without providing any power to the passive network (load) during any disturbances like tripping some of the wind turbines which are verified in chapter 3 or increasing demand load. This is done by putting SMES unit in the DC link in the MTDC transmission system. It also shows the importance of SMES unit in compensating the output power fluctuation which produced from fluctuations in wind speeds. Various problems may be caused from these fluctuations in wind speeds such as frequency and voltage oscillations when a large number of generators wind power are coupled to the main AC grid.
To evaluate the performance of the SMES in the power system, the system in Fig.(4-1) was performed by Matlab /Simulink with four different instability cases: wind speed fluctuation, tripping some of the wind turbines, changing of the load demand and voltage drop. The specification of the SMES unit and the power system in the design system are given in Appendix I. In this paper, the system is designed to be able to supply the required power to the grid during normal or abnormal conditions. This means that the grid is not providing any power during any kind of disturbance that may occur in the power system. This is done by controlling the duty cycle DC-DC chopper connected to the SMES coil as depicted in Fig.(4-6). In this article, the system is designed to supply 150MW to the grid. The simulation studies the performance of the SMES coil integrated with the HVDC system into the grid during steady state and transients.
Case 1: Wind speed fluctuations
The fluctuations of the wind speeds cause fluctuation in the output power of the OWFs as shown in Fig.(4-7) (blue line). These fluctuations are transferred to the AC grid and may cause interruption to the power system. But, using SMES coil it can absorb these fluctuation and supply power to AC grid without any fluctuation as shown in Fig.(4-7) (magenta line) where the DC -DC chopper is controlled to provide 150 MW to the active system. The cyan line in Fig.(4-7) is the power to load demand (passive network). Fig.(4-8) shows the current in the SMES coil. It is obvious from this figure that SMES coil can absorb the fluctuations. So the SMES in DC link can stabilize the power system where the large variations do not reach the grid. In such a way the smoothed powers reach the customers.
Fig.(4-7): Active power in case 1
Fig.(4-8): SMES current during case 1
Case 2: Load change
In this case, the system is tested when increasing the load power supposing that the wind speed is steady. Firstly, we make the load demand is 50 MW and the load is increased to 250MW during 3 Sec at time t= 6 Sec and for 1 sec at time t= 15 Sec. According to the designed system the grid power does not provide any power in this case. This occurs in Fig.(4-9) the grid power is still in 150MW. The SMES coil can supply the load demand during the disturbance period as shown in Fig.(4-10). It absorbs surplus energy and during increasing in load it releases the required power to the load.
Fig.(4-9): Active power in case 2
Fig.(4-10): SMES current during case 2
Case 3: Tripping out some turbines
By using a detector in the system to detect any of wind turbines is tripping out or return back to service during simulation. In the case of tripping out some of the wind turbines, the SMES begin to export power to compensate energy shortages and when the wind turbines become in service the SMES charges. It is assumed that 60 PMSG wind turbines and 45 DFIG wind turbines are tripping out from OWFs for 2 Sec starting at time t=13 Sec as shown in Fig.(4-11). The SMES coil can supply power to demand load and trying to make the grid power stable as shown in Fig.(4-12). It shows that SMES is discharged for 5 second this is due to the time is taken to reach the previous value of wind power and charge again.
Fig.(4-11): Active power in case 3
Fig.(4-12): SMES current during case 3
Case 4: Voltage dip
In this case, the system is tested when the voltage of the AC grid (active network) is a dip. It is supposed that the AC voltage is dropped to 0.5 p.u. for 500 ms at period 10-10.5 Sec. When the voltage dips on the grid side converter terminal bus, the power transfer capability of the grid is reduced. In such a case the wind farm may be commanded to reduce the power generation. Any excess power fed into the dc link would result in DC over voltage as shown in Fig.(4-13) where it shows the system behavior without the SMES. The system behavior with SMES is shown in Fig.(4-14) shows the stability of DC voltage at 1 p.u. It shows the effectiveness of SMES coil in stabilizing the system.
Fig.(4-13): DC voltage without SMES
Fig.(4-14): DC voltage with SMES