Small Signal Stability Analysis of Jepirachi Wind Park.: Y. Ulianov S. Laín
Small Signal Stability Analysis of Jepirachi Wind Park.: Y. Ulianov S. Laín
Small Signal Stability Analysis of Jepirachi Wind Park.: Y. Ulianov S. Laín
Y. Ulianov
Department of Energetic and Mechanical Universidad Autnoma de Occidente Cali, Colombia yuri.lopez@ieee.org
S. Lan
Department of Energetic and Mechanical Universidad Autnoma de Occidente Cali, Colombia slain@uao.edu.co
AbstractThe actual development of wind energy in several countries around the world become it as the fastest growing power generation technology between green and clean technologies. Almost 50% of the installed wind turbines are induction generators. With the increase of induction instead of synchronous generators, electrical networks are experiencing different disturbances, which require permanent transient and small signal stability studies. This article presents an analysis of a small signal stability study in the first wind park installed in Colombia. In this analysis a software VORTEXOFT developed by the authors is used for aerodynamic purposes. A modal analysis and eigenvalues were done considering high level of wind power penetration. The results obtained indicate that the proposed method effectively identifies the stability operational point for a high level of wind power. Index TermsDFIG, modal analysis, wind turbine.
"north-east wind". This is the first ever wind park feeding to the electricity grid in Colombia. This study, allows identifying the impact of this technology both on the connection node and Colombian electric system stability. Up to now, none of those studies has been neither carried out at Jepirachi, nor published. However, the wind power behavior and previous operational experiences has been reported [3]-[5]. Wind energy generation in the most promising location in the Caribbean sea cost in Colombia, has exceeding typical values reported in the wind energy literature [3]. All these factors and the need of developing wind power technology in Colombia, unveils that stability studies are important. II. STABILITY ANALYSIS Small signal stability is defined as the capability of a power system to return to a stable operating point or to the original steady, after the occurrence of a disturbance that leads to an incremental change in one or more of the state variables of the power system [1], [6], [7]. In this work, the small signal stability problems for wind power generation will be researched and how they impact on a power system. All this work begins on the space equation and the output equation which give the necessary information.
I. INTRODUCTION Today high levels of wind power and its penetration on electrical grids, are being considered in several countries, requiring very deep analysis connection. Research about it is being carried out. Knowing previously the dynamic operation of wind turbines makes important to design, model and simulate their behavior prior to install them. By this way, one of the main studies realized for new wind farms are dynamic and stability studies that professionals, researchers and specialists are developing constantly due to the fast growth of the wind turbine industry and the way more wind farms are aggregated to a power systems. For instance, studies about small signal stability analysis on wind turbines (WT) with direct-drive permanent magnet generators (DDPMG), with and without controllers shows system stability [1]. High penetration levels of squirrel cage induction generators (SCIG), and doubly fed induction generators (DFIG), have also been studied, showing that adequately controlled, WTs can contribute to improve the frequency dynamics and power system stability [2]. So far on literature, basically two main small disturbance stability problems have been widely discussed and by that, are still clearly and widely accepted such as rotor angle stability and voltage stability [1]-[4]. But none at Jepirachi wind park. Jepirachi ("Jepirachi" is a word from the language of the Wayuu Indians living long time ago in the area and means
x (t ) = f ( x , u , t ) y(t) = g(x, u , t )
(1)
Where equation (1) has all state variables, u is the input variables, t the time and y the output function. The linearization of (1) allows investigating the response to small variations. To do this, and considering that state functions contains some different variables on its polynomial equation, these equations are developed in Taylors series expansion in which with higher orders of the derivatives omitted because of its minimum interaction to the response. The linear result is a system presented in equation (2) [10]:
x = Ax + Bu y = Cx + Du
(2)
III. CHARACTERISTICS OF JEPIRACHI WIND PARK. Where A, B, C and D are obtained from the Jacobian matrix, containing the partial derivatives of the functions in f and g respectively to the state variables x and the input variables u as: Jeprachi wind park has 15 wind turbines (1.5 MW), for a 19,5MW total power capacity. It was inaugurated on December 21st 2003 and began operation at 2004 with an investment of US$28 million. Nordex was responsible for supplying, installing and commissioning of the turbines. At 2006 the German company Nordex Energy GmbH left the project and the assessment. The farm generates 82 million kWh of energy each year. The electricity is fed into the national grid via the Jepirachi transformer, where it enters the 110 kV high-voltage line located only 700 meters away [13]. At a height of 50 meters the average wind speed is around 10 m/s. Some blades data is shown in Table 1. Table 1. Wind turbine blades data [13].
NORDEX N60 1300 Data. Blades Blade radio Rotational speed of rotor Air density Cut in wind speed Nominal speed Cut out wind speed Blade type 3 30 m 12.819.2 RPM 1.225 kg/m3 34 m/s 15 m/s 25 m/s LM 29.0
f1 f n f1 f1 x ... x u ... u n r 1 1 A = ... ... ... B = ... ... ... f n ... f n f n ... f n x1 u1 xn ur g1 g1 g1 g1 x ... x u ... u n r 1 1 C = ... ... ... D = ... ... ... g m ... g m g m ... g m x1 u1 xn u r
(3)
Where A is the state matrix, B is the input matrix, C the output matrix and D coefficients matrix. These matrices can be transformed using Laplace to find the state equation linearized, in the frequency domain. Finally to obtain the eigenvalues it must be found the values of s that satisfy [10]:
det( sI A) = 0
(4)
IV. MODEL OF THE DFIG Induction generators are the most commonly used machine for wind power generation. The rotor terminals are fed with a symmetrical three-phase voltage of variable frequency and amplitude. To create and regulate this voltage, a voltage source converter is used, usually based and equipped with IGBT. The basic structure is shown in Fig. 1.
The n-solutions of = 1 , 2 , n are the eigenvalues of A. These eigenvalues can be real or complex [7]. In this paper, PSAT simulation software [11], is used to analysis the eigenvalues. Roots positions and its values as they appear on this plot show if system is stable or not. As can be seen all real values are negatives which means system is stable under this conditions. The full eigenvalues report shows results in matrix and also participation factors are presented. The electromechanical modes generated for small signal stability studies occur in the frequency range of 0.1 to 2 Hz. Interarea oscillations, are typically in the frequency range of 0.1 to 1 Hz. The interarea modes are usually complex and associated with groups of machines swinging to other groups across a transmission line. The higher frequency electromechanical modes defined as local modes are in the order of 1 to 2 Hz, typically involve one or two generators swinging against the rest of the power system [12]. The power system models in this project were all built using SIMULINK and the specialized Power Systems Analysis Toolbox, PSAT a software application developed for MATLAB, which performs both the numerical simulation and linearized eigenstructure analysis. PSAT 2.1.6 version by Federico Milano (c) 2002-2010 [11].
Slip rings connected to the wound rotor in this DFIG, take current into or out of the rotor. The crowbar (CB), is a resistance or group of resistances connected to the rotor circuit for a short period for de-energizing the machine while the converter is disconnected and it also protect the machine against voltage limits. The pulse wide modulation block
(PWM), includes the power control system IGBT based. In this type of control, active and reactive currents and thus P and Q of the DFIG are controlled by the rotor side converter independently. The voltage source converter (VSC), is a backto-back system connected in a DC-link.
The total value of the axial thrust, T, and torque, M, are, therefore:
T = K T q ,
q =1
M = K Qqr q
q =1
(9)
V. AERODYNAMIC ANALYSIS OF THE TURBINE A compromise between accuracy and computational cost for prediction the aerodynamic characteristic of a HAWT has been found defining an enhanced non-linear Lifting Surface Method (LSM) blade [17]. LSM is the most advanced technique in Blade Element Hypothesys context (Prandtl hypothesis applied to rotating blades) taking in account chord wise and spanwise circulation and introducing information on airfoil camber line. The main features of the computational strategy are: combined, 3D Interacting twisted Lifting Surface-/2D Vortex Panel Method; prescribed cylindrical wake with constant wake pitch relaxed iteratively; effective incident velocity and angle of attack at each section are found which allows computation of the aerodynamic characteristics of the HAWT [16]. Under the Lifting Surface approach, the absolute nonstationary irrotational velocity field is computed in terms of upstream velocity, V, plus a perturbation due to velocities induced by the vorticity field. It is assumed that vorticity is concentrated at thin blades boundary layers and wakes, modeled as vorticity sheets. Vorticity associated to thin boundary layers can be merged in to a single vortex surface, the Lifting Surface. Lifting Surface geometry is a cambered twisted continuous surface build up by the sequence of "profile cambered middle lines" going from the leading edge to the trailing edge. This approach allows obtaining an effective velocity and angle of attack at each section of the blade. Once the effective relative velocity Wq and angle of attack iq is known for each blade section, assuming two-dimensional irrotational flow, the experimental aerodynamics gives us the expressions for the lift and drag forces on each profile
where r q is the distance from the section to the rotation axis. The corresponding axial thrust, torque and power coefficients are defined as:
CT = 2T
(10) is the tip speed ratio defined as = R/V where is the rotational speed of the rotor and R its radius. The afore described computational method has been implemented in the software VORTEXOFT developed by the authors and the wind turbine behavior presented on Figure 2..
AV2
, CM =
2M
ARV2
, CP = CM =
2P
AV3
Fig. 2. Power coefficient versus tip speed ratio for the NORDEX N60 1300 This computation strategy has been validated previously [16], [17], comparing calculated power coefficients versus manufacturer data for two- and three-bladed horizontal axis wind turbines. The nominal power of these HAWTs ranged from 1 kW to 3 MW. The turbine considered in this study is NORDEX N60/1300 Kw which has a rotor diameter of 60 m and a nominal power output of 1300 kW. The blade is LM 29 whose construction is based on the family of profiles NACA634XX. The actual computations are based on the configuration reported in [5]. Figure 2 shows the comparison between the obtained computational results with VORTEXOFT and the manufacturer data. The agreement is very good in the range of tip speed ratios between 2.5 and 7.3, which covers the most common operating conditions of the turbine. This data are the input for the small signal stability study module. VI. SIMULATION A model of the wind park developed with PSAT is showed in Figure 3. The entire park is simulated as generator connected to an infinite bus sytem.
Lq =
2 1 C LW q b q 2
Dq =
2 1 C DW q b q 2
(7)
Provided that the lift, CL(iq), and drag, CD(iq), coefficients are known for the specific profile considered. Here bq represents the profile chord length. The resultant aerodynamic force Rq is the vectorial sum of the lift Lq and drag Dq. The projection of Rq on the rotation axis gives the contribution to the axial thrust T q and the projection on the rotation plane the corresponding tangential contribution Q q. These can be expressed as:
(8)
the state variable mechanical speed of the DFIG, theta_p are the indexes of the state variable p. Finally, idr and iqr represent direct and quadrature currents of the DFIG model In a full representation, conjugate eigenvalues are still complex numbers ( = j w ), and each pair corresponds to an oscillatory mode, where the real part gives the damping and the imaginary gives the frequency of oscillation [2]. In this case these negative complex eigenvalues represents a damped oscillation. The frequency column on Table 2 is the oscillation frequency (Hz), where
A power flow analysis in 0.057 s with the PF solver: Newton-Raphson method is carried out before the stability analysis is completed. After iterations, a Maximum Convergency Error of 5.3321e-010 is obtained. Then the initialization is ready and this system is complete. Once the eigenvalues is run, five eigenvalues appears and are presented on Table 1. According to the theory, the analysis of those numbers allows to define whether Jepirachi is stable or not with the wind turbine used and the meteorological conditions in the park. After Then, the eigenvalues result is presented graphically on Figure 4. Visually only three (3) points appears on Figure 2, but this is just to show the largest eigenvalue. In this results as can be seen on table 2, eigenvalues that are complex numbers have only the real part. Between the state variables, the next five are relevant for this simulation. The variable name and its values are listed below.
Eigenvalues of 19.5 MW Jepirachi 1 0.8 0.6 0.4 0.2 Imag 0 -0.2 -0.4 -0.6 -0.8 -100 -90 -80 -70 -60 -50 Real -40 -30 -20 -10
f =
w ( Hz ) 2
(5)
This represents damped frequency, while equation 6 gives the damping ratio. [2].
TABLE 2. DAMPING RATIOS. Eig. Most Associated States Real part Eig #1 Eig #2 Eig #3 Eig #4 Eig #5 iqr_Dfig_1 idr_Dfig_1 -99,46 -5,7998 Imag. Damping Freq. Part ratio 0 0 0 0 0 Hz 0 0 0 0 0 1 1 1 1 1
2 + w 2
(6)
Fig. 4. Eigenvalues S-domain of Jepirachi. vw_Wind_1 omega_m_Dfig_1 theta_p_Dfig_1 idr_Dfig_1 iqr_Dfig_1 1.3575 0.9 0 -0.05542 -0.18657
vw_Wind_1 is the state variable that identifies the wind speed at the rotor, omega_m_Dfig_1, represents the indexes of
Due to the fast increasing wind power penetration the demands of grid operators and response of wind turbine generators manufactured, the aggregation of wind turbines in a grid is a very important matter to research. Research on wind energy aggregation is also important for the European Wind Energy Association EWEA, where previous results has showed that wind power plants contribute positively to system stability in the system [14]. Simulation yields with no positive eigenvalues. This result shows that eigenvalues without imaginary component do not produce big influence on small signal analysis which makes an stable behavior. A PSAT dynamic modeling shows different behaviors in important system variables. Simulation is made fast and completed in few second. Finally, graphics presented here shows the results. Rotor mechanical speed (wm). It can be seen on Figure 5, a small declination on imaginary axis, 0.125 p.u. magnitude but,
1.05
mDfig 1
mechanicla speed (p.u.)
represented, interarea and local. It was confirmed, that wind resource variability seems not to create power system instabilities when it has Squirrel Cage or Doubly-Fed Induction generators aggregated on a small grid under certain conditions simulated here. VIII. REFERENCES
[1] F. Wu, X.-P. Zhang, P. Ju. Small signal stability analysis and control of the wind turbine with the direct-drive permanent magnet generator integrated to the grid. Electric Power Systems Research, vol. 79, pp. 16611667, 2009. R.Fernandez, R. Mantza, P. Battaiotto. Impact of wind farms on a power system. An eigenvalue analysis approach. Renewable Energy vol. 32, pp. 16761688, 2007. A. Pinilla, L. Rodriguez, R. Trujillo. Performance evaluation of Jepirachi Wind Park. Renewable Energy. vol. 34, pp.4852, 2009. J. Meja, F. Chejne, R. Smith, L. Rodrguez, O. Fernndez, I. Dyner. Simulation of wind energy output at Guajira, Colombia. Technical note. Renewable Energy 31 (2006) 383399. J. Meja, F. Chejne, R. Smith, L. Rodrguez, O. Fernndez, I. Dyner. Propuesta metodolgica para el diseo de aspas de turbina de viento de eje horitzontal. Energetica, Vol. 33, 2005, pp. 37-45. L. Stephen. Dynamical Systems with Applications using Matlab. Birkhauser. 2004. P. Kundur. Power System Stability and Control. New York: McGrawHill, 1994. IEEE PES Working Group on System Oscillations, Power System Oscillations, IEEE Special Publication 95-TP-101, 1995. J. Slootweg, W. Kling. 'Modelling and Analysing Impacts of Wind Power on Transient Stability of Power Systems', Wind Engineering. Vol. 26, no. 1, pp. 3-20, 2002. K. Ogata. Modern Control Engineering. Prentice Hall. 4th Ed. 2002. F. Milano, An Open Source Power System Analysis Toolbox, IEEE Trans. on Power Systems, Vol. 20, No. 3, pp. 1199-1206, August 2005. R. Farmer. Power Systems Dynamics and Stability. The Electric Power Enginering Handbook. Ed. L. L. Grigsby. Chapter 11.3. Boca Raton CRC Press. LLC, 2001. Nordex. Project profile: Jepirachi, Colombia.case Jepirachi N60 EN. EWEA. The European Wind Energy Association-EWEA. Large Scale Integration of Wind Energy in the European Power Supply: analysis, issues and recommendation. (2005, Dec.). O. Anaya-Lara. Wind Energy Generation: Modelling and Control. Wiley. Ed. 1. September 8, 2009. S. Lan, B. Quintero, Y. Lpez. Aeromechanical evaluation of large HAWTs blades. J. Scientific and Industrial Research, Vol. 69, 2010, pp. 142-145. S. Lan, J.A.Garca and R. Aliod, Development of a lifting surface 2D panel method to compute the three-dimensional pressure distribution over the blade of a HAWT, Wind Engineering, 19, (1995), pp. 21-40. http://www.lm.dk
0.95
0.9
[2]
0.85 0 5 10 time (s) 15 20
Voltaje del Generador (Vbus1). This Figure shows DFIG voltage wave form produced.
1 VBus1 VBus2
[7] [8]
[9]
1
0.5
1 time (s)
1.5
In spite of all this process and simulation behavior is run for 20 s. the small voltage drop before 0.5s considers induction generator machine start process on bus 1. After 2 s, voltages on bus 1 and 2 remain constant showing stability on this case. VII. CONCLUSIONS As has been proved here, PSAT is a powerful power system toolbox, to analyze the small signal stability. It comes with procedures for static and dynamic wind models for power analysis, and a Simulink-based network editor. Its main advantages are: model collection, easy handling, and a good graphic presentation of results. This robust method and its fast convergence, allows determining all system eigenvalues and indicates system stability Small signal stability and power system oscillations show that power systems contain many modes of oscillation due to a variety of interactions among all its components. This analysis shows that the wind turbines have a great influence on the small signal power system stability. Frequencies determine that two possible oscillation modes are
IX. BIOGRAPHIES
Y. Ulianov, received the B.Sc. degree in electrical engineering from Universidad Autnoma de Occidente, in 1995. He received his Ph.D. degree on Renewable Energies and Energy Efficiency at Universidad de Zaragoza, Spain, in June 2010. He works in Department of Energetic and Mechanical at Universidad Autnoma de Occidente, Cali, Colombia, where he is a researcher at the Energies Research Group. His field of interest are the analysis of renewable power systems, focusing on modeling, simulation and power systems stability. S. Lan, received his degree in Physics from Universidad de Zaragoza (Spain) in 1991. He received his Ph.D. degree on Fluid Mechanics at Universidad de Zaragoza, Spain, in July 1997. Recently, he also got the degree of Dr.-Ing. Habil. At Martin-Luther Universitt Halle-Wittnberg, Germany in March 2010. He is Fluid Mechanics Professor at department of Energetic and
Mechanics of Universidad Autonoma de Occidente, Cali, Colombia since January 2008 and he is also head of Fluids Mechanics Research Group at the same university.