1

Power System Small-Signal Oscillation Stability as Affected by Large-scale PV Penetration

W. Du, H. F. Wang, and R. Dunn
aspect, one of the most important issues is the impact of large penetration of PV generation on power system stability, which must be examined carefully [1]. Dynamic of PV generation is considerably different with that of conventional generation involving rotating machines. Hence a thorough study on the case of interaction of PV generation with conventional power generation and transmission systems is an urgent task to be pursued. [8] and [9] have reported the results of case studies about the effect of PV generation on power system dynamics, which have provided good foundation for further investigations. The objective of this paper is to examine the effect of operation of PV generation jointly with that of conventional generator and transmission system on power system small signal oscillation stability. The paper beings with the establishment of a comprehensive model of a single-machine infinite-bus power system integrated with a PV power generation plant. Then an example single-machine infinite-bus power system integrated with a PV power plant is presented. Results of numerical computation and non-linear simulation at different operating conditions are given when the load conditions of conventional power generation and levels of mixture of conventional and FC generation change. Results from the example power system indicate that power system oscillation stability can be affected either positively or negatively. There exists an operational limit of the PV power plant as far as system oscillation stability is concerned. Beyond the operational limit, the PV generation supplies negative damping torque, thus damaging system oscillation stability. Hence for the safe penetration of PV generation into power systems, the operational limit of oscillation stability of the PV power plant must be considered. II. MATHEMATICAL MODEL Figure 1 shows the configuration of a single-machine infinitebus power system, where a large-scale PV power generation plant is connected at busbar s. Typical voltage-current characteristic of PV generation is non-linear, given by the following expression [1].

Abstract This paper investigates the impact of a large photovoltaic (PV) penetration on power system small signal oscillation stability. A comprehensive model of a single-machine infinite-bus power system integrated with a PV power generation power plant is established. Numerical computation of damping torque contribution from the PV power plant is carried out, which is confirmed by the results of calculation of system oscillation model and non-linear simulation. Those results indicate that power system oscillation stability can be affected either positively or negatively. There exists an operational limit of the PV power plant as far as system oscillation stability is concerned. Beyond the operational limit, the PV generation supplies negative damping torque, thus damaging system oscillation stability. Hence for the safe penetration of PV generation into power systems, the operational limit of oscillation stability of the PV power plant must be considered. Index Terms PV generation, power system oscillations, power system small signal oscillation stability, DTA.

I. INTRODUCTION

he fast development of photovoltaic (PV) technology in recent two decades strongly indicates that in medium to long term, PV generation will become one of most attractive renewable energy sources in large-scale applications. Evidence of such indication is that the growth of demand on solar energy has consistently been by over 20% per annum due to the decreasing cost and price. The European Union is well on the track to fulfilling its own target of 3GW PV generation by 2010; While capacity of PV generation in the UK is forecasted to be 30GW by 2050 [1][2]. Grid connection is the operational practice of large-scale PV generation for its best utilization. This has attracted the most R& D interests in the area [3]. So far, majority of work on the grid connection of PV generation has focused on the R&D issues of efficient and effective PV generation itself, such as generation control and design of interface power electronics [4]-[7]. Large scale grid-connected PV generation will significantly affect power transmission and generation systems. In this
The authors would like to acknowledge the support of the EPSRC UKChina joint research consortium (EP/F061242/1), Supergen 1 flexible network consortium, Supergen 3 energy storage consortium, UK, and the Fund of Best Post-Graduate Students of Southeast University, China. W. Du is with the Southeast University, Nanjing China, and working at the Queens University of Belfast, Belfast, UK at present (e-mail: ddwenjuan@googlemail.com). H F Wang is with the Queens University of Belfast, Belfast, UK and R. Dunn is with the University of Bath, Bath, UK.

V pv

N p I sc I r I pv N s nkT 100 = + 1 ln q N p I0

(1)

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on July 24,2010 at 06:06:30 UTC from IEEE Xplore. Restrictions apply.

that the characteristic of PV generation of Eq.(1) is given such that the curve Ppv max can be obtained. Therefore, the function of MPPT can be modeled as, d c = dc 0 + K pv ( s )( Ppv Ppv max )

(2)

no matter what scheme it is implemented in practice, where K pv ( s ) is the transfer function of the MPPT controller. Dynamic of the DC/DC inverter can be modeled simply to be 1 [V pv (1 d c )Vdc ] I pv = (3) Ldc DC/AC inverter usually employs the PWM (pulse width modulation) control to regulate the exchange of active and reactive power between the PV generation and the rest of the power system. This can be achieved by controlling the modulation ratio m and phase of the PWM algorithm through the following AC and DC voltage control function respectively m = m0 + K ac ( s )(Vs Vsref ) (4)

Figure 1 A power system integrated with a PV generation plant

where T is the junction temperature, I r the irradiance, N s and N p number of cells in series and parallel respectively, n ideality factor, k Boltzmanns constant, q charge of the electron, I sc short-circuit current and I 0 saturation current.
Ipv Ppv-max Ir

= 0 + K dc ( s )(Vdc Vdcref )

(5)

where K ac ( s ) and K dc ( s ) is the transfer function of the ac and dc voltage controller respectively.

Ppv Vpv
Figure 2 Non-linear V-I characteristic of PV generation

In Figure 1, the PV power plant is connected to the power system through a two-stage topology of inverter system, which is usually employed by gird-connected PV generation [2][3]. In addition to the function of enhancing output DC voltage, DC/DC inverter is mainly for implementing the maximum power point tracking (MPPT) control. As illustrated by Figure 2, change of PV output current and voltage change are affected by variations of various factors, such as the irradiance, I r . In order to make full use of PV cells (due to their limited life span and high initial cost), MPPT is employed in order to extract as much power from them as possible by controlling the cycle duty of the DC/DC inverter, dc . Design and implementation of MPPT control has been one of the most important issues in developing PV generation. Many methods have been proposed, such as the curve-fitting technique, perturb-and observe method, etc., considering the need of practical operation of PV generation [4]-[7]. Basically, the objective of MPPT is to ensure the operation of PV generation to be on the curve, Ppv max , by controlling dc . Because in establishing the mathematical model of PV generation integrated with the power system of Figure 1, the purpose is to investigate the impact of PV generation on the power system by theoretical analysis, numerical computation and simulation (not experiment). Hence it must be assumed

Figure 3 Phasor diagram of power system of Figure 1

AC voltage at terminal of the DC/AC inverter, VC , can be expressed in the d-q coordinate of the generator (Figure 3) to be [13] Vc = mkVdc (cos + j sin ) = mkVdc (6) where k is the converter ratio dependent of inverter structure, Vdc is the DC voltage across the capacitor Cdc . Active power received by the DC/AC inverter from the power system is Vdc I dc1 = isd vcd + isq vcq = isd mkVdc cos + isq mkVdc sin where subscript d and q denotes the d and q component of the variable respectively. Hence I dc1 = isd mk cos + isq mk sin (7) Because Ppv = I pvV pv = I dc 2Vdc = (1 dc )Vdc I pv (8)
Hence I dc 2 = (1 dc ) I pv such that dynamic of the DC/AC inverter is expressed to be

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on July 24,2010 at 06:06:30 UTC from IEEE Xplore. Restrictions apply.

Vdc = =

1 ( I dc1 + I dc 2 ) Cdc

1 [isd mk cos + isq mk sin + (1 d c ) I pv ] Cdc

(9)

General mathematical model of the synchronous generator can be written as X g = F ( X g , I ts ) , where X g is the state
variable vector associated with generator dynamics and I ts (or itsd and itsq ) is the interface variable between the generator and rest of the system. In this paper, the following generator model is used, which is sufficient for the study of power system small signal oscillation stability [10]. = o ( 1)

PV power plant of Figure 1 are given in Appendix I. The effective technique of damping torque analysis (DTA) [10][12] was used to study the oscillation stability of the power system as affected by the PV power plant. Appendix I presents the details of deriving the linearized model of the power system integrated with the PV power plant of Figure 1, given by (A-2) and (A-8)-(A-11). Computational results of damping torque contribution from the PV power plant, confirmed by the computational results of system oscillation mode, from the complete linearized model of Figure 4 and 5 are given in Table 1 and 2. In Table 1, the total active power supplied by the generator and PV power plant to the load at the infinite busbar is fixed at 1.0 p.u., but level of mixture of conventional and PV power generation varies. In Table 2, the PV power generation is fixed to be 0.3 p.u., but active power supply from the generator changes.
Table 1 Computational results of the example power system when total active power received at the infinite busbar is fixed at 1.0 p.u. ( Pt 0 + Ppv 0 = 1.0 p.u. )
Pt 0 Ppv 0

1 [Pm Pt D( 1)] M 1 ( Eq + E fd ) Eq '= Td 0 ' E fd '=TE ( s )(Vtref Vt )

(10)

0 (degree)
54.3 58.6 62.9 67.4 72.1 76.8 81.5 86.4 91.3 96.1

Tdt
1.80 1.12 0.49 -0.09 -0.62 -1.11 -1.56 -1.97 -2.36 -2.72

Tddt
1.43 0.79 0.21 -0.32 -0.81 -1.27 -1.68 -2.06 -2.42 -2.75

Tdt ac
0.0006 0.0015 0.0021 0.0023 0.0021 0.0017 0.0010 -0.0000 -0.0012 -0.0003

Oscillation mode -0.57 j 3.86 -0.43 j 3.97 -0.31 j 4.09 -0.22 j 4.49 -0.15 j 4.29 -0.09 j 4.39 -0.04 j 4.47 0.01 j 4.55 0.04 j 4.62 0.07 j 4.69

where for simplicity of discussion, transfer function of the AVR (automatic voltage regulator) is taken to be a first-order KA in this paper and system TE ( s ) = 1 + sTA

Pt = Eq ' itsq + ( xq xd ')itsd itsq Eq = Eq ' ( xd xd ')itsd Vt = vtd 2 + vtq 2 = ( xq itsq )2 + ( Eq ' xd ' itsd ) 2 From Figure 1 it can have Vt = jxts I ts + Vs
Vs = jxs I s + Vc Vs Vb = jxsb ( I ts I s ) Those equations above give jxs I s + Vc Vb = jxsb ( Its I s ) (12)

(11)

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

Table 2 Computational results of the example power system when the PV power generation is fixed at 0.3 p.u. ( Ppv 0 = 0.3 p.u. )
Pt 0

0 (degree)
89.6 85.8 81.9 78.2 74. 6 70.9 67.5 64.1 60.8 57.6

Tdt
-2.04 -1.75 -1.45 -1.14 -0.81 -0.46 -0.09 0.33 0.80 1.34

Tddt
-2.07 -1.81 -1.55 -1.27 -0.99 -0.67 -0.33 0.05 0.49 0.97

Tdt ac
-0.0000 0.0001 0.0003 0.0006 0.0010 0.0016 0.0023 0.0031 0.0042 0.0056

Oscillation mode

(13) Vt = jxts I ts + jxsb ( I ts I s ) + Vb In d-q coordinate of the generator, as shown by Figure 3, from Eq.(13) it can be obtained that xsb xs xsb itsq Vc cos + Vb sin = x + x + x xsb isq Vb sin ts sb q (14) xsb xs xsb I tsd Vc sin Vb cos = xsb I sd Eq ' Vb cos xd '+ xts + xsb The complete mathematical model of the power system of Figure 1 thus is established, consisted of the model of PV generation and control of Eq.(1)-(5) and (9), generator of Eq.(10) and (11), and integration of the generator and the PV power plant with the rest of the power system of Eq.(14). III. AN EXAMPLE POWER SYSTEM Parameters and initial operating conditions of an example single-machine infinite-bus power system integrated with a

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

-0.01 j -0.06 j -0.12 j -0.17 j -0.23 j -0.29 j -0.36 j -0.44 j -0.52 j -0.61 j

8.13 8.09 8.05 7.99 7.93 7.85 7.76 7.64 7.51 7.35

From Table 1 and 2 it can be seen that (1) Damping torque supplied by the PV power plant changes at different levels of mixture of conventional and PV power generation, which can be positive or negative. With the fixed load at the infinite busbar, the more power contributed by the PV generation, the worse the impact of PV generation on system oscillation stability, because the more negative damping torque is provided by the PV power plant.

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on July 24,2010 at 06:06:30 UTC from IEEE Xplore. Restrictions apply.

(2) Damping torque supplied by the PV power plant also changes from positive to negative at different levels of conventional power generation even though the PV power generation is fixed. In this case, the lighter the load of conventional power generation (power supplied by the generator), the worse the impact of the PV generation on power system oscillation stability. (3) Damping torque contribution from the PV power plant changes sign between 0 = 62.90 and 0 = 67.40 (Table 1) or 0 = 64.10 and 0 = 67.50 (Table 2). Hence there must exist a critical operating condition critical when the damping torque contribution from the PV power plant changes sign. The operation of the PV power plant should avoid 0 > critical when it provides negative damping to power system oscillation.

the DC/DC and DC/AC inverter of the PV power plant respectively.

1.3 1.2 1.1 1 0.9 Pt 0.8 0.7 0.6 0.5 0.4 0

5 t

10

5 (a) Pt 0 = 0.9 p.u. and Ppv 0 = 0.1 p.u.

0.35

0.8
0.3

0.6
0.25

0.4
0.2

0.2 Pt 0
0.1 0.15

-0.2
0.05

0 0 0.1 0.2 0.3 0.4 0.5 0. 6 0.7 0.8 0.9

-0.4

-0.6

4(a) Set of simulated V-I characteristic of PV generation and tracking of the maximum power point
0.135

5 t

10

5 (b) Pt 0 = 0.1 p.u. and Ppv 0 = 0.9 p.u. Figure 5 Simulation at two different levels of mixture of conventional and PV power generation when the total power received by the load at the infinite busbar was fixed at 1.0 p.u.

0.13

0.125

0.12

0.115

0.11

0.105

0.1

4 t

10

6(b) Power output from the DC/DC and DC/AC inverter of the PV power plant Figure 6 Simulation results when the irradiance changed

Non-linear simulation using the non-linear model given in section II of the paper was carried out to verify the mathematical model and confirm the results of computation from system linearized model. Figure 4 shows the simulation results when the irradiance changed at 1.0 second of the simulation. The change was a step increase of 1% every 0.1 second and lasted for 2 seconds. Figure 4(a) shows the set of V-I curves of PV generation used and tracking of the maximum power point. Figure 4(b) is the power output from

Figure 5 gives the results of simulation (the fault applied was the three-phase to-earth short circuit on the transmission line occurred at 1 second of simulation for 100ms) at two different levels of mixture of conventional and PV power generation when the total power received by the load at the infinite busbar was fixed at 1.0 p.u. Figure 6 is the results of simulation at two different levels of conventional generation and the PV power generation was fixed at 0.3 p.u. Those results of non-linear simulation confirm both the analysis in the above section and computational results presented in Table 1 and 2. (1) When the total power received by the load is fixed, the heavier load condition the SOFC power plant operates at, the more it will damage the damping of power system oscillation; (2) When the SOFC power generation is fixed, the lighter the load condition of the conventional power plant is, the more negative damping torque is provided by the SOFC power plant.

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on July 24,2010 at 06:06:30 UTC from IEEE Xplore. Restrictions apply.

1.3 1.2 1.1 1 0.9 Pt 0.8 0.7 0.6 0.5 0.4

V. REFERENCES
[1] Y. T. Tan, D. S. Kirschen and N. Jenkins, A model of PV generation suitable for stability analysis, IEEE Trans. on Energy Conversion, Vol.19, No.4, 2004, pp748-755 [2] J. M. Carrasco, L. G. Franquelo, etc., Power electronic systems for the grid integration of renewable energy sources: a survey, IEEE Trans. On Industrial Electronics, Vol. 53, No.4, 2006, pp1002-1016 [3] S. B. Kjaer, J. K. Pedersen and F. Blaabjerg, A review of single-phase grid-connected inverters for photovoltaic modules, IEEE Trans. on Industry Applications, Vol. 41, No.5, 2005, pp1292-1306 [4] S. Jain and V. Agarwal, A single-stage grid connected inverter topology for solar PV systems with maximum power point tracking, IEEE Trans. on Power Electronics, Vol. 22, No. 5 2007, pp1928-1940 [5] K. K. Tse, M. T. Ho, H. S. H. Chung and S. Y. Hui, A novel maximum power point tracker for PV panels using switching frequency modulation, IEEE Trans. on Power Electronics, Vol. 17, No.6, 2002, pp980-989 [6] S. Jain and V. Agarwal, Comparison of the performance of maximum power point tracking schemes applied to single-phase grid-connected photovoltaic systems, IET Proc. Electric Power Applications, Vol.5, No.1, 2007, pp753-762 [7] M. T. Ho and H. S. H. Chung, An integrated inverter with maximum power tracking for grid-connected PV systems, IEEE Trans. on Power Electronics, Vol. 20, No.4, 2005, pp953-962 [8] L. Wang and T. Lin, Dynamic stability and transient responses of multiple grid-connected PV systems, Proc. of IEEE PES T&D Conference, 2008, pp1-8 [9] Y. T. Tan and D. S. Kirschen, Impact on the power system of a large penetration of photovoltaic generation, Proc. of IEEE PES General Meeting, 2007, pp1-8 [10] Y. N. Yu, Electric Power System Dynamics, Academic Press Inc., 1983 [11] F. P. deMello and C. Concordia, Concepts of synchronous machine stability as affected by excitation control, IEEE Trans. Power Appar. Syst., Vol.88, No. 4, 1969 pp316-329 [12] E.V. Larsen and D.A. Swann, Applying power system stabilizers Part IIII, IEEE Trans. Power Appar. Syst., Vol. 100, No. 6, 1981, pp3017-3046 [13] Modeling of power electronics equipment (FACTS) in load flow and stability programs, CIGRE T F 38-01-08, 1998 [14] W.G.Heffron and R.A.Phillips, Effect of a modem amplidyne voltage regulator on under excited operation of large turbine generator, AIEE Trans. 71, 1952 [15] H .F. Wang, Phillips-Heffron model of power systems installed with STATCOM and applications, IEE Proc. Part C, Vol.146, No.5, 1999, pp521527 [16] H.F.Wang and F.J.Swift, The capability of the Static Var Compensator in damping power system oscillations, IEE Proc. Part C, May, No.4, 1996

5 t

10

6(a) Pt 0 = 1.0 p.u. and Ppv 0 = 0.3 p.u.

0.3

0.25

0.2

0.15 Pt 0.1 0.05 0 -0.05

5 t

10

6 (a) Pt 0 = 0.1 p.u. and Ppv 0 = 0.3 p.u. Figure 6 Simulation when the generator operated at two loading conditions and the PV power generation was fixed at 0.3 p.u.

IV. CONCLUSIONS The paper investigates the impact of grid-connected PV generation on power system small signal oscillation stability. Conventional technique of damping torque analysis is used for the investigation of a single-machine infinite-bus power system integrated with a PV power plant. Results of damping torque computation, confirmed by calculation of system oscillation mode and non-linear simulation, of an example power system integrated with a PV power plant are presented in the paper. They demonstrate that the effect of PV penetration on system oscillation stability varies with changes of system operating conditions, because the damping torque contribution from the PV generation can be positive or negative. The critical operating condition of the PV power plant is when its damping torque contribution changes sign, which in fact indicates a stability limit of PV operation as far as power system oscillation stability is concerned. The effect of PV generation on system oscillation stability is negative beyond the critical operating condition. When the total load is fixed, a higher PV penetration should be avoided because this increases the chance of its damaging system oscillation stability. When the PV penetration is fixed, its operation can also approach to the critical operating condition with the decrease of conventional power generation.

APPENDIX I
Transmission line: xts = 0.2 p.u., xsb = 0.2 p.u., xs = 0.2 p.u. ; Generator: xd = 1.3 p.u., xq = 0.47 p.u., x 'd = 0.3 p.u., M = 3s., D = 2 p.u., T 'd 0 = 5s. ; AVR: TA = 0.1s. K A = 10 p.u. ; Initial load condition: Vt 0 = 1.0 p.u., Vs 0 = 1.0 p.u., Vb 0 = 1.0 p.u. PV power plant:

APPENDIX II
For the simplicity of analysis, it is assumed that control functions of PV generation are implemented by proportional control law. Hence linearization of Eq.(2), (4) and (5) is dc = K pv Ppv , m = K ac Vs , = K dc Vdc (A-1) From Figure 3 it can be seen that = . Hence it can have

= K dc Vdc
Linearization of Eq.(1) and (8) is V pv = a1 I pv ,

(A-2)

Ppv = V pv 0 I pv + I pv0V pv = a2 I pv
That gives dc = K pv a2 I pv

(A-3)

(A-4)

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on July 24,2010 at 06:06:30 UTC from IEEE Xplore. Restrictions apply.

6 Because Vs = jxs I s + Vc ,Vs = vsd 2 + vsq 2 , it can have

vsd = xs isq + kmVdc cos , vsq = xs isd + kmVdc sin

Hence Vs = b1 + b2 Eq + b3Vdc + b4 m + b5

(A-5)

From Eq.(A-1), (A-2) and (A-5) it can have K ac (A-6) m = ( B6 + B7 Eq '+ B8 Vdc ) 1 b4 K ac By using Eq.(A-2) and (A-6), linearization of Eq.(14) can be obtained to be itsd E ' c i q 11 c12 c13 tsq = C V = c (A-7) c22 c23 Eq ' 21 dc
isd isq m c31 c32 c33 Vdc

Linearization of Eq.(3) and (9) is I pv = C1 I pv + C2 Vdc

Vdc = C3 + C4 Eq + C5 Vdc + C6 I pv + K dm m + K d

(A-8) (A-9)

Linearization of Eq.(10) is = o

1 ( Pt D ) M 1 Eq '= ( Eq + E fd ' ) Td 0 '

E fd '=TE ( s)( Vt )
By using Eq.(A-2), linearization of Eq.(11) can be obtained to be Pt = K1 + K 2 Eq + ( K pdc + K p K vdc )Vdc + K pm m
Eq = K 4 + K 3Eq + ( K qdc + K q K vdc )Vdc + K qm m Vt = K 5 + K 6 Eq + ( K vdc + K v K vdc )Vdc + K vm m

(A-10)

(A-11)

Ldc = 1.5 p.u., Cdc = 1.0 p.u., Vdc 0 = 1.0 p.u., K dc = 0.3, K ac = 0.1, K pv = 5

ACKNOWLEDGEMENTS

The authors would like to acknowledge the support of the EPSRC UK-China joint research consortium (EP/F061242/1), Supergen 1 flexible network consortium, Supergen 3 energy storage consortium, UK, and the Fund of Best PostGraduate Students of Southeast University, China.

Authorized licensed use limited to: MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY. Downloaded on July 24,2010 at 06:06:30 UTC from IEEE Xplore. Restrictions apply.