The Journal of Engineering

The 14th IET International Conference on AC and DC Power

Transmission (ACDC 2018)

Stability analysis of grid-connected PV eISSN 2051-3305

Received on 23rd August 2018

generation with an adapted reactive power

Accepted on 19th September 2018
E-First on 7th December 2018
doi: 10.1049/joe.2018.8485
control strategy www.ietdl.org

Gangui Yan1, Yuru Cai1 , Qi Jia1, Shuai Liang1

1College of Electrical Engineering, Northeast Electric Power University, Jilin, Jilin, People's Republic of China
E-mail: [email protected]

Abstract: The static VAR compensator is widely applied in large-scale grid-connected photovoltaic (PV) generation to
participate in voltage regulation of power system, which ignores the reactive power regulation capability of PV inverter. In this
study, firstly, the constraints of maximum operating current and output voltage's amplitude of PV inverter are considered, and the
reactive power regulation capability of the PV generation is evaluated; secondly, an active participation in voltage regulation and
control strategy of PV generation is proposed considering reactive power demand of system. This control method through the
real-time detection and comparison with reference value of point of common coupling (PCC) voltage, automatically obtain the
reactive demand to maintain PCC voltage by PI controller and realises the dynamic adjustment; finally, a time-domain small-
signal model for PV generation is established, and based on this small-signal model, eigenvalue analysis is employed to study
the influence of the grid strength, operating condition, and the reactive power control strategy on operating stability. Based on
the time-domain simulation example in EMTDC/PSCAD, the validity of the theoretical analysis and the feasibility of the control
strategy are verified.

1 Introduction In summary, existing researches mostly focus on distributed

generation, lack of research on active participating voltage
Benefiting from the significant technical advances in solar cells regulation of large-scale PV generation.
and power electronics, the penetration of large-scale photovoltaic In this paper, firstly, the constraints of maximum operating
(PV) generation is increasing year by year. Due to the low energy current and output voltage's amplitude of PV inverter are
densities and uneven distributions of solar resources, most PV considered, and the reactive power regulation capability of the PV
power plants are deployed in remote areas or even desert with high generation is evaluated. Secondly, an active participation in voltage
solar irradiance [1]. Consequently, the long-distance power regulation and control strategy of PV generation is proposed
transmission lines with low short circuit ratio (SCR) have become considering reactive power demand of system. Finally, a time-
the major bottleneck to effectively transmit the generated power to domain small-signal model for PV generation is established, and
the load centre [2]. Due to the perturbation of the PV array output based on this small-signal model, eigenvalue analysis is employed
power and the disturbance of the power grid and other aspects, the to study the influence of the grid strength and the reactive power
point of common coupling (PCC) voltage fluctuates easily [3], control strategy on operating stability. Based on the time-domain
even worse when connect to weak grid. Large-scale PV generation simulation example in EMTDC/PSCAD, the validity of the
affects the power flow sise and distribution of transmission side theoretical analysis and the feasibility of the control strategy are
grid [4], therefore, large-scale grid-connected PV generation verified.
should have ability to regulate reactive power and participate in
voltage control of grid according to the PCC voltage, so as to meet
the reactive power demand of grid. 2 Main structure and control strategy
To unblock the bottleneck caused by PCC voltage fluctuation, 2.1 Main structure of the system
the scholars have done quite a lot of work and got tremendous
analyse result on this field. Liu et al. [5] show that the voltage In this paper, a single PV unit connected to an infinity system is
regulation capacity of PV generation can completely replace the used to equivalent the discussing system. The structure of PV
voltage regulator capacitor when penetration is >30%. Alam et al. generation through VSC connected to AC grid is shown in Fig. 1.
[6] designed two kinds of voltage source converter (VSC) reactive C is DC filter capacitor, UDC is DC voltage, Lf is filter inductance
power control modes, by using fuzzy adaptive control strategy to of inverter, Cf is filter capacitor of VSC, Rg and Lg are resistance
support grid voltage when PV output or load fluctuated. Research and inductance of grid, Ut and Ug are the PCC voltage and grid
on voltage control for PV generation in China is mostly focused on voltage.
distributed generation [7, 8]. Jahangiri and Aliprantis [7] utilised In Fig. 1, the AC grid strength is generally described by the
energy storage for voltage regulation, which could suppress active short SCR, which is the ratio of the short circuit capacity of the AC
fluctuation but increased the cost and complicates the control system to the rated power of the PV generation. As shown in (1)
system. [8] mainly utilised reactive compensation for local load,
but could not realise the reactive power independent control of the SAC UN2
PV generation. SCR = = (1)
PN Zg ⋅ PN

where Zg is the grid resistance, UN is the rating of the grid line

voltage. Weak AC system usually refers to the AC system
operating at 2 ≤ SCR ≤ 3.

Fig. 1 PV generation access weak grid structure

J. Eng., 2019, Vol. 2019 Iss. 16, pp. 2980-2985 2980

This is an open access article published by the IET under the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/3.0/)
 UDCUPCC 2 2
UPCC UDCUPCC 2
− − P2 − ≤Q≤ − P2
2ωL ωL 2ωL
2 (6)
UPCC
−
ωL

The lower limit of reactive power regulation is Qmin, and the upper
limit of reactive power regulation is Qmax.

UDCUPCC 2 2
UPCC
Fig. 2 Topology of VSC Qmin = min − − P2 − , 2
− uPCC 2
Imax − P2
2ωL ωL
UDCUPCC 2 2
UPCC
Qmax = max − P2 − , 2
uPCC 2
Imax − P2
2ωL ωL
(7)

2.3 Adapted reactive power control strategy

Based on the constraints mentioned above, an active participation
in voltage regulation and control strategy of PV generation is
proposed considering reactive power demand of system. This
control method through the real-time detection and comparison
with reference value of PCC voltage, automatically obtain the
reactive demand to maintain PCC voltage by PI controller and
Fig. 3 PV cell equivalent circuit
realises the dynamic adjustment.
These strategy provide the opportunity to keep Qref = 0 when
2.2 Analysis of the limit of reactive power regulation the voltage is within an acceptable boundary. Fig. 2 illustrates
Specific control; however, Umin and Umax are different for each
The function of reactive power regulation on improving system voltage. Using this method, the maximum capacity of the VSC is
operation voltage level is apparently, especially connected to weak demanded for reactive power support when UPCC ≤ Umin or UPCC 
AC system. The VSC can provide active power and reactive power ≥ Umax; however, the support will be fully provided only if the
to system simultaneously, for maintaining PCC voltage constantly. other voltage is (Umin + ΔU) ≤ UPCC ≤ (Umax − ΔU).
The reactive power generated by VSC is mainly restricted by
The expression equations of the control strategy are as (8).
the following factors:
Qmax UPCC ≤ U1
i. The margin of Imax maximum operating current Imax usually
takes 10% of the rated current value, in order to ensure the Qmax
U − U1 U1 < UPCC ≤ U2
operation safety of VSC. So the amplitude of Imax is U1 − U2 PCC
2 ∗ 1.1IN. Qref = 0 U2 < UPCC ≤ U3 (8)
ii. As for a three-phase inverter circuit using SPWM (sinusoide Qmax
pulse width modulation), the range of modulation ratio m is (0, U
U3 − U4 PCC
− U3 U3 < UPCC ≤ U4
1] to ensure the effectiveness of SPWM control and waveform
quality of output voltage of VSC. When modulation ratio is 1, −Qmax UPCC > U4
the fundamental wave amplitude of the output phase voltage is
Udc/2 maximum. 3 Small-signal modelling
3.1 PV array modeling
The main parameters are discussed as follows, SN, UN, and IN
are rated capacity, voltage, and current, L is a filter reactance, Udc PV array composed of a parallel string of Ns, which formed by Np
is DC busbar voltage, UPCC is voltage of PCC. PV cells is shown in Fig. 3, where UDC and IPV are output voltage
According to the main operation constraints of VSC, the and current of PV array.
reactive power regulation capability is analysed. The maximum Based on the consistency of the operating characteristics of the
current constraint conditions can be obtained by (2). PV cells, the UDC–IPV of PV array port can be obtained by (9).

Id2 + Iq2 ≤ Imax

2
(2) IPV = NPISC 1 − C1 e UDC / C2NsUOC − 1 (9)

The active power and reactive power of VSC are as (3). where Isc is short-circuit current, Uoc is open-circuit voltage, C1
P = Ud ⋅ id and C2 are constant.
(3)
Q = − Ud ⋅ iq 3.2 Model of the VSC and the control system
Equation (3) is brought into (2) and can be obtained: Configuration of VSC as shown in Fig. 4. The model of VSC in dq
frame is shown in (10).
2
− UPCC 2
Imax − P2 ≤ Q ≤ UPCC
2 2
Imax − P2 (4)
did
L = Ud − Utd + ωLiq
The maximum voltage constraint conditions can be obtained in (5). dt
diq
1 2 L = Uq − Utq + ωLiq (10)
Ud2 + Uq2 ≤ UDC (5) dt
4
dUDC
UDC ⋅ C = iPV ⋅ UDC − Utd ⋅ id
Derived from (6). dt

J. Eng., 2019, Vol. 2019 Iss. 16, pp. 2980-2985 2981

This is an open access article published by the IET under the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/3.0/)
 3.3 Model of PV generation connected to grid
The equivalent circuit of AC side of PV generation connected to
AC system is shown in Fig. 6, where P and Q are the active power
and reactive power.
According to the circuit principle

Utx + jUty = Ugx + jUgy + j ix + jiy xg (12)

The grid connection point voltage of VSC and the xy axis

component of grid-connected current, Utx, Uty, ix, iy and dq axis
component Utd, Utq, id, iq satisfy the following relationships.
Fig. 4 Topology of VSC
Utx = Utd ⋅ cos θpll − Utq ⋅ sin θpll
Uty = Utd ⋅ sin θpll + Utq ⋅ cos θpll
(13)
ix = id ⋅ cos θpll − iq ⋅ sin θpll
iy = id ⋅ sin θpll + iq ⋅ cos θpll

In summary, (9)–(13) compose the differential equations of PV

generation, the equations are linearised at the equilibrium point to
get the small-signal model of PV generation connected to AC
system.
Status variables Δx are selected as follow:
T
Δx = Δx1, Δx2, Δx3, Δx5, Δxpll, Δθpll, Δid, Δiq, ΔUDC

4 Eigenvalues analysis
Fig. 5 Control strategy of VSC
Taking the PV generation shown in Fig. 1 as an example to
analysis the small-signal stability, the main parameters of the
system are summarised in Table 1, when SCR = 1.5, the
eigenvalues of the system state matrix are summarised in Table 2.
Fig. 7 shows that all the eigenvalues of the system are distributed
on the left side of the complex plane, and the PV generation is
Fig. 6 PV generation access weak grid structure connected to weak grid with small disturbance stability. There are
five attenuation modes and two oscillation modes in the system.
The vector control strategy based on grid voltage orientation is Considering the influence of grid strength, reactive power
often used in VSC control, as shown in Fig. 5. The current control parameters and reactive loop parameters on small-signal
command values of the d-axis and the q-axis are generated, stability of PV generation, the change regulation of the
respectively, according to the demand of active power and reactive characteristic root is mainly analysed.
power controlled by the power outer loop. The constant DC voltage The eigenvalues locus for SCR varies from 4.5 to 1.5 is shown
control mode is applied to the control of active power while the in Fig. 8. As the grid strength decreases, it can be seen that there
constant terminal voltage amplitude control method is applied to are five eigenvalues changing λ1, λ5, λ6 move to right-half plane;
the reactive power control. Where UDCref and UDC are reference
λ3, λ4 move to right-half plane. Therefore, the stability of the PV
value and actual value of DC voltage.
generation connected to weak grid becomes worse with reduction
The governing equation of the control strategy of the VSC in
of grid strength.
Fig. 5 as follows: (see (11)) where x1, x2, x3 and x5 are state
As the absolute value of reactive power control parameter k
varables; kp1, kp2, kp3, kp4 and kp5 are the proportional coefficients increase, it can be seen from Fig. 9 that there are five eigenvalues
of the controllers; ki1, ki2, ki3, ki4 and ki5 are the integral changing. λ1, λ5 and λ6 move to left-half plane; λ3, λ4 move to
coefficients of the controllers; θpll is the angle produced by the right-half plane. Therefore, the stability of the PV generation
phase-locked loop. becomes better with the absolute value of reactive power control
parameter increasing.

dx1
= UDC − UDCref
dt
dx2
= kp1 ⋅ UDC − UDCref + ki1 ⋅ x1 − id
dt
dx3
= kp5 ⋅ −Utd ⋅ iq − k ⋅ Ut + b + ki5 ⋅ x5 − iq
dt
dx5
= Q − Qref = − Utd ⋅ iq − k ⋅ Ut + b (11)
dt
dxpll
= Utq
dt
dθpll
= kp4 ⋅ Utq + ki4 ⋅ xpll + ω0
dt
Ud = kp2 kp1 UDC − UDCref + ki1 ⋅ x1 − id + ki2 ⋅ x2 + Utd − ωLiq
Uq = kp3 kp5 −UDC ⋅ iq − k ⋅ Ut + b + ki5 ⋅ x5 − iq + ki3 ⋅ x3 + Utd + ωLid
2982 J. Eng., 2019, Vol. 2019 Iss. 16, pp. 2980-2985
This is an open access article published by the IET under the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/3.0/)
Table 1 Main parameters
Parameter Value
rated output power of PV generation P, kW 500
rated DC voltage UDC, V 600
grid voltage Ug, V 270
DC filter capacitor C, F 0.02
inverter filter inductance Lf, mH 0.5
grid inductance Lg, mH 0.3
DC voltage control loop (kp1, ki1) (2, 200)
AC voltage control loop (kp5, ki5) (2, 100)
active current control loop (kp2, ki2) (2, 100)
Fig. 9 Eigenvalues locus for varying control parameters (k)
reactive current control loop (kp3, ki3) (2, 100)
phase-locked loop (kp4, ki4) (12, 1500)

Table 2 Eigenvalues of the system

Mode Eigenvalues Oscillation frequency, Hz Damping ratio
λ1 −1613835.6 0 1
λ2 −3904.4 0 1
λ3,4 −25.5 ± j71.9 11.4 0.33
λ5,6 −3.2 ± j35.4 5.6 0.09
λ7 −50.37 0 1
λ8 −49.65 0 1 Fig. 10 Eigenvalues locus for varying reactive loop proportional
coefficient (kp5)
λ9 −50.22 0 1

Fig. 11 Eigenvalues locus for varying reactive loop integral parameters

(ki5)
Fig. 7 Distribution of the characteristic root

stability. There is no obvious change in dominant characteristic

roots, which indicates that the parameters of the reactive loop have
little effect on small-signal stability.

5 Simulation validation
In order to verify the validity of the theoretical analysis and the
feasibility of the control strategy, a time-domain simulation
example is constructed in EMTDC/PSCAD as shown in Fig. 10.
Due to the limitation of system capacity, the lower limit of reactive
power regulation Qmin and the upper limit of reactive power
regulation Qmax are −0.2 and 0.2 MVar. Other system simulation
parameters are shown in Table 1.
Fig. 8 Eigenvalues locus for varying grid strength
5.1 Simulation of different reactive loop parameters
When SCR = 4 and ki5 = 100, only reactive loop proportional
Fig. 12 shows the PCC voltage UPCC, rated DC voltage UDC, q-
coefficient kp5 changes (0.1–10), the root locus of system are
axis component current iq and output reactive power Qg when SCR 
shown as shown in Fig. 10. It can be seen that λ1, λ5 and λ6 move to
= 4, kp5 = 2, ki5 = 100 at different values of reactive loop parameters
left-half plane; λ3, λ4 and λ9 moves to right-half plane as the value
(in Fig. 12a ki5 = 100 and in Fig. 12b kp5 = 2). From Fig. 12, PV
of kp5 increasing. Therefore, the increasing of kp5 is good for
generation faces grid disturbance at t = 4 s which brings a damped
system stability.
oscillation, but final becomes stable. The larger value of reactive
When SCR = 4 and kp5 = 2, only reactive loop integral
loop can make system back to a stable state faster.
coefficient ki5 changes (100–1000), the root locus of system are
shown as shown in Fig. 11. It can be seen that λ5, λ6 and λ9 move to
left-half plane; λ3, λ4 move to right-half plane as the value of ki5
increasing. Therefore, the increasing of ki5 is good for system

J. Eng., 2019, Vol. 2019 Iss. 16, pp. 2980-2985 2983

This is an open access article published by the IET under the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/3.0/)
 Fig. 14 Response speed of different k

The total time of simulation is 12 s. The green, red and blue
lines indicate that value of reactive power control parameter k are
−10, −20 and −50. When t = 4 s, the increasing of grid impedance
makes PCC voltage drop. When t = 6 s, reactive voltage control
plays a role in increasing PCC voltage. From Fig. 10, it is not hard
to see that proposed control method through the real-time detection
PCC voltage, automatically obtain the reactive demand to maintain
PCC voltage by PI controller and realises the dynamic adjustment,
maintains PCC voltage within an acceptable range. As the absolute
value of reactive power control parameter k increase, the better
reactive power compensation ability of the method is.
The total time of simulation is 8 s. When t = 4 s, the increasing
of grid impedance makes PCC voltage drop, reactive power control
is thrown into system directly. Different k are used to determine
response speeds under the same grid disturbance. From Fig. 14, the
larger absolute value corresponds to the faster response speed.

6 Conclusion
Fig. 12
In this paper, a simple and effective adapted Reactive Power
(a) Response speed of different reactive loop different proportional coefficient kp5, (b)
Control strategy has been proposed and validated using PSCAD/
Response speed of different reactive loop different integral coefficient ki5
EMTDC simulations. It shows that the method utilises all the
available capacity of VSC (when it is needed) without violating
maximum capacity limitation, and provides reactive power support
for the system. A small-signal model for PV generation is
established to investigate its stability with varies grid strength,
reactive loop parameters, and reactive power parameters. The
following conclusions are obtained:

i. With the decrease of grid strength, the stability of the small-

signal model becomes worse.
ii. With the increase of the absolute value of reactive power
control parameters, the small-signal stability becomes better.
iii. Increasing of reactive loop parameters makes the small-signal
stability better, but not obviously.
iv. With the same grid disturbance, the larger absolute value of k
represents the better of the VSC performance to meet the
reactive power demand and the faster response speed.

7 Acknowledgments
This work is supported by Research Program of State Grid
Corporation of China (study on active frequency and voltage
Fig. 13 Reactive power compensation ability of different k
control technologies for second level power disturbance in
photovoltaic power plant).
5.2 Simulation of different adapted control coefficient
Figs. 13 and 14 show the PCC voltage UPCC, rated DC voltage 8 References
UDC, q-axis component current iq and output reactive power Qg [1] Komoto, K., Ehara, T., Xu, H., et al.: ‘Energy from the desert: very large
when SCR = 4, kp5 = 2, ki5 = 100 at different values of k. scale PV power plants for shifting to renewable energy future’ (International
Energy Agency, Paris, France, 2015)

2984 J. Eng., 2019, Vol. 2019 Iss. 16, pp. 2980-2985

This is an open access article published by the IET under the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/3.0/)
[2] Brown, T.: ‘Transmission network loading in Europe with high shares of [6] Alam, M, Muttaqi, K, Sutanto, D.: ‘A multi-mode control strategy for var
renewables’, IET Renew. Power Gener., 2015, 9, (1), pp. 57–65 support by solar PV inverters in distribution networks’, IEEE Trans. Power
[3] Hu, J.B., Huang, Y.H., Wang, D., et al.: ‘Modeling of grid-connected DFIG- Syst., 2015, 30, (3), pp. 1316–1327
based wind turbines for DC-link voltage control stability analysis’, IEEE [7] Jahangiri, P, Aliprantis, D.: ‘Distributed volt/var control by PV inverters’,
Trans. Sust. Energy, 2015, 6, pp. 1325–1336 IEEE Trans. Power Syst., 2013, 28, (3), pp. 3429–3439
[4] Ge, H., Bi, R., Xu, Z.C., et al.: ‘Research on reactive power and voltage [8] Zhang, W., Liu, Z.M., Shen, L.X.: ‘Flexible grid-connection of photovoltaic
control of large-scale photovoltaic power station’, Power Syst. Prot. Control, power generation system with energy storage system for fluctuation
2014, (14), pp. 45–51 smoothing’, Electr. Power Autom. Equip., 2013, 33, (5), pp. 106–111
[5] Liu, Y, Bebic, J, Kroposki, B, et al.: ‘Distribution system voltage performance
analysis for high penetration PV’. IEEE Energy 2030 Conf., Atlanta, GA,
2008, pp. 1–8

J. Eng., 2019, Vol. 2019 Iss. 16, pp. 2980-2985 2985

This is an open access article published by the IET under the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/3.0/)