Voltage Stability of Low Frequency AC
Transmission Systems
Tuan Ngo, Quan Nguyen, and Surya Santoso
The University of Texas at Austin, Austin, TX, USA
Abstract—This paper investigates the voltage stability of a
power system with low frequency AC (LFAC) transmission.
Under low frequency conditions, transmission line reactance is
reduced and thus the voltage drop along the transmission line
is decreased. The LFAC system, therefore, is superior to the
conventional 60-Hz system in terms of power transfer capability
and, more importantly, voltage instability. In other words, the
LFAC can drive a power system further away from instability
mode in comparison to the conventional 60-Hz system. A power
system voltage stability can be quantified and displayed with the
eigenvalues of a Jacobian matrix and the self-sensitivity values
by using the modal analysis method. In this paper, the theoretical
stability of an LFAC system is discussed first and then validated
using several transmission system case studies. The simulation
and analysis results show that the LFAC system outperforms the
conventional 60 Hz system in terms of voltage stability.
Index Terms—Low frequency transmission, Voltage stability,
Self-sensitivity, Modal analysis method.
I. I NTRODUCTION
With the exponential growth of renewable energy, especially
photovoltaics and wind energy, a massive amount of electric
power is generated. This means the power transmission line
system needs to upgrade in order to carry more and more
power. One solution is to increase the high voltage AC
system voltage level to expand the power transfer capability.
This solution, however, can only upgrade the power system
to several voltage levels due to the constraints of installed
transmission overhead lines. An alternative solution is to use
a high voltage DC (HVDC) system to transfer power. In this
case, the AC current is transformed into DC current and is
carried by transmission lines. This solution is beneficial for
transmission system in terms of the amount of power transfer
and distances which can be over 1000 km. However, an HVDC
system requires specialized protection and is not flexible for
multi-terminal connections. Moreover, because onshore and
offshore wind farms are often located close to the main
transmission system, it is not cost-effective to utilize an HVDC
system for wind power transfer. Alternatively, a low frequency
AC (LFAC) has the advantages of both high voltage AC and
DC systems.
The low frequency AC (LFAC) power transmission system
is first introduced by Funaki in [1]. The key concept of LFAC
is to operate a power system at a low frequency, i.e, below 50
or 60 Hz. To do so, power converters change the conventional
60 Hz current into a lower frequency current and transmit this
current through transmission lines. If the operating frequency
978-1-5090-2157-4/16/$31.00 ©2016 IEEE
is very low, i.e. 1 Hz, the transmission line can handle almost
the same amount of power as an HVDC system [1]. With
the LFAC, the transmission losses and the charging current
in the transmission line are significantly lower than a 60 Hz
system, allowing transmissions to occur for longer distances.
More importantly, the AC breakers available in the market
are still applicable for the LFAC system fault clearing [2].
Therefore, the LFAC advantages include being capable of bulk
power transfer while using the same protection system as an
AC system and being flexible for a multi-terminal connection.
Several studies have investigated the power capability and
economic costs of using LFAC configurations to transmit
power from new wind farms to the grid [3]. These studies
show that operating a power system at a low frequency allows
a power system to benefit from both high voltage AC and DC
systems. However, previous investigations have not considered
the voltage stability of a transmission system operating at a
low frequency. For example, to our best knowledge, no one
has examined how stable a power system is at low frequency
or how close the low frequency drives a power system to
instability in comparison to a 60 Hz system. This paper,
therefore, investigates the stability of a LFAC system under
different operating frequencies. The system voltage stability
is first estimated and then compared according to the Q-V
characteristic curves. Then, a thorough examination of the
system stability is presented by examining the eigenvalues and
self-sensitivity values from a power system load flow solution.
This paper begins in Section II with a theoretical analysis
of a two-bus system operating at two different frequencies.
The system transfers power from the sending-end source
to the receiving-end source via two transmission lines. One
line operates at 60 Hz and the other line operates at a low
frequency. The effects of frequency on voltage, and on real
and reactive power are investigated. The results show that
low frequency can improve the power system both power
transfer capability and stability. Consequently, using the nu-
merical modal analysis method, the benefits of low frequency
transmission on voltage stability are examined. Section III
presents a comparison of power systems operating at 60 Hz
versus a low frequency. The eigenvalues and self-sensitivity,
which are useful for estimating voltage stability, are analyzed
in details. The results show that the low frequency transmission
is superior to the conventional 60 Hz transmission. Finally,
Section IV delivers some important conclusions about voltage
stability in the LFAC power system.
 II. L OW F REQUENCY AC T RANSMISSION S TABILITY
In this section, voltage stability in a two-bus system is
discussed according to V -P and Q-V characteristic curves.
Second, using modal analysis, the power system voltage
stability is presented.
A. Voltage stability in LFAC transmission
A two-bus system is used for theoretical analysis, as shown
in Fig. 1. The power is transferred from bus B1 to bus B2
via two transmission lines: line 1 is run with a conventional
60 Hz and line 2 operates at a low frequency. For the sake of
simplicity, the power converters for frequency conversion are
assumed to be ideal.
Fig. 1: A two-bus multi-frequency power system.
example, the voltage is more stable if the electric power system
operates at a low frequency, rather than at 60 Hz, Fig. 2(a).
This is to be expected because the voltage drop depends on
the reactance which is proportional to the operating frequency,
and the active and reactive power as referred to (3). With the
same load demand, the AC system at 5 Hz exhibits a much
smaller voltage drop than the AC system at 60 Hz, resulting
in a broader operating points range. It is also noticing that the
power transfer capability of system is improved by a factor of
more than four if transmission line 2 is operated at 5 Hz.
The system stability is demonstrated in the relationship
between reactive power and voltage. This relationships, known
as the Q-V curve, reveals the reactive reserved margin or
the Q-reserved margin in the system. This stability margin is
defined as the distance from the minimum point (∂V /∂Q = 0)
to the zero-Q line. If a system has large Q-reserve margin, it
is more stable [4]. Fig. 2(b) clearly shows that the stability
margin is higher at a lower frequency. For instance, the Q-
reserve margin at 5 Hz is 1.37 pu, which is four times more
than 0.33 pu of Q-reserve in the 60 Hz system. Interestingly,
the Q-reserve margin is also proportional to the frequency.

Denoting the equivalent impedance and admittance of trans-

mission line as: Z = R + jωL and Y = G + jωC, and let
the sending-end voltage is 1.0 pu, it can be shown that the
receiving-end voltage VR 6 −δ = VR,r + jVR,i is given by
2 2
VR,r + VR,i = VR,r − (RPR + XQR )
VR,i = XPR − RQR (1)
then
r
1 1
VR,r = + − (RPR + XQR ) − (XPR − RQR )2
2 4
VR,i = XPR − RQR (2)
From (2), the voltage magnitude at the receiving-end terminal
is given by
r (a)
2 2 1 1
a VR = + − a(RPR + XQR ) − a2 (XPR − RQR )2
2 4
−a(RPR + XQR ) (3)
where a is denoted as (1 + Y Z/2).
It is obvious from (3) that as the reactance X is reduced,
the receiving-end voltage is increased. In other words, voltage
drop is decreased when the power system operates at a low
frequency.
The voltage versus power characteristic V -P curve at differ-
ent AC system under 0.95 lag of load power factor is plotted
in Fig. 2. The real and reactive power are expressed in term
of the maximum power transfered in the 60 Hz system [4]
1 VS2
PR,max = cos φ (4)
2 + 2 cos (θ − φ) Z
where θ is angle of transmission line and φ is load angle. (b)
The operating frequency of an AC power system signif-
icantly affects the voltage-power V -P characteristics. For Fig. 2: (a) V -P and (b) Q-V curves at different frequencies.

978-1-5090-2157-4/16/$31.00 ©2016 IEEE

 From the above discussion, the LFAC transmission outper-
forms the traditional 60 Hz system in terms of voltage drop,
power transfer capability, and system stability. If a power
system operates at 5 Hz, the voltage drop is smaller along
the transmission line, requiring smaller reactive compensation
equipments. More importantly, the power transfer capability
of the system is increased considerably without extra cost.
In addition, the system stability margin is extended when
the operating frequency is reduced from 60 Hz to 5 Hz.
The power capability and system Q-reserve margin increases
proportionally to the reduction in frequency. In the next sec-
tion, the voltage stability of LFAC transmission is numerically
examined in a specific wind farm system.
B. Voltage stability based on modal analysis method
In this section, the voltage stability of a low operating
frequency power system is investigated with the modal anal-
ysis method. The traditional V -P and Q-V curve analysis
methods do not provide knowledge about system stability,
such as when the power system experiences, or comes close to
experiencing, voltage instability, or where weak-voltage points
exist. The modal analysis method, a numerical approach, can
demonstrate power system voltage stability through the load
flow results of a power system [5]. The key concepts are
summarized in this section.
The real and reactive power at each bus in a power system
is given by
n
X investigated since Bus 1 and Bus 2 are connected to generators.
Pi = Vi Vj Yij cos(δi − δj − θij )
j=1
n
X
Qi = Vi Vj Yij sin(δi − δj − θij ) (5)
j=1
Manipulating (7) and (8), the voltage increment associated
with ith mode is given by
1
∆Vmi = ∆Qmi (9)
λi
It is important to draw several observations from (9). The
voltage increment at each bus depends on the eigenvalue. If
λi is small, the voltage variation caused by an reactive power
increment is large. In other words, a small eigenvalue means
a weak bus voltage. If λi is zero, the system voltage collapses
regardless of any reactive power variation. By investigating
the corresponding eigenvalues, the voltage stability of the
conventional 60 Hz and the low frequency power systems can
be estimated and compared. In the following section, the nu-
merical modal analysis method for estimating and comparing
voltage stability in each system is demonstrated through case
studies.
III. VOLTAGE S TABILITY OF S YSTEM : A C ASE S TUDY
In this section, the voltage stability of LFAC transmission
is investigated in terms of both the eigenvalues and the V -Q
sensibility via modal analysis [4], [6]. A five-bus and six-
bus power systems are used to calculate these quantities in
different cases of low frequencies.
A. A five-bus system operating at a specific frequency
Fig. 3 presents a five-bus power system used to study
voltage stability. Only voltages at Bus 3, Bus 4, and Bus 5 are
The detailed specifications of this system can be found from
[7].

At steady-state operation, the above equation can be linearized

as follow     
∆P JP δ JP V ∆δ
= (6)
∆Q JQδ JQV ∆V
where (∆P, ∆Q) are increments in real and reactive power,
and (∆δ, ∆V ) are increments in voltage angle and magnitude,
respectively. The Jacobian matrix in the linearized equation
is same as the Jacobian matrix from the load flow solved by
Newton-Raphson method [5].
Since the voltage magnitude depends mainly on reactive
power, to investigate the system stability, let ∆P be 0 then
∆Q = [JQV − JQδ JP−1
δ JP V ]∆V Fig. 3: A five-bus power system.
= JR ∆V (7)
The power flow for this system can be solved with the
or the voltage increment can be expressed in terms of reactive Newton-Raphson method. First, the entire system is tested at
power increment a specific operating frequency. Then at each frequency, the
n voltage magnitudes, eigenvalues from the Jacobian matrix, and
−1
X ξi ηi
∆V = JR ∆Q = ξΛ−1 η∆Q = ∆Q (8) self-sensitivity values at each bus are collected in Table. I.
i
λi
The data shows that the voltage of the lower frequency
where JR = ξΛη is decomposed into a diagonal eigenvalue system is more stable than the 60 Hz system. For example,
matrix by right and left eigenvector matrices, and (ξi , ηi ) are the voltage magnitude at Bus 5 is 0.9717 pu for 60 Hz and
the ith column right and row left eigenvectors of JR . 0.9795 pu for 5 Hz. The higher voltage stability in the lower

978-1-5090-2157-4/16/$31.00 ©2016 IEEE

 TABLE I: Bus Data of Five-Bus System. generators, the voltages at these buses are considered to be
constant at 1.0 pu for all studied cases. Bus 5 is connected to
60 Hz 30 Hz 10 Hz 5 Hz
the wind farm and is assumed to be the most sensitive bus.
V3 0.9872 0.9907 0.9935 0.9937
The V -Q self-sensitivities, presented later on, will explain this
V4 0.9841 0.9870 0.9897 0.9897
assumption. It can be observed from Table II that the voltage
V5 0.9717 0.9756 0.9789 0.9795
magnitudes at low frequencies are close to 1.0 pu and much
λ3 72.2566 151.1695 454.8012 907.7094 higher than that in high frequency. Bus 5 voltage V5 at 60
λ4 7.4813 15.2164 45.8326 90.8637 Hz is 0.9563 pu, which is slightly above the lower boundary
λ5 13.7457 27.7789 83.6118 166.2663 value, and it is 0.9717 pu at 30 Hz, 0.9833 pu at 10 Hz, and
V Q3 0.0632 0.0311 0.0103 0.0052 0.9864 pu at 5 Hz, correspondingly. The same observations
V Q4 0.0653 0.0322 0.0107 0.0054 can be seen in voltages at the rested buses V2 , V3 , and V4 .
V Q5 0.0912 0.0450 0.0149 0.0075
TABLE II: Bus Voltage Magnitude.

Vmag 60 Hz 30 Hz 10 Hz 5 Hz
frequency power system is due to smaller impedances, which V1 1.0000 1.0000 1.0000 1.0000
cause smaller voltage drops on the transmission lines. Since
V2 0.9850 0.9926 0.9973 0.9984
Bus 4 is connected to heavy load, it is expected to be the
V3 0.9607 0.9782 0.9893 0.9920
weakest bus in the system. Indeed, the data shows that Bus 4
V4 0.9625 0.9788 0.9896 0.9923
has the smallest eigenvalue of the three buses. However, the
V5 0.9563 0.9717 0.9833 0.9864
eigenvalue is increased more than 12 times, i.e. from 7.4813 to
V6 1.0000 1.0000 1.0000 1.0000
90.8637, if the power system operates at 5 Hz. Additionally,
the self-sensitivity value at each bus is another indicator of
voltage stability. This value indicates the strength of a bus The eigenvalues, which represent the system voltage stabil-
voltage when it is subjected to reactive power variations [6]. ity, are determined via the Jacobian matrix and are shown in
The V -Q self-sensitivity values are selected from the V -Q Table III in descending order. Because the voltages at Bus 1
sensitivity, which is the inverse of the reduced Jacobian. A and Bus 6 are fixed, the reduced Jacobian matrix shows the
small self-sensitivity value indicates a strong bus [4], [6]. four eigenvalues at other buses. The eigenvalues are improved
Therefore, the power system is more robust if it operates at significantly when the frequency is lowered. For instance, the
a low frequency because the self-sensitivity values decrease lowest eigenvalue in 60 Hz AC system 3.0512 is increased to
significantly as the frequency is scaled down. 6.3103 (30 Hz), 19.2789 (10 Hz), and 38.6460 (5 Hz). With
higher eigenvalues, the system is more voltage-stable [6].
B. A six-bus system operating at a specific frequency
As shown in Fig.4, a six-bus system representing the power TABLE III: Eigenvalues and Self-Sensitivities.
transmission for a wind farm is used to calculate stability
λi 60 Hz 30 Hz 10 Hz 5 Hz
conditions. The wind farm is 85 km from Bus 4, and the
λ2 104.6664 212.0111 640.7134 1283.0660
maximum power the wind farm can produce is 140 MW. The
λ3 45.5025 92.8477 281.8955 565.4993
system parameters are detailed in [6].
λ4 13.6468 28.2279 86.2896 173.1871
λ5 3.0512 6.3103 19.2789 38.6460
V Qi 60 Hz 30 Hz 10 Hz 5 Hz
V Q2 0.019318 0.009473 0.003126 0.001560
V Q3 0.044492 0.021303 0.006935 0.003451
V Q4 0.069012 0.033247 0.010856 0.005404
V Q5 0.299730 0.145362 0.047650 0.023783

The voltage at Bus 2 is the most stable since its self-

Fig. 4: A six-bus LFAC system for wind turbine.
The entire system is calculated at a fixed operating fre-
quency to validate the results. The wind farm power output is
assumed to vary between 0 MW and 100 MW, at which the
weakest bus voltage V5 passes over the lower boundary. Table
II indicates the voltage magnitude at each bus when the wind
farm does not generate any output power. The power providing
to the loads, therefore, comes from two generators: Gen 1 and
Gen 6. Since Bus 1 and Bus 6 are connected directly to the
sensitivity is smallest. This is easy to understand because Bus
2 is connected directly to Bus 1 via a transformer. The voltage
at Bus 5, on the other hand, is the most sensitive to the reactive
power since it has smallest value. The self-sensitivity at this
bus is 0.30 for 60 Hz AC transmission. This value is reduced to
0.145 in 30 Hz, 0.048 in 10 Hz, and 0.024 in 5 Hz systems.
It is true that as the operating frequency is decreased, the
self-sensitivity is reduced considerably. In other words, the
dependence of voltage on reactive power is less and the system
is more stable at low frequencies.

978-1-5090-2157-4/16/$31.00 ©2016 IEEE

 The calculation results for Bus 5 at different wind farm
output power cases are represented in Fig. 5. It is worth-
while to summarize some observations here. The bus voltages
increase when the power increases. In the conventional 60
Hz transmission system, the voltage begins decreasing when
the wind farm produces over 30 MW of power output, and
the voltage passes over the lower boundary when the wind
farm produces 80 MW of output power. The voltage quickly
decreases after it passes over the lower boundary because the
capability limit of the system has been reached. Therefore,
a high frequency system is less stable when wind power
penetration is high. In the low frequency system, the voltage
V5 at Bus 5 builds up slowly but never reaches the upper
boundary, i.e 1.05 pu, when the power output is 100 MW,
as shown in Fig. 5(a). This means the low frequency system
can handle more power without losing its stability. The same
(a) conclusions for eigenvalues and self-sensitivity can be drawn
from Fig. 5(b), 5(c).
Another test for the six-bus system involves a portion of
it operating at a low frequency. The longest transmission line
45 runs a low frequency. The technique to solve power flow
for this case is presented in our other paper. Although the
results for this multi-frequency system are not presented, it is
reasonable to assume similar conclusions. The power system
voltage magnitude is higher at each bus. Also, the larger eigen-
value and smaller self-sensitivity value in comparison to those
eigenvalues and self-sensitivity values in the conventional 60
Hz system indicate the multi-frequency system is more stable.
IV. C ONCLUSIONS
This paper has investigated the voltage stability of low
frequency AC (LFAC) transmission. At a low frequency, reac-
tance is reduced and the system is more capable of transferring
power. More importantly, because of lower voltage drops along
(b) the transmission line, the system exhibits a better voltage
profile and its stability has been extended with less dependence
on power variation. The LFAC theory has been applied to
two specific case study systems and has demonstrated great
benefits to the power system voltage stability characteristics.
R EFERENCES
[1] T. Funaki and K. Matsuura, “Feasibility of the low frequency ac transmis-
sion,” in Power Engineering Society Winter Meeting, 2000. IEEE, vol. 4,
pp. 2693–2698 vol.4, 2000.
[2] W. Fischer, R. Braun, and I. Erlich, “Low frequency high voltage
offshore grid for transmission of renewable power,” in Innovative Smart
Grid Technologies (ISGT Europe), 2012 3rd IEEE PES International
Conference and Exhibition on, pp. 1–6, Oct 2012.
[3] PSERC, Low Frequency Transmission. PSERC Publication, 2012.
http://www.pserc.wisc.edu/documents/publications/reports/2012 reports/
Meliopoulos PSERC Report S-42 Oct 2012 ExSum.pdf.
[4] P. Kundur, Power System Stability and Control. McGraw-Hill, Inc, 1994.
[5] B. Gao, G. Morison, and P. Kundur, “Voltage stability evaluation using
modal analysis,” Power Systems, IEEE Transactions on, vol. 7, pp. 1529–
1542, Nov 1992.
(c) [6] H. Le and S. Santoso, “Analysis of voltage stability and optimal wind
power penetration limits for a non-radial network with an energy storage
Fig. 5: Bus 5 profiles with different frequencies: (a) voltage system,” in Power Engineering Society General Meeting, 2007. IEEE,
pp. 1–8, June 2007.
magnitudes, (b) eigenvalues, and (c) self-sensitivities ( 60Hz, [7] S. G. W and A. H. El-Abaid, Computer Methods in Power System
4 30Hz, ◦ 10Hz, 2 5Hz ). Analysis. McGraw-Hill, Inc, 1968.

978-1-5090-2157-4/16/$31.00 ©2016 IEEE