492
1, FEBRUARY 2021
Air Gap Inductance Calculation for Transformer Transient Model
Milad Akbari , Student Member, IEEE, Afshin Rezaei-Zare , Senior Member, IEEE,
Muhammad Ali Masood Cheema, Member, IEEE, and Tomasz Kalicki
Abstract—Zero sequence characteristic and parameters play a
vital role in some transformer transients involving unbalance or
saturation conditions such as inrush, ferroresonance, and Geo-
magnetically Induced Current (GIC). An important parameter of
the zero-sequence characteristic is the inductance of the air gap
between the core and the tank. Based on a Finite-Element Method
(FEM), this letter determines the air gap inductance, considering
fringing flux effects. The comparison of the simulation results with
an analytical method shows a good agreement for small air gaps
but reveals a significant difference for relatively large air gaps.
Thus, an inductance enhancement factor (IEF) is introduced to
correct the analytical results and facilitate the transformer model-
ing parameter estimation from the design data. Furthermore, the
significance of the accurate calculation of the air-gap inductance is
demonstrated in the time-domain simulation of different saturation
levels of the transformer core, when the transformer is subjected
to GIC.
Index Terms—Transformer modeling, electromagnetic
transient, air-gap inductance, Geomagnetically Induced Current
(GIC), Finite Element Method (FEM).
I. INTRODUCTION
T RANSFORMER modeling is immensely complicated due
to the nonlinearity and frequency-dependency of the trans-
former parameters. Although myriad of research have been de-
voted to present an accurate model of transformers [1]–[3], a few
can adequately represent the transformer structural parts and the
tank effects. Modeling these parts is crucial for determination of
the transformer zero-sequence characteristic. Such a characteristic
plays a vital role in the transformer behavior under unbalance
operation and core saturation conditions. Nevertheless, many power
transformers in service do not have sufficient zero-sequence data
that allow an accurate analysis of the transformer behavior [1].
In most transformer models, zero sequence impedance is consid-
ered as a linear inductance parallel with a resistance [2], which
is not accurate due to the nonlinearity of the iron parts. Also,
the models with nonlinear zero-sequence characteristic suffer from
some limitations, pertaining to correct physical interpretation [3].
Manuscript received February 7, 2020; revised May 22, 2020; accepted June
25, 2020. Date of publication July 16, 2020; date of current version January 22,
2021. (Corresponding author: Milad Akbari.)
Milad Akbari and Afshin Rezaei-Zare are with the Department of Electrical
Engineering and Computer Science, York University, Toronto ON M3J 1P3,
Canada (e-mail: akbari@eecs.yorku.ca; rezaei@eecs.yorku.ca).
Muhammad Ali Masood Cheema is with the Northern Transformer, E. Maple
ON L6A4P5, Canada (e-mail: acheema@northerntransformer.com).
Tomasz Kalicki is with Hydro One Networks Inc, Toronto ON M5G 2P5,
Canada (e-mail: tomasz.kalicki@hydroone.com).
Color versions of one or more of the figures in this article are available online
at https://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TEC.2020.3009818
0885-8969 © 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: ULAKBIM UASL - Balikesir University. Downloaded on May 18,2021 at 23:27:37 UTC from IEEE Xplore. Restrictions apply.
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 36, NO.
In [4], a transformer model is proposed, based on enhanced topo-
logical representation of the transformer core, flux air paths, and
tank. For the first time in the technical literature, the model of [4]
introduces the inductance of the air gap between the core and tank
as a crucial parameter of the zero-sequence characteristic, with
significant impacts on the observed behavior of the transformer
under core saturation conditions, such as GIC and ferroresonance.
An analytical method is employed to calculate the air gap inductance
considering nonuniform air gap length and fringing flux effects [4].
For the simulated transformer under study with relatively small air
gap, the model built based on the analytical result shows a good
agreement with the experimental results of the GIC tests [4]. How-
ever, the accuracy of the analytical method requires to be verified for
applying it to any transformer with different dimensions and core
constructions.
In general, modeling the air gap reluctance and inductance is
divided in two different methods, i.e., analytical and numerical.
The analytical methods such as [5] mainly use conformal mapping
approaches to transform the main geometry to a well-defined ge-
ometry, obtain the inductance, and convert the geometry back to
its original shape and obtain its inductance. On the other hand,
numerical solution techniques [6] such as finite element method
(FEM), boundary element method (BEM), and boundary integral
method (BIM) are accurate and form an important part of the
design procedure; however, they require rather elaborate computer
resources and modeling with high computational burden [5].
Furthermore, the transient programs and manufacturers typically
employ analytical formulas for parameter calculations. This study
investigates the accuracy of such analytical methods and presents a
correction factor for accurate parameter estimation. The proposed
method is based on both numerical and analytical techniques.
First, the air gap inductance is calculated using FEM, and then a
factor is proposed to correct the analytical result obtained from the
traditional inductance calculation method. The FEM results are also
compared with the analytical results, deduced from the conformal
mapping [4]. This is the first time that such validations and FEM
analysis are presented, and the existing technical literature lack such
information for building an accurate transformer model for transient
studies. Comparing the results of these methods shows significant
differences especially with the increase of the air gap length. To
investigate the profound impacts of the different parameter values
on the transformer behavior, a detailed transformer model is devel-
oped in the EMTP-RV environment, and the transformer response
to GIC is simulated in time-domain. The study results highlight the
importance of the correct representation of the air gap inductance
for the transformer transients.
II. MAGNETIC CIRCUIT WITH AIR GAP
Among the existing transformer models, duality-based magnetic
circuit equivalent models have been widely used for the analysis of
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 36, NO. 1, FEBRUARY 2021
Fig. 1. Flux paths for a transformer considering core, winding, and the tank.
Fig. 2. Off-core flux paths equivalent (a) magnetic circuit, and (b) dual electric
circuit.
the transformer transients. This model, which is a kind of topology-
based models, is derived from a magnetic circuit model using the
principle of duality [2]. Fig. 1 shows the flux paths and relevant
reluctances for one leg of the transformer, considering the winding,
core, and tank. gap presents the reluctance of air gaps between
the core and tank, tank is the reluctance of tank, and 0−air is
the reluctance of the shown air path flux. The equivalent circuit
associated with the flux paths of Fig. 1 is obtained as Fig. 2(a)
which can be converted to the equivalent circuit of Fig. 2(b), using
the principle of duality. The inductances Lgap , Ltank , and L0
are the equivalent representations of gap , tank , and 0−air , in
the dual circuit respectively. Among these, finding Lgap is rather
complicated because of two reasons. First, it is difficult to precisely
determine the air gap length in non-uniform air gaps [2], and
secondly, the fringing flux effect in the air and around the edges
of the iron core makes the calculation of the effective cross-section
area of the flux path difficult.
In the case of uniform air gap length, the air gap reluctance and
inductance can be calculated by the traditional equations:
lgap
gap =
μ0 Agap
N2
Lgap =
gap
where lgap is the mean length of air gap, Agap is the cross-section
area, μ0 is the permeability of the air, and N is the winding number
of turns. For very small air gaps, (1) and (2) yield acceptable results,
due to low fringing flux effects. However, in transformers, due to
considerable air gap length, fringing flux affects the reluctance.
In addition, the air gap length between the tank and the yoke is
Authorized licensed use limited to: ULAKBIM UASL - Balikesir University. Downloaded on May 18,2021 at 23:27:37 UTC from IEEE Xplore. Restrictions apply.
493
Fig. 3. Fringing flux between a planar and a rectangular iron surface.
Fig. 4. FEM model showing the magnetic flux density (T) including fringing
flux paths.
nonuniform. Hence, as proposed in [4], (1) should be modified to
lgap−ef f
gap = (3)
μ0 Agap−ef f
with the effective air gap length lgap−ef f and effective area
Agap−ef f which are affected by the nonuniform air gap and the
fringing flux, respectively. In order to calculate these effective
variables, the following fringing flux factor (FFF) is defined based
on a conformal mapping approach [4], [5]
  
4lgap−ef f πh
FFF = 1 + 1 + ln (4)
wπ 4lgap−ef f
where w and h, in Fig. 3, are the width and height of the unwound
part of the transformer leg. Although [5] showed the accuracy of
(4) with FEM model, it is acceptable for very short air gap lengths,
however, it is not verified for larger air gaps.
A solution for considering both the nonuniform air gap length
and fringing flux for any air gap length is to employ an inductance
enhancement factor (IEF) based on the FEM model to correct the
air gap inductance calculated by the traditional method. With this
proposed factor, not only the nonuniformity of air gap length and
the fringing flux effects are considered, but also the lengthy setup
and elaborate modeling of FEM is avoided. Hence, the following
IEF is defined as
LF EM
IEF = (5)
Lgap
where LF EM is the air gap inductance based on the FEM model,
and Lgap is calculated based on (1) and (2).
(1)
III. RESULTS AND DISCUSSION
(2) In order to show the limitation of FFF, i.e., equation (4), the
airgap inductance between the planar and a rectangular iron surface,
Fig. 3, is calculated with both (4) and FEM method. Fig. 4 shows
the magnetic flux density and the fringing flux paths in the vicinity
of air gap with white lines, which is constructed and simulated by
the FEM model.
Since the two important factors are air gap distance lgap , and
surface of rectangular iron w, they are considered as variables while
494
Fig. 5. The comparison between IEF and FFF based on different lgap /w ratio.
Fig. 6. Magnetizing current of a 230/27.6 kV three-phase 3-leg transformer
based on two air gap inductances obtained from FFF and IEF, under the neutral
GIC of 40 A.
h is constant. The air gap inductance is calculated for different ratio
of lgap /w to compare IEF with FFF, which is shown in Fig. 5. The
typical relative air gap length (lgap /w) varies from less than 0.1
in large power transformers to more than 1.0 in small distribution
transformers. The results show that FFF matches well with IEF for
very short air gaps, up to about lgap /w = 0.18. However, increasing
the ratio of lgap /w leads to significant difference between FFF and
IEF. For instance, with lgap /w = 0.5 and lgap /w = 1 the IEF is
10.3% and 23.5% higher than FFF, respectively.
This noticeable difference can have profound effects on the low
and mid frequency transients of power transformers such as the
transformer response due to GIC. During geomagnetic disturbance
(GMD), the GIC flows in the system transformers with the grounded
neutrals. The GIC magnitude and polarity in all phases are identical
and constitute a zero-sequence current in the transformer. Hence,
the zero-sequence characteristic of the transformer is of paramount
importance, under the GIC conditions [7]. To investigate the differ-
ence between the estimated air gap inductances from FFF and IEF,
Authorized licensed use limited to: ULAKBIM UASL - Balikesir University. Downloaded on May 18,2021 at 23:27:37 UTC from IEEE Xplore. Restrictions apply.
IEEE TRANSACTIONS ON ENERGY CONVERSION, VOL. 36, NO. 1, FEBRUARY 2021
the GIC injected from the grid into the HV side of a 230/27.6 kV
three-phase, 3-leg transformer is simulated in the EMTP-RV. The
detailed transformer model of [4], [7] is adopted in this study and
the saturation level of the transformer under the neutral GIC of 40
A is investigated by analyzing the core magnetizing current. Fig. 6
compares the phase-A magnetizing currents of the transformer HV
winding, with the air gap inductances obtained from FFF and IEF
and for lgap /w = 0.75. The peak of the currents are 28.07 A and
46.1 A for FFF and EIF, respectively, corresponding to a differ-
ence of 64.23%. These results clearly demonstrate the noticeable
difference of transformer response to the GIC with different air gap
inductances.
IV. CONCLUSION
This paper presents a new inductance enhancement factor (IEF)
for calculating the air gap inductance of the transformers for the
transformer transient analyses. The proposed IEF is derived from
FEM, taking into account the fringing flux effects.
The IEF is also compared with the existing analytical method,
referred to as FFF in this paper. It is shown that the FFF method
is accurate for small air gaps; however, its error increases with the
air gap length, whereas the proposed IEF can be used for any air
gap length. Moreover, manufacturers, power utilities, and users of
transients programs can apply the ready-to-use correction factor
IEF to analytical formulas and accurately calculate the air gap
inductance from design data.
To demonstrate the effects of proposed estimation of the air
gap inductance on the transformer behavior, the response of a
230kV/27.6kV three-phase 3-leg transformer is investigated with
time-domain simulations, under the neutral GIC of 40A. The study
results show that the saturation level of the transformer is noticeably
higher when the air gap inductance is calculated based on the
proposed IEF as compared with the FFF method. The correspond-
ing magnetizing currents reveal 64.23% higher peak current for
IEF-based air gap inductance compared with that of the FFF-based
parameter.
REFERENCES
[1] A. Ramos, J. C. Burgoes, A. Moreno, and E. Sorrentino, “Determination of
parameters of zero-sequence equivalent circuits for three-phase three legged
YNynd transformers based on onsite low-voltage tests,” IEEE Trans. Power
Del., vol. 28, no. 3, pp. 1618–1625, Jul. 2013.
[2] J. A. Martinez and B. A. Mork, “Transformer modeling for low- and mid-
frequency transients—A review,” IEEE Trans. Power Del., vol. 20, no. 2,
pp. 1625–1632, Apr. 2005.
[3] A. Narang and R. H. Brierley, “Topology based magnetic model for steady-
state and transient studies for three phase core type transformers,” IEEE
Trans. Power Syst., vol. 9, no. 3, pp. 1337–1349, Aug. 1994.
[4] A. Rezaei-Zare, “Enhanced transformer model for low- and mid-frequency
transients-Part I: Model development,” IEEE Trans. Power Del., vol. 30,
no. 1, pp. 307–315, Feb. 2015.
[5] A. Balakrishnan, W. T. Joines, and T. G. Wilson, “Air-gap reluctance and
inductance calculations for magnetic circuits using a Schwarz-Christoffel
transformation,” IEEE Trans. Power Electron., vol. 12, no. 4, pp. 654–663,
Jul. 1997.
[6] J. Zhang and W. G. Hurley, “Gapped transformer design methodology
an implementation for LLC resonant converters,” IEEE Trans. Ind. Appl.,
vol. 52, no. 1, Jan. 2016.
[7] A. Rezaei-Zare, “Enhanced transformer model for low-and mid-frequency
transients-Part II: Validation and simulation results,” IEEE trans. Power
Del., vol. 30, no. 1, pp. 316–325, Feb. 2015.