IEEE C57.161-2018-Guide-For-Dielectric-Frequency-Response-Test

IEEE Guide for Dielectric Frequency
Response Test
IEEE Power and Energy Society
Sponsored by the
Transformers Committee
IEEE
3 Park Avenue IEEE Std C57.161™-2018
New York, NY 10016-5997
USA
Authorized licensed use limited to: York University. Downloaded on December 20,2018 at 18:21:24 UTC from IEEE Xplore. Restrictions apply.
IEEE Std C57.161™-2018

Response Test
Sponsor
Transformers Committee
of the
IEEE Power and Energy Society
Approved 27 September 2018
IEEE-SA Standards Board
Abstract: Dielectric Frequency Response (DFR) test methods of liquid immersed transformers are
discussed in this guide. The guide includes recommendations for instrumentation, procedures for
performing the tests, and techniques for analyzing the data.
This guide describes factors affecting the DFR test moisture estimate including winding
configuration and measurement set-up, cellulose material type, low molecular weight acids, and
background ac/dc noise interference.
This guide is applicable to both field and factory applications. The purpose of this guide is to provide
the user with information that will assist in performing Dielectric Frequency Response
measurements and interpreting the results from these measurements.
Keywords: dielectric frequency response, dissipation factor, IEEE C57.161™, moisture
estimation, oil conductivity

The Institute of Electrical and Electronics Engineers, Inc.

3 Park Avenue, New York, NY 10016-5997, USA
Copyright © 2018 by The Institute of Electrical and Electronics Engineers, Inc.

All rights reserved. Published 20 November 2018. Printed in the United States of America.
IEEE is a registered trademark in the U.S. Patent & Trademark Office, owned by The Institute of Electrical and Electronics
Engineers, Incorporated.
PDF: ISBN 978-1-5044-5211-3 STD23337

Print: ISBN 978-1-5044-5212-0 STDPD23337
IEEE prohibits discrimination, harassment, and bullying.

For more information, visit http://www.ieee.org/web/aboutus/whatis/policies/p9-26.html.
No part of this publication may be reproduced in any form, in an electronic retrieval system or otherwise, without the prior written permission
of the publisher.
2
Copyright © 2018 IEEE. All rights reserved.
Important Notices and Disclaimers Concerning IEEE Standards Documents
IEEE documents are made available for use subject to important notices and legal disclaimers. These notices
and disclaimers, or a reference to this page, appear in all standards and may be found under the heading
“Important Notices and Disclaimers Concerning IEEE Standards Documents.” They can also be obtained on
request from IEEE or viewed at http://standards.ieee.org/IPR/disclaimers.html.
Notice and Disclaimer of Liability Concerning the Use of IEEE Standards

Documents
IEEE Standards documents (standards, recommended practices, and guides), both full-use and trial-use, are
developed within IEEE Societies and the Standards Coordinating Committees of the IEEE Standards
Association (“IEEE-SA”) Standards Board. IEEE (“the Institute”) develops its standards through a consensus
development process, approved by the American National Standards Institute (“ANSI”), which brings
together volunteers representing varied viewpoints and interests to achieve the final product. IEEE Standards
are documents developed through scientific, academic, and industry-based technical working groups.
Volunteers in IEEE working groups are not necessarily members of the Institute and participate without
compensation from IEEE. While IEEE administers the process and establishes rules to promote fairness in
the consensus development process, IEEE does not independently evaluate, test, or verify the accuracy of
any of the information or the soundness of any judgments contained in its standards.
IEEE Standards do not guarantee or ensure safety, security, health, or environmental protection, or ensure
against interference with or from other devices or networks. Implementers and users of IEEE Standards
documents are responsible for determining and complying with all appropriate safety, security,
environmental, health, and interference protection practices and all applicable laws and regulations.
IEEE does not warrant or represent the accuracy or content of the material contained in its standards, and
expressly disclaims all warranties (express, implied and statutory) not included in this or any other document
relating to the standard, including, but not limited to, the warranties of: merchantability; fitness for a
particular purpose; non-infringement; and quality, accuracy, effectiveness, currency, or completeness of
material. In addition, IEEE disclaims any and all conditions relating to: results; and workmanlike effort. IEEE
standards documents are supplied “AS IS” and “WITH ALL FAULTS.”
Use of an IEEE standard is wholly voluntary. The existence of an IEEE standard does not imply that there
are no other ways to produce, test, measure, purchase, market, or provide other goods and services related to
the scope of the IEEE standard. Furthermore, the viewpoint expressed at the time a standard is approved and
issued is subject to change brought about through developments in the state of the art and comments received
from users of the standard.
In publishing and making its standards available, IEEE is not suggesting or rendering professional or other
services for, or on behalf of, any person or entity nor is IEEE undertaking to perform any duty owed by any
other person or entity to another. Any person utilizing any IEEE Standards document, should rely upon his
or her own independent judgment in the exercise of reasonable care in any given circumstances or, as
appropriate, seek the advice of a competent professional in determining the appropriateness of a given IEEE
standard.
IN NO EVENT SHALL IEEE BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO:
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
OTHERWISE) ARISING IN ANY WAY OUT OF THE PUBLICATION, USE OF, OR RELIANCE UPON
ANY STANDARD, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE AND
REGARDLESS OF WHETHER SUCH DAMAGE WAS FORESEEABLE.
3
Translations
The IEEE consensus development process involves the review of documents in English only. In the event
that an IEEE standard is translated, only the English version published by IEEE should be considered the
approved IEEE standard.
Official statements
A statement, written or oral, that is not processed in accordance with the IEEE-SA Standards Board
Operations Manual shall not be considered or inferred to be the official position of IEEE or any of its
committees and shall not be considered to be, or be relied upon as, a formal position of IEEE. At lectures,
symposia, seminars, or educational courses, an individual presenting information on IEEE standards shall
make it clear that his or her views should be considered the personal views of that individual rather than the
formal position of IEEE.
Comments on standards
Comments for revision of IEEE Standards documents are welcome from any interested party, regardless of
membership affiliation with IEEE. However, IEEE does not provide consulting information or advice
pertaining to IEEE Standards documents. Suggestions for changes in documents should be in the form of a
proposed change of text, together with appropriate supporting comments. Since IEEE standards represent a
consensus of concerned interests, it is important that any responses to comments and questions also receive
the concurrence of a balance of interests. For this reason, IEEE and the members of its societies and Standards
Coordinating Committees are not able to provide an instant response to comments or questions except in
those cases where the matter has previously been addressed. For the same reason, IEEE does not respond to
interpretation requests. Any person who would like to participate in revisions to an IEEE standard is welcome
to join the relevant IEEE working group.
Comments on standards should be submitted to the following address:
Secretary, IEEE-SA Standards Board

445 Hoes Lane
Piscataway, NJ 08854 USA
Laws and regulations

Users of IEEE Standards documents should consult all applicable laws and regulations. Compliance with the
provisions of any IEEE Standards document does not imply compliance to any applicable regulatory
requirements. Implementers of the standard are responsible for observing or referring to the applicable
regulatory requirements. IEEE does not, by the publication of its standards, intend to urge action that is not
in compliance with applicable laws, and these documents may not be construed as doing so.
Copyrights
IEEE draft and approved standards are copyrighted by IEEE under U.S. and international copyright laws.
They are made available by IEEE and are adopted for a wide variety of both public and private uses. These
include both use, by reference, in laws and regulations, and use in private self-regulation, standardization,
and the promotion of engineering practices and methods. By making these documents available for use and
adoption by public authorities and private users, IEEE does not waive any rights in copyright to the
documents.
4
Photocopies
Subject to payment of the appropriate fee, IEEE will grant users a limited, non-exclusive license to photocopy
portions of any individual standard for company or organizational internal use or individual, non-commercial
use only. To arrange for payment of licensing fees, please contact Copyright Clearance Center, Customer
Service, 222 Rosewood Drive, Danvers, MA 01923 USA; +1 978 750 8400. Permission to photocopy
portions of any individual standard for educational classroom use can also be obtained through the Copyright
Clearance Center.
Updating of IEEE Standards documents
Users of IEEE Standards documents should be aware that these documents may be superseded at any time
by the issuance of new editions or may be amended from time to time through the issuance of amendments,
corrigenda, or errata. An official IEEE document at any point in time consists of the current edition of the
document together with any amendments, corrigenda, or errata then in effect.
Every IEEE standard is subjected to review at least every ten years. When a document is more than ten years
old and has not undergone a revision process, it is reasonable to conclude that its contents, although still of
some value, do not wholly reflect the present state of the art. Users are cautioned to check to determine that
they have the latest edition of any IEEE standard.
In order to determine whether a given document is the current edition and whether it has been amended
through the issuance of amendments, corrigenda, or errata, visit the IEEE-SA Website at
http://ieeexplore.ieee.org/xpl/standards.jsp or contact IEEE at the address listed previously. For more
information about the IEEE-SA or IEEE’s standards development process, visit the IEEE-SA Website at
http://standards.ieee.org.
Errata
Errata, if any, for all IEEE standards can be accessed on the IEEE-SA Website at the following URL:
http://standards.ieee.org/findstds/errata/index.html. Users are encouraged to check this URL for errata
periodically.
Patents
Attention is called to the possibility that implementation of this standard may require use of subject matter
covered by patent rights. By publication of this standard, no position is taken by the IEEE with respect to the
existence or validity of any patent rights in connection therewith. If a patent holder or patent applicant has
filed a statement of assurance via an Accepted Letter of Assurance, then the statement is listed on the IEEE-
SA Website at http://standards.ieee.org/about/sasb/patcom/patents.html. Letters of Assurance may indicate
whether the Submitter is willing or unwilling to grant licenses under patent rights without compensation or
under reasonable rates, with reasonable terms and conditions that are demonstrably free of any unfair
discrimination to applicants desiring to obtain such licenses.
Essential Patent Claims may exist for which a Letter of Assurance has not been received. The IEEE is not
responsible for identifying Essential Patent Claims for which a license may be required, for conducting
inquiries into the legal validity or scope of Patents Claims, or determining whether any licensing terms or
conditions provided in connection with submission of a Letter of Assurance, if any, or in any licensing
agreements are reasonable or non-discriminatory. Users of this standard are expressly advised that
determination of the validity of any patent rights, and the risk of infringement of such rights, is entirely their
own responsibility. Further information may be obtained from the IEEE Standards Association.
5
Participants
At the time this guide was completed, the Dielectric Frequency Response Working Group had the following
membership:
Ali Naderian, Chair

Peter Werelius, Vice Chair
Poorvi Patel, Secretary
Emil Bercea Nathan Jacob Thomas Prevost

Arup Chakraborthy Rudra Kharel Kirk Robbins
Dinesh Chhajer Egon Kirchenmayer Diego Robalino
Michael Franchek Mani Kumar Subhas Sarkar
George Frimpong Mark Lachman Daniel Sauer
Detlev Gross Mario Locarno Charles Sweetser
Ismail Guner Shawn Luo Ajith Varghese
Thomas Hartmann David Murray Rogerio Verdolin
Ronald Hernandez Ali Naderian Sukhdev Walia
Jose Izquierdo Arturo Nunez Baitun Yang
Mark Perkins
The following members of the individual balloting committee voted on this guide. Balloters may have voted
for approval, disapproval, or abstention.
Thomas Barnes Nathan Jacob Christopher Petrola

Barry Beaster John John Donald Platts
W.J. (Bill) Bergman Laszlo Kadar Alvaro Portillo
Wallace Binder Gael Kennedy Thomas Prevost
Thomas Bishop Sheldon Kennedy Iulian Profir
Thomas Blackburn James Kinney Ulf Radbrandt
William Bloethe Zan Kiparizoski Oleg Roizman
W. Boettger Neil Kranich Zoltan Roman
Chris Brooks Jim Kulchisky Thomas Rozek
Kent Brown John Lackey Dinesh Sankarakurup
Demetrio Bucaneg, Jr William Larzelere Daniel Sauer
Paul Cardinal Aleksandr Levin Bartien Sayogo
John Crouse J. Dennis Marlow Ewald Schweiger
Gary Donner Joseph Melanson Hyeong Sim
Jorge Fernandez Daher Daniel Mulkey Jerry Smith
Namal Fernando Jerry Murphy Steve Snyder
Bruce Forsyth Ryan Musgrove Brian Sparling
George Frimpong Ali Naderian P. Michael Spurlock
Ramsis Girgis K. R. M. Nair David Tepen
Jalal Gohari Kris K. Neild James Van De Ligt
Randall Groves Michael Newman Roger Verdolin
Bal Gupta Joe Nims John Vergis
John Harley Lorraine Padden Jane Verner
Roger Hayes Bansi Patel David Wallace
Werner Hoelzl Dhiru Patel David Wallach
Gary Hoffman Branimir Petosic Joe Watson
Richard Jackson Peter Werelius
6
When the IEEE-SA Standards Board approved this guide on 27 September 2018, it had the following
membership:
Jean-Philippe Faure, Chair

Gary Hoffman, Vice Chair
John D. Kulick, Past Chair
Konstantinos Karachalios, Secretary
Ted Burse Xiaohui Liu Robby Robson

Guido R. Hiertz Kevin Lu Dorothy Stanley
Christel Hunter Daleep Mohla Mehmet Ulema
Joseph L. Koepfinger* Andrew Myles Phil Wennblom
Thomas Koshy Paul Nikolich Philip Winston
Hung Ling Ronald C. Petersen Howard Wolfman
Dong Liu Annette D. Reilly Jingyi Zhou
*Member Emeritus
7
Introduction
This introduction is not part of IEEE Std C57.161-2018, IEEE Guide for Dielectric Frequency Response Test.
DFR testing has been used worldwide for several years. This guide provides a review of the DFR
methodology and guidance in the practical application and interpretation of the results. This guide also helps
users to perform DFR testing in a recommended consistent manner to allow for a better trending and
comparison of the data.
This guide is applicable to the methods of Dielectric Frequency Response (DFR) of liquid-immersed
transformers. The guide includes recommendations for instrumentation, procedures for performing the tests,
and techniques for analyzing the data. This guide can be used in both field and factory applications. The
purpose of this guide is to provide the user with information that will assist in performing Dielectric
Frequency Response measurements and interpreting the results from these measurements.
8
Contents
1. Overview .................................................................................................................................................. 10
1.1 Scope ................................................................................................................................................. 10
1.2 Purpose .............................................................................................................................................. 10
2. Normative references................................................................................................................................ 10
3. Definitions ................................................................................................................................................ 10
4. Dielectric frequency response (DFR) measurement overview ................................................................. 11

4.1 DFR test ............................................................................................................................................. 11
4.2 Use of DFR ........................................................................................................................................ 11
4.3 Differences between DFR and single-frequency PF testing techniques ............................................ 15
4.4 DFR moisture application .................................................................................................................. 16
4.5 General DFR measurement test practice............................................................................................ 17
5. Making a DFR measurement .................................................................................................................... 18

5.1 Safety ................................................................................................................................................. 18
5.2 Test environment preparation ............................................................................................................ 18
5.3 Test object preparation ...................................................................................................................... 19
5.4 Test set ............................................................................................................................................... 19
5.5 Measurement configurations ............................................................................................................. 19
5.6 Measurement ..................................................................................................................................... 20
6. Test records and report ............................................................................................................................. 23

6.1 Introduction ....................................................................................................................................... 23
6.2 Ancillary test information .................................................................................................................. 23
6.3 Test data............................................................................................................................................. 24
7. Measurement analysis and interpretation ................................................................................................. 27

7.1 Introduction ....................................................................................................................................... 27
7.2 Modeling............................................................................................................................................ 27
7.3 Factors influencing moisture estimate ............................................................................................... 29
7.4 Examples ........................................................................................................................................... 33
Annex A (informative) Dielectric frequency response analysis—Theory and validation ............................ 37

A.1 Theoretical basis for DFR analysis ................................................................................................... 37
A.2 Modeling dielectric response of a power transformer....................................................................... 43
A.3 Validation—Summary of reports ...................................................................................................... 52
Annex B (informative) Non-moisture related factors influencing the DFR measurements.......................... 60

B.1 Introduction ....................................................................................................................................... 60
B.2 Case 1—Influence of corrosive sulphur contamination .................................................................... 60
B.3 Case 2—Influence of carbon contamination ..................................................................................... 63
B.4 Case 3—Influence of contamination in tap changer ......................................................................... 65
B.5 Case 4—High resistance due to un-shorted core-to-ground resistance ............................................. 67
B.6 Case 5—Influence of poor grounding of electrostatic shield ............................................................ 68
B.7 Conclusion ........................................................................................................................................ 69
Annex C (informative) Examples of typical measurement challenges ......................................................... 71
Annex D (informative) Bibliography ........................................................................................................... 73
9
Response Test
1. Overview
1.1 Scope
This guide is applicable to the methods of Dielectric Frequency Response (DFR) of liquid-immersed
transformers. The guide includes recommendations for instrumentation, procedures for performing the tests, and
techniques for analyzing the data. This guide can be used in both field and factory applications.
1.2 Purpose
The purpose of this guide is to provide the user with information that will assist in performing Dielectric
Frequency Response measurements and interpreting the results from these measurements.
2. Normative references
The following referenced documents are indispensable for the application of this document (i.e., they must
be understood and used, such that each referenced document is cited in text and its relationship to this
document is explained). For dated references, only the edition cited applies. For undated references, the latest
edition of the referenced document (including any amendments or corrigenda) applies.
IEEE Std C57.152™-2013, IEEE Guide for Diagnostic Field Testing of Fluid-Filled Power Transformers, Regulators,
and Reactors.1, 2
3. Definitions
The IEEE Standards Dictionary Online should be consulted for terms within this standard.3
1
IEEE publications are available from the Institute of Electrical and Electronics Engineers (http://standards.ieee.org/).
2
The IEEE standards or products referred to in Clause 2 are trademarks owned by the Institute of Electrical and Electronics Engineers, Incorporated.
3
IEEE Standards Dictionary Online is available at: http://dictionary.ieee.org.
10
IEEE Std C57.161-2018
IEEE Guide for Dielectric Frequency Response Test
4. Dielectric frequency response (DFR) measurement overview
4.1 DFR test
DFR measurement is a non-intrusive, non-destructive off-line testing technique developed as an advanced diagnostic
tool. It can be performed in the frequency domain to determine the moisture content of the transformer’s solid
insulation and conductivity of the liquid insulation.
All dielectric response methods reflect the same fundamental polarization and conduction phenomena in transformer
insulation. The oil-paper insulation system is a composite of two different dielectric media, where an insulating liquid
with ionic conduction is mixed with a less conducting impregnated solid (pressboard or paper). The insulation system
has its own dielectric response, which not only reflects the properties of each material but also the way they are
combined. Therefore, the geometrical arrangement of the insulation system, its thermal condition, the condition of the
solid insulation, and he condition and the type of oil, all have a significant impact on the dielectric response.
The representation of dielectric response in the frequency domain is termed dielectric frequency response (DFR). The
results can be displayed as parameters associated with various equivalent circuits including capacitance and power
factor (or dissipation factor) as a function of frequency. The test is also known as Frequency Domain Spectroscopy
(FDS).
4.2 Use of DFR
The DFR measurement is, in principle, similar to the well-known capacitance and power factor/dissipation factor test
obtained by the measurement of complex impedance at one specific frequency value (close to line frequency
50/60 Hz). For a DFR measurement, the capacitance and power factor/dissipation factor is measured over a wide
frequency range, typically from an upper limit of 1 kHz down to a lower limit that ranges between 10 mHz and
0.1 mHz. The frequency range may be subject to change based on the physical and thermal conditions of the insulation
under test.
The resultant dielectric spectrum is a unique representation of the condition of the combined insulation system inside
the transformer.
4.2.1 Theoretical basis for power factor (PF) and dielectric frequency response (DFR) testing
techniques
A transformer can be modeled as a network of capacitances, inductances and resistances. There is a need for a
convenient technique by which insulation may be tested. An electric field (E) is applied to an insulating material
represented in electric circuits as a capacitor (C). A potential difference (V) can be maintained by means of a voltage
source. This creates a positive charge on one plate and a negative charge on the other plate of the capacitor. The charge
on the plates is proportional to the potential difference between them. See Equation (1).
Q
C= [F] (1)
V
where
C is the capacitance in Farads (F)

Q is the charge in Coulombs (C)
V is the potential difference in volts (V)
For parallel capacitors, the electric flux density (D) will be given at the surface of the plate, therefore
11
IEEE Std C57.161-2018
Q  C 
D=  2 (2)
A m 
where
A is the surface area of the capacitor’s plate
For a capacitor where “vacuum” is in-between plates, the correlation between flux density (D) and the electric field
(E) can be expressed as follows in Equation (3):
D V
E= m (3)
ε0  
where
ε0 is the permittivity of vacuum and its magnitude is 8.854 × 10−12 (F/m)
Assuming infinitely long capacitor’s plates, the field is uniform and thus the flux density is the same everywhere
between plates. The electric field is related to the charge through geometrical characteristics and dielectric properties
of the capacitor as presented in Equation (4).
Q  V × m −1 
=E (4)
ε0 × A  
The potential difference V is the work needed to move a charge, in the case of a capacitor, i.e., it is the work needed
to move the charge from one plate to the other, thus
Q×d
V = ∫ − E × dl = E × d = (5)
ε0 × A
where
d is defined as the distance between capacitor’s plates
Combining Equation (1) and Equation (5) defines capacitance as shown in Equation (6).
ε0 × A
C= (6)
d
By replacing vacuum with another dielectric material, the capacitance value will change.
The term ε r is now introduced as the relative permittivity of the insulating material representing the ratio of the
amount of electrical charge stored in a material by an applied voltage, relative to that stored in a vacuum.
ε0 × εr × A
C= (7)
d
The absolute permittivity ε is given by the following product in Equation (8):
ε=
ε0 × εr  F × m −1  (8)
 
12
IEEE Std C57.161-2018
The term ε r is a dimensionless number and is obtained from the ratio of two measured capacitances: capacitance with
the insulating material to free-space capacitance. Therefore, the capacitance depends on the ratio of free charge to total
charge. At this stage, Equation (2) can be modified for the condition where an insulating material is present between
the electrodes of a capacitor and the flux density becomes as in Equation (9).
D = ε0 × εr × E C × m −2  (9)
 
Table 1 —Relative permittivity values of materials
Relative permittivity
Material
εr
Vacuum 1.0
Air 1.00055
Mineral Oil 2.2
Silicone Fluid 2.7
Natural Esters 3.2
Synthetic Esters 3.2
Oil Impregnated Paper 3.8−4.2
Oil Impregnated Pressboard
Low Density 3.8
High Density 4.6
Water (20 °C) 80
Capacitive current (Ic) is directly related to the capacitance.
I c = V × ω × C = 2π × f × C × V [A] (10)
From Equation (10), it is clear that capacitive current is dependent on frequency (f) and therefore for repetitive
measurements, the specimen (capacitor) tested at similar voltages and the same frequencies should provide similar
responses (assuming no deterioration of the insulation material). This important correlation between current and
capacitance also provides information regarding the mechanical condition of the insulation system. Following basic
circuit theory, it is obvious that the capacitance is increased by connecting capacitors in parallel and it is reduced by
connecting capacitors in series. This is fundamental for the proper understanding of the capacitance values obtained
by measurement during a typical dissipation factor/power factor test.
The dielectric insulating material is not perfect; therefore, losses are to be expected in the system and those should be
measured. The approach is taken by measuring dielectric loss. In order to represent the losses of the dielectric material,
the current producing the dielectric loss of the insulating material is in phase with the applied test voltage. This current
(IR) is created by polar contaminants in a dielectric that are influenced by electrical stress. This resistive component is
very small in the transformer insulation and its increase is due to a variety of factors such as aging by-products,
chemical contamination, moisture in the cellulose, high oil conductivity, carbon tracking, or high resistance in the core
(Perkins, et al. [B19]). 4
The relationship between the total current (IT), resistive current (IR) and capacitive current (IC) is shown in Figure 1
and Figure 2. Capacitance (C) [Equation (10)], power factor (PF) [Equation (11)], and dissipation factor (DF)
[Equation (12)] are measured identically on a single-frequency power factor test or in a multiple frequency DFR test.
4
The numbers in brackets correspond to those of the bibliography in Annex D.
13
IEEE Std C57.161-2018
Figure 1 —Complex Impedance measurement (simplified diagram)
IR
Power factor
= cos=θ (11)
IT
IR
Dissipation factor
= tan=δ (12)
IC
Figure 2 —Power factor/dissipation factor vector diagram
In a simple way, power factor and dissipation factor (tan δ ) are very close in magnitude only if the angle δ is very
small (<5 degrees) or tan δ is below 10%. The power factor correlates to the dissipation factor as follows in Equation
(13):
14
IEEE Std C57.161-2018
Power factor  Cos   PF ;

Dissipation factor  tan   DF ; and,
   90    ; (13)
sin  sin  90    sin(90)  cos( )  cos(90)  sin( ) cos 

 DF  tan      ;
cos  cos(90   ) cos(90)  cos( )  sin(90)  sin( ) sin 
where
  is the angle between IR and IT, Figure 2
Replacing PF and DF within the expression above, yields the following:
cos 
DF  tan   ;
sin 
PF PF PF (14)
DF   
sin  2
1  cos  1  ( PF )2
It is important to mention here that because DFR is a different application of the same power factor technique; both
test procedures are highly dependent on frequency and temperature.
4.3 Differences between DFR and single-frequency PF testing techniques
Power factor is a trending value usually monitored to evaluate the condition of the insulation system and typically
performed in the field as a routine test at 10 kV or at a voltage not greater than the rated voltage of the insulation
system under test. The purpose is to determine the degree of contamination or deterioration of the insulating material
without stressing the system significantly. The great majority of oil-impregnated transformers in good condition show
the power factor at line frequency not exceeding 0.5%. (IEEE Std C57.152™-2013 [B13]) at 20 ºC. Values showing
accelerated increase of losses or beyond standard limits require further investigation.
Table 2 enumerates some of the differences between single frequency power factor testing and DFR.
Table 2 —Differences between typical power factor and DFR tests
Parameter Power factor test DFR test

Variable frequency from 1 kHz down
Line frequency or close to line
Frequency to values in range of
frequency (50/60 Hz)
10 mHz to 0.1 mHz
10 kV or below apparatus rated
Voltage 140 VRMS a
voltage
Time is dependent on stop frequency
Measurement time <1 min (mHz), typically takes several
minutes to several hours
Analysis Trending, standard references Instantaneous modeling
Used to identify general Primary focus of moisture content
Objective of test
contamination of insulation estimate and oil conductivity
a
On very low capacitances and in high EMI locations, HV DFR may be used. See 7.3.4.
15
IEEE Std C57.161-2018
4.4 DFR moisture application
DFR displays the fundamental polarization and conduction phenomena in transformer winding insulation, which is a
combination of oil gaps and solid insulation dielectric responses. As a result, it is influenced strongly by both solid
insulation moisture content and oil condition, and also, but less strongly, by the geometry of the solid and liquid
insulations. Mathematical modeling was seen as the key to determining how measured responses are affected by oil
conductivity and moisture content.
On this basis, the complex relative permittivity εˆ (ω ) at the frequency of the applied field, assuming a capacitive test
object, can be found in Equation (15). It is important to notice that the imaginary part of the complex relative
permittivity, ε ′′ (ω ) , (loss part) contains both the resistive (dc conduction) losses and the dielectric (polarization)
losses. Another way of presenting the measured information of DFR is to use the dissipation factor as shown in
Equation (16).
ˆ (ω ) ε ′ (ω ) − jε ′′ (ω )
ε= (15)
ε " (ω )
DF (ω ) tan
= = δ (ω ) (16)
ε ' (ω )
Figure 3 —Dielectric frequency response ( ε ′ , ε ′′ , DF, and C) of a typical inter-winding insulation with
0.5% moisture at 25 °C in full measurement spectrum from 1 kHz down to 0.1 mHz
The example presented above, for reference only (Figure 3), shows the dielectric response of a typical inter-winding
insulation with 0.5% moisture in the solid insulation and a liquid conductivity of 1 pS/m. In the frequency range from
1000 Hz down to about 10 Hz, the losses in the cellulose, and the relation between cellulose and liquid, determine the
DF level. In the frequency range from 0.1 Hz down to 0.01 Hz, the liquid conductivity dominates the losses. At
frequencies below 0.01 Hz, the effect of liquid losses are less dominating and the relation between liquid and solid
16
IEEE Std C57.161-2018
insulation gives this section its characteristic response. Cellulose losses also affect these lowest frequencies, more
pronounced if cellulose has high moisture content. See Annex A for a more detailed description.
The variable and fixed input parameters are the following:
a) Normally fixed:
 Liquid insulation relative permittivity [omitted in Equation (17)]
 Thermal activation energy for solid and liquid insulation [omitted in Equation (17)]
 Temperature (T), [omitted in Equation (17)]
b) Variable:
 Liquid insulation conductivity ( σ )
 Moisture in solid insulation (m)
c) Fixed or variable within limits:

 Geometry parameters (X, Y)
 Geometrical capacitance (C0)
See Annex A for further details on the variable and fixed parameters.
The thermal activation energy of the liquid insulation is only needed when the oil conductivity is measured directly
from an oil sample and used as a known parameter in the mathematical model.
In the moisture analysis, a mathematical algorithm finds the best fit between the modeled and measured dielectric
response by computing the minimum error as follows:
2
min ∑  F ( C0 , X , Y , σ , m, fi ) − datai 
C0 , X ,Y ,σ , m i
(17)
where the function F represents the modeled response; datai the measured response; and fi the frequency; all of them
in logarithmic form and parameters as defined previously.
4.5 General DFR measurement test practice
The dielectric response measurement uses the same preparation setup and execution procedure as traditional (50 Hz
or 60 Hz) capacitance (C) and power factor (PF)/dissipation factor (DF) measurements as illustrated in Figure 4.
In general, the insulation region of focus for DFR measurement is the one containing the biggest amount of solid
insulation within the specimen under test. For better interpretation, a typical two-winding transformer is represented
in Figure 4 where the inter-winding insulation between high-voltage (HV) and low-voltage (LV) windings (CHL) is
not influenced by other components such as bushings and clearly represents the majority of the solid insulation in the
overall transformer dielectric system. In addition, the high-voltage side insulation to ground and the low-voltage side
insulation to ground can also be measured. The objective of these additional measurements (CH and CL) is to identify
possible non-typical responses on those insulation areas.
In Figure 4, CHL represents the capacitance between HV and LV windings, CH represents the capacitance between
high-voltage winding and ground; and, CL represents the capacitance between low-voltage winding and ground. In the
case of a three-winding transformer, CHT represents the capacitance between high (H) and tertiary winding (T) and CT
represents the capacitance between the tertiary winding (T) and ground.
17
IEEE Std C57.161-2018
Figure 4 —DFR test set connected to a two-winding transformer
It is recommended to record the transformer nameplate data and insulation temperature. Temperature is critical for
accurate assessment of the insulation condition. See 5.6.1 for temperature recording method. It is important to
emphasize that thermal equilibrium is quite difficult to obtain on a unit that has been recently removed from operation.
Running the test during fast dynamic thermal changes will generate incoherent results and confusion, for example
immediately after de-energizing the unit. Recommendations with regard to test records are provided in Clause 6.
5. Making a DFR measurement
5.1 Safety
As with any electrical test, making a dielectric frequency response measurement should be performed in a safe and
controlled manner regardless of test location. The measurement should be carried out by qualified and authorized
personnel. Considerations for electrical safety in testing apply not only to personnel, but also to the transformer and
test equipment. Prior to testing, involved personnel should discuss the test procedure and environment for ensuring
that the work to be performed and any safety precautions are clearly understood. Other safety aspects are covered in
the industry standards, company, or local regulations and manufacturer’s instruction manual. Any transformer under
test should be completely isolated from any high-voltage source or power system source. The transformer tank should
have a solid ground. During the measurement, there should be strict adherence to local safety regulations and
guidelines.
5.2 Test environment preparation
Environmental factors may influence the measurement under test, and must be considered during the interpretation
and analysis of the results. Under ideal conditions, the temperature of the insulation should remain constant, and close
to 20 °C. It is not unreasonable to have temperatures as low as 10 °C or as high as 40 ºC. Measurement of an
ungrounded specimen test such as CHL to assess moisture content is very robust against influence of weather and
moisture. However, the grounded specimen tests, such as CH and CL, are sensitive to environmental conditions.
Specifically, humidity combined with bushing geometry and cleanliness of its surface condition are known to
influence the results of the dielectric loss measurements. Experience shows that at 50/60 Hz, with bushing surfaces
cleaned, the losses can be influenced when humidity exceeds 85−90%.
18
IEEE Std C57.161-2018
5.3 Test object preparation
It is recommended that the transformer be in as close to “in service” condition as possible. All bushing connections,
including the neutrals, should be disconnected from the grid and be short-circuited prior to performing the test. The
jumpers used to short the winding should be bare conductors, and shall be isolated from ground surfaces. If CH and
CL measurements are performed, it is imperative to make sure the bushings are dry and clean. Accessible tertiary
winding typically has bushings to allow the tertiary winding to be used during testing. In such case, the DRF test can
be performed similar to a 2-winding (or a 3-winding) transformer. After the test is completed, restore the grounding
to the original connection.
Whenever possible all test lead connections should be made directly to the bushing terminals. If the bus bar is
connected, its support insulation will influence the CH (high side winding to ground) measurement, and the CL (low
side winding to ground) and CT (tertiary side winding to ground) measurement. Therefore, as a general guideline,
external bus bar connections should be removed whenever possible.
The low-voltage DFR measurements can be affected by substation interference such as electrostatic and
electromagnetic energy. Therefore, it is important to ensure proper ground connections of all floating and un-energized
terminals as well as the transformer tank during the measurements to provide a controlled path for interference.
Voltages higher than typical 140 V RMS for DFR test, improve the signal to noise ratio helping to reduce the effect
of substation EMI (Werelius, et al. [B20]).
Special considerations should be given to safety when testing a transformer without oil so excessive voltages are not
applied or induced in a combustible environment. DFR modeling input parameters will have to reflect air in lieu of
oil.
If internal current transformers are present, they should be configured in the in-service conditions. If the core is
grounded through resistor(s), the resistor should be shorted during measurements.
5.4 Test set
The test set should be grounded in accordance with the recommendations of the test equipment manufacturer, or to
the same point as the transformer under test, in the absence of the equipment manufacturer’s recommendations.
Dielectric response test equipment should be furnished with selectable test modes that simplify the testing of the
complex insulating systems. The two basic test modes are grounded specimen test (GST) and ungrounded specimen
test (UST). Test equipment should also include additional guard circuitry that allows for variation on these two modes,
thus allowing each section of the complex insulation systems to be tested separately if possible. For simplicity,
instrument manufacturers may use the same test modes as those used for single frequency power factor testing, these
are: UST, GST-GND, and GST-GUARD. As with power factor measurements, the use of shielded test cables is
recommended.
The test equipment should always be calibrated within the recommended calibration interval. When possible, prior to
use, a self-check of the operation of the test equipment using a standard test object with a known DFR response may
be employed as a means of assuring correct operation of the equipment.
5.5 Measurement configurations
In estimating the moisture content of the paper/pressboard insulation for power transformers, emphasis is given to the
major insulation, the inter-winding insulation such as that between the high-voltage and the low-voltage windings in
a two-winding transformer. This insulation is commonly denoted as CHL. Supplementary measurements can be
performed, such as measuring the winding insulation to ground, commonly referred to as the CH (high side winding
to ground) and the CL (low side winding to ground) insulation. The importance of these supplementary measurements
is to identify possible non-typical responses of those insulation system areas.
19
IEEE Std C57.161-2018
5.5.1 UST measurement
The UST configuration is used for measurements between two terminals of a test specimen that are isolated from
ground. In the UST configuration, only the current flowing in the insulation between the two terminals is measured.
Any current path to ground, bypasses the metering circuit, therefore this current is not measured. A connection diagram
for a UST measurement on a power transformer is shown in Figure 5.
Figure 5 —Measurement of CHL using UST configuration
5.5.2 GST measurement with guard
The grounded specimen test (GST) measurement is performed between two terminals of a test specimen, of which
one terminal is connected to ground. If other terminals are present and associated with the specimen, the use of the
guard circuit eliminates these undesired current paths from the measurement. A high-voltage winding to ground (CH)
connection diagram for a GST guard measurement on a power transformer is shown in Figure 6.
Figure 6 —Measurement of CH using GST with guard configuration
5.6 Measurement
5.6.1 Record the temperature
The temperature of the insulation system should be recorded. Dielectric response measurements are influenced by
temperature and is preferably to perform after the transformer has been de-energized for many hours or several days,
or else soon after de-energizing. Temperature is not homogeneous within transformer insulation. Typically,
temperature is the highest towards the top of the transformer winding, and tends to decline towards the bottom of the
20
IEEE Std C57.161-2018
transformer winding. Several methods can be applied to estimate the insulation temperature. The methods include
direct oil measurements, top oil temperature gauge readings, measurement using a thermal gun, or by comparing a
winding resistance measurement against factory test data or resistance at a known temperature. It is recommended that
this test be done in a stable insulation temperature condition, preferably after the transformer has been de-energized
for many hours or several days. Since the temperature of a transformer can drop rapidly immediately after de-
energization, it is not recommended to perform this measurement right after the transformer has been de-energized.
Further, it is important to record the temperature, just prior to the measurement and again after completion. The follow
up ensures that the temperature has not changed by more than ±5 °C during the measurement. Simulations have shown
that the estimated moisture does not change significantly within this temperature range of the measured temperature.
The longer the time spent in acquiring the DFR response, the higher the possibility of thermal fluctuations and erratic
assessment.
5.6.2 Environmental condition
Record the ambient temperature, relative humidity, and weather conditions of the test environment.
5.6.3 Record the transformer nameplate information
See 6.2.1.
5.6.4 LTC tap position
While for the great majority of transformers, the DFR measurements are not influenced by the position of the LTC,
there are exceptions. A known example corresponds to designs that include resistor in the path to ground. This resistor
in the Neutral position contributes to measured losses increasing the power factor value. These units should be tested
in the off-Neutral position to remove the influence of the resistor.
5.6.5 DFR test procedure
The DFR test is typically performed by applying a frequency spectrum energy signal to the insulation system under
test and measuring the applied voltage, current, and phase angle to determine the specimen capacitance and power
factor over the frequency range of interest. The test connections are the same as those used for the standard power
factor and capacitance measurements. Refer to Figure 4, Figure 5, and Figure 6.
The measurement can be performed in either frequency domain or time domain. Data presentation and analysis for
both methods are carried out in the frequency domain.
DFR test is performed using low voltage. Selecting the test voltage level is a compromise—too high a voltage may
introduce significant voltage dependent effects at low frequencies; however, too low a voltage may make it impossible
to suppress interference currents that may be several orders of magnitude greater when compared to insulation loss
current at low frequencies. Test voltage levels typically used today are 140 V RMS, but higher voltages may be used
in high interference environments (Werelius, et al. [B20]).
The full frequency range used during the measurement is 1000 Hz to 0.1 mHz. The test starts at 1000 Hz to reduce
the effects of the remnant charge introduced by test voltage at lower frequencies. The need for the measurements in
the 10 mHz to 0.1 mHz range is influenced by temperature, insulation dryness, and condition of the oil (Cheng, et al.
[B2]). In cases where, during the test, a review of the data determines that sufficient information is present (e.g., the
loss peak, known as “the hump” of the power/dissipation factor versus frequency curve is observed), the test can be
stopped since there is not a need for measurements at lower frequencies.
5.6.6 Test connections
Connect the ground of the instrument to the ground of the transformer tank. Connect the instrument cables to the
instrument then to the transformer.
21
IEEE Std C57.161-2018
All windings should be shorted, using un-insulated jumpers, including the neutral connection. Note the neutral shall
be disconnected from ground. Table 3, Table 4, Table 5, and Table 6 describe the test connections for the various
winding configurations commonly found in transformers.
Table 3 —Two-winding transformer tests

Measurement Insulation Moisture content
Test Source outputa Guard
inputa measured assessment
High-voltage Low-voltage Transformer tank
CHL CHL Recommended
winding winding ground
High-voltage Transformer Low-voltage
CH CH Supplementary
winding tank ground winding
Low-voltage Transformer High-voltage
CL CL Supplementary
winding tank ground winding
aThe output/input are referred to test equipment.
Table 4 —Autotransformer without tertiary

Measuring Insulation Moisture content
Test Source output Guard
input measured assessment
CH Autotransformer Transformer CH Recommended
winding (H-X) tank ground
Table 5 —Autotransformer with a tertiary

Source Measuring Insulation Moisture content
Test Guard
output input measured assessment
CHT Autotransformer Tertiary Transformer tank CHT Recommended
winding (H-X) winding ground
CH Autotransformer Transformer Tertiary winding CH Supplementary
winding (H-X) tank ground
CT Tertiary Transformer Autotransformer CT Supplementary
winding tank ground winding (H-X)
Table 6 —Three-winding transformer

Source Measuring Insulation Moisture content
Test Guard
output inputs measured assessment
CHLX High-voltage Low-voltage
Transformer CHLX Recommended
(see NOTE) winding (H) winding (X)
tank ground and
LVY winding
CHLY High-voltage Low-voltage Transformer CHLY Recommended
(see NOTE) winding (H) winding (Y) tank ground and
LVX winding
CLXLY Low-voltage Low-voltage Transformer CLXLY Recommended
(see NOTE) winding (X) winding (Y) tank ground and
HV winding
CH High-voltage Transformer LVX and LVY CH Supplementary
winding (H) tank ground windings
CLX Low-voltage Transformer High-voltage and CLX Supplementary
winding (X) tank ground LVY windings
CLY Low-voltage Transformer High-voltage and CLY Supplementary
winding (Y) tank ground LVX windings
NOTE—Minimum one of the three recommended measurements is required for moisture estimation.
22
IEEE Std C57.161-2018
6. Test records and report
6.1 Introduction
It is important that measurements are made in a consistent manner; details about the specimen and measurement
arrangement must be recorded so that accurate assessment of the insulation condition (moisture content) is possible.
Any test parameter that could affect moisture estimation in the post-analysis should be identified in a clear, concise
format such that the DFR signatures can be reconstructed for analysis and interpretation without the use of proprietary
software. The presence of external environmental interferences should also be noted as a potential source of error in
test records.
6.2 Ancillary test information
This subclause refers to test information that is recorded to aid in the interpretation of the test data displayed in DFR
measurement curves. A number of different elements of a particular test need to be recorded—these help ensure better
accuracy in the measurement analysis, interpretation, and reporting.
The ancillary test information given here is separated in items relating to the test specimen, test equipment used, test
organization (performer), test measurement set-up, and assessment parameters. Under each informative item, the
suggested minimum information is listed in the following subclauses. Additional data can be inferred from asset
management systems within the transformer owner’s organization.
6.2.1 Transformer (test specimen) information
The following is the “static” data suggested to be part of the test record for identifying the transformer for use in data
management, which does not vary between tests on a particular transformer:
a) Manufacturer
b) Year of manufacture
c) Serial number
d) MVA rating, and cooling modes (ONAN, OFAF, ODAF, etc.)
e) Winding configuration and voltage ratings
f) Substation/Location
g) Asset ID
h) OLTC and DETC type and position
6.2.2 Test equipment information
Record the following data of the test equipment used for the measurement:
a) OEM name, equipment name, make/model number
b) Equipment serial number
c) Year of manufacture
d) Calibration date
6.2.3 Test organization (performer)
This records by whom the testing was performed, as follows:
23
IEEE Std C57.161-2018
a) Company name
b) Instrument operator—Name of technician or test engineer
6.2.4 Test set-up information
The following are elements that may vary between tests on a given oil filled transformer on a given day.
a) Date and time of measurement
b) Ambient temperature
c) Weather conditions (indoor, rain, snow, fog, etc.)
d) Ambient relative humidity
e) Measurement configuration (insulation component under test: CH, CHL, CL, etc.)
f) Transformer oil level (no oil, partially filled, without oil)
g) Supply voltage magnitude
h) Start/Stop frequency values
6.2.5 Assessment parameters
The following are parameters that directly affect the analysis and moisture estimation of a DFR measurement. Properly
knowing and using these parameters may directly affect the quality and accuracy of the analysis. Whereas, temperature
and fluid type are critical, all of the other parameters are used to better support the analysis if they are available.
Required parameters are as follows:

 Top oil temperature
 Winding temperature (if available)
 Type of insulating fluid (mineral oil, natural ester, etc.)
Optional parameters are as follows:

 Oil conductivity
 Insulation geometry (X/Y—see Figure 10)
 Transformer construction—Core form, shell form
 Type of material used in solid insulation (if known)
6.3 Test data
The measured data collected in a DFR test include the following: capacitance, percentage power factor and/or
percentage dissipation factor ( tan δ ) at a constantly applied test-voltage over a wide range of frequencies.
6.3.1 Data curves
The number of data points collected in a measurement should be sufficient such that the data may be fitted with a
continuous curve when the power factor/dissipation factor is plotted as a function of frequency (on a logarithmic
scale).
24
IEEE Std C57.161-2018
Because the curve is plotted on a logarithmic horizontal axis, the concentration of data points per decade in the lower
frequency range is the same as that used for the medium and higher frequencies. This allows proper curve fitting and
easy visualization. The example provided in Figure 7 shows the frequency response of a dry specimen. The frequency
range for DFR measurement is discussed separately in other subclauses.
Figure 7 —Data points on the DFR curve representing 0.5% moisture, 1pS/m oil conductivity
at 25 °C
6.3.2 Data format
The measured data obtained from a DFR test should be able to be exported to most standard applications to allow
flexible analysis of the results, comparison between results from different test systems, and inclusion in reports.
6.3.3 External conditions
External influences in the test environment at the time when measurements are being performed can adversely impact
the accuracy of measurement results. Many times these influences can be observed as a distortion in the DFR
measurement curve. Whenever such external influences exist, they should be noted in the test record.
Some external influences are due to problems in the measurement circuit such as poor condition of grounding
connections, or lack of a stable power supply. Other external influences may be caused by excessive ambient relative
humidity and moisture combined with contamination on bushing porcelain surfaces. This could result in measurement
errors due to elevated losses not caused by the internal insulation. Refer to 5.2 in this document.
Another category of external influences is elevated electromagnetic interference, which can be coupled into the
measurement circuit (Werelius, et al. [B20]). This interference can be produced from nearby overhead transmission
lines, or power electronic equipment such as motor drives or high-voltage direct current (HVDC) converters.
25
IEEE Std C57.161-2018
6.3.4 Test results
Once accurate and reliable test data has been acquired during the test, mathematical algorithms and a reference
database (see for more detail in Annex A) are used to process the data and estimate the percentage moisture
concentration weight/weight in the solid insulation and the conductivity value of the liquid insulation (pS/m). These
values are the final results of the testing process and shall be included in the final report.
6.3.5 Test report
An example DFR measurement test report is provided in Table 7. The reported data shown is only suggested, and the
user may choose not to include some of the information shown, or they may choose to include additional information.
An example of a dielectric frequency response graph to include in the test report is shown in Figure 7.
Table 7 —Example DFR test report

Test Instrument Information
Serial # 2305
Software version: v. D.1.2
Calibration date: 2012-01-23
Operator: TP
Supply voltage: 140 V RMS
Frequency range: 1 kHz−1 mHz
Test Object Information
Asset Owner: Utility XYZ Serial #: 1234567.89
Asset ID: Bank 123 Voltage rating: 138 kV/33 kV
Location: South Switchyard Winding Configuration: Dyn1
Manufacturer: Transformer Company Cooling Insulating Fluid: Mineral oil
Manufacturing year: 1970 Class: ONAF
Measurement configuration: UST
Test Results
Test date: 2012-04-24 Xa: 20%
Test time: 10:25 am Ya: 20%
Moisture in cellulose calculated
Air temperature: 20 °C 0.5%
[%]:
% Relative humidity: 40%RH Conductivity [S/m]: 1.0E−12
Insulation temperature: 20 °C
aX and Y are defined in 7.2.
26
IEEE Std C57.161-2018
7. Measurement analysis and interpretation
7.1 Introduction
In this clause, the necessary information regarding the analysis of the dielectric response in the frequency domain is
described. It is important to understand the DFR fundamentals introduced in Clause 4, and the way the modeling
system works to estimate moisture in the solid insulation and conductivity of the liquid insulation. Further details are
provided in the discussions included in Annex A through Annex C.
7.2 Modeling
The insulation system of a liquid immersed, cellulose insulated power transformer consists of two material groups,
the liquid insulation and the solid insulation.
The liquid insulation dielectric properties can be described by a simple parallel RC-circuit, the capacitance
representing the liquid permittivity and the resistance the liquid conductivity. The DFR response for a parallel RC-
circuit, with a capacitance of 22 nF, representing a relative permittivity ( ε r ) of 2.2 and a geometrical capacitance (C0)
of 10 nF, and a resistance equivalent to a conductivity ( σ ) of 1 pS/m looks as shown in Figure 8a). The capacitance
is constant and does not vary with frequency whereas a constant conductivity results in a dissipation factor (DF) that
increases as frequency decreases.
a) b)
Figure 8 —a) DFR of only liquid insulation with ε r = 2.2, σ = 1 pS/m, C0 = 10 nF

b) DFR of only solid insulation at 20 °C with 0.5% moisture concentration, C0 = 10nF
The dielectric properties of solid insulation are more complex, both capacitance and losses vary with frequency in a
manner that cannot be described by a simple parallel or series RC-circuit. Attempts have been made to describe the
dielectric properties by many RC-circuits. The best approach known today is to describe the response using a set of
other equations. For more details, see Annex A. The DFR response for a high density, mineral oil impregnated
pressboard at 20 °C and 0.5% moisture in an insulation system with geometrical capacitance of 10 nF is plotted in
Figure 8b). The capacitance at 60 Hz is about 42 nF and capacitance decreases slightly as frequency increases. The
dissipation factor (DF) behaves in a different way, DF decreases with frequency down to about 1 Hz and below 1 Hz,
the DF increases.
27
IEEE Std C57.161-2018
For new, un-aged, well-dried (<0.5% moisture content) liquid impregnated cellulose insulation, the dielectric response
depends mainly on the density of the cellulose material. Higher density gives higher losses. Some differences based
on raw material also exist but are assumed to only have minor effects in estimating the moisture content.
Figure 9 —Basic structure of CHL duct insulation of core type transformer (top view)
The insulation system between the windings of a core form power transformer can be illustrated as in Figure 9. The
insulation consists of solid insulation and cooling ducts where the insulation liquid flows. The relationship of
liquid/solid insulation system in radial and circular direction is described by the parameters X and Y, respectively. A
simplified X-Y model, Figure 10, is then used to calculate the total insulation system response consisting of the liquid
and solid insulation. The total response then depends on the geometry (C0) and the relative amount of liquid and solid
insulation (CIGRE TB 254 [B3], CIGRE TB 414 [B4]) where “r” in equation for Y is the average radius. Equation
(18) describes the relative permittivity illustrated by the simplified X-Y model presented in Figure 10.
Figure 10 —Simplified X-Y model for duct insulation
   
   
 Y   1− Y 
=ε (ω , T )  + (18)
  1− X   X     1− X   X 
  ε Spacer  +    ε + 
   ε Barrier     oil   ε Barrier  
28
IEEE Std C57.161-2018
where
Total thickness of barriers
X × 100%
width of duct
Total width of spacers
Y × 100% (for core form)
2π r
Y = Total spacer block coverage in washers (for shell form)
Figure 11 shows the liquid insulation, the solid insulation and the total response with X = 20% and Y = 20%. At higher
frequencies, the liquid insulation losses are very low and the total losses are determined by solid insulation losses and
the geometry; the more liquid insulation the lower the losses in the combined insulation system.
When the frequency is lowered, the losses caused by the liquid insulation dominate the response. At the lowest
frequencies, the DF is highly influenced by the geometrical parameters. For dry insulation and low conductivity liquid,
the pressboard moisture does not affect the lowest frequencies. In cases where the geometrical parameters are not
known, it is recommended to perform sensitivity analysis.
Figure 11 — Dielectric response a) Dissipation factor; b) Capacitance of only liquid insulation with
ε r = 2.2, σ = 1 pS/m (red), of only solid insulation at 20 °C with 0.5% moisture concentration (blue)
and of both materials combined with X = 20% and Y = 20% (black)
7.3 Factors influencing moisture estimate
There are many things that may influence the estimate of moisture content. The most important are discussed in detail
in the subclauses that follow.
29
IEEE Std C57.161-2018
7.3.1 Winding configuration and measurement set-up
Measurement of insulation between windings (CHL, CHT, and CLT) is recommended for moisture assessment. In
cases where the CH or CL measurement configurations are used, the influence of bushings and possible environmental
contamination are added, e.g., bushing creep currents. Also, the tap changer may influence the reading. Usually the
estimate of moisture from measurements of winding-to-ground insulation tends to overestimate the moisture content.
7.3.2 Influence of cellulose material type
Generally, both the real part of permittivity (capacitance) and the complex part of permittivity (loss) increase if density
of oil-impregnated cellulose material increases. Also, the dissipation factor increases with material density, however,
the effect is smaller. The dissipation factor at 20 °C, 60 Hz for oil-impregnated cellulose with 0.5% moisture content
typically ranges from 0.3% (crepe paper) up to about 0.6% (high density pressboard).
There is naturally a difference in properties of winding insulation paper, e.g., use of Kraft paper versus thermally
upgraded paper. However, no publication indicate that this effect has any significant effect on the dielectric response
of the insulation system between the windings, CHL, where the pressboard barriers, sticks, and the insulation liquid
are located.
7.3.3 Influence of low molecular weight acids
Low and high molecular weight carboxylic acids (LMWA and HMWA, respectively) are polar organic compounds
that appear in transformers as by-products of the aging of the oil and cellulosic materials. The LMWA are mostly
absorbed by the cellulose and to a much lesser degree in the oil. LMWA are highly soluble in water. The HMWA are
mostly absorbed by the insulating liquid and to a much lesser extent in the paper. The HMWA are not significantly
water soluble. This allows for extraction of LMWA from oil using water, which will leave the HMWA in the oil phase.
Due to its polar nature and easy absorption by the cellulose, the presence of LMWA in cellulose affects the dielectric
frequency response (DFR) in a manner similar to the influence of water. These similarities in the dielectric response
make it difficult to separate the influence of either LMWA or water on the measured data. As a result, the DFR
moisture assessment on aged units with high amount of LMWA can be overestimated. Therefore, in the cases where
the insulation system is assessed with high moisture content, further analysis, e.g., analysis of an oil sample, may
improve the information to decide the best maintenance action. Alternatively, a correlation between the LMWA
content in solid insulation and shifts in DFR traces will allow for automatic compensation for the presence of LMWA
in the cellulose once their content in oil is determined.
The error in moisture content when the concentration of LMWA in the pressboard is in the order of 1 mg KOH/gcellulose
has been estimated to be between 0.08% to 0.15% (CIGRE TB 414 [B4]). It is important to note that concentration of
LMWA in the pressboard is what contributes to the moisture estimate and not what is in the oil. For a neutralization
value 1 mg KOH/gcellulose of LMWA in the cellulose, the corresponding equilibrium neutralization value of just the
LMWA in the oil is estimated to be in the order of magnitude of 0.01 mg KOH/goil (i.e., 1:100 partitioning of LMWA
between oil and cellulose) (CIGRE TB 414 [B4]). For transformers rated <69 kV, the suggested limit of total acid
content for continued use of mineral oil is 0.2 mgKOH/goil, while for transformers rated ≥230 kV, the suggested limit
is 0.1 mgKOH/goil (IEEE Std C57.106™-2006 [B12]). Using an average content of LMWA in aged oil (20%) derived
from measurements from 41 oil samples from aged transformers (Hernández, et al. [B10], [B11]) and 1:100
partitioning of LMWA between oil and cellulose, it is estimated that the maximum percent error in moisture content
at the limit of suggested use of oil for transformers rated <69 kV is in the range 0.32%−0.60%. The equivalent range
at the limit of oil use for transformers rated ≥230 kV is 0.16%−0.30%.
Figure 12 shows an example where the DFR data obtained from an oil-impregnated pressboard sample with no acids
and 2.8% moisture is quite similar to the data obtained from a sample with 1.9% moisture content and acidity of
2.46 mg KOH/gcellulose (associated with a LMWA content of 0.06 mg KOH/goil in the oil). In other words, if this was
a real transformer with 1.9% moisture in the presence of 2.46 mg KOH/gcellulose of LMWA content in the cellulose, the
DFR model without compensation for LMWA content would erroneously assess moisture content to be 2.8%.
However, note that 0.06 mg KOH/goil of LMWA content corresponds to an equivalent 0.3 mg KOH/goil total acid
30
IEEE Std C57.161-2018
content (assuming 20% of total acid content is LMWA), which is above the acceptable limit for continued use in
power transformers and should trigger further investigation of the moisture estimated by the DFR model.
Figure 12—%DF vs. frequency of oil-impregnated pressboard showing influence of LMWA
7.3.4 Influence of ac and dc interference
Performing measurement of capacitance and dissipation factor of transformer insulation in a substation means that
interference currents are added to measurements. In terms of magnitude, switching noise in HVDC stations may reach
tenths of mA, ac interference and network frequency (50/60 Hz) and its harmonics are usually below 1 mA, typically
ranges from 10 µA to 100 µA. In addition to ac components with frequency components of network frequency and
above, there is interference of dc currents, and low frequency ac currents. Figure 13 is an example of measured dc
current, using 1 second of integration time, from a nearby 50 Hz corona source. In this specific case the measured ac
interference was a few hundreds of µA and the average dc current almost −0.02 µA (−20 nA). It is clearly observed
that the dc measurements give fluctuating results, and also a low-frequency component exists.
31
IEEE Std C57.161-2018
Figure 13 —Consecutive 1 s, dc interference measurements during ac corona interference
The high-frequency part of the interference is mostly taken care of by not applying a signal of frequencies of known
interference frequencies, i.e., 50 Hz, 60 Hz, 100 Hz, 120 Hz, 150 Hz, 180 Hz, etc. combined with signal processing.
However, low-frequency ac interference may interfere with the measurements, especially when measuring low
capacitances (<1 nF), since the losses are low at low frequencies. For example, the capacitive current produced in a
1.0 nF capacitor at 1 mHz by a 140 V RMS source is only 0.008 uA (0.8 nA). By visual inspection of Figure 13, it is
easily understood that this is not an easy task. One alternative, to get better signal to noise ratio, is simply to increase
the applied voltage, see 5.3.
The dc component of the interference has less influence, if any at all, and in cases of performing DFR measurements
using an ac signal, the frequency component of a dc-current is zero. However, in cases of performing polarization
current measurements, a dc component will be transformed into losses that will increase the measured losses above
the actual insulation losses, see Figure 14. This effect is especially important at lower frequencies, in cases of
measurement at 0.1 mHz, the actual capacitive current in a 1.0 nF sample (at 140V RMS) is only 0.08 nA.
32
IEEE Std C57.161-2018
Figure 14 — 1 nA dc current bias added to DFR response using an ac signal (C’_DFR and C’’_DFR) and
DFR response transformed from a dc measurement (C’_PDC and C’’_PDC)
7.4 Examples
7.4.1 Aged transformer with very low moisture content
A 400 kV Auto transformer, YNa0d1, from 1985 was measured in 2011. The insulation temperature was 19 °C, and
the insulation between HV+LV windings to tertiary windings was measured in the frequency range from 1000°Hz
down to 2 mHz. See Figure 15.
The oil conductivity was very low, only about 0.007 pS/m, meaning that solid insulation losses dominate from 1000 Hz
down to about 0.5 Hz. However, since the solid insulation dominates the response down to a frequency as low as
0.5 Hz at a temperature as high as 19 °C, the moisture could be determined to be very low, 0.4%, as for a newly
manufactured transformer, using data down to 2°mHz. To illustrate that higher moisture should have resulted in a
significant different response; a model curve representing 1% moisture content is added in Figure 16.
In Figure 16, the model response is set according to best match to measurement 1 kHz to 2 mHz with X = Y = 20%,
i.e., moisture content of 0.4% an oil conductivity of 0.0074pS/m) is extended down to 1 µHz (0.001 mHz). The model
data indicates that in this case not much further information is achieved by extending the measurement to 0.1 mHz.
33
IEEE Std C57.161-2018
Figure 15 —Capacitance (right-hand side Y-axis) and %DF (left-hand side Y-axis) of measurement and
model using X = 20% and Y = 20%
Figure 16 —Capacitance and %DF measurement: a) two alternative model curves (0.4% and 1%
moisture content) b) measurement plotted with 0.4% model curve down to 1 µHz
7.4.2 Aged transformer in reasonably good condition
An old auto transformer with a tertiary winding, YNa0d1 400/121/10 kV 250 MVA, was measured between the
HV+LV windings and the tertiary delta-winding. The transformer was de-energized weeks prior to test, and had an
insulation temperature of about 18 °C.
34
IEEE Std C57.161-2018
Initially, a model match was made setting X=20%, Y=20%, a geometry reasonably valid for many core form
transformers. The model fit gave an estimated moisture of 0.9% and an oil conductivity of 1.5 pS/m (at 18 °C).
It is obvious from Figure 17 that the measurement loss curve is above the model at high frequencies and below the
model at the lowest frequencies. This is an indication that the actual amount of solid insulation (especially in X-
direction) is higher than that used in modeling.
Since the measured data shows a strong influence of geometry (the hump at low frequencies is clearly visible), a
matching with X and Y as free parameters has to be performed. See Figure 17. The best match resulted in X = 44%
and Y = 23%. Later, a design expert was consulted. He confirmed that the transformer design included more solid
insulation in the insulation measured than usual. The model fit gave estimated moisture of 1.0% and an oil conductivity
of 1.8 pS/m (at 18 °C).
Figure 17 —Capacitance and %DF measurement and model using: a) X = 20%, and Y = 20% and b) X
and Y as free parameters
7.4.3 Aged transformer in poor condition
An old traction transformer, 25 MVA, 16 2/3 Hz two winding transformer was measured between high-voltage and
low-voltage windings (CHL) The transformer had an insulation temperature of about 18 °C and the power factor at
60 Hz was 1.48%.
DFR assessment using fixed X=Y=20% gave a moisture estimate of 3.4% with an oil conductivity of 71 pS/m. See
Figure 18.
In cases of aged and wet transformers, where losses are very high at low frequencies, the influence of geometric
parameters, X and Y, is usually very low. At lowest frequencies, the oil ducts are all shorted and only the solid
insulation is measured. An auto match including X and Y will estimate X and Y inaccurately but as mentioned, since
X and Y does not influence the result, it does not matter. Figure 18 shows that the moisture content with X and Y
parameters included in the match gave almost the same result, 3.5% as compared to 3.4%.
35
IEEE Std C57.161-2018
Figure 18 —Real and imaginary part of complex permittivity of measurement (orange) and model
(black). Moisture assessed to 3.4% and the oil conductivity to 71 pS/m
36
IEEE Std C57.161-2018
Annex A
(informative)
Dielectric frequency response analysis—Theory and validation
A.1 Theoretical basis for DFR analysis
If an electric field E(t) is applied to dielectric material that is isotropic and homogeneous, the current density, J(t),
through the surface of the material can be written as follows in Equation (A.1) (Jonscher [B16], Ildstad, et al. [B15]):
dD  t 
J t     E t    A  m 2  (A.1)
dt  
where
 is the dc conductivity and D(t) is the electric displacement
The first term in Equation (A.1) represents the conduction current, which is caused by the movement of free charges
in the material. The second term represents the displacement current, which is caused by changes in polarization in
the material due to the time varying nature of the applied electric field. D(t) is the dielectric displacement, which is
made up of vacuum polarization and polarization of the dielectric material, P(t):
D t   0  E t   P t  C  m2  (A.2)
 
For dielectric materials used in transformers, the polarization can be considered linear and is therefore proportional to
the electric field intensity. When the electric field is turned on at time t = 0, the polarization of the material takes some
time to react.
Because of the assumption of linearity of the dielectric material, one can define the dielectric response function, f(t),
as the proportionality factor between an electric field applied at time =  of infinitesimal duration t , and the
remaining polarization, P  t  .
P  t    0  E    t  f  t    (A.3)
When all contributions from all times are summed together and one assumes the electric field is zero at time t = 0 (i.e.,
the insulation has been shorted for a long time prior to the measurement), the expression for the total polarization
becomes:

P  t    0  E  t     f    d (A.4)
0
The dielectric response function is influenced by any effect that alters the polarization processes of the dielectric
material, for example moisture. As a result, dielectric response measurements can be used as a means to monitor
changes in the polarization processes of the insulating material. In the time domain, the response function can be
obtained by measuring the polarization and depolarization currents respectively during and after application of a dc
voltage.
37
IEEE Std C57.161-2018
Since the scope of this guide covers dielectric response analysis in the frequency domain, Fourier transforms are used
to derive the frequency domain equivalent of the response function and ultimately the current density:
  σ 
J (ω )= jωε 0 × ε r + χ ' (ω ) − j + χ " (ω )   × E (ω ) (A.5)
 ω ⋅ε0 
where the complex dielectric susceptibility, χ ′ (ω ) − j χ ′′ (ω ) is the Fourier transform of the response function, f(t)
(Helgeson and Gäfvert [B9]). If for the purpose of dielectric response analysis, one defines a complex relative
permittivity,
= ε ε ′ (ω ) − jε ′′ (ω ) , Equation (A.5) becomes:
J (ω=) jωε 0 × ε × E (ω ) (A.6)
where
ε ′ (ω=
) ε r + χ ′ (ω ) is the real part of the complex permittivity, and represents the real capacitance of the
dielectric material
σ
(ω )
ε ′′= + χ ′′ (ω ) is the imaginary part of the complex permittivity, and represents the losses in the dielectric
ωε O
material
From the expression for the loss, the effect of conductivity will be more pronounced at low frequencies than at high
frequencies.
As mentioned above, in the time domain the response function can be measured using polarization and depolarization
currents and then transformed into frequency domain for analysis in the frequency domain (DFR). In the frequency
domain, the response function can be measured directly by measuring the current through the insulation upon applying
a sinusoidal signal.
I (ω ) =ω × C0 × ε " (ω ) + jε ' (ω )  × V (ω )

(A.7)
= I R (ω ) + jI C (ω )
Where C0 is the geometric capacitance of the sample, IR and IC are the loss and capacitive components of the current.
Equation (A.7) can be represented in vector form as shown in Figure A.1.
38
IEEE Std C57.161-2018
Figure A.1—Vector diagram of voltage and current through insulating material
The ratio of loss to capacitive current is called the dissipation factor or loss tangent and represented as:
I R (ω ) ε " (ω )
δ (ω )
DF tan=
= = (A.8)
I C (ω ) ε ' (ω )
Note that for dielectric materials used in transformers, the range of values of power frequency dissipation factor are
the same as for the power factor (Cosine of the complementary angle to δ in Figure A.1).
To measure the components of the response function in the frequency domain, sinusoidal voltages of fixed frequencies
are sequentially, or simultaneously (Gäfvert and Nettelblad [B8]), applied to the insulation system. The applied voltage
and the current through the insulation are measured. The capacitance and loss at each frequency are then calculated
from the ratio of current to voltage and the phase difference between the current and the voltage. Knowledge of the
geometric capacitance allows for determining the real and imaginary parts of the complex permittivity. The results are
typically represented as a log-log plot of the real and imaginary parts of the permittivity, capacitance or the dissipation
factor versus frequency.
A.1.1 Effect of moisture on dielectric response of oil-impregnated cellulose
The addition of moisture to the solid insulation increases the polarization processes in the insulation system. This is
shown in Figure A.2 by the increase in dispersion (change with frequency) of especially the low-frequency region of
both the real and imaginary parts of the permittivity of pressboard samples as moisture content is increased.
39
IEEE Std C57.161-2018
Figure A.2—Permittivity of mineral oil-impregnated pressboard at 50 °C

a) real component; b) imaginary component
A.1.2 Effect of temperature on dielectric response in oil-impregnated cellulose
For oil impregnated pressboard with a certain amount of moisture content, the shape of the permittivity curves does
not change with temperature (CIGRE TB 254 [B3]). Instead, there is merely a shift in frequency of the entire curve
from one temperature to the next. The shift in the permittivity curves as a function of temperature follows an Arrhenius
− Ea
equation of the type, e kT with activation energy, Ea. This relationship allows for creating “master curves” with
extended frequency ranges by shifting measured permittivities at different temperatures to one reference temperature.
The temperature dependence of oil-impregnated pressboard samples with 1% weight moisture is shown in Figure A.3.
40
IEEE Std C57.161-2018
Figure A.3—Permittivity of mineral oil-impregnated pressboard with 1% (by weight) of moisture

a) real component; b) imaginary component
The master curve technique (Linhjell, et al. [B17]) is useful for creating a database of measured permittivities of oil-
impregnated pressboard with various moisture contents at different temperatures. This database is used as the basis
for analysis of moisture in the solid insulation of transformers. An example of such a database that was created as part
of a CIGRE working group on dielectric response is included in Ekanayake [B5].
A.1.3 Other effects on dielectric response of oil-impregnated cellulose
In the subclauses above, the effect of moisture and temperature on the dielectric response of oil-impregnated
pressboard has been demonstrated. However, there are other parameters, for example low molecular weight acids
(LMWA), that show a similar response as moisture on the dielectric response of oil impregnated pressboard
41
IEEE Std C57.161-2018
(Hernandez, et al. [B10], Linhjell, et al. [B17]). Further analysis of the composition of acids in the oil and their partition
coefficients between oil and pressboard may be necessary for distinguishing between the effects of moisture and
LMWA.
A.1.4 Dielectric response of mineral oil
Mineral oil is non-polar and non-dispersive and as such its dielectric losses under low electric field stresses are mainly
caused by the conduction of ionic impurities. The complex permittivity of the oil is composed of its dielectric constant
and dc conductivity (Gäfvert [B6]) and Equation (A.6) for the current density in mineral oil becomes:
 σ 
J (ω )= jωε 0 × ε r − j × E (ω ) (A.9)
 ω × ε 0 
Where ε r is the dielectric constant (2.2 for mineral oil) and 𝜎𝜎 is the dc conductivity of the oil, which is mainly
determined by impurities and aging products. Its temperature dependence is reasonably well described by an Arrhenius
law with activation energy around 0.5 eV.
Under low electric stress, oil has a linear response (no voltage dependence in permittivity), the real and imaginary
parts of the permittivity vary with frequency on log-log plot as shown in Figure A.4.
Figure A.4—Permittivity of mineral oil with dc conductivity 1E-12 pS/m
In reality, new mineral oil with low conductivity (~10−13 S/m) displays a non-linear response at low frequencies
depending on the voltage applied across the oil gap (Neimanis, et al. [B18]). Transformer liquids behave like weakly
dissociated electrolytes (Gäfvert [B6]) and display a perfectly linear Ohmic response when exposed to low voltages.
However, at higher voltages, non-linear effects are observed. When an ac voltage is applied across the liquid, the ions
will oscillate back and forth with amplitude given as shown in Equation (A.10):
Vµ
∆x = (A.10)
π fd
where
42
IEEE Std C57.161-2018
V is the applied voltage

µ is the ion mobility
f is the frequency of the applied voltage
d is the width of the oil gap
In order for the liquid to behave as a linear dielectric, the oscillation amplitude must be much smaller than the width
of the oil gap. This results in a critical ac voltage that can be applied before the liquid transitions into a non-linear
dielectric:
π fd
V << Vcrit = (A.11)
µ
It is clear from Equation (A.11) that for a given insulating liquid, this effect is more pronounced for small gaps and at
low frequencies, where it is likely to alter the response of the oil as a linear dielectric shown in Figure A.4. It also
suggests using the lowest possible voltages that can overcome the effects of noise when making dielectric response
measurements for DFR analysis.
It is also observed from Equation (A.11) that nonlinearities in the response of the complex permittivity of the oil will
have an impact on its conductivity. According to Equation (A.11), this effect is more pronounced at lower frequencies.
In particular, for a given insulating liquid, the minimum frequency before altering the linearity is proportional to the
voltage stress. Oil conductivity data for a naphthenic mineral oil is presented in Figure A.5. It is observed that the
minimum frequency is influenced by the voltage stress, e.g., in the data shown, the higher the stress the higher the
minimum frequency.
Figure A.5—Influence of voltage stress on naphthenic mineral oil dc conductivity
A.2 Modeling dielectric response of a power transformer
The insulation system of a power transformer is made up of oil and cellulose, both of which are affected by moisture
and contaminants. Mineral oil and oil-impregnated cellulose materials used in the insulation system of transformers
are isotropic, homogeneous and linear dielectric materials. Thus, dielectric response analysis may be applied to the
combined insulation geometry made up of insulating liquid and cellulose. To analyze the measured dielectric response
of a power transformer requires information about the geometrical design of the transformer. The geometrical
parameters needed for the analysis may be derived from the design drawings or estimated within limitations by
modeling software. The measurement would be made across one or more insulation ducts and the geometries of
interest are those across the measured ducts.
43
IEEE Std C57.161-2018
A.2.1 Winding configuration
In order to define a proper model, knowledge of the measured winding configuration is needed. Figure A.6 (a) and (b)
show two typical power transformer configurations. In a typical measurement for moisture estimation, voltage is
applied to the HV winding and current is sensed at the LV winding, which stresses the insulation between LV and
HV. In the case of winding configuration shown in Figure A.6 (a), there is one insulation duct between HV and LV
windings. However, in the case of winding configuration shown in Figure A.6 (b), there are two parallel ducts between
HV and LV windings. This shows the importance of having knowledge of what part of the transformer insulation is
included in the measurement when defining a proper model.
Figure A.6—Examples of power transformer configurations

(a) with one duct between HV and LV, (b) with two ducts between HV and LV
Equation (A.8) shows the need to know the geometric capacitance of the insulation being measured from the voltage
and current. For the transformer configurations shown in Figure A.6, the geometric capacitance is simply the
capacitance of the measured ducts between the windings assuming they were filled entirely with just air or vacuum.
For a core-type transformer the geometric capacitance of each duct is well estimated by the cylindrical capacitance:
ε 0 × 2π h
C0 = (A.12)
( r)
ln r2
1
where
r2 is the inner radius of the cylinder forming the outermost winding of the duct
r1 is the outer radius of the cylinder forming the innermost winding of the duct
h is the average winding height
Calculation of the geometric capacitance of shell type transformers is a bit more challenging and requires intimate
familiarity with the insulation layout of shell type designs.
A.2.2 Duct structure and model for analysis
The insulation between the windings is typically made up of a series of oil-impregnated pressboard barriers and oil
gaps in the radial direction of the windings. Spacers are used to support the barriers at equally spaced distances along
the circumference of the oil gaps. Figure A.7 shows the basic principle of the insulation structure between windings
of a core type transformer. This insulation structure is represented by a simplified X-Y model (Gäfvert, et al. [B7])
(Figure A.8), where X is the relative barrier amount and calculated as the ratio of the sum of all the thicknesses of
barriers in the duct to the width of the duct; Y is the relative spacer coverage along the perimeter of the oil ducts and
44
IEEE Std C57.161-2018
is the average of the ratio of total width of all spacers in an oil duct to the average length of the perimeter of that duct.
The relative amount of barriers (X) and the relative spacer coverage (Y), together with the geometric capacitance, Co
are used to characterize the duct geometry. Note that there have been other variations to the X-Y model that attempt
to address the influence of edge effects at the corners of the interface between spacers and oil (Cheng, et al. [B2]).
The range for the relative barrier amount (X) is typically 10% to 50% and the relative spacer coverage (Y) is typically
15% to 25% but could also depend on manufacturer design preferences and voltage class.
Figure A.7—Basic structure of duct insulation of core type transformer (top view)
Figure A.8—Simplified X-Y model for duct insulation
The effective complex permittivity of the duct structure shown in Figure A.8 can be expressed as shown in Equation
(A.13).
   
   
 Y   1− Y 
=ε (ω , T )  + (A.13)
  1 − X   X     1− X   X 
  ε Spacer  +    ε + 
   ε Barrier     Oil   ε Barrier  
The permittivities ε Spacer ,ε Barrier , which are made of oil-impregnated pressboard are represented by the complex
permittivity functions given in Equation (A.5) and ε Oil is represented by the complex permittivity function given in
Equation (A.9).
All parameters needed for analysis of DFR measurements are now in place—the X and Y quantities of the measured
duct are derived or estimated from the design drawings; the oil conductivity can be measured directly from an oil
sample or estimated during the analysis; and the average insulation temperature may be estimated from an average of
45
IEEE Std C57.161-2018
measurements of the top and bottom oil temperatures. For the pressboard components in the duct, a parameterized
version of permittivity measurements with various moisture contents and temperatures (see A.1) is used via an
algorithm to find the best fit of the real and imaginary parts of Equation (A.13) to the measured permittivity.
A.2.3 Simulation of various parameter influences
In this subclause, several simulations are made to show the influence of the main insulation parameters on dielectric
frequency response of a transformer duct. For these simulations, the assumed temperature is 30 °C, the oil conductivity
is 1 pS/m, the barrier content, X = 25%, spacer coverage, Y = 15% and the geometric capacitance is 10 nF.
A.2.3.1 Influence of moisture
Figure A.9 shows the influence of moisture on the capacitance ( C ′ ( ω ) = C0 × ε ′ ( ω ) ) and loss
( C " ( ω=
) C0 × ε " ( ω ) ) calculation of the transformer with the specified duct composition. Two moisture contents of
the pressboard material are simulated, 0.5% for a “dry” transformer and 3.0% for a “wet” transformer. It is observed
that the loss part of the response increases over the whole frequency range with higher moisture content in the
pressboard.
46
IEEE Std C57.161-2018
Figure A.9—Influence of moisture on DFR of a power transformer

a) Capacitance C’; and, b) Loss C”
Figure A.10 shows the loss tangent of the transformer insulation duct over the entire frequency range including the
power frequency value (also power factor). Again, it is observed that the dissipation factor ( tan δ ) increases over the
entire frequency range for higher moisture content in the pressboard insulation.
47
IEEE Std C57.161-2018
Figure A.10—Influence of moisture on DFR (DF) of a power transformer
A.2.3.2 Influence of oil conductivity
Figure A.11 shows the influence of oil conductivity on the capacitance and loss of the transformer with the specified
duct configuration. The pressboard insulation is assumed to be “dry”, with a moisture content of 0.5%. Two oil
conditions are simulated: new oil with conductivity of 1 pS/m and aged oil with conductivity of 10 pS/m. The increase
in oil conductivity is apparent only in the low frequency region of the capacitance curve, but it is evident in the loss
curves at all frequencies except the “high” frequencies. It should be noted that the profile of the response curves for
higher oil conductivity and higher pressboard moisture are quite distinctive.
Figure A.12 shows the loss tangent of the simulation with new and aged oils and the loss tangent changes for all
frequencies except at high frequencies. The power frequency values are shown on the curves and indicate the power
factor increases as well with aged oil in a “dry” transformer. It is again noted that the profiles for the dissipation factor
curve for higher oil conductivity and higher pressboard moisture are quite distinct.
48
IEEE Std C57.161-2018
Figure A.11—Influence of oil conductivity on DFR (capacitance and loss) of a power transformer
49
IEEE Std C57.161-2018
Figure A.12—Influence of oil conductivity on DFR (DF) of a power transformer
A.2.3.3 Influence of duct geometry and composition
The designs of insulation systems have varied over the long history of transformer manufacturing. Some old designs
may contain considerably more solid insulation than newer transformers. Figure A.13 shows simulations of
transformers with three different compositions of barriers and spacers in the duct between windings: X=10%, Y=15%;
X=25%, Y=15%; and X=50%, Y=25%. The solid insulation is considered “dry” with a moisture content of 0.5% and
the oil is considered “new” with conductivity of 1 pS/m. Since the composition of the duct affects the capacitance, the
differences, though small, are seen over the whole frequency range. On the other hand, the effect of different duct
composition on the loss is seen mostly on the lower and higher end of the frequency range.
50
IEEE Std C57.161-2018
Figure A.13—Influence of duct composition on DFR of a power transformer

a) Capacitance C’; and, b) Loss C”
Figure A.14 shows the dissipation factor for the transformer simulations with different duct compositions. The results
show the geometry affects the dissipation/power factor in the lower and higher ends of the frequency range including
power frequency.
51
IEEE Std C57.161-2018
Figure A.14—Influence of duct composition on DFR (DF) of a power transformer
A.3 Validation—Summary of reports
As part of the work performed by the IEEE Transformers Committee task force [B14] to investigate the viability of
DFR as a method of assessing moisture content in transformer insulation, the following table was assembled. The
table provides a summary of measurements and analysis of dielectric response parameters performed on power
transformers both in the factory and in the field having various years of service. The methods of validation were
performed by taking samples of cellulose from the transformers or making moisture estimation from moisture in oil
samples. Both positive and negative aspects of the dielectric methods are addressed. There is in general good
agreement between the dielectric response measurements and Karl Fischer moisture titration on paper samples, but
inconsistent agreement with moisture in oil samples.
52
IEEE Std C57.161-2018
Table A.1—Moisture prediction related DFR papers

Ref # Authors (Title) Source Method of validation Results
1 U. Gafvert, G. Frimpong, J. CIGRE General Quantitative analysis using a model based on Showed good agreement between (+)
Fuhr (Modelling of Conference, Paris, 1998, linear dielectric response and comparing the the PDC and RVM methods.
Dielectric Measurements on Paper 15-103 model to the measured curve of a 700 MVA Verified the results of both PDC
Power Transformers) generator step up transformer comparing the PDC and RVM can be quantitatively
and recovery voltage methods. modeled based on linear response
method.
2 Gafvert, U. Adeen, L. Properties and Four power transformers (1970). Tested FDS, Described possible effects of noise (−)
Tapper, M. Ghasemi, P. Applications of Dielectric PDC and RVM. Results compared against on the measurement process. The
Jonsson, B. (Dielectric Materials, 2000. moisture concentration in oil. tan  values reflect the different oil
Spectroscopy in Time and Proceedings of the 6th conductivities but the moisture
Frequency Domain Applied International Conference, levels in the oil, deduced from Karl
to Diagnostics of Power Issue Date : 2000, Fischer Titration seem unrelated to
Transformers) Volume : 2, On the moisture in the board deduced
page(s): 825 from the dielectric measurements.
3 Ekanayake, C. Gubanski, Electrical Insulation FDS test carried out on 105 transformers (0.5 up Good agreement is achieved (+)
S.M. Mularachchi, K.D. Conference and Electrical to 71 MVA, 0−25 years of service) and between the evaluations resulting
Fernando, M.A.R.M. Manufacturing & Coil comparison of measurements on a model rig of from FDS measurements and from
(Diagnostic of Power Winding Conference, oil-paper insulation at 85 °C, 65 °C, 41 °C and evaluations based on other analysis
Transformers in Sri Lanka 2001. Proceedings , Issue 20 °C. FDS before and after oil purification. (electrical - chemical).
Application of Dielectric Date: 2001, On
Spectroscopy in Frequency page(s): 593, Meeting
Domain) Date: 16 Oct 2001−18 Oct
2001
4 Perkins, M. Fazlagic, A. Electrical Insulation, 2002. Shell form and core form transformers The DFR is applicable to detect: high (+)
Frimpong, G. (Dielectric Conference Record of the signatures of the measured DFR responses of core ground resistance problems,
Frequency Response 2002 IEEE International Shell and Core form transformers are compared the amount of moisture in the
Measurement as a Tool for Symposium, Issue Date : with modeled responses of “normal” insulation insulation, chemical contamination
Troubleshooting Insulation 7−10 Apr 2002 , On structures (containing only moisture) and a library of windings. Authors have found
Power Factor Problems) page(s): 162 of signatures of known defects. the DFR method to be more
accurate than existing industrial
methods, for example, dew point
test and moisture/oil equilibrium
method.
Table continues
53
IEEE Std C57.161-2018

5 CIGRE Task Force D1.01.09 Report of CIGRE Task RVM, PDC, and FDS on insulation model The results of this report confirm (+)
Task Force members: Force D1.01.09, CIGRE evaluation of thermodynamic moisture that the dielectric response
S.M. Gubanski (chair), P. Report #254, 2002 equilibrium between paper and oil and; oil measurements provide valuable
Boss, G. Csépes, V. Der conductivity. Summary of study cases in (National information on the state of oil-paper
Houhanessian, J. Filippini, P. Grid Company) UK and (ABB & Vattenfall) insulation in power transformers, in
Guuinic, U. Gäfvert, V. Sweden. Tested 2 mm Transformer board Type T particular the moisture content.
Karius, J. Lapworth, G. IV impregnated with Technol US 3000 mineral Regarding the influence of
Urbani, P. Werelius, W. oil, SS electrodes, in an oil-filled glass vessel. geometry, it has an influence on the
Zaengl Moisture % in aged samples measured by KFT. response but not as significant as
(Dielectric Response Study cases presented in Section 6 (Pancake the effect of the oil conductivity.
Methods for Diagnostics of model - Kraft thermo 70 + Nynas Nitro 10GBN
Power Transformers) and Shell Diala D, and UK, Sweden, Switzerland
and Germany real transformers).
6 Neimanis, R. Arvidsson, L. Electrical Insulation FDS and 3-terminal cell for measurement on Relative permittivity defined to 2.2 (+)
Werelius, P. Conference and Electrical liquids with cell heater. Several oil samples were in mineral oil, frequency
(Dielectric Spectroscopy Manufacturing & Coil analyzed. dependence, activation energy, and
Characteristics of Aged Winding Technology voltage dependence.
Transformer Oils) Conference, 2003.
Proceedings , Issue Date:
23−25 Sept. 2003, On
page(s): 289
7 Linhjell D., Gafvert U., Electrical Insulation and FDS on paper samples: Kraft paper, Munskjo Apparently, it is not the DP of the (−)
Lundgaard L.E. Dielectric Phenomena, Termo 70, diameter 76 mm. Acidity and moisture cellulose, but the presence of aging
(Dielectric Response of Oil- 2004. CEIDP '04. 2004 content is measured on paper strips. Aging byproducts that influence the
Impregnated Paper Annual Report Conference performed in glass jars 70 g (dry weight) paper dielectric response. Results indicate
Insulation: Variation with on , Issue Date : 17−20 and 1.45 L of oil. Calculated amount of water was that chemical oil analysis is
Humidity and Ageing Level Oct. 2004 , On page(s): admitted and kept for 4 days at 45 °C, later required to be able to separate aging
[Power Transformer 262 impregnated with new dried degassed oil. from humidity.
Applications]) Samples were tested for moisture in oil and paper
by means of coulometric (electrochemical) KFT.
8 Gielniak, J. Ekanayake, C. Electrical Insulating Paper samples 1.5 mm thickness and 160 mm The results presented in this paper (+)
Walczak, K. Graczkowski, Materials, 2005. (ISEIM diameter. After drying samples, these were left at show the dependence of dielectric
A. Gubanski, S. 2005). Proceedings of 2005 the lab environment to absorb moisture from air. response of oil-impregnated
(Dielectric Responses of International Symposium, Weight changes were monitored. Immersed later pressboard on moisture content,
New and Aged Transformer Issue Date: 5−9 June 2005, in NYNAS Nytro 10 GBN. Fifteen samples with temperature and aging. It is difficult
Pressboard in Dry and Wet Volume: 2 , On page(s): the same moisture intake were placed into one to distinguish between aging by-
States) 386 container. Thermal aging and DP sampling. products and water content.
Coulometric KFT using Metrohm 756 KF
Coulometer instrument with 832 Thermoprep.
Table continues
54
IEEE Std C57.161-2018

9 J. Blennow, A. Dernfalk, In Proc. 14th International An old 40 MVA transformer was used; a The FDS method estimated (+)
K. Walczak, C. Ekanayake, Symposium on High comparison of FDS, KFT of paper samples and moisture content in the paper
P. Koestinger, L. Karlsso, Voltage Engineering, 2005, moisture in oil samples was performed at various insulation of 1.4−2.4%, which
U. Gäfvert, S. Gubanski, Page 505. stages: 1) Day transformer removed from service, compared well with the paper and
A. Bartnicki 2) Eleven weeks out of service, 3) After repair and pressboard of 1.0 to 2.5%. The
(Sequential Comparative short vapor phase, 4.) At retrofill and vacuum moisture estimation in oil
Study of Dielectric Response filtering of oil on site, 5) Transformer in service comparison with FDS is
and Analyses of Oil and for 6 weeks. questionable due to temperature.
Paper from a Power The results using Oommen curves
Transformer Undergoing was high; however, the temperature
Repair) used was questionable, so author
made an assumption on proper
temp, and the moisture content was
2.4%.
10 Blennow, J. Ekanayake, C. This paper appears in: FDS only. 3-ph 50/20 kV, 40 MVA transformer. Different moisture distribution at (+)
Walczak, K. Garcia, B. Power Delivery, IEEE Samples taken from different positions and tested different levels of the winding duct,
Gubanski, S.M. Transactions, Issue Date: KFT. Oommen’s curves were used to estimate good understanding of the factors
(Field Experiences with April 2006, Volume : 21, moisture content in blocks, the conductivity of the that may influence the results and
Measurements of Dielectric Issue:2, On page(s): 681 oil at different temperatures were derived from therefore also the interpretation of
Response in Frequency real measurements done in lab, dielectric response dielectric response measurements
Domain for Power of pressboard for the given moisture at given on power transformers in field
Transformer Diagnostics) temperatures were obtained from lab conditions.
measurements and/or recalculated.
11 L. S. Gubanski, J. CIGRE Conference, 2006, a) DFR on model samples: FDS on precise and Results of the investigations (+)
Blennow, l. Paper D1-207 reliable data from well-controlled oil impregnated performed allow believing that the
Karlsson, K. Feser, S. samples of pressboard and paper. Pressboard dielectric response measurements,
Tenbohlen, C. Neumann, H. samples 1.5 mm thickness and 160 mm in when properly performed and
Moscicka-Grzesiak, A. diameter. Paper samples 60 μm and 160 mm in interpreted, providing more
Filipowski, Tatarski diameter. Sample preparation included: drying, accurate information on moisture
(Reliable Diagnostics of HV moisturizing, impregnation with transformer oil content in paper and pressboard in
Transformer Insulation for and in some cases thermal aging. Transformer oil transformers than the use of
Safety Assurance of Power NYNAS Nytro 10 GBN. Five groups of samples conventional equilibrium curves.
Transmission System, were kept in containers for 6 months. Moisture estimation based on
REDIATOOL—a European b) Measurements on Model transformer: dielectric response measurements,
Research Project) Measurements performed with Coulometric KFT as well as on KFT analysis of oil
(as per IEC 60814). Test with RVM, PDC and samples in combination with the use
FDS. of the equilibrium curves, yield
c) Field Measurements: 80 transformers in similar results as the direct KFT
Europe, 17 in Sri Lanka. analyses of paper samples, if
d) Units sent for repair: paper and pressboard appropriate temperature values were
samples were collected from different areas and used.
analyzed with KFT.
Table continues
55
IEEE Std C57.161-2018

12 M. Koch, S. Tenbohlen, M. MatPost 07 Comparative study among RVM, PDC and FDS. Commercial software for PDC and (+)
Kruger, A. Kraetge Used pancake Model with constant 1% moisture, FDS are able to compensate for the
(Improved Moisture Analysis measured on paper and pressboard at 160 °C. Oil influences of insulation geometry,
of Power Transformers type Shell Diala D (  = 1.57 pS/m) and later 25- insulation temperature and oil
Using Dielectric Response year-old aged transformer oil (  = 16.5 pS/m). conductivity. RVM method too
Methods) simplistic. Generated new data pool
and new device.
13 Saha, T.K. Purkait, P. This paper appears in: Pancake Transformer model consisting of three Interpretation of RVM and PDC test (−)
(Investigations of Power Delivery, IEEE windings insulated with oil and cellulose. results remains a difficult task as it
Temperature Effects on the Transactions, Issue Date: Geometric detail of model is presented. PDC and is influenced by insulation aging
Dielectric Response Jan. 2008 RVM tests were performed. condition, geometry of insulation,
Measurements of moisture content, and also operating
Transformer Oil-Paper temperature. On-site test results
Insulation System) presented in the paper indicate the
necessity of careful understanding
of the effect of temperature on the
dielectric response measurement for
correct analysis and interpretation.
14 P. Patel, M. Perkins 2008 Weidmann Annual DFR (FDS) only. 1-ph, 2W GSY transformer Anomaly in DFR signature was (+)
(DFR – An Excellent Diagnostic Solutions producing combustible gases Hydrogen and used to detect abnormality on the
Diagnostic Tool for Power Technical Conference Acetylene. HV to ground path. Performed
Transformer) internal inspection on HV H1
bushing shield. Shielding tube was
loose and did not provide proper
contact. Performed on-site repair.
Fixed unit put back to service.
15 Jadav, R.B. Ekanayake, C. This paper appears in: Three sets of pressboard samples (density 1.21 PDC and FDS have significant (+)
Saha, T.K. IPEC, 2010 Conference g/cm3) with 0.2%, 1.8%, and 2.3% moisture level impact on moisture and
(Dielectric Response of Proceedings, Issue Date: were prepared. Response measurements taken at temperature. It is observed that
Transformer Insulation - 27−29 Oct. 2010, On 40 °C, 60 °C, and 80 °C. almost similar level of accuracy can
Comparison of Time Domain page(s): 199 be achieved while calculating dc
and Frequency Domain conductivity from both methods.
Measurements)
Table continues
56
IEEE Std C57.161-2018

16 Ekanayake, C. Saha, T.K. Power and Energy Society FDS with X-Y model and PDC on two groups of The presented quantitative analysis (+)
Ma, H. Allan, D. General Meeting, 2010 transformers: three in Sri Lanka and one AT in of FDS response through X-Y
(Application of Polarization IEEE , Issue Date: 25−29 Australia. Testing before and after oil model shows that geometrical
Based Measurement July 2010 refurbishment. parameters and oil conductivity is
Techniques for Diagnosis Of always interlocked. Comparison of
Field Transformers) the results from PDC and FDS
shows that estimated moisture
content from both techniques are
not equal but reasonably close. The
study shows the ability of
polarization techniques to identify
the status of transformer insulation
through appropriate quantitative and
qualitative analyses.
17 Ohlen, M. Werelius, P. Electrical Insulation Temperature variation on different paper samples, Advantages of DFR (FDS): (+)
(Dielectric Frequency (ISEI), Conference Record transformers and bushings. Measured 60 Hz PF Capability of performing ITC of
Response and Temperature of the 2010 IEEE values at different temperatures vs. new individual 50/60 Hz DF/PF, capability of
Dependence of Power International Symposium, temperature correction algorithm validated. estimating moisture in cellulose,
Factor) Issue Date : 6−9 June dissipation factor/power factor at
2010 , On page(s): 1 operating temperature, and to
investigate increased dissipation
factor in power components.
18 J. Blennow, K. Walczak , A. Nordic Insulation FDS performed on a 19.5/6.8/6.8 kV 40 MVA FDS measurements gave more (+)
Dernfalk, B. Garcia, A. Symposium—Nord-IS-05 - transformer. Results compared with oil and paper accurate estimates of moisture
Bartnicki Trondheim, Norway, June samples (KFT). levels than the estimates based on
S. Gubanski, O. Samuelsson, 13−15, 2005 oil analyses and equilibrium curves
U. Karlsson when compared to the paper
(Comparative Study of samples taken from the unit and
Dielectrric Spectroscopy moisture estimated by KFT.
Measurements and Analyses
of Oil and Paper Samples
from a Power Transformer)
19 R. Hernandez, 80th Annual International Influence of Low Molecular Weight Carboxylic Results show that the effect of (+)
M. Lachman, Doble Client Conference. Acids (LMWA) on DFR of oil-impregnated LMWA on the DFR of oil-
P. Griffin, April 7−12, 2013. Boston, pressboard samples is investigated by using impregnated cellulose is similar to
L. Lewand MA. different moisture contents and levels of acidity. the effects of water. Thus, it
(Progress Report on Moisture contents of samples are measured by imposes a positive bias on the
Experience with Dielectric KFT. estimation of water in cellulose via
Frequency Response DFR.
Measurements on
Transformers)
Table continues
57
IEEE Std C57.161-2018

20 CIGRE Working Group CIGRE Report # 414, FDS compared to KFT of paper and oil samples KFT of oil and paper samples and (+)
D1.01 (TF 14) April 2010 on two transformers that had to be opened for FDS on a 40 MVA, 50/10 kV
(Dielectric Response inspection. One of the transformers was seriously transformer. FDS gave moisture in
Diagnosis for Power aged and designated for scrapping. The influence the range 1.4−2.4%, KFT of paper
Transformer Windings) of high acidity in the oil and paper is seen in samples gave moisture in range of
results of FDS analysis without compensation for 1.0−2.5% and oil samples gave
acids. moisture result of 2.4%. On a
severely aged transformer
(30 MVA, 104/23.4kV), KFT of
paper samples gave 2.6% moisture,
FDS without compensation for
acids gave 3.8−4% moisture, which
FDS with compensation for acids
gave 2.9% moisture. Moisture
estimation from oil sample gave
6%, while relative saturation gave
2.5% moisture in paper.
58
IEEE Std C57.161-2018
A.3.1 Validation statistics
At least thirty-one articles on the subject of dielectric response analysis were reviewed. Of these, twenty were deemed
to have attempted some form of scientific validation of the moisture estimated via the dielectric response method. In
the figures that follow, the results are summarized according to the following:
a) Type of test object used for validation

b) Type of comparative methods used to validate the dielectric response method
c) Effectiveness of the measurement in estimating moisture in the test object
The numbers represented in the charts correspond to the number of articles containing each of the identified elements.
The number of transformers represented in the articles is at least 218.
Figure A.15—Dielectric response moisture validation statistics
59
IEEE Std C57.161-2018
Annex B
(informative)
Non-moisture related factors influencing the DFR measurements
B.1 Introduction
The dielectric frequency response (DFR) test method is primarily used for moisture estimation in cellulose insulation.
There are other factors, however, that can influence DFR measurements. In the case of a non-typical DFR response
or unexpected curve, the situation needs to be investigated further for problems other than moisture. These include
but are not limited to carbon tracking, contamination, presence of corrosive sulphur, effect of stray and creepage
currents, broken static shield and high resistance in the core to ground circuit. In addition there have been cases where
DFR curves showing a non-typical response along with higher dielectric losses, the presence of a hump in certain
frequency ranges, and the inability of being able to get a match against typical oil-paper DFR response, have led to
incorrect moisture estimation.
In these cases, it may be useful to know the design and construction of the transformer to better investigate unusual
DFR traces. Knowledge of the presence of a static shield, location of main and regulating windings, core ground
resistance, LTC design and mechanism and information on other diagnostic test results will help in diagnosing causes
of unexpected DFR response.
There are many internal and external factors (including connection errors) that can affect the DFR response. The user
should consider those factors that could influence the inter winding and winding to ground insulation response to help
in diagnosing issues that might not be identified otherwise. The following case studies cover some of the situations
that show the influence of various non-moisture related factors that explain the unusual DFR behavior. These case
studies below are somewhat specific and should not be generalized. Any DFR response that looks more different from
normal needs to be investigated individually by considering the design and operating conditions of the transformer
and the influence of factors external to the transformer.
Generally, it is not recommended to use a DFR measurement to evaluate moisture content where the measured
capacitance is very low, typically below 1 nF, or has a curve shape that poorly matches the XY-model assessing the
moisture content of the cellulose insulation.
In cases were the capacitance is reasonably large (more than 1 nF) but the XY-model gives a poor match, this usually
indicates that other effects besides moisture are influencing the DFR response. This annex presents a few cases
illustrating possible effects on DFR measurements by different kinds of degradation of transformer insulation systems.
B.2 Case 1—Influence of corrosive sulphur contamination
A three-phase, two winding 138kV/13.8 kV, 45/60/75 MVA, OA/FA/FA generator step up transformer, manufactured
in 1997 was taken out from service for standard electrical testing. The main insulation power factor was found to be
abnormal, with negative power factor results and high losses. In many cases, the physics for negative power factor
results, obtained in the UST measurement can be explained by having knowledge of the winding design or the presence
of an inter-winding shield. The negative power factor can also be caused by the resistive leakage to ground (e.g.,
carbon tracking). Also, as noted above, in cases where the measured capacitance/current is very low this can result in
negative losses given the effect of stray currents in the subject area. However, in this case, additional measurements
on other parts of the transformer windings showed elevated losses for reasons unknown at the time. DFR testing was
suggested to assist in diagnosing the cause of the abnormal 60 Hz power factors listed as follows:
 CH = −0.22%
 CL = 5.02% plus a 17% increase in capacitance
 CHL = 1.11%
60
IEEE Std C57.161-2018
The transformer had a winding configuration as shown in Figure B.1. For the DFR measurements, three configurations
were measured—CHL, CH, and CL.
Figure B.1—Winding configuration
The DFR analysis showed that the unit had an average bulk moisture level below 1% in all three measured duct regions
(CHL, CL, and CH).
In Figure B.2, for the measured main insulation duct, CHL, DFR results are shown with red marks and the model in
blue lines (the modeled data is based on the transformer being in normal condition). It should be noticed that the
dielectric losses are elevated in this unit indicating the insulation has significantly higher losses than in normal paper-
oil insulation.
Figure B.2—DFR results for Case 1
Based on the DFR results the customer decided to take the unit to a factory location for investigating the source of the
elevated power factor and abnormal DFR results. The unit was un-tanked and a coating of dark deposits was observed
61
IEEE Std C57.161-2018
on the windings, (as seen on Figure B.3). Wipe samples of the deposits were taken and analyzed. The analysis indicated
traces of sulfur and copper deposits.
Figure B.3—Layer of contamination on the windings
Further tear down of the windings showed more Cu2S (Copper Sulfide also known as Corrosive Sulfur) deposits and
also discolored copper, see Figure B.4.
Figure B.4—Teardown of the windings
62
IEEE Std C57.161-2018
B.3 Case 2—Influence of carbon contamination
A three-phase, two winding 24 kV/6.9 kV/6.9 kV, 30/40 MVA (ONAN/ONAF) transformer with dual low voltage
windings and LTC connected to the HV manufactured in 1999 was taken out of service after a Bucholz relay alarm
operated indicating a severe increase in dissolved gasses. It was observed that the winding resistance in phase three
was approximately 32% higher than the original factory levels. Other test results such as TTR, insulation resistance,
core resistance and SFRA did not indicate any issues. A DFR test was performed to assess insulation condition.
The transformer had a winding configuration as shown in Figure B.5. For the DFR measurements, the following
configurations were measured: HV to XV1, HV to XV2, HV to ground, XV1 to ground and XV2 to ground.
Figure B.5—Winding configuration of Case 2
The DFR analysis showed that the unit had an average bulk insulation moisture below 1% between the HV-windings
and XV1 and HV-windings and XV2.
Figure B.6 shows the XV1 to ground measurements (red marks) and calculated data (blue lines). It should be noticed
in Figure B.6 that there is a deviation between the measured and calculated dielectric losses, which could be due to
contamination, such as carbon tracking.
63
IEEE Std C57.161-2018
Figure B.6—DFR results for CL case
Based on the DFR results, high winding resistance values and elevated dissolved gasses, the customer decided to drain
the unit to perform an internal inspection. During the internal inspection, a severely carbonized XV lead connection
was found as seen in Figure B.7. The unit was repaired in the field.
Figure B.7—Layer of contamination on the XV exit leads
64
IEEE Std C57.161-2018
B.4 Case 3—Influence of contamination in tap changer
A three phase, two winding 132 kV/13.8 kV, 65.3 MVA generator step up transformer manufactured in 2007 was
stored as a spare unit for two years before being prepared for service. During the preparation, ice particulates were
discovered inside the unit. A dry-out was performed and DFR measurements were carried out to evaluate the bulk
insulation moisture content.
The transformer had a winding configuration as shown in Figure B.8, a regulating winding is connected to the low-
voltage winding and the high-voltage winding has a de-energized tap-changer (DETC).
For the DFR measurements, three configurations were measured—CHL, CH, and CL.
Figure B.8—Winding configuration of case 3
The DFR analysis showed that the unit had an average bulk insulation moisture below 1% and the oil conductivity
was as it would be for new oil. Figure B.9 shows the measured (red marks) and calculated dielectric losses (blue lines)
are aligned with each other. No evidence of any abnormality was observed in this path.
65
IEEE Std C57.161-2018
Figure B.9—DFR results for CLH
In Figure B.10, the high voltage to ground insulation duct (CH) was measured. The DFR results are shown with red
marks and the modeled in blue line (the modeled data is based on the transformer being in normal condition). It should
be noticed that the dielectric losses are higher in the frequency span between 0.1 Hz to 100 Hz, indicating the insulation
possibly has a slightly higher surface conductivity, (such as for carbon tracking) than in normal paper-oil insulation.
The CH path includes the HV bushings, and insulation on the end of HV winding to ground, and the winding/leads to
tank/ground.
Figure B.10—DFR results for CH
A series of measurements were done to try to locate the source of contamination. An internal inspection was also
performed.
66
IEEE Std C57.161-2018
The investigation and measurements pointed out that the problem may lie in the tap changer shaft and motor
mechanism assembly area of the high voltage de-energized tap-changer.
DFR was performed after disconnecting the glass fiber shaft shown in Figure B.11 (b) from the tap changer
mechanism. During disassembly, it was noticed that the motor drive had layers of rust on the surface, see Figure B.11
(a). With the tap changer dissembled from the motor mechanism, the DFR HV to ground measurement looked normal,
the abnormality “hump” was not observed.
(a) (b)
Figure B.11—(a) Rust on the motor mechanism, (b) glass fiber shaft disconnected from the
mechanism
The glass fiber shaft was replaced and the DFR was re-measured. Figure B.12 shows the high voltage to ground path
with the new tap-changer shaft. The measured (red marks) and calculated dielectric losses (blue line) were now aligned
with each other. No evidence of any abnormality was observed in this path after the repair.
Figure B.12—Capacitance and losses for the HV winding to ground case-post repair
B.5 Case 4—High resistance due to un-shorted core-to-ground resistance
While performing a typical DFR test, it is recommended that all the accessories of the transformer including the core-
to-ground resistance should be outside the measurement circuit zone. The core-to-ground resistance is shorted during
67
IEEE Std C57.161-2018
the DFR test. Otherwise, the power factor (in the range of 100 Hz to 1000 Hz) increases significantly as shown in
Figure B.13. The DFR test results of such cases could be misleading—both when estimating the moisture and finding
insulation deficiencies.
The reason is the core-to-ground resistance that is in series with the insulation system, disturbs the resistor part of the
measured impedance, adding an artificial component to the imaginary part of the measured capacitance (C”).
Figure B.13—Comparison of two DFR measurements (DF): Blue: core-ground resistor is not
shorted; Red: Core-ground resistor is shorted
B.6 Case 5—Influence of poor grounding of electrostatic shield
Noting failures of similar units, (three-phase substation transformers 230 kV/28 kV, 83 MVA) questions were raised
that led to checking other sister units for abnormalities. A DFR test was conducted primarily to estimate the moisture
content in the cellulose.
During the DFR test, it was noticed there was a winding shield between HV and LV, which was not documented on
the nameplate or any drawing. A utility representative verbally confirmed there was a modification 15 years after the
original manufacturing date to add static shielding.
DFR results, as seen in Figure B.14, showed high power factor at higher frequencies (100 Hz to 1000 Hz). This is
usually due to core-to ground resistance that is supposed to be shorted during the DFR test. That was not the case here
as the core was solidly grounded during the test (resistance was shorted). One possibility to have such a high resistance
on the shielding electrode could have been due to poor lead connections (inside or outside). In this case the shielding
external connection to ground was burned off (Figure B.15) and after internal inspection, it was confirmed that
shielding lead inside was also partially burned away.
68
IEEE Std C57.161-2018
Figure B.14—DFR results with high line-frequency dissipation factor
Figure B.15—Terminal with static shields grounded and one phase open
B.7 Conclusion
As stated Dielectric Frequency Response (DFR) is a diagnostic test method used primarily for moisture estimation in
cellulose insulation. The intent of this section was to show some specific examples of non-moisture related issues
affecting insulation systems and how they can influence the DFR measurement. It is by no means a comprehensive
list of possible defects affecting DFR responses.
It is important to recognize that many internal and external factors can affect the measurement. A strange DFR
response and unexpected curves need to be investigated for problems other than moisture. These include but are not
limited to carbon tracking, contamination, presence of corrosive sulphur, effect of stray and creepage currents, broken
static shield, and high resistance in core to ground circuit. It can also be useful to know something about the design
and construction of the transformer, core ground resistance, and LTC design along with information from other
diagnostic test results to better investigate unusual DFR traces.
69
IEEE Std C57.161-2018
Table B.1 is a chart showing the relative areas of influence within a particular part of a transformer insulation system
where non-moisture related issues are thought to have an effect.
Table B.1—Non-moisture factors and the suggested area of influence

Test setup
Factors
CHL CLT CHT CL CH CT
Static shield X X X
Bushing surface X X X
Non-homogeneous oil X X X X X X
Dirty tap-changer X X X
Stacked winding design X X
Non cellulosic material X X X X X X
Fluids other than mineral oil X X X X X X
Improper grounding X X X
Bushing gaskets X X X
As seen in Table B.1 above, the chart shows some typical examples of where non-moisture related factors that can
influence a DFR trace. (Please note this chart is informative only and is not to be considered absolute.) The information
provided in Annex B was included in the DFR guide to provide some examples of cases where non-moisture related
issues could affect the DFR response. The user would be well served not to consider it a complete list.
70
IEEE Std C57.161-2018
Annex C
(informative)
Examples of typical measurement challenges
Diagnostic testing of any sort brings with it a host of measurement challenges. Whether it is performed in the factory
or the field, a number of factors come into play.
A thorough understanding of the test instrument hardware and software is critical to the process of obtaining valid test
data. Equally important is a fundamental understanding of the test methodology. Additionally, test configurations,
temperature, and the surrounding environment will play a role in the eventual outcome. Subject material found in
other subclause of this guide discusses in more detail the effects of design, aging, time, and temperature on the
insulation system.
Therefore, this subclause presents examples of typical measurement challenges.
Several documents to date (IEEE Std C57.152-2013 [B13], CIGRE TB 414 [B4]) have stated the importance of
completely disconnecting the transformer from the system.
The measurements are often affected by interference in the substation, such as parasitic leakage currents and induced
electromagnetic disturbances. It is therefore important to secure proper grounding connections of the transformer tank
during the measurements to reduce the ground interference. Proper placement of connecting leads can also reduce the
influence of capacitive coupled or radiated noise. Use of shielded connecting cables is recommended in noisy
environments (CIGRE TB 414 [B4]). Field measurements have also been performed at higher voltages (>140 Vrms)
in substations where noise may be influencing the DFR readings (Werelius, et al. [B20]).
Care should be taken however to try and remove the presence of other parts of any system that is connected to the
whole unit; otherwise, stray capacitances can unduly influence the results. Most modern dielectric loss/PF test sets are
equipped with selectable test modes that simplify the testing of complex insulating systems. The two basic test modes
are grounded specimen test (GST) and ungrounded specimen test (UST). Test equipment usually includes additional
guard circuitry (GST-Guard) that allows for variations on these two modes, thus allowing each section of complex
insulating systems to be tested separately. It is important for individual sections of insulation to be tested separately if
possible, to prevent large sections from concealing the deterioration in small sections (IEEE Std C57.152-2013 [B13]).
The temperature of the insulation system in a transformer has a great influence on the results of dielectric response
measurements. Therefore, in order to perform reliable modeling and interpretation, correct temperature values are
needed. The most favorable measuring condition would appear at long-lasting steady and elevated temperature when
the equilibrium of moisture partition between oil and pressboard is reached. The reality is however different.
Transformer temperature often varies and it is not equal inside the tank—higher at the top than at the bottom. The
time available for the measurements is often also limited and it is not possible to wait the time required to attain steady
state. If the measurements are made just after taking the transformer out of operation, the temperature will be slowly
decreasing. Another practical problem related to the temperature determination is a question of which temperature
value, i.e., top-, middle-, or bottom-tank temperature, should be used in the analysis (Blennow, et al. [B1]).
Temperature should be always noted from as many sources as possible in order to estimate the correct temperature.
One recommendation is to register the temperature of oil just before starting the dielectric response measurements.
The most accurate way to determine the oil temperature is to take top and bottom oil samples and measure the
temperatures directly onsite in the sampled oil. After opening the tap, cold oil flows out first, thus waiting for sufficient
time is recommended in order to get a representative sample. If oil sampling involves too much effort, indications
from the built-in temperature gauges may be used, though their readings depend on location of the temperature probe.
Alternatively, the average winding temperature can be calculated from a comparison of the actual winding resistance
71
IEEE Std C57.161-2018
to a measurement at ambient temperature. It is also recommended to register the ambient temperature, relative
humidity and weather conditions in the station (CIGRE TB 414 [B4]).
Recent field measurements from a North American utility showed some difficulty with performing measurements on
transformers within an energized station due to elevated noise/interference.
While performing DFR measurements at an HVDC station some low frequency interference in the range of
0.2−0.05 Hz was noted. There was also 60 Hz noise of approximately ~8uA for UST measurements and ~85uA for
GST measurements. Further measurements have recorded similar low frequency interference in the range between
0.2−0.05 Hz. The opinion was that the interference is worse when the measurement leads are connected to the larger
HV bushings. This example supports the theory that there is capacitive charging/discharging current in this frequency
range coupled from overhead lines and bus-work in the station.
As with any diagnostic test, safety cannot be overemphasized when working on or around high-voltage electrical
apparatus. Companies that produce, generate, transmit, or use high-voltage electricity have precise rules for safe
practices and procedures. These practices are important for personnel whose responsibilities involve testing and
maintaining high-voltage apparatus.
72
IEEE Std C57.161-2018
Annex D
(informative)
Bibliography
[B1] Blennow J., Ekanayake C., Walczak K., Garcia B., and Gubanski S., “Field experiences with
measurements of dielectric response in frequency domain for power transformer diagnostics,” IEEE
Transactions on Power Delivery, vol. 21, no. 2, pp. 681−688, 2006.
[B2] Cheng J., Robalino D., Werelius P., and Ohlen M., “Improvements of the transformer insulation XY
model including effect of contamination,” IEEE International Symposium on Electrical insulation (ISEI),
2012, pp. 169−174.
[B3] CIGRE TB 254, Dielectric Response Methods for Diagnostics of Power Transformers, Task Force Report
D1.01.09, Paris, 2002.
[B4] CIGRE TB 414, Dielectric Response Diagnoses for Transformer Windings, Task Force report D01.01,
Paris, 2010.
[B5] Ekanayake C., “Diagnosis of Moisture in Transformer Insulation - Application of Frequency Domain
Spectroscopy,” Chalmers University of Technology, Gothenburg, Sweden, Thesis For The Degree of
Doctor of Philosophy, January 2006.
[B6] Gäfvert U., “Dielectric Response Analysis of Real Insulation System,” IEEE International Conference
on Solid Dielectrics, Toulouse, France, 2004.
[B7] Gäfvert U., Frimpong G., and Fuhr J., “Modeling of Dielectric Measurements on Power Transformers,”
CIGRE Session, Paris, 1998, Paper 15-103.
[B8] Gäfvert U., and Nettelblad B., “Measurement techniques for dielectric response characterisation at low
frequencies,” Nordic Insulation Symposium, NORD-IS, Lyngby, Denmark, 1990, paper 7.1.
[B9] Helgeson A., and Gäfvert U., “Dielectric Response Measurements in Time and Frequency Domain on
High Voltage Insulation with Different Response,” International Symposium on Electrical Insulating
Materials (ISEIM), Toyohashi, Japan, 1998, pp. 393−398.
[B10] Hernandez R., Lachman M., Griffin P., and Lewand L., “Experience with DFR-Based Moisture
Estimation in Transformers and Impact of Aging By-Products,” in 81st Annual International Conference
of Doble Clients, 2014.
[B11] Hernández R., Lachman M., Griffin P., and Lewand L., “Progress Report on Experience with Dielectric
Frequency Response Measurements on Transformers,” in 80th International Conference of Doble Clients,
Boston, 2013.
[B12] IEEE Std C57.106™-2006, IEEE Guide for Acceptance and Maintenance of insulating Oil in Equipment.
[B13] IEEE Std C57.152™-2013, IEEE Guide for Diagnostic Field Testing of fluid-filled Power transformers,
Regulators, and Reactors.
[B14] IEEE Transformers Committee, “DFR Task Force Final Report,” Task Force report PC57.161, 2012.
[Online]. http://www.transformerscommittee.org/subcommittees/dielectric_test/F13-C57.161-DFR-TF-
Final%20Report.pdf
[B15] Ildstad E., Thärning P., and Gäfvert U., “Relation between Voltage Return and Other Methods for
Measurements of Dielectric Response,” IEEE International Symposium on Electrical Insulation, 1994.
[B16] Jonscher, A. K., Dielectric Relaxation in Solids, Chelsea Dielectrics Press, 1984.
73
IEEE Std C57.161-2018
[B17] Linhjell D., Lundgaard L., and Gäfvert U., “Dielectric Response of Mineral Oil Impregnated Cellulose
and the impact of Aging,” IEEE Transactions on Dielectrics and Electrical Insulation, vol. 14, no. 1, pp.
156−169, February 2007.
[B18] Neimanis R., Arvidsson L., and Werelius P., “Dielectric Spectroscopy Characteristics of Aged
Transformer Oils,” IEEE Conference on Electrical Manufacturing and Coil Winding Technology, 2003,
pp. 289−293.
[B19] Perkins M., Fazlagic A., and Frimpong G., “Dielectric Frequency Response Measurement as a Tool for
Troubleshooting Insulation Power Factor Problems,” IEEE International Symposium on Electrical
Insulation, Boston, pp. 162−165, 2002.
[B20] Werelius P., Ohlen M., and Skoldin J., “Dielectric Frequency Response Measurement Technology for
Measurements in High Interference AC and HVDC Substations,” Techcon Asia-Pacific, 2011.
74
I
EEE
st
andards
.i
eee.
org
Phone:+17329810060 Fax:+17325621571
©IEEE

IEEE C57.161-2018-Guide-For-Dielectric-Frequency-Response-Test

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

IEEE C57.161-2018-Guide-For-Dielectric-Frequency-Response-Test

Uploaded by

Copyright:

Available Formats

IEEE Guide for Dielectric Frequency

IEEE Power and Energy Society

IEEE Guide for Dielectric Frequency

Approved 27 September 2018

IEEE-SA Standards Board

The Institute of Electrical and Electronics Engineers, Inc.

Copyright © 2018 by The Institute of Electrical and Electronics Engineers, Inc.

PDF: ISBN 978-1-5044-5211-3 STD23337

IEEE prohibits discrimination, harassment, and bullying.

Notice and Disclaimer of Liability Concerning the Use of IEEE Standards

Comments on standards should be submitted to the following address:

Secretary, IEEE-SA Standards Board

Laws and regulations

Updating of IEEE Standards documents

Ali Naderian, Chair

Emil Bercea Nathan Jacob Thomas Prevost

Thomas Barnes Nathan Jacob Christopher Petrola

Jean-Philippe Faure, Chair

Ted Burse Xiaohui Liu Robby Robson

4. Dielectric frequency response (DFR) measurement overview ................................................................. 11

5. Making a DFR measurement .................................................................................................................... 18

6. Test records and report ............................................................................................................................. 23

7. Measurement analysis and interpretation ................................................................................................. 27

Annex A (informative) Dielectric frequency response analysis—Theory and validation ............................ 37

Annex B (informative) Non-moisture related factors influencing the DFR measurements.......................... 60

Annex C (informative) Examples of typical measurement challenges ......................................................... 71

Annex D (informative) Bibliography ........................................................................................................... 73

4. Dielectric frequency response (DFR) measurement overview

4.1 DFR test

4.2 Use of DFR

C is the capacitance in Farads (F)

A is the surface area of the capacitor’s plate

ε0 is the permittivity of vacuum and its magnitude is 8.854 × 10−12 (F/m)

d is defined as the distance between capacitor’s plates

The absolute permittivity ε is given by the following product in Equation (8):

Table 1 —Relative permittivity values of materials

Capacitive current (Ic) is directly related to the capacitance.

Figure 1 —Complex Impedance measurement (simplified diagram)

Figure 2 —Power factor/dissipation factor vector diagram

Power factor  Cos   PF ;

sin  sin  90    sin(90)  cos( )  cos(90)  sin( ) cos 

  is the angle between IR and IT, Figure 2

Replacing PF and DF within the expression above, yields the following:

4.3 Differences between DFR and single-frequency PF testing techniques

Table 2 —Differences between typical power factor and DFR tests

Parameter Power factor test DFR test

4.4 DFR moisture application

The variable and fixed input parameters are the following:

c) Fixed or variable within limits:

4.5 General DFR measurement test practice

Figure 4 —DFR test set connected to a two-winding transformer

5. Making a DFR measurement

5.2 Test environment preparation

5.3 Test object preparation

5.4 Test set

5.5 Measurement configurations

5.5.1 UST measurement

Figure 5 —Measurement of CHL using UST configuration

5.5.2 GST measurement with guard

Figure 6 —Measurement of CH using GST with guard configuration

5.6.1 Record the temperature

5.6.2 Environmental condition

5.6.3 Record the transformer nameplate information