Promising Technologies for the

Inspection of Polymer Insulators:

Technology Review and Evaluation

Technical Report
Promising Technologies for the
Inspection of Polymer Insulators:
Technology Review and Evaluation

1002025

Final Report, October 2003

EPRI Project Manager

A. Phillips

EPRI • 3412 Hillview Avenue, Palo Alto, California 94304 • PO Box 10412, Palo Alto, California 94303 • USA
800.313.3774 • 650.855.2121 • [email protected] • www.epri.com
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES
THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN
ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH
INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE
ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM:

(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I)

WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR
SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS
FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR
INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL
PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S
CIRCUMSTANCE; OR

(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER

(INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE
HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR
SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD,
PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.

ORGANIZATION(S) THAT PREPARED THIS DOCUMENT

EDM International, Inc.

Southwest Research Institute, Inc.

ORDERING INFORMATION
Requests for copies of this report should be directed to EPRI Orders and Conferences, 1355 Willow
Way, Suite 278, Concord, CA 94520, (800) 313-3774, press 2 or internally x5379, (925) 609-9169,
(925) 609-1310 (fax).

Electric Power Research Institute and EPRI are registered service marks of the Electric Power
Research Institute, Inc. EPRI. ELECTRIFY THE WORLD is a service mark of the Electric Power
Research Institute, Inc.

Copyright © 2003 Electric Power Research Institute, Inc. All rights reserved.
CITATIONS

This report was prepared by

EDM International, Inc.

4001 Automation Way
Fort Collins, CO 80525-3479

Principal Investigator
A. Stewart

Southwest Research Institute

6220 Culebra Road
San Antonio, TX 78228-0510

Principal Investigator
E. Franke

This report describes research sponsored by EPRI.

The report is a corporate document that should be cited in the literature in the following manner:

Promising Technologies for the Inspection of Polymer Insulators: Technology Review and
Evaluation, EPRI, Palo Alto, CA. 2003. 1002025.

iii
PRODUCT DESCRIPTION

This report documents an investigation into feasible inspection technologies to identify high-risk
polymer insulators—also called non-ceramic insulators (NCI) or composite insulators—in-
service. The concepts and technologies identified and evaluated form a platform from which
future EPRI research initiatives will be launched.

Results & Findings

The study identified a range of technologies. These technologies include imaging methods and
processing, mechanical response, chemical species, and products; hyperspectral imaging and
spectroradiometry; and, millimeter wave sensor applications. The results showed that a number
of the identified techniques are feasible and warrant further investigation. In fact, both the EPRI
base-funded enhanced NCI project and supplemental opportunities have been developed to
investigate and develop the concepts further.

Challenges & Objectives

Polymer insulators, composed of fiberglass rods covered with polymer sheds, have been in use
for nearly 40 years. Polymer insulators offer many advantages over ceramics, but field
experience has shown that they also have unique failure modes. Accumulated experience also
has shown that damaged insulators cannot always be detected by existing methods and that an
NCI can fail with no advance indication. Insulator failure without any prior indication can pose a
hazard to public and utility personnel. In addition, premature failures represent a significant
threat to circuit and system reliability.

Applications, Values & Use

As the population of installed polymer insulators ages, and as the use becomes more widespread,
the need for an efficient and effective inspection method grows. If no inspection technologies are
developed to assess the condition of polymer insulators, the future of these lightweight and cost-
effective insulators may be limited to only applications where their traditional ancestors, ceramic
insulators, have significant disadvantages (for example, high contamination or vandalism areas.)
To make polymer insulators a truly feasible option, an effective and efficient inspection
technique must be developed that can be performed from a remote location and have a high
success rate.

v
EPRI Perspective
EPRI has taken a unique position in identifying and evaluating a range of technologies that have
not been traditionally applied in the power industry. Although this is a comparatively high-risk
approach, it has the potential to solve a large issue that has been eluding the industry for years.
Based on this study’s results, EPRI has launched the following four initiatives:
1. A supplemental project to develop a tool for identifying high-risk polymer insulators by
evaluating the dynamic mechanical response.
2. A supplemental project to develop low-cost modifications in the design of polymer
insulators to make them self-diagnosing (units having a high inspectability).
3. A base-funded task to investigate imaging techniques for identifying high-risk NCI in-
service.
4. A base-funded task that is continually identifying the state-of-the-art in inspection and
assessment techniques.
If you have any interest in any of these initiatives, contact Dr. Andrew Phillips at
[email protected] or 704.717.6438.

Approach
A team of technology experts met to identify novel, unexplored concepts for detecting a
damaged or defective NCI well in advance of failure. The experts conducted a brainstorming
session to identify concepts, evaluate the concepts’ potential effectiveness, and assess the
likelihood of successfully developing the more promising concepts into viable inspection
technologies.

Keywords
Polymer insulator
Non-ceramic insulator
NCI
Composite insulator
Polymer insulator
Inspection
Assessment

vi
CONTENTS

1 PROJECT BACKGROUND....................................................................................................1-1
Background and Summary....................................................................................................1-1

2 BRAINSTORMING SESSION ................................................................................................2-1

3 SUMMARIES OF CONCEPTS FOR INSPECTING NCI ........................................................3-1

Concept 1: Imaging Methods ...............................................................................................3-2
Technology Description ....................................................................................................3-2
Potential Benefit ...............................................................................................................3-3
Implementation Unknowns ...............................................................................................3-4
Brief Technology Evaluation and Development Plan ......................................................3-5
Phase 1. Feasibility Investigation ................................................................................3-5
Phase 2. Prototype Development ................................................................................3-5
Concept 2: Image Processing of Insulator End Fitting and Rod Alignment..........................3-6
Technology Description ...................................................................................................3-7
Potential Benefit ...............................................................................................................3-7
Implementation Unknowns ...............................................................................................3-8
Brief Technology Evaluation and Development Plan .......................................................3-8
Phase 1. Feasibility Investigation ................................................................................3-8
Phase 2. Prototype Development ................................................................................3-8
Initial Testing ....................................................................................................................3-9
Concept 3: Mechanical Testing ..........................................................................................3-13
Technology Description ..................................................................................................3-13
Potential Benefit .............................................................................................................3-14
Implementation Unknowns .............................................................................................3-15
Brief Technology Evaluation and Development Plan .....................................................3-15
Phase 1. Feasibility Investigation ..............................................................................3-15
Phase 2. Prototype Development ..............................................................................3-16

vii
 Initial Testing ..................................................................................................................3-16
Concept 4: Chemical Species and Products ......................................................................3-20
Technology Description ..................................................................................................3-20
Potential Benefit .............................................................................................................3-21
Implementation Unknowns .............................................................................................3-21
Brief Technology Evaluation and Development Plan .....................................................3-21
Phase 1. Feasibility Investigation ..............................................................................3-21
Phase 2. Prototype Development ..............................................................................3-22
References - Image Processing of Insulator End fitting and Rod Alignment ..................3-22
Concept 5: Hyperspectral Imaging and Spectroradiometry Methods.................................3-23
Technology Description ..................................................................................................3-23
Potential Benefit .............................................................................................................3-24
Implementation Unknowns .............................................................................................3-24
Brief Technology Evaluation and Development Plan .....................................................3-24
Phase 1. Feasibility Investigation ..............................................................................3-24
Phase 2. Prototype Development ..............................................................................3-25
References – Hyperspectral Imaging and Spectroradiometry Methods .........................3-25
Concept 6: Millimeter Wave Sensor Applications...............................................................3-26
Technology Description ..................................................................................................3-26
Potential Benefit .............................................................................................................3-27
Risks and Implementation Unknowns ............................................................................3-27
Brief Technology Evaluation and Development Plan .....................................................3-28
Phase 1. Feasibility Investigation ..............................................................................3-28
Phase 2. Prototype Development ..............................................................................3-28
References - Millimeter Wave Sensor Applications........................................................3-29
Related Concept: Active Microwave, Radiofrequency, or Maser Pumping
Combined with Thermal Imaging....................................................................................3-29
Brief Description ........................................................................................................3-29
Potential Benefit.........................................................................................................3-30
Implementation Unknowns ........................................................................................3-30
Brief Technology Evaluation and Development Plan.................................................3-30

viii
LIST OF FIGURES

Figure 3-1 Spectrophotometer Measurements of Spectral Absorption......................................3-3

Figure 3-2 Rod Section with Fracture Under Load and No Load...............................................3-6
Figure 3-3 Angular Displacement of a Rod when a Fracture Occurs and the Unit is
Under Unrestrained Loading of a Sufficient Magnitude......................................................3-7
Figure 3-4 Insulator Under No Load and With No Fracture. ......................................................3-9
Figure 3-5 Photograph of Entire Insulator with 2500 lbs Applied and a Fracture Depth
of 10 mm (Note: The Information on the Yellow Board is Incorrect–the Load was
2500 lbs) ..........................................................................................................................3-10
Figure 3-6 Close-up Photograph of 4 mm Cut with the Rod under 2500 lbs Load ..................3-11
Figure 3-7 Increase in Gap Size as a Function of Cut Depth (Note: Increase in Gap Size
is Between Insulator Unloaded and Loaded with 2500 lbs) .............................................3-12
Figure 3-8 Schematic of Mechanical Testing Concept ............................................................3-13
Figure 3-9 Impulse Testing of Gas Turbines Blades to Detect Shifts in Natural Frequency
Due to Progressive Airfoil Cracking .................................................................................3-14
Figure 3-10 Impulse Testing Equipment: 1) Dual Channel Spectrum Analyzer
(HP3560A) used to Record Transfer Function (vibration/force) Spectrums; 2)
Instrumented Impulse Hammer Delivers Measured Impulses to Bottom Attachment;
3) Accelerometer Attached to Bottom Attachment Piece Measures Vibration
Response; 4) Microphone Located Between Sheds Measures Sound Signal
Related to Vibration..........................................................................................................3-17
Figure 3-11 Impulse Response Spectra of 6-ft Insulator using Accelerometers and
Microphone Pickups.........................................................................................................3-18
Figure 3-12 The Mid-wave Infrared IMSS has a Spectral Resolution of 0.01 Microns and
Over 200 Spectral Bins Between 3 to 5 Microns (from Pacific Advanced
Technology) .....................................................................................................................3-24

ix
LIST OF TABLES

Table 3-1 Details of insulator under test ....................................................................................3-9

Table 3-2 Gap Size Both With and Without Load Applied as a Function of Cut Depth
(Note: The No Load Values were Calculated as an Average of the Gap Size Both
Before and After the Load was Applied) ..........................................................................3-11
Table 3-3 Potential Applications of MMW for NCI Defect Detection ........................................3-27

xi
1
PROJECT BACKGROUND

Background and Summary

Non-ceramic insulators (NCI, alternatively called polymer or composite insulators – terms will
be under interchangeably in this report), composed of fiberglass rods covered with polymer
sheds, have been in use for nearly 40 years. Polymer insulators offer many advantages over
ceramics but field experience has shown that they also have unique failure modes. Accumulated
experience has also shown that damaged insulators cannot always be detected by existing
methods and a NCI can fail with no advance indication. Insulator failure without any prior
indication can pose a hazard to the public and utility personnel. In addition, premature failures
represent a significant threat to circuit and system reliability.

EPRI has recognized that the electric utility industry needs an inspection method for detecting a
damaged or defective NCI well in advance of failure. As such, they contracted Southwest
Research Institute (SwRI) and EDM International, Inc. (EDM) to work together with EPRI as
part of a Project Team to identify novel, unexplored concepts for defect detection and to detect a
damaged NCI before failure. Visual inspection, widely used for ceramic insulators is not
effective for a NCI since the polymer sheath can cover severe degradation of the rod. The goal
of this project was to identify several concepts with promise for detecting this degradation.
Therefore, a team of technology experts conducted a brainstorming session to identify concepts,
evaluated the potential effectiveness of the concepts, and assessed the likelihood of successfully
developing the more promising concepts into viable inspection technologies. SwRI and EDM
worked cooperatively on this project. SwRI’s contribution was primarily one of scientific and
technical expertise. EDM supplemented SwRI’s scientific and technical expertise and provided
knowledge of electric utility inspection procedures.

SwRI had responsibility for organizing and conducting the brainstorming meeting, which
included staff from EPRI, SwRI and EDM. A summary of the concepts that were selected
during a Project Team brainstorming session is included herein.

1-1
2
BRAINSTORMING SESSION

A creative brainstorming session was held at SwRI offices on July 9, 2002. A total of 15
representatives from EPRI, SwRI and EDM participated in the brainstorming session that was
organized to identify novel, unexplored concepts for defect detection in NCI. Numerous potential
concepts were identified during the brainstorming session; the concepts were than prioritized
based on the consensus of the experts and the top six concepts were identified for further
investigation. Each of the six most promising technologies was assigned to the appropriate
“technology experts” who were then asked to conduct a more in-depth analysis to estimate the
potential effectiveness of the methods, to identify concepts with fundamental flaws that should
not be pursued and to assess the likelihood of successfully developing the more promising
concepts into viable inspection technologies. In addition, the technology experts were asked to
develop summaries of the concepts including a description of the technology, potential benefits,
implementation unknowns, a program plan that includes feasibility investigation and prototype
development. Estimates of costs associated with these various concepts have not been included
in this report. Preliminary feasibility tests of two of the concepts were conducted shortly after the
brainstorming session and the results are summarized in this report. The six concepts identified
as warranting further examination were:
• Imaging Methods
• Image Processing of Insulator End Fitting and Rod Alignment (Note: This concept is a subset
of the broader “Imaging Methods.” Preliminary testing of this concept was performed shortly
after the brainstorming session. The results of this testing are summarized in this report.)
• Mechanical Testing (Note: Preliminary testing of this concept was performed shortly after
the brainstorming session. The results of this testing are summarized in this report.)
• Chemical Species and Products
• Hyperspectral Imaging and Spectroradiometry
• Millimeter Wave Sensor Applications

The summaries prepared for each of the concepts are provided in Section 3 of this report.

2-1
3
SUMMARIES OF CONCEPTS FOR INSPECTING NCI

Summaries of each of the top six most promising concepts identified during the brainstorming
session are provided below. These summaries include an overview of the inspection technology
concept and a brief outline of the major tasks that may be required to evaluate the feasibility of
the concept and develop a prototype inspection device. (Note: The concept summaries were
prepared by different authors. Therefore, the formats and contents may vary. Every effort was
made to merge styles into a consistent format whenever possible.)

3-1
Summaries of Concepts for Inspecting NCI

Concept 1: Imaging Methods

This approach is based on the use of a sensor to produce an image of the insulator and computer
image processing to enhance the image for viewing or to automatically detect a failing NCI.
Different sensors could be considered to obtain images in the visible, infrared or x-ray spectra, or
a combination. Visible images could identify problems such as splits or tears in the sheath or
defective seals on end caps. X-ray images could detect cracks in the fiberglass rod or water
intrusion under the sheath or in the end cap.

Technology Description

Imaging methods include using the electromagnetic spectrum for developing images from the
infrared to visual to x-ray. Each of these methods can produce an image that may be analyzed to
detect defects/damage. Brief descriptions of different imaging technologies that hold promise
are provided below

♦ X-Ray and Neutron Imaging: The x-ray method will provide the ability to observe cracks in
the fiberglass rod underneath the end fittings. Cracks in the fiberglass rod indicate an NCI
that is at high risk of failure. X-ray imaging would utilize an isotopic source (properly
collimated to minimize exposure to the x-ray beam) in conjunction with a real-time imaging
x-ray camera. A slightly different approach would utilize a Californium 254 neutron source.
Neutrons are preferentially absorbed by hydrogen atoms and would allow early and positive
detection of moisture in the end fitting. Commercially available sources (x-ray or neutron)
and cameras could be used if they can be placed within 3 to 4 inches of the insulator during
inspection. An inspection system based on this approach would consist of a horseshoe frame
holding the source and camera so they could be positioned with the insulator between them.
This assembly would consist of a source that would be a cylinder of approximately 2 inches
in diameter and 4 inches long and a detector that would be a cylinder approximately 4 inches
in diameter and 10 inches long. The complete assembly would weigh approximately 6-8
pounds. It would be positioned at the end fittings of an insulator or scanned along the axis of
the insulator by a boom, a lineman or a robotic line crawler.

♦ Visible Spectrum, Spectral Band Filtered, and Infrared Imaging: A video camera would be
used to obtain monochrome or color images (with or without spectral band filters) of a
complete insulator or an infrared camera to obtain thermal images. The camera(s) would be
placed within 10 to 20 feet of the insulator and in a position to obtain a clear view of critical
areas. Visible Spectrum Imaging -Image processing software would analyze the insulator
shape to identify insulators where partial weakening or fracture of the rod causes the end
caps of the insulator to be out of alignment. (Note: The concept of using imaging of NCI
alignment to detect damage is further explored in the following section, “Concept 2 - Image
Processing of Insulator End fitting and Rod Alignment”). Image feature analysis could
identify cracks and splits in the sheath, damaged sheds and broken seals on the end caps.
Spectral Band Filtered Imaging -The ability to detect damage could potentially be enhanced
by using filters. Spectrophotometer measurements show that the sheath material absorbs
more at the green/blue spectral bands than in the red and near-infrared regions (see Figure

3-2
 Summaries of Concepts for Inspecting NCI

3-1). On the other hand, the fiberglass rod absorbs more in the red/infrared region. This can
be exploited by using optical filters to capture images in several different spectral bands.
Image processing techniques would then color map the images to enhance any areas where
the fiberglass rod is exposed. Infrared Imaging - Infrared imaging has been used to identify
problems in substations and power plants but has been less successful for transmission line
inspection. One of the reasons is that inspections have generally been done from a relatively
long distance from the insulators. If an infrared camera is placed 10 to 20 feet from an
insulator, the sensitivity will be much greater and smaller temperature differences can be
detected.

Spectral Absorption

1.3
Relative Absorption

0.8

0.3

-0.2
445 531 614 695 773 849
Wavelength
Seal 01 Polymer01 Polymer02
Rod 01 Rod 02

Figure 3-1
Spectrophotometer Measurements of Spectral Absorption

♦ Image Processing and Analysis: Image processing and feature analysis would be used to
automatically detect critical indicators: broken fibers, water under the sheath or in the end
caps, cracks in the sheath or broken seals and heat caused by leakage currents. If more than
one type of sensor is included in an inspection system, data fusion techniques would be used
to confirm indications of insulator failure.

Potential Benefit

♦ X-Ray and Neutron Imaging: A The x-ray real-time imaging capability would allow
detection of cracking in the fiberglass rod under the sheath and even while looking through
the end fitting steel components. If this were sensitive enough to locate a few broken fibers,
this would be a reliable method for early detection of brittle fracture, which is responsible for
over 50% of the failures in the EPRI NCI failure database. Neutron imaging would detect
moisture under the sheath and in the end fitting, which is the cause of over 75% of the
failures investigated (brittle fracture, electrical discharge and flashunder). This would
provide a method for detecting high-risk insulators even before any mechanical changes
occur.

3-3
Summaries of Concepts for Inspecting NCI

♦ Visible Spectrum, Spectral Band Filtered, and Infrared Imaging: At long distance or in
field conditions it is difficult to see where the sheath is damaged, split or unclipped to expose
the rod. Multi-spectral imaging could be used to produce images where any exposed rod is
marked with a distinctive color.

♦ Image Processing and Analysis: Image analysis of the condition of the seals, sheds and
sheath and the alignment or shape of insulators could be used to detect cases where there is
danger of water ingress or where some mechanical changes have occurred. If the change is
detectable, this inspection could be performed from a helicopter or from a safe region on a
tower.

Implementation Unknowns

♦ X-Ray and Neutron Imaging: To produce an adequate image of the fitting region
using x-rays, it is necessary to have an understanding of the material dimensions,
especially the thickness of the end fitting metal areas. The attenuation of x-rays is
greatly affected by the metal thickness and to develop the optimum defect detection
technique for the fiberglass rod, it is necessary to optimize the x-ray energy. It is
believed that there are a minimal number of designs with different metal thicknesses
in the end-fitting region. If neutron radiography is evaluated, then the issue of most
concern is the neutron absorption of fiberglass versus the attenuation of moisture.
The attenuation of neutrons by the metal is negligible.

♦ Visible Spectrum, Spectral Band Filtered, and Infrared Imaging: Dust, corrosion
or aging effects may mask visual indications of splits in the sheath, damaged seals or
thermal patterns of leakage current. The size and contrast of these features must be
determined in order to specify camera resolution, camera mount vibration limits,
requirement for illumination and other factors that will affect the limits of
damage/defect detectability.

Based on similar, previous investigations, there is a good probability (greater than 75%) that the
methods under discussion will successfully detect:
• Broken fibers, if a significant percentage of the bundle is broken
• Moisture under the sheath, but the detectable amount of moisture is unknown at this time
• Splits, tears, and other types of physical damage in the sheath larger than .05” width and .25”
length
• Thermal patterns caused by temperature differences greater than .1° C

Based on information from the EPRI Guide to Visual Inspection of Non-Ceramic Insulators,
these methods would have detected a significant number of the recorded insulator failures. The
actual sensitivity/detectability of the various defects/conditions that will be practically achievable
is an unknown that will need to be assessed through a feasibility study.

3-4
 Summaries of Concepts for Inspecting NCI

Brief Technology Evaluation and Development Plan

The following tasks would be required to confirm the feasibility of this concept and conduct
preliminary tests.

Phase 1. Feasibility Investigation

♦ X-Ray and Neutron Imaging: The initial investigation to assess the feasibility of using x-
ray would include obtaining values of the end fitting thickness from the various
manufacturers, developing preliminary x-ray procedures (including evaluation of x-ray
energy and real-time imaging parameters) and verifying that cracking in the fiberglass rod
can be detected using the procedure. If the results are positive, additional investigations will
determine the minimum defect that can be detected, standoff distance, and other quantitative
parameters.

A similar approach would be used for a neutron source, including development of

preliminary inspection procedures (including energy and moderation techniques and real-
time imaging parameters), verification that moisture can be detected in the end fitting of the
NCI and investigation of the minimum amount of moisture that can be detected. Both cases
would include an initial evaluation of how the corona/grading rings would interfere with
obtaining the x-ray or neutron image and the analysis of the image.

♦ Visible Spectrum, Spectral Band Filtered, and Infrared Imaging: Phase 1 tasks to develop
radiography procedures and verify the concept would require about six months. Feasibility
investigation of visible and infrared imaging methods would require about five months if
currently available cameras and imaging processing computers were used.

♦ Image Processing and Analysis: Aged NCI insulators with defects in the sheath and seals
would be used to determine how surface conditions affects the detectability of defects.
Image processing methods such as edge enhancement and curve fitting would be tested to
determine how well this can enhance the image features of cracked fibers, moisture, and/or
other visually detectable surface defects/damage. Pattern recognition techniques would be
developed and tested to provide automatic detection of defects above a specified threshold.

Phase 2. Prototype Development

♦ X-Ray and Neutron Imaging: If Phase 1 results are positive, a prototype inspection device
will be developed. The prototype x-ray sensor would consist of a source and a real-time
imager. The scenario is to have the source emit x-rays (or neutrons) that pass through the
NCI end and then into the imaging system. This prototype must be attached either to a hot
stick system or a line crawler.

♦ Visible Spectrum, Spectral Band Filtered, and Infrared Imaging: A prototype visible or
infrared inspection sensor could be included on the same deployment system or could be
developed as a separate system since it could be used at greater stand-off distances than
radiography imagers.

3-5
Summaries of Concepts for Inspecting NCI

Concept 2: Image Processing of Insulator End Fitting and Rod Alignment

This approach is a subset of the Imaging Methods described previously. It involves

using/processing high-resolution images of NCI under load to determine whether the end fittings
are aligned and hence whether the rod has been partially damaged by a “brittle fracture” type
failure.

This method assumes that if a fiberglass rod has a fracture perpendicular to the rod axis and an
unrestrained tension load is applied to the rod the distance between the fracture surfaces will
increase as shown in Figure 3-2.

Applied Load

x x+∆

No Load Applied Load

Figure 3-2
Rod Section with Fracture Under Load and No Load

If the distance between the fracture planes, x, increases when an unrestrained load in applied
then the sides of the rod will no longer be parallel.

In order to ensure good mechanical performance, manufacturers put a considerable amount of

effort into the attachment of end fittings. Rod diameters have a high tolerance; the inside of the
end fittings are machined and the compression process is well controlled. This may allow the
end fittings to be used as measurement points to determine whether the rod has a fracture (Figure
3-3).

3-6
 Summaries of Concepts for Inspecting NCI

θ α

load
load

α
θ

load load

Figure 3-3
Angular Displacement of a Rod when a Fracture Occurs and the Unit is Under
Unrestrained Loading of a Sufficient Magnitude

The angles of the edges of the end fittings are measured to some reference resulting in angles, α
and θ. If β = θ - α, a significant value for β may indicate a fractured rod.

(Note: Following the brainstorming session described in Section 2, exploratory testing of this
concept was performed to assess whether this approach is feasible. A summary of this initial
testing is presented in the Section, “Initial Testing”).

Technology Description

High-resolution digital images of an insulator are taken and digitally processed. Edge detection
is used to determine the values for α and θ and the difference between them, β, is calculated. A
probability (or confidence index) for the condition of the insulator is then calculated.
Alternatively, the alignment of the rod itself can be examined.

The technology will require a high-resolution digital camera, image processing equipment, etc.

Related ideas include:

• Using a laser to make the measurement
• Outlining the entire insulator to see if alignment anomalies can be detected.

Potential Benefit

The following benefit statements may be made:

• The technique utilizes known technologies that are readily available.
• The technique can be applied from either an airborne platform or ground location.

3-7
Summaries of Concepts for Inspecting NCI

• When fully implemented a device can be developed that can provide an instantaneous answer
to the inspector.
• The concept maybe implemented in the EPRI APS type of airborne inspection system where
high-resolution images of each structure are taken at high speed and the images are processed
after the inspection. This concept may be developed and implemented as an automated
processing tool for such an inspection.

Implementation Unknowns

The following unknowns exist:

• The speed of brittle fracture growth may be so fast that by the time any deformation can be
accurately determined the insulator may only have hours or days of life left.
• The sufficiency of end fitting tolerances to allow meaningful measurements.
• The impact of extraneous factors such as grading rings and line hardware on the ability to
make measurements.
• The sufficiency of rod deformation under everyday loads to enable detection of an anomaly.
• The sensitivity of the inspection technique to the viewing angle.

Brief Technology Evaluation and Development Plan

The following tasks would be required to confirm the feasibility of this concept and conduct
preliminary tests.

Phase 1. Feasibility Investigation

An initial investigation would be conducted to assess the amount by which the angle between the
two end fittings will deform and whether this is measurable for:
• Normal everyday loads
• Normal rod sizes

In addition, a determination will be made concerning whether the tolerance of end fittings is
sufficient to allow the measurement and the influence of corona rings.

Phase 2. Prototype Development

A set of insulators with different types of defects would be fabricated and images of the units
captured while the units are under load. The images may be captured using standard high-
resolution digital cameras. The images may then be processed using standard desktop computers
with image processing software. The system may be developed such that the digital camera
could be directly linked to a laptop and the image processed in this manner.

3-8
 Summaries of Concepts for Inspecting NCI

Based on the success of the prototype a development plan for a stand-alone device may be
developed.

Initial Testing

Initial testing of the concept was performed at the EPRI center in Lenox, MA. An NCI was hung
from a beam and a weight of 2500 lbs was hung from the end fitting. The rubber weathershed
system was removed in order to simplify measurements. Figure 3-4 shows the insulator setup
when the unit was intact and under no load.

Figure 3-4
Insulator Under No Load and With No Fracture.

The insulator information is shown in the table below:

Table 3-1
Details of insulator under test

Insulator Ohio Brass Hi-Lite XL Cat. # 511010-1201

Length SML 25,000 lbs

Rod diameter 15.95 mm (0.63 in.) RTL 12,500 lbs

Comments: Rubber sheds removed and aluminum end fittings installed.

A cut, of increasing depth, was made orthogonal to the rod axis close to the live end to simulate a
brittle fracture in progress. The cut was made with a hacksaw. The depths of the cut were 2, 4, 8,
and 10 mm.

3-9
Summaries of Concepts for Inspecting NCI

The following measurements/recordings were performed both with and without the load applied
for the intact rod and each cut depth:
• Photograph of entire insulator
• Close-up photograph of cut
• Feeler gauge measurement of distance between fracture surfaces. (Measurements were made
under no load conditions both prior to and after the load was applied).

Figure 3-5
Photograph of Entire Insulator with 2500 lbs Applied and a Fracture Depth of 10 mm
(Note: The Information on the Yellow Board is Incorrect–the Load was 2500 lbs)

3-10
 Summaries of Concepts for Inspecting NCI

Figure 3-6
Close-up Photograph of 4 mm Cut with the Rod under 2500 lbs Load

Table 3-2
Gap Size Both With and Without Load Applied as a Function of Cut Depth (Note: The No
Load Values were Calculated as an Average of the Gap Size Both Before and After the
Load was Applied)

Gap size (mm)

Cut Depth Delta
(mm) (mm)
No load 2500 lbs load

0.000

2.000 0.851 0.915 0.064

4.000 0.940 0.991 0.051

8.000 0.928 1.143 0.216

10.000 0.915 1.194 0.279

3-11
Summaries of Concepts for Inspecting NCI

0.300

0.250

Increase in gap size (mm)

0.200

0.150

0.100

0.050

0.000
0 2 4 6 8 10 12
Cut Depth (mm)

Figure 3-7
Increase in Gap Size as a Function of Cut Depth (Note: Increase in Gap Size is Between
Insulator Unloaded and Loaded with 2500 lbs)

Notes:
• The accuracy with which the cuts were made was poor. The 2, 4, 8 and 10 mm sizes should
be used as reference points only.
• The feeler gauge measurements were done using standard “spark plug” feeler gauges. The
accuracy was limited especially for the smaller gap sizes.
• The load applied was 2500 lbs. This is a lower than normal everyday load for a dead-end
insulator and a higher than normal load for a suspension (i.e. I- or vee-string) unit.
• These tests were a first “quick” pass at the concept. More accurate measurements need to be
made.
• The images from each test are available for analysis in JPEG format.

3-12
 Summaries of Concepts for Inspecting NCI

Concept 3: Mechanical Testing

This approach is based on the concept that degradation of the insulator fiberglass rod will cause
measurable changes in insulator stiffness and deflections under load (see Figure 3-8). In this
case, insulators with partially failed rods will exhibit changes in length, bending flexure, or
torsional flexure due to load variations such as cable oscillations resulting from wind vorticity,
cable sway, excitation caused by helicopter blade wash, or other external forces. These
variations would be detected by remote instruments such as a laser vibrometer or by contact
instruments such as accelerometers and microphones. A microprocessor would analyze the
signals to automatically determine if the rods are intact or if brittle fracture or some other
mechanism has broken some of the fibers.

Brittle Fracture
Crack

Applied Force
or Impact

Mechanical
Resonant Response

Figure 3-8
Schematic of Mechanical Testing Concept

(Note: Following the brainstorming session described in Section 2 exploratory testing of this
concept was performed to assess whether this approach is feasible. A summary of this initial
testing is presented in the Section “Initial Testing”).

Technology Description

Dynamic measurements are commonly used to detect flaws in structures; for example, tapping
on a plaster wall to find areas that are not bonded securely has been used for generations. An
elaborate device to automate this process is adapted to detect the quality of bonded tiles on the
space shuttle lifting body. Dynamic characteristics are much more readily detectable than static
characteristics of mechanical devices. Inexpensive vibration sensors are available with very high
sensitivities and vibration testing is well established. One dynamic method uses an applied
impulse and simultaneous measurement of both vibration and the applied impulse to produce a
transfer function (vibration/force) spectrum. The transfer function is relatively forgiving with

3-13
Summaries of Concepts for Inspecting NCI

respect to the impulse level and location. While this is the case, location will affect the frequency
content and only modes that are not a node at that point will appear in the spectrum.

The progressive cracking of turbine blades can be detected by a simple shift in natural frequency
(Figure 3-9). More comprehensive data analysis techniques are available to assess structural
stiffness and damping based on these measurements. Dynamic modeling techniques provide a
means to assess the contributions of specific structural elements to the measured stiffness and
damping parameters, and can predict how those parameters are affected by defects in the
structural elements.

Blade Stress Response

25,00
5X
4X

ML
20,00 ML
0 8
ML
10
ML
ML

15,00 AH
C k d
Stress, psi

10,00

5,00

0
1,00 1,10 1,20 1,30 1,40 1,50 1,60
Frequency, Hz

Figure 3-9
Impulse Testing of Gas Turbines Blades to Detect Shifts in Natural Frequency Due to
Progressive Airfoil Cracking

Potential Benefit

This concept may lead to three possible inspection instruments. The first is a contact or impact
device for quickly verifying the integrity of an insulator. This could be as simple as adding
accelerometers to hotsticks so that a lineman could tap an insulator and determine if it has
experienced brittle fracture before placing any weight on the conductor or beginning work. This
would be relatively inexpensive and could reduce the chance of an insulator breaking when line
work is conducted.

The second possibility includes systems developed with non-contact excitation and measurement
methods (acoustic, interferometer or electromagnetic) which would allow non-contact inspection
from a distance so that failing insulators could be replaced before complete failure.

3-14
 Summaries of Concepts for Inspecting NCI

Third, a line crawler system could be adapted with resonance identification instruments to
address the task of surveillance/inspection of a large number of insulators as a predictive
maintenance activity. In this case, it would probably be more effective to use a small
electrodynamic shaker to create the extension forces and a self-contained accelerometer and or
microphone to detect response signals.

In addition to the potential for assessing the condition of insulators based on stand alone
measurements this method holds potential for assessments based on comparative measurements
to detect changes in insulator condition. For example, changes in the condition of an individual
insulator could be assessed based on comparison of a measurement at given time to a baseline
measurement made earlier in an insulator’s life, or the method could be used as the basis for a
system to make go, no-go decisions for like insulators in a long section of the line, i.e., by
making measurements on several identical insulators it might be feasible to identify insulators
with different and potentially unacceptable frequency responses.

Implementation Unknowns

SwRI conducted brief, preliminary tests to confirm that adequate quality response signals can be
acquired using intrusive accelerometers attached to the insulator end cap; these tests are
described in the Section “Initial Testing.” Using this technique, it is reasonably certain that
substantial damage can be detected. It is less certain, however, that incipient damage can be
quantified at an early-stage especially with non-contact measurement and excitation techniques.
Preliminary tests indicate that it should be possible to detect significant natural frequencies using
a properly placed microphone and that changes in stiffness can be detected by measuring natural
frequency shifts. One uncertainty is the sensitivity of signal strength for different microphone
placements and shielding in the presence of conflicting noise such as wind or other vibrations.
Another factor is the ability to detect frequency shifts due to damage in the presence of normal
natural frequency variants due to manufacturing variables and insulator configurations e.g.,
length, number of sheaths and sheds, presence of corona ring etc.

Brief Technology Evaluation and Development Plan

Phase 1. Feasibility Investigation

Laboratory tests will be performed to identify natural frequencies for different configurations of
NCIs and to quantify detectable shifts in natural frequencies for progressive damage to the
insulator rod. The tests will also investigate various vibration sensor types and excitation
mechanisms that could be applicable to insulator inspection for a lineman’s structural integrity
confirmation and for predictive maintenance surveys. The sensitivity of different microphones
and accelerometers will be compared and data will be analyzed to determine the capability of
this approach for detecting failing insulators. It will take approximately four months for this
phase of study.

If initial tests demonstrate that failing insulators can be identified by mechanical measurements
additional tests will be used to extend the concept to longer distance, non-contact techniques.

3-15
Summaries of Concepts for Inspecting NCI

This will include the use of a laser vibrometer or interferometer for measurement of vibration
due to artificial or natural stimulus sources. An estimate of time for this portion of the study
would be eight months.

If the laboratory tests show that the mechanical measurement concept is feasible, additional
testing and development will be done in a transmission line environment. Procedures and
instruments will be selected, based on results from the laboratory tests. Similar tests will be
performed on insulators suspended from a simulated transmission tower. Six months will be
required for this portion of Phase 1.

Phase 2. Prototype Development

Several different prototype systems could be developed, depending on the results of the Phase 1
investigations. One might be a Lineman's Toolkit for contact testing insulators before work is
begun. Another could be a non-contact, optical method for analyzing the mechanical vibrations
of insulators that could be applied from a tower, or possibly from a helicopter or from the
ground. A third possibility could be an integrated wire-crawler system that includes a dynamic
shaker, accelerometer and data analysis components to identify failing insulators. It is
impossible to give an accurate estimate of the development without the results from the Phase 1
testing, but this development is likely to require a year or more.

Initial Testing

Because there was concern that the neoprene sheds and sheath would damp out vibration induced
into the insulator rod, SwRI conducted impulse testing of a high voltage insulator to evaluate the
data quality. The test was conducted by simply supporting the insulator with small blocks at
both ends to allow the rod and sheds to move freely. A series of tests was conducted to compare
two different types of sensors shown in Figure 3-10, an accelerometer (item 3) and a microphone
(item 4). For both measurement sensors, impulses were delivered in the lateral and in the axial
direction using an impulse hammer (item 2). The portable dual channel HP 3560A, spectrum
analyzer (item 1) was used to acquire the vibration transfer functions presented in Figure 3-11.
Lateral impulses are expected to produce more bending action, which should be more sensitive
to progressive rod damage.

3-16
 Summaries of Concepts for Inspecting NCI

Figure 3-10
Impulse Testing Equipment: 1) Dual Channel Spectrum Analyzer (HP3560A) used to
Record Transfer Function (vibration/force) Spectrums; 2) Instrumented Impulse Hammer
Delivers Measured Impulses to Bottom Attachment; 3) Accelerometer Attached to Bottom
Attachment Piece Measures Vibration Response; 4) Microphone Located Between Sheds
Measures Sound Signal Related to Vibration

3-17
Summaries of Concepts for Inspecting NCI

X:67.5 Hz Y:3.670804 Freq Resp: Hyx 2:1 104.TRC

Freq Resp: Hyx 2:1 102.TRC 1
10

Mag (Log)
Mag (Log)

Lateral Impulse at Lower Connector Axial Impulse at Lower Connector

Accelerometer Measurement Accelerometer Measurement

0.01 0.001
0 Hz 500 Hz 0 Hz 500 Hz

X:1.88125 kHz Y:1.006957 X:2.3625 kHz Y:38.25506 m

Freq Resp: Hyx 2:1 106.TRC Freq Resp: Hyx 2:1 110.TRC
10 0.1

Mag (Log) Mag (Log)

Axial Impulse at Lower Connector Lateral Impulse at Lower Connector

Accelerometer Measurement Microphone Measurement at
Lower Connector
0.01 0.001
0 Hz 5 kHz 0 Hz 5 kHz

X:4.925 kHz Y:140.7071 m X:4.925 kHz Y:69.45056 m

Freq Resp: Hyx 2:1 111.TRC Freq Resp: Hyx 2:1 112.TRC
1 Lateral Impulse at Lower Connector 0.1
Microphone Measurement at
Second Shed

Mag (Log) Mag (Log)

Lateral Impulse at Lower Connector

Microphone Measurement at
Second Shed
0.0001 0.0001
0 Hz 5 kHz 0 Hz 10 kHz

Figure 3-11
Impulse Response Spectra of 6-ft Insulator using Accelerometers and Microphone Pickups

3-18
 Summaries of Concepts for Inspecting NCI

The two transfer function spectrums in the upper left of Figure 3-11 represent the response of
lateral and axial impulses using an accelerometer to measure vibration. These plots show a
dominant natural frequency and several lower-level natural frequencies that can be clearly
identified which implies that using this technique with an accelerometer should produce high
confidence in detection of changes due to fiberglass rod element damage. The other four
spectrums in Figure 3-11 represent the response of lateral and axial impulses using a microphone
to measure the noise response to impulses in the axial and lateral direction. From this data,
damping and stiffness can be estimated and changes can be detected by developing a suitable
algorithm. Similar tests with a non-contact microphone produced irregular or noisy data.

3-19
Summaries of Concepts for Inspecting NCI

Concept 4: Chemical Species and Products

Several experimental studies have identified chemical changes within the NCI that cause brittle
fracture failure. This concept is based on determining the resulting chemical products and then
developing a means for detection. It will also be necessary to determine whether analyses can be
done remotely, without contacting the NCI.

Technology Description

NCI brittle fracture failures have been linked to infiltration of water into the fiberglass/glass
reinforced plastic (GRP) rod. The failure mechanisms have been attributable to either acid or
water leaching of the metallic ions in the glass fibers resulting in stress corrosion cracking.
[“Stress corrosion cracking (SCC) is the cracking induced from the combined influence of tensile
stress and a corrosive medium.” (Reference 1)] If the metallic ion concentration manifests itself
in the protective polymeric water shed, then a remote technique such as x-ray fluorescence
(XRF) might be able to detect the increasing metal ion concentration. Furthermore, it may be
possible for the XRF to see through the external polymer water shed and directly measure the
ionic concentration of the glass fiber.

Brittle fracture theories require the accumulation of water within the GRP rod. The water can be
aided by acids (either resident within the GRP from hydrolysis of the epoxy hardener (Reference
2) or from corona-created nitric acid) to corrode the glass fiber. The theories all require a
permeation pathway for water though the external polymer water shed.

It is well known by polymer chemists that solvents swell polymers. Furthermore, some polymer
characterization tests are based upon the degree of swelling of an unknown polymer in different
solvents; the swelling ratios can fingerprint the polymer. When polymers swell, diffusion
pathways are expanded permitting increased chemical flux (both larger molecule sizes and
greater mass flux for smaller molecules) through the polymer. There may be some
environmental chemicals that swell the outer polymer protective coating allowing more water to
diffuse into the GRP rod. It is recommended that the database of failed NCIs be analyzed for
locale-specific chemical environments that could change the water barrier properties of the
water-shed polymer. If unique environmental chemicals are found that would increase the
diffusion rate of water, then a monitoring technique could be targeted to high-risk environments.

Several external, remote techniques can be envisioned that could assess the moisture absorption
rate into the polymeric water shed or the presence of higher acid concentrations. For example,
one might spray an indicator chemical (such as a pH indicator that changes color if the acid
concentration is high) or a tracer chemical (similar to a dye penetrant to concentrate in water
diffusion pathways). Then using remote spectral measurement instruments, one could identify
whether the indicator has changed color or if the dye penetrant has an increased localized
concentration on the NCI. If either situation is found, the NCI could be removed from service or
inspected more closely.

3-20
 Summaries of Concepts for Inspecting NCI

Potential Benefit

The chemical species and products concept summary presented herein focuses on a more
fundamental aspect of NCI inspection than the other concept summaries. Results would be used
to guide investigations and development of instruments to detect the specific chemical species
that can be used to identify failing insulators. The NCI inspection program would benefit by
learning which types of detection instruments are likely to be successful and which are likely to
be unsuccessful because they are not detecting the correct products of chemical reactions. The
potential inspection instruments include non-contact, close range (several feet) as well as longer-
range detection devices.

Implementation Unknowns

The studies referenced have evaluated the chemistry within the NCI, i.e., the changes within the
GRP. For a remote inspection technique, the GRP will not be visible since it is covered by the
polymeric water shed. It is unknown whether the chemical changes within the GRP during stress
corrosion cracking conditions manifest themselves on the surface of the NCI. For example, if
water or acid migrates to the GRP and metallic ions are leached from the glass fiber, the metallic
ions may concentrate in the polymeric water shed. Burnham et al (Reference 3) identify brittle
fracture by the loss of calcium and aluminum ions in the GRP. By destructive chemical
analytical techniques [e.g., atomic absorption (AA) spectroscopy or ion chromatography/mass
spectrometer (IC/MS)] the concentration of calcium and aluminum ions in the polymeric water
shed from new and failed (samples from both the failed portion and a large distance away in an
intact segment) NCIs can be measured. Technical risk is present since these ions may not
migrate into the outer polymer shed or they may be washed away by rainwater.

Brief Technology Evaluation and Development Plan

The following tasks would be required to confirm the feasibility of this concept and conduct
preliminary tests.

Phase 1. Feasibility Investigation

Two matched sets of new and failed NCIs (matched by manufacturer and materials of
construction) would be supplied from industry. Approximately 8 small samples (about one
milligram each) will be cut from each NCI and analyzed for metal ion content, focusing on the
calcium and aluminum concentrations. Composition of the chemical species will be documented
and the viability of using metallic ion concentrations as an indicator of SCC will be assessed.
(Approximate timeframe: 2 months).

A parallel effort would include experimenting with XRF to determine its ability to measure metal
ions in both the polymer water shed and the glass fiber of the matched sets of NCIs. This effort
will include tuning the XRF to see through the polymer water shed (3 months).

3-21
Summaries of Concepts for Inspecting NCI

To study the impact of locale-specific chemical environments that could change the water barrier
properties of the water shed polymer, a stand-alone effort could be done. The database of failed
NCIs would be analyzed to extract locality of failure, and a field survey to determine the
environmental chemicals in those locations would be conducted. Unique chemicals or cocktails
would be theoretically assessed for potential swelling using solubility parameters (3 months).

In another independent effort, a series of indicator chemicals (such as a pH indicator that changes
color if the acid concentration is high) or tracer chemicals (similar to a dye penetrant to
concentrate in water diffusion pathways) would be evaluated. First, a theoretical assessment
would be made to rank order the chemicals. Secondly, the highest probable solutions would be
sprayed onto the matched sets of NCIs to evaluate the feasibility of this approach (4 months).

Phase 2. Prototype Development

At this time there is a wide range of possible prototype inspection instruments that might be
developed. These could range from technologies based on x-ray fluorescence to spray
application of an indicator. After the Phase 1 tests, a better definition of the complexity, cost and
time required for prototype development will be made.

References - Image Processing of Insulator End fitting and Rod Alignment

1. http://www.corrosion-doctors.org/Forms/scc.htm on 8/16/02.

2. “Brittle Fracture” of Composite Insulators: Why and How they Occur, by Claude de Tourreil,
Laurent Pargamin, Guy Thévenet, and Sandrine Prat; 2000 IEEE report 0-7803-6423-6/00.

3. IEEE Task Force Report: Brittle Fracture in Non-ceramic Insulators [PE-504PRD (02-2002)]
by J.T. Burnham, Chairman and Principal Author; Members in Alphabetical Order: Tony
Baker, Al Bernstorf, Claude de Tourreil, Jean-Marie George, Ravi Gorur, Ralf Hartings, Bob
Hill, Arjan Jagtiani, Terry McQuarrie, David Mitchell, Don Ruff, Herman Schneider,
Jennifer Yu, Jack Varner.

3-22
 Summaries of Concepts for Inspecting NCI

Concept 5: Hyperspectral Imaging and Spectroradiometry Methods

This approach is based on the use of a hyperspectral imager or an optical spectrometer to

"image/detect the whereabouts of various chemical species that may be present on or around an
NCI. By identifying chemical species associated with degradation, useful inspection techniques
could be developed. These species may include a variety of reaction products possibly driven by
unwanted water that has encroached into the NCI material, or possibly chemical reaction
products, which result from the unusually high corona discharge, such as ozone-induced
chemical reactions.

Technology Description

The technology is based on the concept that there may be various chemical species on or around
an NCI that indicate a problem. An imaging spectrometer would be used to indicate this.
Although the midwave (MWIR) and longwave (LWIR) infrared regions of the spectrum are very
important (since those are the regions that are usually used to identify certain chemical species),
in concept this can apply, in theory, to any region of the spectrum, including near infrared (NIR),
shortwave infrared (SWIR), and even the ultraviolet (UV). The costs for optics and focal plane
arrays will vary widely between the different wavelength regions.

A simple imaging spectrometer could be thought of as a camera with a color filter in front of it
taking a picture of something using the sun as the light source. In this simple example, not much
chemical information would be obtained; however, if the camera were to be in the infrared, and
the filter very narrow in bandwidth, then some chemical information could be obtained. To
improve the situation, a very narrow filter could be used. To continue to improve the situation,
several filters could then be used in sequence. Recently, commercial hyperspectral imagers have
been produced (Reference 4), which very rapidly process information in the visible and infrared
region of the spectrum. (There are several patented technologies, which make this happen, such
as a moving diffractive element front-end for an infrared camera.) Thus, chemical information
can be obtained by correlation of the absorption or emission regions (dark or light areas of the
image) with the corresponding wavelengths related to chemicals of interest. (Note: this is not
thermal information, as is obtained from standard broadband infrared cameras, but rather
chemical information, obtained due to the narrow bandwidths used.)

It is possible that these imagers could be operated from the ground or helicopters. (The degree of
platform stabilization will influence image quality.) Additionally, it seems possible that these
imagers could be imaging the surface contamination, as well as the surrounding vapors, which
result from the degradation processes.

Note: While there are obvious advantages associated with the flexibility, capabilities and ease of
use of a hyperspectral imager and this technology is the focus of the concept summary presented
herein, an optical spectrometer could be used as an alternative basis for an inspection system
based on the concept of detecting chemical species.

3-23
Summaries of Concepts for Inspecting NCI

Figure 3-12
The Mid-wave Infrared IMSS has a Spectral Resolution of 0.01 Microns and Over 200
Spectral Bins Between 3 to 5 Microns (from Pacific Advanced Technology)

Potential Benefit

The benefit is in obtaining images that represent chemical species, which may result from
undesirable chemical reactions, which degrade the NCI.

Implementation Unknowns

It is unknown at the present time whether the chemical species present on or around a faulty NCI
are concentrated enough to get reproducible data using the hyperspectral imaging technique. As
with any other spectral technique, concentration is of major importance. Additionally, the
characteristics of the species of interest will determine exactly what basic wavelength band
would be of use. If the contamination is on the surface, it will be easier to find, because of the
increased concentration, and hence increased contrast ratio.

Brief Technology Evaluation and Development Plan

The following tasks would be required to confirm the feasibility of this concept and conduct
preliminary tests.

Phase 1. Feasibility Investigation

Determine the chemical species of interest, based on probable chemical reactions, such as
oxidation etc. on or inside the NCI. Borrow, rent, or subcontract to Pacific Advanced
Technology (PAT) for the use of a hyperspectral imager to test the concept. (SwRI currently has
a teaming arrangement with PAT on another project, in which a similar concept is being
explored to image potential chemical/biowarfare agent liquids on surfaces for the U.S. Army.)
For the first step it would be important to have an assortment of "bad" NCIs to investigate, and a
rigorous literature search associated with various chemical species and possible reactions. For a

3-24
 Summaries of Concepts for Inspecting NCI

second step, it would be very advantageous to investigate NCIs that are operational, so that any
possible out-gassing products might be investigated. Several wavelength bands are possible,
from the visible to the longwave infrared. This concept applies to both the insulator surface
itself, as well as the possible chemical vapor around the insulator.

It is estimated that first tests in the Phase 1 Feasibility Investigation to determine if the concept is
feasible at all would last about three months.

Phase 2. Prototype Development

This approach will be highly dependent on the wavelength band (determined by the chemical
species of interest) and operational needs (i.e. ground-based or helicopter-based), therefore a
plan for development of a prototype cannot be specified at this point. Hyperspectral imagers
have been flown in helicopters, operated on missile trackers, off tripods, and even handheld.
Prototype development seems very possible for this application, if there is enough chemical
concentration for detection.

References – Hyperspectral Imaging and Spectroradiometry Methods

4. Pacific Advanced Technology: http://www.patinc.com.

3-25
Summaries of Concepts for Inspecting NCI

Concept 6: Millimeter Wave Sensor Applications

Note: Alternative approaches that are somewhat similar to the use of MMW but are based on
active microwave, radio frequency, or maser pumping are described in the Section, “Related
Concept: Active Microwave, Radiofrequency or Maser Pumping Combined with Thermal
Imaging.”

This approach is based on exploiting the unique ways that millimeter waves interact with matter.
A passive MMW detector could be used to detect moisture migration under the sheath or in the
end cap fittings. An active MMW imaging sensor would generate images of the internal
composition of NCI’s. Any resulting instrument would probably have to be placed fairly close
(within several feet) to the insulator to be inspected. Microprocessor analysis of the resulting
signal or image would provide an indication of the condition of the insulator.

Technology Description

The millimeter wave (MMW) spectrum falls between microwave and infrared frequencies, from
30 to 300 GHz. Use of MMWs holds promise for detecting specific molecular structures, such
as water, and for high resolution imaging of dielectric structures. Although it may be possible to
detect broken fibers with high resolution MMW imaging, it is more likely that detecting the
molecular precursors or byproducts of fiber failure will identify failing insulators. In contrast
with infrared or optical techniques, MMWs penetrate deeper into non-metallic structures, up to
10 cm, to provide better internal viewing. The electromagnetic signals will not pass through the
metallic end-cap fittings; however, MMW could be directed along the rod to detect moisture or
other conditions inside the end-cap.

A standoff sensor can be either passive or active and detectors can take several forms: spot beam
(for mechanical scanning), an imaging array of antennas or a focal plane detector. Passive
detectors sense variations in black body thermal radiation from objects, while active detectors
respond to a transmitted signal, either imaging a reflection from (i.e., radar) or measuring
transmission through an object. Low-power MMW radars are used for security applications such
as sensing people through walls, or creating 3D holographic images of an object hidden beneath
clothing. Passive radiometers are used in many roles such as measurement of air turbulence,
cloud profiling, motion detection or reading the amount of moisture in crops. Multi-channel
radiometers can detect molecular resonance profiles and identify the presence of specific
molecules such as water or pipeline gases (methyl chloride, hydrogen sulfide, etc.).

In addition to standoff sensors, a contact sensor may also be used to measure the relative
transmissibility of MMW energy in a dielectric structure, which can indicate internal problems
with the material.

3-26
 Summaries of Concepts for Inspecting NCI

Potential Benefit

Table 3-3 addresses promising applications of MMW technology to the problem of insulator
failures caused by seal penetration and moisture ingress. In all cases, the sensor is used to see
beneath the rubber sheath of an insulator and inspect the fiberglass rod beneath. With the
exception of the down-rod signal measurement, all approaches are non-contact.
Table 3-3
Potential Applications of MMW for NCI Defect Detection

Technique Pre-Failure Detection Indication

Down-rod signal measurement Flash-under, Brittle Fracture Loss of signal through structure
(contact method) indicating changes in material

Passive Spot Radiometer Flash-under (This is a non- Water resonance line or Raster-
(Multi-channel) imaging method and the scan image
resolution is too low to “see”
cracked fibers.)

Passive Array Imager Flash-under, Brittle Fracture Interpretable Image

Active Array Imager Flash-under, Brittle Fracture Interpretable Image

Active Chirp Radar Flash-under, Brittle Fracture Interpretable Image

(3D holography)

Risks and Implementation Unknowns

Based on tests and investigations of similar materials, a contact sensor, measuring transmission
and reflection in fiberglass, has a high probability (approximately 75%) of successfully detecting
fiber problems. Non-contact methods using either passive or active millimeter wave imaging are
estimated to have lower probability (approximately 50%) of success, largely because of
uncertainty of the MMW penetration through the metal fixture. Detection of trapped water is
considered a low probability, primarily because the resonance of water is not very strong in this
band and because there is little data on the characteristics of trapped water.

Several questions need to be answered before implementing one or more MMW approaches:

1. Which method provides the best detection for the various failure modes (direct
transmissibility measurement, passive radiometer, active radar)?

2. Is imaging of the structure required, or can a spot-beam radiometer sense water or corrosive
chemicals associated with coronal interaction?

3. If an imaging system is used, which is the preferred method (phased array, synthetic aperture
or mechanical raster scan)?

3-27
Summaries of Concepts for Inspecting NCI

4. For standoff methods, electromagnetic energy will not penetrate the metal end cap of the
insulator. Can the technique chosen “see” enough of the metal-fiberglass junction region to
detect brittle fractures in progress?

5. If water has penetrated the fiberglass structure, is it of sufficient amount to be easily

detected? When it is bound within the rod, will it still have a characteristically distinctive
resonance?

Brief Technology Evaluation and Development Plan

The following tasks would be required to confirm the feasibility of this concept and conduct
preliminary tests.

Phase 1. Feasibility Investigation

The first phase of investigation should strive to answer the questions above and narrow the
options to one or two best approaches. Both laboratory tests and paper analyses will be
conducted (5 months).

1. Devise a method of coupling MMW energy into structure for down-rod contact
measurement. Characterize rods to compare effects of different faults.

2. Measure swept frequency (multi-channel) resonance spectra of “typical” amounts of water at

23 GHz

3. Purchase and evaluate a COTS MMW handheld imaging scanner

4. Tradeoff imaging methods with respect to standoff distance, platform stability, processing
time and image resolution.

Phase 2. Prototype Development

A prototype system will be developed to prove the best concept selected in Phase 1 (this will
take approximately 10 months).

1. Design and develop a MMW sensing system.

2. Build and test key components.

3. Integrate and test sensor system.

4. Evaluate sensor against sample targets.

3-28
 Summaries of Concepts for Inspecting NCI

References - Millimeter Wave Sensor Applications

Passive MMW Sensors

5. http://www.trexenterprises.com/environ.html

6. http://millivision.com:8071/hscanner.html

Active MMW (Radar) Sensors

7. http://www-ee.eng.hawaii.edu/~delisio/delisio/RESEARCH/ACTIMG.HTM

8. http://www.gtri.gatech.edu/res-news/SPFlash.htm

MMW Holography

9. http://www.pnl.gov/nsd/commercial/body/description.html

MMW Spectroscopy

10. http://www.netl.doe.gov/scng/publications/t&d/tsa/microwave_radar.pdf

11. http://spec.jpl.nasa.gov/

12. http://physics.nist.gov/PhysRefData/MolSpec/index.html

Related Concept: Active Microwave, Radiofrequency, or Maser Pumping

Combined with Thermal Imaging

Brief Description

Additional concepts that are somewhat related to those described earlier are elaborated on here.
They are separated since they are hybrid concepts, not easily integrated into the preceding
information on MMW, and may be applicable for greater distances away from the insulator than
the MMW concept. (1) Active microwave or radiofrequency pumping (from a distance) of the
contained water molecules would be used to heat up trapped water or chemical compounds
within the NCI, which would then be imaged with a thermal imager. This concept could also be
expanded to excitation of the water with a maser, tuned to absorption lines of water or chemical
compounds of interest. (2) Active microwave, radiofrequency, or maser pumping into the power
line structure (upstream or downstream of the NCI) and then imaged with a thermal imager as
mentioned above. (3) These techniques could also be expanded to looking for the re-radiated
energy from the water molecules, which could be in the RF or MMW region. Note: It may be
preferable to avoid tuning the excitation source exactly to the water resonances, since they may
be shifted due to interactions with the polymer, etc. Additionally, as in a microwave oven, one

3-29
Summaries of Concepts for Inspecting NCI

does not necessarily want all the absorption to occur near the surface, so some detuning may be
highly desirable.

Potential Benefit

Concepts (1) and (2) described in the Brief Description above could be "add-ons" to existing
thermal imaging techniques. All three concepts could be looked on as "resonance" phenomena,
in which case good differentiation from the background may be available.

Implementation Unknowns

The power levels for the pumping are unknown at this time, although some rough estimates
could be made based on the known background heating of the NCIs from the sun. It is not
known if enough water is present inside the NCI to make this concept desirable. Further, even if
the heating occurs as desired, the thermal diffusion of the heat back out may be so blurred that a
clear image cannot be obtained.

Brief Technology Evaluation and Development Plan

Laboratory experiments could be conducted with NCIs that were purposely spiked with water.
Basic laboratory experiments of this type could be conducted with existing laboratory
equipment.

3-30
Program: About EPRI
Overhead Transmission Systems EPRI creates science and technology solutions for
the global energy and energy services industry. U.S.
electric utilities established the Electric Power
Research Institute in 1973 as a nonprofit research
consortium for the benefit of utility members, their
customers, and society. Now known simply as EPRI,
the company provides a wide range of innovative
products and services to more than 1000 energy-
related organizations in 40 countries. EPRI’s
multidisciplinary team of scientists and engineers
draws on a worldwide network of technical and
business expertise to help solve today’s toughest
energy and environmental problems.
EPRI. Electrify the World

© 2003 Electric Power Research Institute (EPRI), Inc. All rights

reserved. Electric Power Research Institute and EPRI are registered
service marks of the Electric Power Research Institute, Inc.
EPRI. ELECTRIFY THE WORLD is a service mark of the Electric
Power Research Institute, Inc.

Printed on recycled paper in the United States of America

1002025

EPRI • 3412 Hillview Avenue, Palo Alto, California 94304 • PO Box 10412, Palo Alto, California 94303 • USA
800.313.3774 • 650.855.2121 • [email protected] • www.epri.com