Detection of Internal Arcing Faults in Distribution Transformers

Paul Henault, EE
IFD Corporation
Prepared for Presentation at the ESMO 2011 Technical Session
Providence, RI, May 16-19, 2011

Abstract
Utilities are continually looking for ways to increase safety, reliability, and line crew productivity in the
distribution system. One area that is now capable of meeting these goals is the rapid detection of internal
faults in pole top and pad mounted distribution transformers.
This paper will cover a technical overview of internal faults in distribution transformers, and current
industry practices used to troubleshoot and identify faulted transformers. It will also describe recent
advances in internal fault detection technology for distribution transformers, as well as the bottom line
economic impact of using rapid fault detection to provide a smarter distribution system in terms of safety,
reliability, and productivity.

Keywords
Distribution transformer, transformer faults, internal arcing, sudden pressure, transient pressure rise,
transformer protection, reenergization, fault current, tank rupture, pressure relief valve, safety, linemen,
troubleshooting, SAIDI.

1. Introduction
A great deal of effort has been invested over many years to ensure the safety of linemen conducting manual
operations of distribution transformers installed on power supply systems. However, sporadic accidents are
a perpetual reminder of the shortcomings of current methods. In addition to the need for improved safety
for line crews, there is also an industry need for faster and less expensive methods to determine the
presence of faulted transformers, in order to improve system reliability and operational efficiency.
Unless the consequences of an internal fault are clearly visible on the exterior of the transformer, it may be
very difficult to reliably detect and confirm its presence without cumbersome and time consuming
procedures. And yet, the positive detection of the presence of an internal fault is of crucial importance since
during the re-energizing of a faulted unit, the hazards escalate significantly as the fault evolves.
This paper is focused on current methods of detection of internal faults in pole top distribution
transformers, as well as an improved technology that is now available to utilities.

978-1-4577-0567-0/11/$26.00 2011 IEEE

Page 1

2. An overview of distribution transformer faults

a) Causes of Distribution Transformer Faults
Lightning is one of the most common causes of transformer faults. Other causes include equipment
damage, repeated transformer overloading, internal defects, and other causes.
Industry wide,
approximately 15-20% of transformer fuse operations are due to faulted transformers. The cause of the
remaining 80-85% of transformer fuse operations include temporary faults due to animals, wind, and tree
branches, and other system components, such as cutouts, fuses, connectors, and low voltage wiring.
b) Characteristics of Distribution Transformer Faults
Insulation failures in distribution transformers result in arcing faults between turns and layers in the
transformer windings. The short circuit current causes intensive heating of the affected area and further
escalation of the dielectric failure. At the instant of arc ignition, a gas bubble is formed around the arc. As
the gas continues to be heated by the energy dissipated in arc, the oil vapor inside the bubble begins to
decompose and, at the same time, more vapor is generated at the interface between the gas bubble and the
surrounding oil. The volume of the gas bubble expands while the pressure inside it continues to increase as
long as the rate at which the gases are generated exceeds the rate at which the bubble is expanding. The
expanding bubble volume causes displacement of the oil and its level in the transformer tank rises. This
process is illustrated in Fig. 1 by the sequence of pictures taken with the high-speed camera during staged
internal arcing faults in a specially prepared transformer with a transparent tank.

Fig. 1: Doming of Oil Caused by an Arcing Fault

The test results show that the pattern of the resulting transient pressure rise in the air space invariably
follows closely that of the fault current lagging behind it by a few milliseconds required to overcome the
inertia of the oil above the arc. This is illustrated in the oscillograms recorded during staged internal arcing
faults and shown in Fig.2.

Page 2

Arc Voltage

Fault Current
(500 Amps)

Pressure Rise

Fig.2: Arc Voltage, Fault Current and Transient Pressure Rise

The peak transient pressure in the air space depends on many factors including the location of the fault, the
magnitude of the fault current, the volume of the air space, fault duration etc. Even in tests repeated under
nearly identical conditions the peak pressure varied randomly over a wide range
However, regardless what peak pressure was reached, the waveform of the pressure surge displayed fairly
constant rise time to the peak between 5 and 15 milliseconds for the symmetrical and asymmetrical fault
current in the range from a few hundreds to 8000A. Since all other fluctuations of pressure in the air space
of a transformer tank occur at 1000 or more times slower rate, the rate of rise of the pressure surge during
internal arcing faults can be used to reliably identify occurrence of an internal arcing fault.
c)

Operational Issues/Why Internal Fault Detection is important

The process of re-energizing of pole top transformers containing an undetected internal fault can result in
explosive failures. Such an event is illustrated in Fig.3.

Fig.3: Energizing a Faulted Pole Top Transformer

As demonstrated above, there is an industry need to provide linemen with a quick and reliable method to
detect internal faults before they are called on to re-energize the transformer in order to prevent this

Page 3

situation from occurring, as well as to save troubleshooting time and reduce the duration of outages. This
situation is compounded by a variety of operational realities, such as storms, pressures on line crews to
restore power quickly, and inexperienced linemen. There is a risk anytime lineman recloses on a
transformer without knowing if the transformer has faulted or not.

3. Current operational practices to address faulted transformers

When a transformer fuse operates, it is usually not immediately obvious what the cause of the blown fuse
is. It may be a secondary fault, or a temporary fault caused by an animal, tree branch or wind, or it may be
the transformer itself. In order to determine the source of the fault, utilities have traditionally taken one of
three approaches:
a) Automatic Replacement
One approach taken by some utilities, if they cannot immediately identify the cause of the fault, is to
assume that the cause is the transformer, and automatically replace it. However, industry statistics show
that in the event of a transformer fuse operating, the transformer is the cause only 15-20% of the time.
Therefore, automatic replacement is very wasteful and results in many good transformers being removed
from service unnecessarily. This can cost the utility between $3000 and $4000 to remove a transformer
from service, transport it back to the shop, test it, and return it to stock.
b) Field Testing
Another method that some utilities use to determine if a transformer is faulted is to use portable testers to
test the condition of the transformer. This requires climbing the pole or setting a bucket truck,
disconnecting the secondarys of the transformers and applying a test voltage to transformer.
c)

Trial and Error

This is the most common approach that utilities use now to determine if the transformer is faulted. In this
case, the lineman will visually inspect the transformer for any obvious external signs of fault damage, and
if none is found, he will re-fuse the cutout and attempt to re-energize the transformer. There are several
potential issues surrounding this approach:
Due to recently improved enclosure integrity standards, todays distribution transformers usually do
not provide any visual indication that they are faulted, such as deformation to the tank, discoloration of
the paint, oil leaks, etc. This may give the lineman a false sense of security that the transformer is
unfaulted when it is actually faulted.
If the transformer is faulted, re-energizing can be potentially hazardous to the lineman, the
environment, and the public. Due to cumulative internal damage caused by the original fault, it is
more likely that a faulted transformer will experience an eventful failure the second time it reenergized, which is the time that the lineman is present and at risk.
If the transformer is faulted and does not experience an eventful failure upon re-energization, and the
fuse blows, which is usually the case, the lineman will have:
Wasted a fuse, using it as a troubleshooting tool as opposed to its original purpose as a
transformer protective device,
Wasted time getting set up to perform this operation. Depending on the situation, this might entail
setting up a bucket truck, climbing the pole, donning personal protective equipment, and other
preparatory tasks. All this must be done before the condition of the transformer is even
determined.
Possibly tripped another upstream device, such as a recloser on its instantaneous setting, which
will cause an unnecessary momentary outage, which may affect the utilitys SAIFI measure.

Page 4

4. An improved method of detecting and indicating distribution

transformer faults
As mentioned earlier, several fault studies regarding the characteristics of faulted transformers have
revealed that there is a characteristic dynamic pressure signature associated with an internal transformer
fault. Based on this consistent physical principle, efforts were undertaken to use this property to develop a
device to detect and indicate internal transformer faults.
Until the concept of an internal fault indicator was considered, we must remember all other technologies
examine or control external components of the system, and therefore were all best guess options for the
utility. The one exception is off-line testing technology which is somewhat complex and time consuming to
use when the line worker is on the pole. This provides the opportunity for innovation to identify the internal
fault directly, inside the transformer.
The first work on this technology occurred in the late 1980s when engineers were conducting transformer
withstand tests. It was determined there is a unique pressure signal that occurs only in internal fault
circumstances. All other functions of pressure in the air space of a transformer tank occur at a rate 1000 or
more times slower than the rate of rise of the pressure surge during internal arcing faults, so the rate of
pressure rise can be used to reliably identify the occurrence of an internal arcing fault.
a) Development
In consultations with several utilities the following set of requirements were defined for the design and
development of an internal fault detector:
It should provide a clearly visible external signal indicating the occurrence of an internal fault for
short circuit current exceeding 500A and causing a peak transient pressure in the airspace of a
minimum of 0.5 psi over 5-7 ms. It should not be activated by any other fluctuation of the pressure
in the air space that can occur under diverse operation conditions.
It should install in the airspace of distribution transformers without affecting their mechanical and
electrical integrity, and provide reliable and maintenance free service during an expected
transformer life of 30 years. Its reliability should not be affected by the fluctuations in the ambient
temperature between - 40 and +365 F.
Be cost effective to manufacture, and easy to assemble and install in the airspace of the standard
oil filled pole top transformers.
Incorporate standard pressure relief functionality into the device.
One detection method that accomplishes these goals is to use the pressure differential across a flexible
membrane to activate a fault signaling mechanism. A device that incorporates this principle is the Internal
Fault Detector (IFD), developed by IFD Corporation. The IFD was in development for over a decade,
and represents the result of a collaborative R&D effort involving technical support by a group of utilities
with the objective of improving lineman productivity, enhancing customer service, and, increasing lineman
safety associated with transformer failures.
The IFD has been designed to be sensitive only to transformer internal faults. The purely mechanical design
provides inherent advantages over other sensing schemes. It requires no power supply, is insensitive to its
electrical environment and does not depend on the electrical status of the transformer to operate. It does not
react to high electrical or electromagnetic fields or even temperatures within the limits of the transformer
design. It operates reliably in an environment that would destroy most electronic devices. Moreover,
because it is designed to respond to the phenomena that cause catastrophic transformer failure, it does not
carry the sensitivity vs. selectivity design compromises inherent with thermal and electrical over current or
electrical signature sensing devices.
On a more practical note, the simple mechanical design and construction of the IFD make it economic to
manufacture and install in distribution transformers, where large volumes make unit cost an important
selection factor for utilities.

Page 5

b) Operation
The IFD has 2 distinct and separate functions. First, it includes a static pressure relief device (PRD) that
operates the same as standard PRDs in the industry today. Customers asked that this function be included
in the IFD in order that the solution did not add another hole to the tank. This function also provides the
opportunity to reduce the total tank cost due to elimination of the standard PRD and welded boss solution
that is currently standard on most transformer designs.
Second, the IFD is a mechanical sensing device which activates when a transformer faults internally. The
sensor is a specialized membrane that is sensitive only to a rapid rise of pressure in the air above the oil.
This rapid pressure rise occurs only during an internal fault. When the membrane moves in response to the
rapid pressure rise, it releases the spring-loaded, readily visible indicator on the outside of the tank.
The IFD cannot be reset. This feature ensures the IFD on a faulted transformer should not be reset and reenergized without first removing and inspecting the transformer in the shop.
Figures 1 and 2 show a cross sectional drawing and mode of operation.

Figure 1. Prior to operation, the pressure detecting membrane and

its trigger shaft (red and blue vertical rod) are in the lowered
position, locking the indicator (yellow) in place. The large spring
on the right side stores the energy to push the indicator out; the
small coaxial assembly on the left is the pressure relief device
(PRD).

Figure 2. When the membrane reacts to the pressure pulse, it

moves up, carrying the trigger shaft with it (red and green). This
releases the indicator, which is pushed out by the spring.

Rapid Pressure

c)

Installation

The IFD is easily installed in all poletop and padmounted distribution transformers, and is now commonly
installed as a replacement for Pressure Relief Devices by all major distribution transformer manufacturers.
d) Business Impact
Challenges for utilities to continue to improve cost/performance and service have never been greater. The
distribution system must integrate new and old technologies in an evolutionary, cost effective way.
Strategically, utilities have to consider the value of technical solutions over the life of the system, where a
key element of the future is a changing experience and skill level in line workers and a drive to reduce costs
and improve service levels. Importantly, safety references are always top referenced corporate mission and
objectives statements due to the risks inherent in high power electrical systems.

Page 6

These considerations were all a key part of looking for low cost, simplicity and reliability in a transformer
fault detection toolto assist in the journey of doing it right the first time.
The IFD demonstrates a positive financial impact to utilities in the following areas:
Reduced outage durations By enabling the linemen to immediately see if the transformer is or is
not faulted, this allows him to eliminate troubleshooting time and immediately replace a faulted
transformer (if the IFD is activated) or safety re-energize an unfaulted transformer (if the IFD is not
activated). The result is a shorter duration of the outage in either case, and improved SAIDI scores.
Productivity Improvement The IFD enables line crews to be more efficiently utilized, especially
during storms. Since the IFD can be easily seen from the ground, a single troubleman can be
dispatched to determine the cause of the outage vs. a two-man crew with a bucket truck.
Material savings The IFD enables the utility to keep more unfaulted transformers in the air and
waste fewer fuses using them as a troubleshooting tool.
Increased job safety As mentioned in an earlier section, inevitably transformer failures do occur.
These events can be very costly to a utility depending on the impact of the event. Each utility has its
own experience, will likely know their own cost of accidents and injuries, and continually assign a
value to protection and prevention. Most utilities have enough actual experience to make this
calculation. Certainly, even the cost of reporting an incident with the potential for injury, or where a
spill has required and environmental cleanup, can quickly reach thousands of dollars. Accidents that
result in injuries cost much more and the costs of investigation, treatment, and increased insurance
premiums can significantly impact the bottom line. From this angle, any improvement in worker safety
yields a direct bottom line benefit. The opportunity for accident cost reduction is significant. An
event can quickly move into the million-dollar cost category. In this era of intense competition and
escalating insurance rates, a reduction in this class of workers compensation loss will generate
noticeable bottom line savings and directly addresses a utilitys competitive advantage.

5. Conclusion
The presence of an internal fault is difficult to detect without disconnecting the load and testing the
integrity of the insulation.
Re-energizing a faulty transformer may cause further damage due to the escalation of the fault; the
transformer can fail violently during this operation, increasing the safety risk to the line crew, and
increasing utility costs and service delays.
The transient pressure rise in the airspace of the distribution transformers is a finger print common to
all internal arcing faults.
Detecting and signaling the presence of an internal fault by incorporating a built-in detector can:
Eliminate long interactive diagnostic procedures that are prone to human error,
Accelerate restoring of service to the customers, and
Improve line crew safety.
The internal fault detector (IFD) has been co-developed directly with utilities; to enable crews to quickly
identify transformers with potentially dangerous internal faults. With over 300,000 in service, it has proven
itself dependable for this purpose.
More strategically, the IFD represents another small step on the journey to a distribution system where,
more and more, information is used to improve customer service and the effective utilization and safety of
more valuable resources.

Page 7