Effectiveness of Direct Bonding of Gas Piping
in Mitigating Damage from an Indirect Lightning Flash
Prepared By:
Brian Kraft
Rising Tide Consulting LLC
Palmer, MA
Robert Torbin
Cutting Edge Solutions LLC
Framingham, MA
August 2007
© Kraft/Torbin 2007
Reproduction of this document, in whole or in part, without express, written consent is strictly prohibited.
Effectiveness of Direct Bonding of Gas Piping
in Mitigating Damage from an Indirect Lightning Flash
Brian Kraft and Robert Torbin
i
Executive Summary
Damage to gas piping systems during a lightning event may occur due to arcing between the gas
piping and some other conductive metallic pathway that is in close proximity. The presence of
different levels of electrical potential in these pathways creates the impetus for the arc to occur.
The different levels of potential are directly attributable to unequal or non-bonding of the
individual pathways to the premise grounding system. A series of physical tests were performed
by Lightning Technologies Incorporated for the Titeflex Corporation to simulate a corrugated
stainless steel tubing (CSST) gas piping system that has been energized by an indirect lightning
flash. This paper summarizes the interpretation of this data and presents recommendations
concerning the direct bonding of gas piping systems.1 For testing purposes, a standard
residential building was modeled including the bonding (and non-bonding) of the gas piping
system after being energized by an indirect lightning flash. The evaluation of the testing data
indicates that the risk of damage to the gas piping system due to arcing is significantly reduced
when the system is directly bonded to the premise electrical grounding system. A non-bonded
piping system would be subject to 97 percent of the charge from an indirect lightning strike,
while a direct bonded piping system would be subject to only 20 percent or less of the same
charge. Based on the data and the experience of the authors, a set of installation guidelines for
the direct bonding of CSST systems was developed. The recommended direct bonding should
not be considered a definitive protective measure against all lightning strikes nor construed as an
optimum level of bonding.
1.0
Study Objectives
The objective of this reported study was to perform an independent engineering evaluation of
data collected during an experimental test program to evaluate the effectiveness of direct
bonding of the gas piping system to the grounding system of the premise electrical system used
in typical residential construction. The product of this evaluation was the development of
installation guidelines for an effective bonding method.
1 The testing results discussed herein are copyrighted by Titeflex Corporation. Reproduction of information covered by
Titeflex’s copyright without the express, written consent of Titeflex Corporation is strictly prohibited.
© Kraft/Torbin 2007
1
2.0
Background
2.1
Piping Systems and Conductive Metallic Pathways
A residential gas system is composed of several distinct parts including the gas utility service,
the piping system within the building and the pipe and fittings connecting these two parts. The
gas utility service is generally comprised of underground polyethylene (PE) piping leading from
the distribution network/main in the street to the exterior of the building where the piping is
brought above ground via a gas meter riser that connects to the utility gas meter. The meter riser,
which provides the transition from underground PE piping to above ground metal piping, is
almost always electrically conductive. However, these risers have only a minimum length of
steel pipe in direct contact with the earth and are not considered part of the grounding electrode
system. Section 250.52 (B) of the NEC prohibits the use of any underground metallic gas piping
as a grounding electrode.
In the past, the utility underground gas service piping was commonly all steel pipe, and was
often cathodically protected. For this reason, and to prevent inadvertently creating a second
electrical ground point, a dielectric union has been traditionally placed by the utility in the gas
line prior to the connection point with the gas meter. The dielectric union essentially electrically
isolates the underground portion of the gas delivery network from the interior gas piping system.
Even though non-conductive polyethylene piping is now used underground in virtually all new
construction, the dielectric union continues to be installed for protection from stray currents. The
meter riser, regulator and meter are commonly installed outdoors on the side of the building for
protection and other practical considerations. Other meter locations were not considered as part
of the study, nor were the service connections for LP systems.
House gas piping may consist of rigid steel pipe, copper tubing, corrugated stainless steel tubing
or a combination of these three materials. Routing within the building varies greatly depending
on the floor plan, gas appliances installed, and installation methods employed by the piping
installer. Depending on when the gas piping is installed during the construction schedule and
given the limited space available to install utilities, it is not unusual for interferences to occur
between different utilities. Therefore, it is not uncommon for gas piping to be routed in close
proximity to, or even in direct contact with electrical wiring, heating/cooling ducts, water piping,
steel structural supports and appliance exhaust vents. This proximity can have a contributory
effect on the impact of lightning strikes to these various metallic pathways.
All gas piping systems (regardless of material) are required by the electrical and fuel gas code to
be electrically continuous. All gas piping is installed per the requirements of the locally adopted
fuel gas code, most commonly the National Fuel Gas Code (NFPA54), International Fuel Gas
Code (IFGC) and the Uniform Plumbing Code (UPC) and, in the case of CSST, in accordance
with the manufacturer’s instructions. Gas piping systems must also be bonded to the grounding
system in accordance with both Section 250.4(A) (4) and Section 250.104(B) of the NEC.
Traditionally, the means for bonding of gas piping has been the equipment grounding conductor
when the piping is attached to equipment that is electrically powered. However, the requirement
for bonding when the gas equipment is not electrically powered is not clearly stated in the NEC.
© Kraft/Torbin 2007
2
Similar to the gas utility service, a residential electrical system consists of the utility electrical
service mains, the house wiring within the building and an electrical panel/enclosure that forms
the interface between the two. Utility mains are generally run via overhead power lines.
However, newer housing developments often employ underground distribution systems. Both
overhead and underground systems are locally earth-grounded to provide protection from stray
voltage, voltage imposed by lightning, line surges and unintentional contact with higher voltage
lines as stated in Section 250.4 (A) (1) of the NEC. The premise electrical panel also
incorporates a grounding system, commonly through a grounding electrode conductor and buried
electrodes/rods. This provides the electrical ground specific to that building, and is referred to as
the house ground.
Other than a three-foot minimum spacing that may be required in some areas, few requirements
exist for the location of the gas and electric utility service entrances in relation to each other. It
is generally left up to the two utilities to determine the placement of their facilities based on their
own requirements and other considerations. While it is most common for the two utility
entrances to be located in close proximity, installations resulting in the electrical panel being on
the opposite side of the building from the gas meter are not uncommon. The separation distance
between the electric panel and/or the grounding electrode system and the gas meter is an
important factor in the design of an effective bonding connection.
In the event of an electrical fault, stray voltage or fault current could be imposed onto any
conductive metallic pathway within the structure. If there is no effective path to ground, a
significant electrical hazard to the occupants of the building would exist. This hazard is largely
negated by the bonding of these pathways to the electrical ground system of the building. This
bond provides a safe low-impedance path to neutral earth. When the equipment grounding
conductor is used as the bonding means, no additional effort is required by the piping installer or
the electrical contractor to achieve this condition. Thus, we refer to this as “self-bonding”. In a
situation where the gas piping system is not connected to a permanently wired gas appliance, it is
unlikely that the system is bonded to the electrical system ground at all based on conventional
plumbing/electrical construction practices. An un-bonded metallic system increases the potential
for a hazard (to the building occupants) to exist.
2.2
Direct Bonding
It is not uncommon for piping systems, including steel gas piping, to be bonded to the electrical
ground system through the use of a dedicated wire and clamp. As this method provides a direct
connection between the piping system and the electrical ground system, it is referred to as “direct
bonding”. For direct bonding, the clamp is attached to a single location on the piping system to
provide electrical contact. The wire is then connected to the clamp and routed to either the
electrical panel ground or directly to the grounding electrode system.
There are numerous advantages to this method. First, it provides the safety of bonding without
reliance on the connection of the gas piping system to a grounded appliance where bonding
could be compromised if the appliance is taken out of service, replaced with a differently fueled
© Kraft/Torbin 2007
3
unit, or there is a wiring system modification. Additionally, the direct bond method can be
visually inspected for continuity, as the clamp and wire connections must be visible. A self-bond
connection requires the use of a continuity tester to verify a proper connection at the appliance.
Finally, the direct method of bonding provides a low impedance electrical path to ground that is
significantly shorter, less electrically resistive, and therefore, more effective than is a self-bond at
safely conveying stray voltages and ground-faults to ground.
Direct bonding of all metallic supply lines entering a building is a critical but often overlooked
approach when considering protection of a building and its contents during an electrical storm.
Lightning strikes near buildings (with ground current transfer distances of up to 1 to 3-km from
the strike center) can induce differences in potential between the electrical system and any nonbonded mechanical system. Direct bonding of these systems (using a 6 AWG copper wire) to
the building grounding electrode system allows the mechanical systems to be energized at (or
near) the same rate as the electrical system and in unison with the voltage wave induced by the
lightning strike.
Use of the equipment grounding conductor (EGC) as the bonding means does not achieve the
same effect. The EGC (which is typically a 12 or 14 AWG copper wire) does not allow that
mechanical system to be energized at the same rate as the electrical system. The path to ground
through the EGC is typically much longer (and with greater impedance) than the direct bonding
distance (near the service entrance) between each mechanical system and the grounding
electrode system. When energized by lightning, this situation permits the electrical potential in
the many conductive pathways to become unbalanced, and thus arcing is more likely to occur.
Bonding through the equipment grounding conductor is only intended for personnel safety in the
event of an electrical fault occurring in the premise wiring system, and has been shown to be
inadequate when dealing with lightning energy. That being said, it should be noted that the NEC
does not require direct bonding for all metallic systems, nor does it effectively address lightning
protection which is considered out-of-scope.
2.3
Lightning Energy
Lightning is, by its very nature, a massively powerful natural phenomenon, capricious and
impossible to predict. While significant research has been performed on the lightning
phenomenon, general engineering solutions to mitigate its impact have not been widely
implemented in the residential market. The occurrence of lightning varies from intense (2
million strikes per year) to light (2,000 strikes per year) on a state-by-state basis. Given that the
magnitude, frequency and duration of the energy involved with each lightning strike varies
widely, the design of the most effective protection should take local conditions into
consideration. Lightning energy is transferred to a structure whether the strike is directly to the
building or indirectly through another medium (such as the earth). A direct strike involves both
direct hits to the structure or strikes from a branch off of the main bolt. An indirect strike occurs
when the energy from the main bolt enters the structure through another pathway such as the
overhead power line or through the ground under the house.
© Kraft/Torbin 2007
4
Due to the high levels of electrical power inherent in a lightning flash, no single point or path can
provide it with an adequate connection to the earth which it seeks. Practically speaking, this
means that a direct lightning flash to the ground or to a structure will saturate its immediate path
to ground at the object struck. The lightning strike will proceed to conduct its overflow energy
indirectly to ground by generating very high voltages and currents in the surrounding earth,
objects and structures, commonly to very large distances. In this way, lightning will use all
available paths to disperse this electrical energy to neutral earth. In the case of a direct strike to a
structure, energy will travel indiscriminately through all of the conductive systems within the
building including electrical wiring, communications cables, ventilation ducts, water piping, gas
piping, steel structural members and exhaust vents, and often with a destructive effect.
In the case of an indirect flash near to the structure, lightning energy will travel into the structure
and through the aforementioned systems if they provide a significant path to neutral earth. The
energy level in this situation is such that systems or objects that would not normally be
considered electrically continuous and/or conductive can and are used as paths to ground. For
example, water in trees or wooden framing members can provide lightning with a path to ground,
but with significant damage to the insulating cellulose. Re-bar in unburied footings will respond
in the same fashion, with similar explosive damage to the surrounding concrete. The dielectric
insulator at the gas service entrance is commonly arced over during a lightning event, with the
lighting using the underground riser as a ground point. Any of the conductive systems within the
residence that pass near a steel structural support, such as a Lally column in a New England
basement, will provide a potential path to ground, even if a short arc through air is required to
make the connection.
Under certain circumstances (such as close proximity), CSST can be damaged by arcing to/from
other conductors as they are energized by a lightning strike. This arc causes localized heating at
the arc termination points, sometimes creating enough heat to melt through the steel pipe wall.
This effect is similar to damage caused to aircraft skins by direct lightning flashes. The existing
research shows that this type of damage is most closely related to charge delivered, that is, the
amount of charge transferred by the arc/flash and the duration over which the arc is active.
Charge delivered, or the area under the curve of amperes-versus-seconds, is highly predictive of
material damage. Lightning testing commonly utilizes current-versus-time wave shapes
identified by values for time-to-peak current amplitude and time-to-current decay of 50-percent
peak amplitude (as measured in microseconds) to model both direct and indirect lightning
flashes.
© Kraft/Torbin 2007
5
3.0
Laboratory Evaluation Program
3.1
Test Objectives
A testing protocol was developed, in consultation with Lightning Technologies Incorporated
(LTI), to model an indirect lightning flash striking near the house and the associated energy
entering the premise (and the gas piping system) through a pathway provided by the electrical
service entrance.2 LTI (Pittsfield, MA) is a leading developer of, and test facility for, lightning
protection systems, and was under contract with the Titeflex Corporation to perform these tests.
The testing program was designed to evaluate different bonding approaches and techniques for
gas piping systems consistent with the requirements of the NEC and the National Fuel Gas Code.
The laboratory testing program was conducted using so-called “typical” lightning profiles, and
was only intended to allow relative comparisons of alternative bonding techniques. The results
from the testing program provide general guidance on improving the performance of bonding as
it affects the ability of the gas piping to absorb lightning energy without damage and to divert
this energy safely to ground. However, the recommended bonding technique should not be
considered a definitive protective measure against all lightning strikes nor construed as an
optimum level of bonding.
3.2
Simulation of Physical Piping System
In order to develop a test method to achieve the stated objective, the complex physical system
represented by a residential building (including the nearby lightning strike) had to be analyzed
and simplified to allow for laboratory simulation. Factors that were considered in the
development of the laboratory model and testing procedure included:
•
Basic residential wood-framed construction
•
Low-pressure metallic gas piping systems and their installation configurations
•
Premise electrical wiring systems and their configurations
•
Electrically conductive connections between the electrical and gas piping systems
•
Lightning, its behavior and typical energy profiles as determined by industry sources
The model used for the evaluation program represented a standard residential building of wood
construction with the gas and electrical service entrances of varying distances apart and
energized under the following conditions:
•
Lighting striking a nearby utility pole and entering through the electrical service mains
and saturating the premise electrical grounding system
2 The testing results discussed herein are copyrighted by Titeflex Corporation. Reproduction of information covered by
Titeflex’s copyright without the express, written consent of Titeflex Corporation is strictly prohibited.
© Kraft/Torbin 2007
6
3.3
•
Lightning energy traveling through the building via the gas piping and electrical wiring
with the gas service entrance as a primary path to neutral earth via a “failed” dielectric
fitting
•
Polyethylene jacket of the CSST in direct contact with the house wiring
•
Polyethylene jacket of the CSST scuffed through normal installation techniques
•
Either a non-bonded or direct bonded piping system installed
Model of Physical System
In order to model the physical system comprising a gas piping system installed in a typical
single-family residential structure, test samples were setup as shown in Figure 1. A hole was
made in the polyethylene jacket (using a standard soldering iron) to represent damage caused to
the jacket during installation. This hole was placed over the crown of a corrugation to expose the
stainless steel beneath. The electrode was then placed in contact with the PE jacket, entirely
covering, but not touching, the exposed stainless steel. If an arc was developed during the test, a
new hole was prepared and the electrode moved to this new location before the next test was
conducted. New coating holes were spaced approximately one inch apart.
Figure 1: Test Sample Assembly
The generator consisted of a large bank of capacitors connected in parallel, with a
resistor/inductor assembly at the capacitor bank output to control output wave shape. The
generator was then connected to the sample assembly via a switch. The generator was also
connected to the facility ground plane to provide a full circuit. Firing of the generator consisted
of applying power to the capacitor bank (switch open) to develop a charge across this bank.
When the desired voltage across the bank was reached, the power cycle was discontinued. The
switch was then closed, thus subjecting the test sample assembly to the stored charge.
© Kraft/Torbin 2007
7
Two current probes and one voltage probe were connected to an oscilloscope to record current
output of the generator, current passing into the electrode via an arc to the CSST, and voltage in
the CSST. Data time plots allowed the calculation of charge output from the generator and
charge traveling down the electrode via an arc from the CSST. Various generator charges were
used to explore the relationship between the measured values, arc generation, and the generation
of a hole in the CSST for various bonding configurations. Tests were performed for the
following direct bond configurations:
•
10 feet of 6 AWG copper wire
•
20 feet of 6 AWG copper wire
•
40 feet of 6 AWG copper wire
•
20 feet of 8 AWG copper wire
•
40 feet of 8 AWG copper wire
Tests were also performed with no direct bond in order to generate some baseline data. A
resistor bank, highly resistive relative to the wiring in the sample assembly, was supplied as a
direct path to ground for the generator charge in order to model other available, low quality
ground paths. This direct path to ground was also necessary for equipment and personnel safety.
3.4
Lightning Profile
Due to the chaotic nature of lightning, it is impossible to re-create in an experimental setting.
Natural lightning displays an extremely wide range of electrical characteristics for which modern
physics is incapable of providing a definitive model. However, simplified models of the
electrical impulse from a lightning strike have been developed and can be found in various
technical papers and standards. The model chosen for the evaluation of lightning mitigation
techniques and materials must be determined by the researcher based on the physical system
being studied. The model used for this testing was based on lightning current traveling via the
overhead utility wires and internal conductors to the test site (a residential house). This resulted
in the selection of an impulse model intended for indirect effects as the most appropriate.
There are commonly used current impulse models for indirect lightning effects found in both the
aircraft and surge suppressor industries. An example of one of these waveforms can be seen in
Figure 2. Further details and examples of these waveforms can be found in SAE ARP-5412 [ref
1] and various IEEE standards. The majority of these waveforms are similar in shape with the
greatest variation found in the magnitude and duration of the lightning energy. These values are
expressed in time to rise to peak current and time to fall to 50 percent of peak current.
© Kraft/Torbin 2007
8
Figure 2: Waveform Example [ref 1]
A review of surge suppressor specifications found three common waveform durations: 8µs x
20µs, 10µs x 300µs and 10µs x1000µs. Preliminary testing was performed using all three of
these waveforms. This testing determined that the 10µs x 300µs waveform caused damage at
significantly lower charge levels than the other forms. In order to provide the most conservative
results, the 10µs x 300µs waveform was selected for all subsequent testing.
3.5
Testing Results
Using the test protocol previously described, 40 generator firings were performed on non-bonded
test samples. Generator settings were selected to explore the conditions required to create an arc
to the test sample, and thus firings were performed at settings where arcing was only
intermittent. Arcing was generated in only 12 firings. Arcing occurred intermittently at voltages
of 2.35 kV to the maximum setting of 4.75 kV. All data was collected to provide a baseline for
subsequent testing of bonded test samples. A summary of this data can be found in the Test
Results Summary, Table 1 and Table 2, as well as in Appendix C.
Based on the same test protocol, 49 generator firings were performed on bonded test samples.
Generator settings were selected to initiate arcing so as to explore current and charge transfer to
the test sample while in a bonded assembly. Arcing was generated in 37 firings. A summary of
this data can be found in the Test Results Summary, Table 1 and Table 2, as well as in Appendix
C.
© Kraft/Torbin 2007
9
Table 1: Test Results Summary, Generator Output
Test
Set-Up
No Bond
10-ft, 6AWG
20-ft, 6AWG
40-ft, 6AWG
20-ft, 8AWG
40-ft, 8AWG
Generator Current (kA)
Min
Max
0.28
4.85
2.00
9.50
3.05
9.15
1.95
13.30
5.15
9.25
5.10
10.00
Generator Charge (C)
Min
Max
0.78
1.90
0.75
3.70
1.20
3.80
0.61
3.20
2.10
3.40
2.10
3.45
CSST Voltage (kV)
Min
Max
0.47
4.75
1.30
5.25
4.00
8.20
2.40
8.50
4.65
6.95
4.70
5.60
©Titeflex Corporation 2007 (reproduced herein by consent of Titeflex Corporation)
Table 2: Test Results Summary, Test Sample
Test
Set-Up
No Bond
10-ft, 6AWG
20-ft, 6AWG
40-ft, 6AWG
20-ft, 8AWG
40-ft, 8AWG
Results
Arced
Hole
i
No
i
i
Yes
Yes
Yes
No
No
No
No
No
CSST Current (kA)
Min
Max
% of
Generator
2.30
4.65
96
------1.40
3.95
40
1.15
7.80
57
2.30
4.25
45
3.10
5.60
60
CSST Charge (C)
Min
Max
% of
Generator
1.10
1.95
98
------0.08
0.37
8
0.10
0.88
21
0.25
0.49
13
0.37
0.77
20
Legend: i: intermittent
©Titeflex Corporation 2007 (reproduced herein by consent of Titeflex Corporation)
© Kraft/Torbin 2007
10
4.0
Analysis
4.1
Interpretation of Results
From the test data (see Appendix C) it is clear that the correlation between arc initiation and
voltage at the CSST is poor. Arc initiation occurred and did not occur within the range of
2.35kV to 4.75kV. Conversely, the bond configuration of 10-ft of 6 AWG showed no arcing at
voltages as high as 5.25kV. As would be expected of a phenomenon as complex as arcing, the
data shows that more factors (than were measured during these tests) are significant to arc
initiation. Some of these factors may be the size of the hole in the PE jacket, distance from the
electrode to the stainless steel of the CSST, time versus voltage, and the condition of the
electrode. While further research could be done into these factors, the wide variation in
construction techniques would seem to minimize the relevance. The only conclusion that can be
reached is that fairly high voltage levels, in the order of thousands of volts, are required to
initiate an arc.
It is clear from the non-bonded configuration results that when a poor quality connection to
ground is present, an initiated arc to a well grounded conductor will carry the majority of the
current and charge. Lightning entering a building through the electrical system decreases the
efficacy of the electrical grounding system via charge saturation and travels to neutral earth
through the gas service entrance via an arc-over of the dielectric union. The non-bonded
configuration was the only configuration examined where the arc generated perforations through
the tubing wall.
It is very interesting to note the total lack of arcing with the 10-ft of 6 AWG bond configuration.
This bond configuration presented the lowest impedance path to ground of any of the
configurations examined in this study. This allowed the discharge to occur in a significantly
shorter time period, thus minimizing the exposure of the sample to peak voltages. It can be
inferred that insufficient time was available at peak voltage for the intervening dielectric, in this
case air, to breakdown and allow an arc to initiate. This effect is of great interest in the
prevention of damage to the piping and further studies are recommended.
The most significant finding of an analysis of the testing performed to date is the fraction of
generator current and charge that is conducted by the gas piping via an initiated arc and the total
lack of arc generated perforations due to this decrease. Each configuration shows a fairly stable
percentage regardless of generator settings or output. Table 3 summarizes these percentages.
While the variation in charge transfer due to generator waveform appears to be significant, there
is insufficient data at this point to come to a firm conclusion. Further research could be
undertaken to examine this variation. Contrasting the data between bonding configurations
shows a good correlation between resistance of the bond wire and percent charge transferred.
Doubling the length of the bond wire effectively doubles its resistance. This doubling effect is
shown in an equivalent doubling of percent charge transfer for the same wire gauge.
© Kraft/Torbin 2007
11
Table 3: Percentage of Electrical Impulse Transmitted via Arcing
Bonding Configuration
Peak Current
Charge Transfer
92%
97%
10-ft, 6 AWG
No Arcing
No Arcing
20-ft, 6 AWG
No Bond
40%
8%
1
57%
21%
2
40-ft, 6 AWG
56%
18%
20-ft, 8 AWG
45%
13%
40-ft, 8 AWG
60%
20%
40-ft, 6 AWG
©Titeflex Corporation 2007 (reproduced herein by consent of Titeflex Corporation)
1: Generator waveform 18μs x 215μs rather than normal 13μs x 300μs.
2: Single data point.
The various direct bond configurations modeled represented a variation in distance from the gas
service entrance to the electrical service entrance typical of residential construction. Other
pathway lengths are possible, but have not been studied at this time. The selection of either a 6
or 8 AWG copper bond wire has its precedence in the NEC based on several factors and
operating conditions as discussed in Section 4.2. The implementation of direct bonding of gas
piping for lightning mitigation places additional requirements on the bonding conductor and
connection that exceed the conventional bonding practices for the electrical protection of the gas
service.
4.2
Review of NEC Requirements
Ultimately, the effectiveness of a direct bond connection between the gas piping system and the
grounding electrode system will depend on many inter-related factors. Some of these factors
include the size and ampacity of the bonding jumper; the bonding jumper material; the location
of the gas piping relative to the grounding electrode system; the attachment method; and the
point of attachment to the gas piping system. The electrical system grounding system is assumed
to be designed and installed in accordance with the 2005 Edition of the National Electrical Code
(NEC). The NEC permits the gas piping to be bonded to either the service equipment enclosure,
the grounded conductor at the service, the grounding electrode conductor (if of sufficient size) or
to the one or more grounding electrodes used. The NEC under Section 250.4 (A) (4) requires
that all electrically conductive materials (which potentially includes all metallic piping, ducts,
vents, coax cable etc.) be bonded together in a manner that establishes an effective (lowimpedance) ground-fault current path.
© Kraft/Torbin 2007
12
The National Electrical Code contains many references to the requirements for bonding of
electrically conductive materials which include wiring, piping, ducts, vents and structural steel.
These requirements are specified throughout Section 250 of the NEC and all have the common
goal of protecting the public safety from electrical faults within the premise wiring system by
establishing an effective ground-fault current path. The objectives of this study were focused on
establishing more effective bonding of gas piping systems that may become energized by
indirect lightning strikes near the building. Towards that end, the authors conducted a study of
the NEC to review its coverage for the use and type of bonding for other conductive materials
and pathways to ground, and to compare those requirements with the conventional bonding
practices used for gas piping. Appendix B contains a listing of references within the NEC for
use of direct bonding and the application of a 6 AWG copper wire as the bonding means. It is
quite clear that the use of a 6 AWG copper bond wire is a well established approach for other,
similar conductive metallic systems; that a 6 AWG copper wire will be an effective means for
diverting (to ground) much of the energy associated with an indirect lightning strike; and the use
of direct bonding should be a familiar and straightforward solution to implement in the field by
electrical contractors.
4.3
Installation Instructions for Direct Bonding
Based, in part, on the testing results and subsequent analysis, and the general practice of bonding
established through the NEC, a separate set of installation instructions has been developed for
direct bonding of gas piping systems. These generic instructions have been included as
Appendix A. In essence, the instructions require the installation of a 6 AWG copper wire from a
single attachment point on the gas piping system near the service entrance of the building
directly to the grounding electrode system. All of the recommended practices and sizing criteria
are consistent with the requirements of Section 250 of the NEC and the National Fuel Gas Code.
When approved by each CSST manufacturer, these bonding practices will become part of their
published installation instructions, and a mandatory requirement for all new installations. In
accordance with NEC Section 110.3(B), listed equipment shall be installed and used in
accordance with the manufacturer’s instructions and the terms of the listing. In some
jurisdictions, the manufacturer’s instructions are permitted to take precedent over the affected
code when they are more conservative or restrictive in practice.
© Kraft/Torbin 2007
13
5.0
Conclusions and Recommendations
5.1
Conclusion
A test assembly and testing protocol were developed that represented a typical residential piping
system subjected to an indirect lightning flash via the electrical power lines. This model did
incorporate various configurations of bonding including no bonding and direct bonding with
copper conductors of varying lengths and gauges. The test results show a good correlation
between bonding configuration and percent charge transferred via an initiated arc to the gas
piping. Based on a subsequent review of the test results, it is estimated that a non-bonded system
would be subject to 97 percent of the charge from an indirect lightning flash entering the
building. By comparison, any of the direct bond configurations studied would only be subjected
to 20 percent or less of the same charge. In the case of a direct bond supplying a sufficiently low
impedance path to ground, there is a distinct possibility that arc initiation can be halted all
together. This effect requires further confirmation.
In conclusion, it is clear that the use of direct bonding with a gas piping system, specifically
CSST, can significantly decrease the risk of damage and increase the safety of the gas
installation. This conclusion leads to the recommendation that direct bonding be required for all
new gas installations. To implement direct bonding methods in the field, a set of installation
instructions have been developed and will be used to provide guidance to electrical contractors
who generally install the bonding means and to electrical inspectors who are responsible for
insuring that the bonding is properly installed.
5.2
Recommendations
Based on the results of the testing and the analyses of that data, the appropriate fuel gas and
electrical codes should be amended to require the direct boding of all metallic piping in
residential construction. This process has been initiated with a proposal to the National Fuel Gas
Code. The code change process is long and complex, and it is recommended that a committee of
interested parties be formed to monitor the process to its conclusion. Many advantages have
been highlighted in this report including ease of inspection, robustness (avoiding appliance and
wiring changes), and a significant improvement in providing a low impedance path to ground.
The testing and analyses also highlight a compelling reason to require direct bonding: There is a
significant reduction in the level of imposed current and a lower potential for arcing in the event
of an indirect lightning flash near the building. Further research is required to fully evaluate the
effects of direct bonding conductor length and gauge, and the ability of this bond configuration
to mitigate or eliminate arcing damage. However, there is sufficient evidence to recommend the
adoption of direct bonding by the code community without any further delay.
© Kraft/Torbin 2007
14
6.0
Bibliography and References
1.
----. ARP5412, Aircraft Lightning Environment and Related Test Waveforms,
Warrendale, PA: SAE, 2005.
2.
----. NFPA 54, National Fuel Gas Code, National Fire Protection Association, Quincy,
MA 1999, 2002, 2006.
3.
----. NFPA 70, National Electrical Code, National Fire Protection Association, Quincy,
MA 1999, 2005.
4.
----. NFPA 780, Standard for the Installation of Lightning Protection Systems, National
Fire Protection Association, Quincy, MA 2000, 2004.
5.
Anderson, R.B. and Eriksson, A.J., Lightning Parameters for Engineering Applications,
Suceava, Roumania: Colloquium and Study Committee Meeting, Cigre Study
Committee 33, 1979.
6.
Cianos, N. and Pierce, E.T., A Ground-Lightning Environment for Engineering Usage,
Menlo Park, California: Stanford Research Institute, 1972. Figure 25, pp. 66.
7.
Crouch, K.E., Lightning Technologies Incorporated, Pittsfield, MA 2007.
8.
Kithil, R., Lightning Protection for Engineers, Louisville, CO: National Lightning
Safety Institute, 2005.
9.
Martzloff, F. D. and Crouch, K. E., “Coordination de la Protection Contre les
Surtensions dans les Reseaux Basse Tension Residentiels”, Proceedings 1978 IEEE
Canadian Conference on Communications and Power, 78CH1373-0, pp. 451-454.
10.
Rupke, E., Lightning Technologies Incorporated, Pittsfield, MA 2004-2007.
© Kraft/Torbin 2007
15
APPENDIX A
GUIDELINES FOR DIRECT (ELECTRICAL) BONDING OF CSST SYSTEMS
This guideline document describes the requirements for the direct bonding of corrugated
stainless steel tubing (CSST) gas piping systems. The bonding of CSST shall be installed by a
licensed contractor recognized by the local jurisdiction as capable of performing such work.
Direct bonding is required for all CSST natural and LP gas piping systems whether or not the
connected gas equipment is electrically powered. These guidelines are applicable to all new
CSST installations as well as partial retrofits of CSST to existing steel pipe and copper tubing
systems. These guidelines are applicable for typical single-family and multi-family dwellings
and certain commercial buildings. An engineer knowledgeable in electrical system design and
the local electrical code should review the bonding requirements for each commercial application
to determine if additional bonding and/or larger conductor sizes are required.
CSST installed inside or attached to the exterior of a building or structure shall be electrically
continuous and directly bonded to an effective ground-fault current path. The gas piping system
shall be considered to be directly bonded when installed in accordance with the following:
•
A bonding jumper is permanently and directly connected to the electrical service
grounding system. This shall be achieved through a connection to the electrical service
equipment enclosure, the grounded conductor at the electrical service, the grounding
electrode conductor (where of sufficient size) or to the one or more grounding electrodes
used. (NEC Section 250.104 B)
•
A single bond connection shall be made to the building gas piping downstream of the
utility meter, but near the service entrance (either outdoors or indoors) of the structure, or
downstream of the gas meter of each individual housing unit within a multi-family
structure. A “daisy chain” configuration of the bonding conductor shall be permitted for
multi-meter installations. A bonding connection shall not be made to the underground,
natural gas utility service line.
•
The bonding conductor shall not be smaller than a 6 AWG copper wire. The bonding
conductor shall be installed and protected in accordance with the NEC Section 250.64.
•
The bonding conductor shall be attached in an approved manner in accordance with NEC
Article 250.70 and the point of attachment for the bonding conductor shall be accessible.
•
Bonding/grounding clamps shall be listed to UL 467 or other acceptable national
standards.
© Kraft/Torbin 2007
16
•
The bonding clamp shall be attached at only one point within the piping system to either
the CSST brass fitting (if so listed) or to a rigid steel pipe component located between the
meter and the first downstream CSST fitting. See Figures 1, 2 and 3 for guidance. The
corrugated stainless steel tubing portion of the gas piping system shall not be used as the
point of attachment of the bonding clamp at any location along its length. The bonding
clamp shall be attached such that metal to metal contact is achieved with the steel pipe
component or CSST fitting.
The National Electrical Code recommends the bonding of all metallic piping systems and
metallic air ducts within the premise as a means for providing additional safety. Avoiding direct
contact between these metallic pathways has also been shown to reduce the possibility for arcing
between systems with different levels of electric potential.
Bond wire
Bond wire
Clamp
Clamp
Steel pipe
Figure 1: Bonding Attachment to Pipe
© Kraft/Torbin 2007
CSST fitting
Figure2: Bonding Attachment to CSST Fitting
17
Bond wire
Clamp
Manifold
Figure 3: Bonding Attachment to Uncoated (Customer-Owned) Manifold
© Kraft/Torbin 2007
18
APPENDIX B
NEC REFERENCES FOR THE USE OF 6 AWG COPPER BOND JUMPER
NFPA 70-2005 Edition of the National Electrical Code
250.53: Grounding Electrode System Installation.
(C) Bonding Jumper. The bonding jumper(s) used to connect the grounding electrodes together
to form the grounding electrode system shall be installed in accordance with 250.64(A), (B), and
(E) shall be sized in accordance with 250.66, and shall be connected in the manner specified in
250.70.
250.66: Size of Alternating-Current Grounding Electrode Conductor. The size of the
grounding electrode conductor of a grounded or ungrounded ac system shall not be less than
given in Table 250.66, except as permitted in 250.66(A) through (C).
(A) Connections to Rod, Pipe, or Plate Electrodes. Where the grounding electrode conductor
is connected to rod, pipe, or plate electrodes as permitted in 250.52(A)(5) or (A)(6), that portion
of the conductor that is the sole connection to the grounding electrode shall not be required to be
larger than 6 AWG copper wire or 4 AWG aluminum wire.
Table 250.66: Grounding Electrode Conductor for A-C Systems:
Size of Service Entrance Conductor:
2 or smaller => 8 AWG Copper
1 or 1/0
=> 6 AWG Copper
2/0 or 3/0
=> 4 AWG Copper
250.64: Grounding Electrode Conductor Installation. Grounding Electrode Conductors shall
be installed as specified in 250.64(A) through (F).
250.64(B): Securing and Protection Against Physical Damage. When exposed, a grounding
electrode conductor or its enclosure shall be securely fastened to the surface on which it is
carried. A 4 AWG or larger copper or aluminum grounding electrode conductor shall be
protected where exposed to physical damage. A 6 AWG grounding electrode conductor that is
free from exposure to physical damage shall be permitted to be run along the surface of the
building construction without metal covering or protection where it is securely fastened to the
construction; otherwise it shall be in rigid metal conduit, intermediate metal conduit, rigid
nonmetallic conduit, electrical metallic tubing, or cable armor. Grounding electrode conductors
smaller than 6 AWG shall be in rigid metal conduit, intermediate metal conduit, rigid
nonmetallic conduit, electrical metallic tubing, or cable armor.
© Kraft/Torbin 2007
19
250.104 Bonding of Piping Systems and Exposed Structural Steel
(A) Metal Water Piping. The metal water piping system shall be bonded as required in (A)(1),
(A)(2), or (A)(3) of this section. The bonding jumper(s) shall be installed in accordance with
250.64(A), (B), and (E). The points of attachment of the bonding jumper(s) shall be accessible.
(1) General. Metal water piping systems(s) installed in or attached to a building or structure
shall be bonded to the service equipment enclosure, the grounded conductor at the service, the
grounding electrode conductor where of sufficient size, or to the one or more grounding
electrodes used. The bonding jumper(s) shall be sized in accordance with Table 250.66 except
as permitted in 250.104(A)(2) and (A)(3). {(A)(2): Buildings of Multiple Occupancy; (A)(3):
Multiple Buildings or Structures Supplied by a Feeder(s) or Branch Circuit(s).}
(B) Other Metal Piping. Where installed in or attached to a building or structure, metal piping
system(s), including gas piping, that is likely to become energized shall be bonded to the service
equipment enclosure, the grounded conductor at the service, the grounding electrode conductor
where of sufficient size, or to one or more grounding electrodes used. The bonding jumper(s)
shall be sized in accordance with 250.122, using the rating of the circuit that is likely to energize
the piping system(s). The equipment grounding conductor for the circuit that is likely to
energize the piping shall be permitted to serve as the bonding means. The points of attachment
of the bonding jumper(s) shall be accessible.
FPN: Bonding all piping and metal air ducts within the premises will provide additional safety.
250.122: Sizing of Equipment Grounding Conductors.
(A) General. Copper, aluminum, or copper-clad aluminum equipment grounding conductors of
the wire type shall not be smaller than shown in Table 250.122 but shall not be required to be
larger than the circuit conductors supplying the equipment.
Table 250.122: Minimum Size Equipment Grounding Conductors for Grounding
Raceways and Equipment:
100 AMP => 8 AWG Copper
200 AMP => 6 AWG Copper
250.120: Equipment Grounding Conductor Installation. An equipment grounding conductor
shall be installed in accordance with 250.120(A), (B) and (C).
(C) Equipment Grounding Conductors Smaller Than 6 AWG. Equipment Grounding
Conductors Smaller Than 6 AWG shall be protected from physical damage by a raceway or
cable armor except where run in hollow spaces of walls or partitions, where not subject to
physical damage, or where protected from physical damage.
© Kraft/Torbin 2007
20
800.100: Cable and Primary Protector Grounding (Communications)
(D) Bonding Electrodes. A bonding jumper not smaller than 6 AWG copper or equivalent shall
be connected between the communications grounding electrode and power grounding electrode
system at the building or structure served where separate electrodes are used.
810.21: Grounding Conductors – Receiving Stations (Radio and Television Equipment)
(J) Bonding of Electrodes. A bonding jumper not smaller than 6 AWG copper or equivalent
shall be connected between the radio and television equipment grounding electrode and power
grounding electrode system at the building or structure served where separate electrodes are
used.
820.100: Cable Grounding (Community Antenna Television and Radio Distribution
Systems)
(D) Bonding of Electrodes. A bonding jumper not smaller than 6 AWG copper or equivalent
shall be connected between the community antenna television system’s grounding electrode and
power grounding electrode system at the building or structure served where separate electrodes
are used.
830.100: Cable, Network Interface Unit, and Primary Protector Grounding (NetworkPowered Broadband Communications Systems)
(D) Bonding of Electrodes. A bonding jumper not smaller than 6 AWG copper or equivalent
shall be connected between the network-powered broadband communications system grounding
electrode and power grounding electrode system at the building or structure served where
separate electrodes are used.
NFPA 780-2004 Standard for the Installation of Lightning Protection Systems
4.19.2.2: Conductors used for the bonding of grounded metal bodies or isolated metal bodies
requiring connection to the lightning protection system shall be sized in accordance with bonding
conductor requirements in Table 4.1.1.1 (A) and Table 4.1.1.1 (B).
4.1.1.1 (A): Ordinary structures not exceeding 23-m (75ft) in height shall be protected with
Class I materials as shown in Table 4.1.1.1 (A).
Table 4.1.1.1 (A) Minimum Class I Material Requirements: Bonding Conductor, cable (solid
or stranded): Cross section area = 26,240 cir mils.
NOTE: 6 AWG wire has a cross sectional area of 26,251 cir mils.
© Kraft/Torbin 2007
21
APPENDIX C
TEST RESULTS FROM LABORATORY PROGRAM
No Bond
Generator
CSST
Charge Level
Current
Charge
Voltage
Current
Charge
(kV)
(kA)
(C)
(kV)
(kA)
(C)
1
0.28
NM
0.47
No Arc
2
0.56
NM
0.86
No Arc
3
0.86
NM
1.40
No Arc
4
1.15
NM
1.80
No Arc
5
1.30
0.78
2.35
No Arc
5
1.30
0.78
2.35
No Arc
5
1.30
0.78
2.35
No Arc
6
1.60
0.98
2.80
No Arc
6
1.60
1.00
2.80
No Arc
32
6
1.64
1.04
2.80
No Arc
37
6
1.64
1.04
2.85
No Arc
43
6
1.64
1.04
2.85
No Arc
48
6
1.65
1.15
2.85
No Arc
13
6
1.70
NM
2.80
No Arc
54
7
1.85
1.15
3.25
No Arc
62
7
1.85
1.15
3.25
No Arc
33
7
1.88
1.20
3.30
No Arc
36
7
1.88
1.20
3.00
3.25
1.30
38
7
1.88
1.20
3.30
No Arc
44
7
1.88
1.20
3.30
No Arc
50
7
1.95
1.30
3.25
No Arc
55
8
2.15
1.35
3.75
No Arc
63
8
2.15
1.35
3.75
No Arc
34
8
2.18
1.40
3.80
No Arc
35
8
2.18
1.40
3.80
No Arc
39
8
2.18
1.40
3.75
No Arc
51
8
2.25
1.50
3.75
No Arc
56
9
2.45
1.55
4.20
No Arc
40
9
2.45
1.55
4.25
No Arc
42
6
2.80
1.22
2.35
2.70
1.10
47
6
2.85
1.20
2.85
2.70
1.15
46
6
2.95
1.25
2.85
NM
NM
14
8
3.85
1.65
3.75
2.30
31
8
3.85
1.65
2.45
3.70
1.50
45
8
3.85
1.65
3.55
3.60
1.50
52
9
4.25
1.80
4.10
4.20
NM
53
9
4.35
1.70
3.50
4.20
1.70
64
9
4.35
1.75
3.95
4.25
1.75
57
10
4.85
1.90
4.60
4.65
1.95
41
10
4.85
1.90
4.75
4.45
1.85
©Titeflex Corporation 2007 (reproduced herein by consent of Titeflex Corporation)
Test
No.
9
10
11
12
58
59
60
61
© Kraft/Torbin 2007
22
Direct-Bond: 20 feet of 8 AWG
Generator
CSST
Test
Charge Level
Current
Charge
Voltage
Current
Charge
No.
(kV)
(kA)
(C)
(kV)
(kA)
(C)
80
10
5.15
2.10
4.65
2.30
0.25
81
12
6.15
2.50
4.75
2.75
0.28
82
14
7.25
2.95
5.70
3.30
0.37
83
16
8.20
3.30
6.30
3.75
0.40
85
18
9.25
3.35
6.95
4.20
0.45
86
18
9.25
3.35
6.95
4.20
0.43
84
18
9.25
3.40
5.85
4.25
0.49
©Titeflex Corporation 2007 (reproduced herein by consent of Titeflex Corporation)
Direct-Bond: 40 feet of 8 AWG
Generator
CSST
Test
Charge Level
Current
Charge
Voltage
Current
Charge
No.
(kV)
(kA)
(C)
(kV)
(kA)
(C)
87
10
5.1
2.1
4.9
3.1
0.37
89
12
6.1
2.5
4.7
3.65
0.45
90
14
7.15
2.9
6.85
4.35
0.58
91
16
8.1
3.15
6.85
4.95
0.64
94
18
9.1
3.45
7.35
5.6
0.73
93
18
9.15
3.45
7.8
5.6
0.73
92
18
10
3.35
7.8
5.6
0.77
©Titeflex Corporation 2007 (reproduced herein by consent of Titeflex Corporation)
Direct-Bond: 10 feet of 6 AWG
Test
No.
Charge Level
(kV)
Generator
Current
(kA)
Charge
(C)
Voltage
(kV)
CSST
Current
(kA)
Charge
(C)
2
4
2.00
0.75
1.30
No Arc
3
4
2.10
0.70
1.70
No Arc
4
5.75
3.00
1.00
2.30
No Arc
5
8
4.30
1.40
3.40
No Arc
6
10
5.30
1.80
3.30
No Arc
7
14
7.25
2.40
4.75
No Arc
8
18
9.50
3.70
5.25
No Arc
©Titeflex Corporation 2007 (reproduced herein by consent of Titeflex Corporation)
© Kraft/Torbin 2007
23
Direct-Bond: 20 feet of 6 AWG
Generator
CSST
Test
Charge Level
Current
Charge
Voltage
Current
Charge
No.
(kV)
(kA)
(C)
(kV)
(kA)
(C)
22
6
3.05
1.30
4.25
No Arc
28
6
3.15
1.20
4.00
No Arc
27
6
3.20
1.20
4.05
No Arc
25
7
3.65
1.40
4.50
1.40
0.09
26
7
3.65
1.45
4.15
1.40
0.08
24
7
3.70
1.50
4.25
1.45
0.10
23
8
4.20
1.70
5.35
1.65
0.12
21
10
5.30
2.15
5.80
2.05
0.15
20
12
6.30
2.55
7.05
2.50
0.21
17
14
7.20
2.85
7.30
3.05
0.25
19
14
7.25
2.95
7.25
3.05
0.25
16
14
7.25
NM
6.25
NM
NM
18
18
9.15
3.80
8.20
3.95
0.37
©Titeflex Corporation 2007 (reproduced herein by consent of Titeflex Corporation)
Direct-Bond: 40 feet of 6 AWG
Generator
CSST
Charge Level
Current
Charge
Voltage
Current
Charge
(kV)
(kA)
(C)
(kV)
(kA)
(C)
4
1.95
0.70
3.40
No Arc
3
2.15
0.61
2.75
1.15
0.10
3
2.15
0.62
2.90
No Arc
4
3.00
0.85
3.25
1.65
0.16
4
3.00
0.85
2.40
1.65
0.17
5
3.70
1.05
3.45
2.15
0.20
5
3.75
1.05
3.25
2.15
0.20
6
4.50
1.60
5.20
2.35
0.25
8
5.95
1.65
4.45
3.45
0.35
10
7.50
NM
4.50
4.40
0.47
12
8.80
2.30
5.70
5.20
0.55
14
10.45
2.85
6.95
6.10
0.67
16
11.80
3.10
6.30
6.90
0.76
18
13.30
3.20
8.50
7.80
0.88
Change of Generator Output viz. Waveform
79
18
9.15
3.30
8.10
5.10
0.60
©Titeflex Corporation 2007 (reproduced herein by consent of Titeflex Corporation)
Test
No.
65
71
72
69
70
68
67
66
73
74
75
76
77
78
© Kraft/Torbin 2007
24