Sie sind auf Seite 1von 6

IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 15, NO.

2, APRIL 2000 635

IEC Method of Calculation of Minimum Approach


Distances for Live Working
George Gela, Senior Member, IEEE, Paul W. Hotte, Member, IEEE, and Marc Charest, Member, IEEE

Abstract—After 10 years and as many drafts, proposed IEC which could be well-justified based on stand-alone IEC publica-
standard 61472 is now in the final voting stage. The standard, when tions and experience. In addition, while some regional standards
published, will define the IEC method of calculating Minimum Ap- do exist, IEC/TC78 “Live working” found that an international
proach Distances (MAD) for live working for use world-wide.
The standard 61472 provides a method of calculating MAD for standard was needed also for its own future work. This standard
voltages between 1 kV and 800 kV AC. In common with other stan- should to be able to provide generally conservative MAD values
to assure worker’s safety, while at the same time allowing the
D D
dards, the MAD is taken to consist of two components: an electrical
component, U , and an ergonomic component, E . The compo-
D
nent U is calculated to minimize the risk of breakdown for all
more experienced users to employ their knowledge base to relax
voltages expected at the worksite. The rationale for calculating U D some of the conservative criteria and assumptions to suit partic-
is based on IEC Publications 71-1 and 71-2. The component E D ular individual needs.
The standard 61 472 provides a method of calculating MAD
depends on work procedures, training, workers’ skill, type of con-
for system voltages 1 kV to 800 kV AC. In common with some
D
struction, contingencies such as inadvertent movement and errors
in judging distances. Guidelines for selection of E are included regional standards, the MAD consists of two components: an
in IEC 61472, but without giving specific values. Atmospheric con- D D
electrical distance, U , and an ergonomic distance, E . The
ditions and specific details of the worksite (such as the effects of
D
electrical distance, U , is derived from the sparkover charac-
broken insulators and electrically floating conductive objects) are
accounted for through the application of correction factors. Guide- D
teristics of the worksite air gap. The value of U is selected in
lines for selecting the values for these factors are included. order to minimize the risk of breakdown for all voltages likely
This paper provides an overview of the IEC method of calcu- D
to occur at the worksite. The ergonomic distance, E , is used
lating MAD, its background and theoretical foundations, and gives
comparisons with the IEEE/ESMOL method, which is defined in
D D
in addition to the electrical distance, U . The distance E de-
IEEE Std 516-1995. The paper concentrates on systems voltage pends on work procedures, training, workers’ skill, type of con-
levels above 72.5 kV, and on phase-to-ground distances. struction, and contingencies such as inadvertent movement and
errors in judging distances.
Index Terms—Live working, minimum approach distance, min-
imum clearance, air gaps, dielectric strength, withstand voltage. The rationale for calculating MAD is primarily based on IEC
71-1 and 71-2 [2] and on CIGRÉ Publication 72 [3]. Conserva-
tive values are suggested for various parameters in order to as-
I. INTRODUCTION sure that the calculated MAD values are safe in situations where
FTER 10 years and as many drafts, proposed IEC standard access to more precise data is not available.
A 61 472 is now in the final voting stage [1]. Reaching a con-
sensus during the development stages was expected to be diffi-
The IEEE method, defined in IEEE Standard 516-1995 [4],
is based on an experimentally derived relationship between the
cult. Power companies world-wide that do live work have their U l
50% sparkover voltage, 50 , and the length, , of the air gap.
own well-established minimum approach distances (MAD), but Similarly to the IEC method, an ergonomic distance, E , is D
those distances often vary considerably, even for similar voltage added to the electrical component to obtain the total MAD. The
levels. In Canada alone, various power utility companies in the IEEE method is very simple to use and is supported by several
10 Canadian Provinces have adopted different MAD values. decades of successful field experience.
The methods used by power companies to calculate MAD This paper concentrates on system voltages above 72.5 kV
also vary throughout the world, and they are frequently based and phase-to-ground electrical distance calculations.
on a judicial blend of years of experience, traditional approaches
and strictly scientific principles. II. EXPECTED WORKSITE SWITCHING OVERVOLTAGES
With such a heuristic approach to this very complex and cru- (STRESSES)—IEC METHOD
cial issue of determination of MAD to assure safety of workers, When determining the required length of a worksite air gap
countries which are just starting live working and power utility which will avoid a disruptive discharge (sparkover), the critical
companies which want to review their MAD found it very diffi- factor is expected stress at the worksite, i.e., the magnitude of
cult to adopt the available procedures without a critical review. expected switching overvoltages likely to occur across the air
Hence, a need arose for the development of rational guidelines gap. In live working, the magnitude of such overvoltages can
be reduced at the worksite by the common practice of either
Manuscript received May 19, 1998. preventing reclosure during work or by using portable protective
G. Gela is with EPRI Power Delivery Center, Lenox, MA, USA. gaps. The effect of lightning can be minimized by disallowing
P. W. Hotte is a Consultant in Toronto, Canada.
M. Charest is with AMB Charest Consultants, Inc., Montreal, Canada. work when lighting or thunder is discernible within a distance
Publisher Item Identifier S 0885-8977(00)03484-1. of about 10 km.
0885–8977/00$10.00 © 2000 IEEE
636 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 15, NO. 2, APRIL 2000

TABLE I terms of U50RP of the rod-to-plane gap. The 90% statistical


TYPICAL u2 VALUES, PHASE-TO-GROUND
withstand voltage, U90 , of the worksite gap is related to U50 of
this gap in analogy to (3). Thus, U90 can be expressed in terms
of the known U50RP . Combining (2) and (3), one obtains the
desired expression for U90 of the worksite gap in (4):
U90 = ks 2 U50 = ks 2 kg 2 a2 f 2 i2
k k k RP
U50

=K t2 U50 RP (4)
where
The 2% statistical overvoltage, U2 , that is likely to occur at the U90 = statistical withstand voltage of the worksite gap
worksite (i.e., the electrical stress across the worksite air gap) is U50 = 50% sparkover voltage of the worksite gap
determined from the highest system voltage, Us , in the system K t = ks kg ka kf ki
to be worked on and the per unit overvoltage, u2, as in (1). The k s = conventional deviation factor with the value ks =
phase-to-ground values of U2 can be taken from Table I or they 0:936 for the typical suggested normalized conven-
can be determined from studies.
p p tional deviation of 5%
U2 =( 2= 3) 2 Us 2 U2 (1) k g = gap factor; suggested conservative value of 1.2, un-
less a more precise value can be chosen
If induction is assumed to be the only mechanism for ener- k a = atmospheric factor, to allow for a range of elevations
gization of the equipment being worked, the estimated induced within a service area
voltage is used. k f = factor to allow for the effects (in addition to their
reduction of the air gap length) of electrically floating
III. DIELECTRIC STRENGTH OF A ROD-TO-PLANE AIR objects in the worksite gap, with a suggested values of
GAP—IEC METHOD 0.85 for objects that are stationary, moving, or either:
The switching impulse (SI) dielectric strength of a rod-to-
plane air gap is given by the formula (2) developed by CRIEPI
[5], which relates the 50% sparkover voltage, U50RP , of the
rod-to-plane air gap to the gap length, d:
otherwise kf = 1
U50 RP = 1080 2 ln(0 46: d + 1) (2) k i = factor to be applied to determine the required min-
where: U50RP = the 50% sparkover voltage of the air gap, kV imum length of a defective insulator string near which
and d = the length of the air gap, m live work is to be done:
Eq. (2) applies to positive polarity Sl and assumes that the SI
has the critical front time. Eq. (2) is now recommended by IEC
k i =1 008 d d
: k n =n0 (5)
[2] for general use since it covers a larger voltage and distance where:
range than previous formulae. n d = the number of defective insulators
The concept of the 50% sparkover voltage, U50RP , of the n0 = the number of sound insulators
rod-to-plane air gap means that a positive polarity SI with peak k d = 1 for glass and 0.75 (average) for porcelain
value equal to U50RP (and with the critical front time), when
applied 100 times to the gap, will result in 50 break-downs and V. REQUIRED WORKSITE WITHSTAND VOLTAGE
50 withstands. An SI with a lower peak value will result in less (STRENGTH)—IEC METHOD
than 50 breakdowns (and more than 50 withstands) of the gap in The 90% statistical withstand voltage of the worksite air gap,
100 consecutive applications. In particular, the statistical with- U90, is calculated from U2 in accordance with (3):
stand voltage, U90RP , of the gap is defined as that peak value of
the Sl which will result in 10 breakdowns (and 90 withstands) in U90 = s 2 U2
K (6)
100 applications. Assuming Gaussian distribution of sparkover where: Ks = statistical safety factor with the suggested value of
voltages, the relation between U90RP and U50RP is: 1.1
U90RP = s 2 U50RP
k (3) The 90% statistical withstand also means that the probability
of breakdown is 10% when a switching impulse with peak value
where: ks = conventional deviation factor. of Ks 2 U2 is applied. It should be recalled that U2 means that
only 2% of the overvoltages actually occurring at the worksite
IV. DIELECTRIC STRENGTH OF WORKSITE AIR GAPS—IEC exceed the peak value equal to U2 . Hence, the resulting proba-
METHOD bility of worksite sparkover is small, but this probability is not
Conditions of a typical air gap at the worksite differ in many 0%. While the nonzero probability of sparkover can still be con-
respects from those of the simple rod-to-plane air gap. These sidered to be unacceptable from the worker’s safety viewpoint,
differences are treated in IEC 61 472 by introducing a set of the following mitigating factors combine to reduce the overall
multiplying factors. Values for these factors are suggested in risk of worksite breakdown to an acceptable level:
IEC 61 472. This allows obtaining U50 of the work-site gap in • The actual system voltage is not always maximal
GELA et al.: IEC METHOD OF CALCULATION OF MAD FOR LIVE WORKING 637

TABLE II
IEC ATMOSPHERIC FACTOR, ka , FOR DIFFERENT ALTITUDES AND VALUES OF U90

• The location of the worksite is unlikely to coincide with Generally, a constant value between 0.1 m and 0.2 m is used
that of the maximum overvoltage due to switching for DE (for voltage above 72.5 kV), but the value of 1 ft (0.31 m)
• The stress due to the switching overvoltage with the actual is common in North America.
wave front is less than that due to the overvoltage with the It should be observed that the overall risk of a worksite break-
critical front used to derive (2) down is in general lower when the additional ergonomic dis-
• About half the switching overvoltages are of the less se- tance DE is used, because an overvoltage is unlikely to arise at
vere negative polarity the worksite at the instant when the ergonomic distance is en-
• The frequency and amplitude of switching overvoltages tirely breached by inadvertent movement of the worker. In view
are reduced by reclosure blocking of this mitigating fact, IEC 61 472 allows the factor Ks to be re-
The resulting risk of worksite sparkover is not greater than the duced to 1.0 when a defined ergonomic distance DE is included
risk encountered by an average person performing other typical and DE is large enough that the total value of DA is always
industrial work. greater than the value calculated with Ks = 1:1, i.e., when the
Combining (1), (2), and (4) through (6), the withstand condition in (11) is satisfied:
voltage, U90 , of the worksite characterized by an air gap of
length, d, is given by (7): DU (Ks =1:0) + DE > DU (Ks =1:1) (11)

U90 = Kt 2 1080 2 ln(0 46


: d + 1) (7)
VII. CORRECTION FOR ALTITUDE—IEC METHOD

VI. MINIMUM APPROACH DISTANCE (MAD)—IEC METHOD IEC 61 472 incorporates the most recent information on cal-
culation of the atmospheric correction factor, ka , that is cur-
Rearranging (7) to express d in terms of U90 , and interpreting rently available in and accepted by IEC. The values of ka are
the gap length d as the electrical component, DU , of the Min- calculated using the method of IEC 60-1 and the altitude cor-
imum Approach Distance (MAD) for live working, one obtains rection for pressure given in IEC 71-2 which assumes standard
the expression for DU as in (8): temperature and humidity values. The ka values are given in
DU
U =(1080Kt )
= 2:17(e 90 0 1) (8)
Table II as function of altitude and the U90 value at standard
temperature, pressure and humidity (STPH).
Eq. (8) applies to cases without electrically floating conduc- However, temperature and humidity vary with altitude, [3],
tive objects in the gap. If floating electrodes, such as tools and [6] and they are different from the standard values (20 C and
end fittings of tools, are present, (8) is modified to obtain the 11 g/m3 ) at different altitudes. This results in overly conserva-
general expression for DU in (9): tive values of ka .
It is hoped that the next edition of the IEC 61 472 standard
DU
U =(1080Kt)
= 2:17(e 90 0 1) + F (9) will include the results of further investigations on this topic.

where: F = total length (along gap axis) of all floating objects. VIII. SAMPLE CALCULATION OF DE —IEC METHOD
The Minimum Approach Distance, MAD or DA , is then:
The simplest example to demonstrate the IEC 61 472 method
DA = DU + DE (10) of calculating the electrical component, DE , of MAD is one
with Ks = 1:1, ka = 1:0 (sea level), kf = 1:0 and F = 0:0 (no
where: DE = the ergonomic distance floating objects), nd = 0:0 (no broken insulators). The values
Selection of the value of DE is left to the user in IEC 61 472. of the remaining parameters are: ks = 0:936 (for the assumed
Three approaches are common: conventional deviation of 5%), kg = 1:2 (the recommended
• DE independent of the system voltage (DE constant) value), and ki = 1:0. In this case:
• DE increases with system voltage
• DE decreases as system voltage increases Kt = 0:936 2 1 2 2 1 0 2 1 0 2 1 0 = 1 12
: : : : : (12)
638 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 15, NO. 2, APRIL 2000

TABLE III TABLE IV


D
EXAMPLES OF CALCULATED U VALUES FOR VARIOUS CASES IEEE ALTITUDE CORRECTION FACTORS

For Us = 550 kV and U2 = 2:3 from Table I, U2 = 1033 kV


from (1), and U90 = 1136 kV from (6). Using the above values in
(9), the value of DU is calculated to be DU = 3.38 m or 11.9 ft. The relation between the truncation value, ut = T , and the 2%
Table III summarized the DU values for additional examples for value, u2 is given in (15):
a range of parameter values with Ks = 1:1, ks = 0:936, kg 1:2,
ka = 1:0 (i.e., H = 0 m) and kd = 0:75. u2 = (ut + 0:25)=1:25 (15)

For valid comparison between the IEC and the IEEE methods,
IX. MINIMUM APPROACH DISTANCE—IEEE METHOD ut should be used in the IEEE method and (15) should be used
The calculation of the approach distance, D , used by the IEEE to compute the corresponding value of u2 , for use in the IEC
method is defined in IEEE Std 516-1995 [4]. The calculation method. Of course, the converse is also possible, i.e., one can
employs an experimentally determined curve relating the air gap select the u2 value in the IEC method and then calculate the
distance and the withstand voltage for a selection of air gaps, corresponding value of ut for the IEEE method.
which was determined by tests in 13 laboratories over 30 years
ago. Details of the exact nature of the tested gaps and test con- X. CORRECTION FOR ALTITUDE—IEEE METHOD
ditions have, however, been lost with time. Adjustment for altitude is in the form of multiplying factors
The approach distance (or the insulation distance), D , de- for the calculated distance for locations above 900 m (3000 ft.)
pends on the actual rms system phase-to-ground voltage and the in altitude. These multiplying factors are given in Table IV.
maximum anticipated per-unit switching overvoltage factor, T .
The value of D can be calculated from (13), which is designed XI. SAMPLE CALCULATION OF D —IEEE METHOD
to fit the experimental curve mentioned earlier:
The IEEE method is quite easy to use for simple cases such as
D = (C1 2 C2 + a) 2 2
T kV LG (13) those without floating objects and broken insulators. For the ex-
ample considered earlier (sea level, no broken insulators, Us =
where: 550 kV, i.e., kVLG = 320 kV), the saturation factor. a =
D = insulation distance, ft. 2:9=1000 and truncation value T is calculated by rearranging
C1 = 0.01, or 1% of the line-to-ground system voltage (15): T = 2:63. Then, using (13), D is:
= 1.0 for cases without toolspin the air gap
C2

a = saturation factor for crest 2 2 T 2 kVLG voltages


D = (0:01 2 1 0 + 0 0029) 2 2 63 2 320 = 10 86 ft or 3 31 m
: : : : :

(16)
of 630 kV and above
The above value of D in (16), calculated using the IEEE
T = maximum anticipated per-unit switching over-
method, is about 2% smaller than that in (12) which was cal-
voltage, truncation value
culated using the IEC 61 472 method. Table V shows the com-
kVLG = rms system phase-to-ground voltage—actual
plete set of values of D calculated with the IEEE method for all
The factor “a” can be approximated to within 2% by (14):
voltages above 72.5 kV and for a range of overvoltage factors,
2 LC 0 4:75) 2 1003
at sea level.
a = (0:0075 kV (14)

The meaning of the overvoltage factor T in (13) needs to be XII. COMPARISON OF IEC AND IEEE METHODS
considered carefully. The IEC method uses the 2% statistical Since the IEEE method does not take into account directly
per unit factor, u2 . In the IEEE method, however, the factor T the effects of broken insulators and floating objects, it is not
is not the same as u2 , but it is interpreted instead as the max- possible to develop a table of IEEE calculated distances that
imum anticipated overvoltage, i.e., the truncation value of the would correspond exactly to all cases in Table III. Still, it is
overvoltage distribution beyond which no overvoltages occur. possible to compare the results of the two methods for base
GELA et al.: IEC METHOD OF CALCULATION OF MAD FOR LIVE WORKING 639

TABLE V
IEEE MINIMUM AIR INSULATION PHASE-TO-GROUND DISTANCE, AC ENERGIZED WORK

TABLE VI TABLE VII


VALUES OF DU (IEC) AND D (IEEE) FOR H = 0 m VALUES OF DU (IEC) AND D (IEEE) FOR VARIOUS H VALUES

parameters in Table VII are the same as in Table VI. The IEC
altitude correction factors can be taken from Table II, but, for
cases without broken insulators and without floating objects
(ki = 0:0; kf = 1:0 and F = 0 m in the IEC method). Also,
enhanced accuracy, the factors were calculated for the specific
U90 values in Table VII (the difference in the results is less than
the following parameter values are used in the IEC method for
4%). The IEEE altitude correction factors were taken from Table
this comparison: Ks = 1:1; ks = 0:936, and kg = 1:2. The
IV.
ut values were taken from [4] and the corresponding u2 values
The ratios of IEC to IEEE values are consistently greater than
were calculated from (15). The DU and D values calculated for
1.00 for all cases shown in Table VII. In several cases, the dif-
these bases cases with the IEC and the IEEE methods, respec-
ference between the IEC and IEEE values exceeds 10%. At the
tively, are compared in Table V for altitude H = 0 m.
same time, however, a similar trend is apparent as in Table V,
The bottom row in Table VI gives the ratios of the MAD
i.e., agreement between the two methods tends to improve for
values calculated using the IEC method (DU ) and the IEEE
higher values of u2 .
method (D ). Analysis of these ratios indicates good overall
agreement (within about 10%) between the two methods. Also,
in general it appears that better agreement is obtained at higher XIII. DISCUSSION OF CALCULATION METHODS OF DU AND D
overvoltage factors (u2 ).
Table VII contains the calculated values of DU and D for the For cases where meaningful numerical comparisons of the
same cases as in Table VI, but for various altitudes. All other two methods can be performed, the results are comparable. It
640 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 15, NO. 2, APRIL 2000

appears, however, that the new IEC method with Ks = 1:1 rec- [4] IEEE Guide for Maintenance Methods on Energized Power Lines, IEEE
ommends somewhat larger electrical distance values than the Std 516-1995.
[5] I. Kishizima, K. Matsumoto, and Y. Watanabe, “New facilities for
IEEE method. Also, the two methods tend to agree better at phase-to-phase switching impulse tests and some test results,” IEEE
higher overvoltage levels at the worksite (higher u2 ). Thus, the Transactions on Power Apparatus and Systems, vol. PAS-103, no. 6,
electrical distance values in the regions of greatest interest are pp. 1211–1216, 1984.
[6] Paul W. Hotte, “Atmospheric and surge wavefront effects on limits of ap-
confirmed by two independent methods of computing the dis- proach,” Canadian Electrical Association, Report 330 T 673, December
tances. This provides the needed reassurance that the previously 1992.
used IEEE values are adequate, which has also been supported [7] IEC, Publication 71-2, “Insulation Co-ordination Part 2: Application
Guide,”, 2nd ed., 1976.
by several decades of field experience. At the same time, it ap-
pears that the simple IEEE altitude correction method should be
revised.
It should also be mentioned that in situations in which George Gela. See IEEE Transactions on Power Delivery, Vol. 12, no. 3, July
sufficient physical distances at the worksite are not available 1997, p. 1193 for the biography of Dr. George Gela.
to meet the calculated requirements, external temporary means
of reducing the worksite overvoltages (such as use of portable
protective gaps) can be employed. These overvoltage reduction
methods are beyond the scope of this paper, however. Paul W. Hotte (M’75) received his B.E. (Elect.) from the University of Canter-
bury, New Zealand, in 1964 and his M.Eng. from the Department of Electrical
Engineering and Institute of Biomedical Engineering, University of Toronto, in
XIV. CONCLUSIONS 1982. He worked for firms that are now part of ABB, Joslyn and AGRA Indus-
tries, and for Ontario Hydro before going into private practice in 1982. He spe-
The analyzes and comparisons presented in this paper con- cializes in HV transmission, electric field effects, live line technology, environ-
firm that the IEC and IEEE methods of calculating Minimum mental and health issues and electrical safety. For 20 years, he has participated
Approach Distances in general, and the electrical distance in in the work of IEC/TC 78 - Live Working, the IEEE ESMOL Subcommittee,
and the Canadian Electrical Association. He is a Designated Consulting Engi-
particular, for live working application agree quite well. Some neer in the Province of Ontario.
discrepancies between the two methods are identified and
some shortcomings of the two methods are described. The IEC
method is quite powerful and flexible, and may be applied to a
wide range of worksite situations. The IEEE method is simple Marc Charest (S’55) obtained B.A. Sc. degree in 19S5 in Electrical Engi-
and supported by several decades of experience. In general, neering from the University of Toronto. He joined the Shawinigan Water and
however, the two methods produce equivalent results for base Power as a trainee for one year and then became in turn Engineer, District Engi-
neer, District Superintendent in various districts all related with the maintenance
cases considered in this paper. of the Electrical Transmission System. He became after 9 years of field expe-
rience the Head of Maintenance Division at the Head Office of Hydro-Québec
and finally the Manager of Transmission Line Maintenance of Hydro-Québec
REFERENCES
Transmission Lines Department. Mr. Charest was Chairman of Transmission
[1] IEC, document 78/204/CDV, “Live working: Minimum approach Committee of CEA, the founding Chairman of IEEE/ESMOL Subcommittee
distances—Method of calculation,”, IEC document 78/ /RVC, August and finally the Secretary of IEC/TC 78 since 1976. Mr. Charest has authored
1996. and co-authored many IEC, Electra, and CIGRÉ papers and made many presen-
[2] IEC, Publication 71-1, “Insulation Co-ordination Part 1: Definitions, tations related to the Maintenance of Electrical Systems Dead and Live Line.
Principles and Rules,”, 7th ed., 1993. He contributed greatly in the development the ESMO Conferences namely in
[3] CIGRÉ, Working Group 33.07, “Guidelines for the Evaluation of the setting up the first ESMO Conference in 1977 in Montréal (Canada). He is a
Dielectric Strength of External Insulation,”, CIGRT Document 72, 1993. registered Professional Engineer in the Province of Québec (Canada).

Das könnte Ihnen auch gefallen