You are on page 1of 11

IEEE Transactions on Power Delivery, Vol. 3, No.

2, April 1988
LOW-VOLTAGE-SIDE CURRENT-SURGE PHENOMENA IN SINGLE-PHASE DISTRIBUTION TRANSFORMER SYSTEMS Roger C. Dugan Senior Member, IEEE Canonsburg, PA Stephen D. Smith Member, IEEE Zanesville, OH

63 7

McGraw-Edison Power Systems Division Cooper Industries, Inc. Abstract L i g h t n i n g c u r r e n t s u r g e s e n t e r i n g t h e secondary windings of distribution transformers can be a cause of t r a n s f o r m e r failure. Proposed solutions have included i n t e r l a c i n g t h e s e c o n d a r y w i n d i n g s and a p p l y i n g low-voltage arresters. T e s t s have been proposed t o verify t h e ability of a transformer t o withstand these surges. T h i s p a p e r s h o w s t h a t t h e amount of c u r r e n t varies s i g n i f i c a n t l y f o r d i f f e r e n t s i z e s a n d d e s i g n s of transformers, loads, and secondary cables. I t is also shown that t h e e n t i r e secondary c i r c u i t must b e treated as a s y s t e m . Measures t a k e n t o p r o t e c t t h e t r a n s f o r m e r generally increase t h e s u r g e voltage s t r e s s 011 t h e load equipment. The s o u r c e of t h e problem is t h e voltage drop along t h e secondary cable. Minimizing that voltage can e f f e c t i v e l y alleviate t h e problem a t both t h e trans- former and load. These f a c t s must be taken i n t o con- sideration b e f o r e developing transformer t e s t standards t o address t h e low-voltage-side current s u r g e problem. INTRODUCTION T h e coupling of lightning s u r g e c u r r e n t s d i r e c t l y i n t o t h e secondary c i r c u i t of a distribution transformer through t h e grounding leads has been recognized f o r many y e a r s [ 1 , 2 1 . More recently, r e p o r t s on t h e susceptib i l i t y of distribution transformers t o t h e s e surges have been published with the conclusion t h a t f a i l u r e s a s s o c i a t e d w i t h t h e s e s u r g e s can be minimized by i n t e r l a c i n g t h e s e c o n d a r y w i n d i n g s o r by a p p l y i n g low-vol t a g e a r r e s t e r s [ 3 , 4 1 . Subsequently, t e s t s have b e e n p r o p o s e d t o e v a l u a t e t h e s u s c e p t i b i l i t y of distribution transformers t o these surges. One proposed t e s t is t o i n j e c t a 5 k A c u r r e n t s u r g e i n t o t h e midpoint of t h e secondary of t h e distribution t r a n s f o r m e r , resulting in 2 . 5 k A in each half of t h e winding. After conducting several of t h e s e t e s t s , i t w a s o b s e r v e d t h a t t h e r e is a l a r g e d i f f e r e n c e in impedance between d i f f e r e n t designs of transformers. In p a r t i c u l a r , transformers with non-in t er lac ed secondaries o f f e r several times more impedance t o s u r g e c u r r e n t s t h a n d o t r a n s f o r m e r s with interlaced secondaries and transformers p r o t e c t e d by low-voltage arresters. Also, t h e r e were significant differences in impedances between d i f f e r e n t ratings of transformers. Because t h e higher impedance transformers required an unusually high s u r g e g e n e r a t o r voltage t o obtain a 5 k A c u r r e n t surge, i t was suspected t h a t they would see much l e s s s u r g e c u r r e n t i n normal service. Therefore, a study was undertaken t o analyze t h e typical distribution transformer system and d e t e r m i n e r e a s o n a b l e s u r g e c u r r e n t l e v e l s f o r each transformer. The r e s u l t s of t h i s analysis a r e presented in t h i s paper. T h e distribution transformer system is a Complex, tightly coupled system with respect t o low-voltage-side c u r r e n t surges. That is, changes in one component in t h e s y s t e m c a n have a s i g n i f i c a n t e f f e c t on t h e o t h e r components. T h e c i r c u i t must be considered as an i n t e r c o n n e c t e d s y s t e m o f e l e m e n t s i n which t h e In f a c t , t h e distribution transformer is only o n e part. s e c o n d a r y c a b l e is, perhaps, more influential in t h i s c i r c u i t than t h e transformer. Any proposed distribution transformer t e s t standard must recognize t h i s situation. T h e model f o r t h e simulation is presented f i r s t , f o l l o w e d b y a d i s c u s s i o n of t h e r e s u l t s of t h e simulations and their consequences.

Fig. 1.

Single-phase overhead transformer installation THE SYSTEM

The system under consideration is shown in Figure 1. Lightning s u r g e c u r r e n t s can e n t e r t h e secondary c i r c u i t through either strokes t o t h e primary l i n e o r strokes d i r e c t l y t o t h e secondary circuit. F o r strokes t o t h e primary, t h e c u r r e n t is conducted t o t h e secondary as shown in Figure 2.

LIGHTNING PRIMARY LINE

,I,
TRANSFORMER

xi

SECONDARY CABLE

LOAD

86 T&D 553-2 A p a p e r recommended a n d a p p r o v e d hy t h e I E E Z T r a n s f o r m e r s C o m m i t t e e o f t h e I E E F Power E n g i n e e r i n g S o c i e t v f o r p r e s e n t a t i o n at t h e IF:E?/PFS 1986 T r a n s m i s s i o n a n d D i s t r i h u t i o n C o n f e r e n c e , ,Anaheim, C a l . i f o r n i a , S e p t e m h e r 14 - 1 9 , 1936. ! l a n u s c r i p t s u h m i t t e d larch 2 1 , 1986; made a v a i l a h l e f o r p r i n t i n g June 18, 1986.

I
HOUSE GROUND

GROUND

1 -

Fig. 2. S u r g e c u r r e n t entering secondary c i r c u i t d u e to lightning s t r i k e t o primary. 0885-8977/88/0400-0637$01 .WO 1988 IEEE

638
The bulk o f t h e c u r r e n t is conducted through t h e s u r g e a r r e s t e r t o t h e junction of t h e pole ground lead, t h e X2 b u s h i n g on t h e t r a n s f o r m e r , and t h e n e u t r a l on t h e secondary cable. Some of t h e c u r r e n t is conducted on t h e primary line. The c u r r e n t reaching t h e junction of t h e g r o u n d s a t t h e t o p of t h e pole divides between t h e various paths according t o t h e impedances. As shown, some of t h e c u r r e n t will flow i n t o t h e transformer secondary and, also, i n t o t h e load. This is t h e c u r r e n t that causes the transformer voltages described by McMillen, et. al. [31. I t a l s o causes high v o l t a g e s t r e s s a c r o s s t h e load and, presumably, f a i l u r e s of secondary load equipment. However, t h e a u t h o r s a r e not aware of any work t h a t has been done t o c o r r e l a t e l i g h t n i n g f a i l u r e s of customer appliances with these s u r g e phenomena. Essentially t h e same phenomena o c c u r for s t r o k e s t o t h e secondary c i r c u i t as shown in F i g u r e 3. The exact c u r r e n t division is dependent on t h e s t r o k e location. Transformer Model T h r e e d i f f e r e n t s i z e s of t r a n s f o r m e r s w e r e simulated: 10, 25, and 50 kVA, a l l rated 7200/12470Y 120/240 volts. So t h a t t h e voltage distribution within t h e t r a n s f o r m e r c o u l d be o b s e r v e d , t h e transformer w i n d i n g s w e r e broken i n t o e i t h e r 9 o r 1 0 groups o f winding layers and represented as coupled inductors. The c a p a c i t a n c e s between t h e primary winding layers were A sensitivity analysis showed t h a t a l s o represented. t h e capacitances in t h e secondary had minimal e f f e c t on t h e waveforms f o r low-voltage-side c u r r e n t surges and these capacitances were neglected. Figure 5 shows t h e cross s e c t i o n of t h e c o n s t r u c t i o n of t h e t r a n s f o r m e r s and Figure 6 shows t h e t y p i c a l model schematic.

CORE

P * P R I M A R Y WINDING

S1 = SECONDARY WINDING 1
5 2 = SECONDARY WINDING 2

H 2 L - ---A

Fig. 5. Cross-section of s h e l l - t y p e distribution transformer c o n s t r u c t i o n .

HOUSE

GROUND

GROUN[

Fig. 3. Surge c u r r e n t entering secondary c i r c u i t due t o lightning s t r i k e t o secondary cable. I n t h i s a n a l y s i s , b o t h t y p e s of s t r o k e s w e r e simulated. However, most of t h e work was d o n e with t h e s t r o k e t o t h e primary and t h e r e s u l t s reported h e r e come from t h a t work. It is believed t h a t this is t h e more common way f o r low-voltage-side c u r r e n t surges t o occur. SYSTEM MODEL The model f o r t h e system is shown in F i g u r e 4. This s y s t e m was r e p r e s e n t e d on t h e METAP p r o g r a m , a g e n e r a l - p u r p o s e p r o g r a m f o r t h e s i m u l a t i o n of e l e c t r o m a g n e t i c t r a n s i e n t s [51. A description of t h e model of each element of t h e system follows.

x2'

Fig. 6. Model of transformer used in t h i s s t u d y ; non- i n t e r laced con nec t ion shown. Because of limited space, we will present only t h e 10-kVA model here. T h e problem is more pronounced with t h i s s i z e of t r a n s f o r m e r a n d t h e general principles apply t o t h e o t h e r sizes as well. The primary winding was b r o k e n i n t o 10 w i n d i n g g r o u p s . E a c h 1 2 0 - v o l t s e c o n d a r y winding was broken i n t o two p a r t s so t h a t interlacing could be represented simply by a l t e r i n g t h e connections of t h e terminals. B r i e f l y , t h e general procedure f o r developing t h e coupled inductance model is as follows:
1.

LIGHTNING CURRENT SOURCE ( 8 x 20)

SECONDARY

SURGE

Compute t h e s h o r t c i r c u i t (leakage) r e a c t a n c e , Xs,, between each pair of winding groups (45 d i f f e r e n t values f o r 1 0 winding groups). C h o o s e one winding group a5 a r e f e r e n c e and build t h e so-called "Zbus" matrix [SI, which we w i l l c a l l x b because i t is purely r e a c t i v e in this case. If winding group one is chosen as a r e f e r e n c e , t h e elements of t h e matrix can be determined as follows: Diagonal elements:

(See Fig. 6 )

(See Fig

2.

ARRESTER POLE GROUND LEAD POLE GROUND RESISTANCE HOUSE GROUND

HOUSE GROUND RES I STANCE

N = number of groups

Fig. 4. System model f o r simulations

Subscripts on X S c r e f e r t o winding group.

639
O f f-di agonal el emen t s:

Xbij
3.

= 2L (Xbii

+ Xb,j

- x ~ C i + l , j+l)

48

MIDPOINT OF PRIMARY WINDING

Convert t h e x b matrix t o nodal a d m i t t a n c e form for use in general-purpose transients program. T h i s i s d o n e by a s i m p l e p o w e r - i n v a r i a n t transformation of r e f e r e n c e frame [61.

The matrix will be of order N-1 while t h e nodal a d m i t t a n c e form will be of order N . One must be careful t o compensate for t u r n s ratios throughout this process. The a u t h o r s prefer t o build x b in per unit quantities and then compensate each element of t h e nodal a d m i t t a n c e form. T h e gamma m a t r i x , which is t h e inverse of t h e inductance matrix, for t h e 1 0 kVA transformer model is given in t h e appendix. This can be used in a generalpurpose electromagnetic transients program t o allow t h e interested reader t o d u p l i c a t e t h e si mulat ions presented in this paper. The i n d u c t a n c e s and c a p a c i t a n c e s determined by conventional transformer design formulae were found t o be q u i t e a c c u r a t e f o r t h e t y p e of transient analysis performed. However, t h e apparent winding resistance a t t h e transformer's natural frequency must be determined e m p i r i c a l l y . T h i s was d o n e by comparing s u r g e t e s t r e s u l t s with c a l c u l a t e d r e s u l t s and a d j u s t i n g t h e resistance for t h e best match. The model verification f o r t h e 10-kVA transformer was performed using t h e t e s t c i r c u i t shown i n F i g u r e 7 , which r e s u l t e d i n t h e transient voltages shown i n Figure 8a. The best match was obtained with approximately 6000 ohms of resistance in t h e primary winding. The resulting waveform is shown i n F i g u r e 8b. As can be seen, t h e model very a c c u r a t e l y d u p l i c a t e s t h e m e a s u r e d t r a n s i e n t r e s p o n s e of t h e t r a n s f o r m e r . T h e o n l y e x c e p t i o n s a r e t h e minor o s c i l l a t i o n s i n t h e f i r s t p a r t of t h e l o w - v o l t a g e w i n d i n g waveform. T h e s e o s c i l l a t i o n s a r e due t o c a p a c i t a n c e s i n t h e secondary windings o r in t h e s u r g e t e s t generator, which were not modeled.
ziaJ--

&

1 3 . 8 VOLTS

k u 1
0

P
-X3 TERMINAL
I

A i

1
75

25 YS

50

( a ) Measured transformer response.

-I
50
JJS

25

75

(b) Computed response of t h e model.

Fig. 8. Comparison of measured and computed w a v e f o m f o r t h e t e s t c i r c u i t in F i g u r e 7.


0

-+
TRIPLEXED CABLE

2
V E R T I C A L LY -SPAC E D O P E N - W I R E CABLE

I
TEST WAVEFORM 1 AMPERE

Fig. 9. Load Model

Two types of secondary c a b l e considered in this study.

Fig. 7. Test c i r c u i t for transformer verification. Secondary Cable Model B o t h t h e c o n s t r u c t i o n and t h e l e n g t h of t h e secondary cable were found t o have significant influence on t h e results. Only t h e series impedance of t h e cables was found t o be important. Our investigations did not show any significant influence of t h e c a b l e c a p a c i t a n c e and, therefore, it was neglected. Perhaps, t h e two most common types of secondary c a b l e used in overhead transformer installations a r e t h e t r i p l e xed c a b l e and t h e v e r t i c a l 1 y-spaced open-wir e t y p e s shown in F i g u r e 9. Both of these types were modeled using a coupled s e t of three lossy inductors. The impedances were computed in t h e same manner as a t h r e e - p h a s e t r a n s m i s s i o n l i n e with earth return [71. R e s i s t a n c e and i n d u c t a n c e matrices for two types of c a b l e used with 10-kVA transformers a r e given in t h e appendix. In addition. a shielded c a b l e was investieated. It was assumed t h a t t h e c a b l e would have ympedance c h a r a c t e r i s t i c s similar t o t h e triplexed cable except that t h e s e l f i n d u c t a n c e of t h e neutral, or shield, would be equal t o t h e mutual i n d u c t a n c e between each phase c o n d u c t o r and shield. The load was assumed t o be 50 percent motor load and 50 percent resistive load connected as shown in F i g u r e 10. The motor load was represented as an i n d u c t a n c e in series w i t h a resistance. At t h e frequency of i n t e r e s t , t h e i n d u c t a n c e dominates t h e impedance and is such a high value w i t h respect t o t h e other inductances in t h e c i r c u i t t h a t i t e f f e c t i v e l y blocks t h e flow of s u r g e currents. The t o t a l load was varied from 1 0 percent t o 100 percent of t h e transformer's rating. At times, t h e load was assumed t o be s h o r t circuited by insulation flashover.

t
240V

t
120VI

1
RES1S T l V E

L
MOTOR

120vi

RESISTIVE LOAD

MOTOR
LOAD

Fig. 10.

Load model used in this study.

640
Other Models T h e g r o u n d l e a d s w e r e r e p r e s e n t e d as l o s s y inductors, having an inductance of 0.4 microhenries per foot and a resistance of 260 micro-ohms per foot. A 15 f t (4.6 m ) pole ground lead and a 5 f t (1.7 m) house ground lead were assumed. 'The grounding resistances at the pole and the house were both assumed t o be 25 ohms for most of the investigations. The surge arrester was assumed t o have a discharge v o l t a g e of 39 k V and was represented by a simple, two - s 1ope v o 1 t a g ecu r r en t characteristic. The primary distribution line was represented by a surge impedance of 400 ohms. A mid-line stroke was assumed, which would see this impedance in both directions. Hence, a 200-ohm value was used in t h e simulations. The results were not very sensitive t o variations in t h e surge arrester and line models, but both models should be present in some form t o provide a path for t h e small amount of current t h a t flows i n t h e transformer primary winding as a result of slight mismatches in the ampere-turn balance i n t h e s e c o n d a r y windings. Also, t h e surge arrester discharge voltage biases t h e voltage across t h e primary winding layers so that t h e ground-end layer is stressed higher than t h e line-end layer in transformers without inter laced secondaries. SIMULATION RESULTS Effects of Interlacing Transformer Secondaries Figure 11 shows a comparison of voltage and current w a v e f o r m s c o m p u t e d f o r both i n t e r l a c e d and non-interlaced transformers. The transformer was rated 1 0 k V A and was f u l l y loaded using t h e load model previously described. A 10-kA 8 X 20 microsecond l i g h t n i n g s t r o k e c u r r e n t was applied t o the primary circuit. Note that t h e peak current in t h e interlaced transformer is more than twice that in the n o n - i n t e r l a c e d transformer. For this particular case, the non-interlaced transformer would be expected t o see approximately t h r e e percent of t h e lightning stroke current i n each half of the secondary winding.
VOLTS

TRANSFORMER SECONDARY WINDING INTERLACED


131rH

NON-INTERLACE0
170pH

SECONDARY CABLE
90-250pH

LOAD 5-50 OHMS

90-250pH !5 OHMS 25 OHMS

Pig. 12. Redrawn equivalent circuit showing the surge current paths more clearly. This is a t y p i c a l r e s u l t of comparisons between inter 1aced and non- inter 1aced t r an sf or mew. The current division d u r i n g t h e c r i t i c a l part of t h e transient is governed by t h e inductances of t h e circuit and the differences in response of t h e two types of transformers can be easily explained in terms of these inductances. The circuit is redrawn in Figure 1 2 t o help visualize t h e phenomena. The l i g h t n i n g surge current flowing toward t h e house ground in this problem sees three parallel paths: t h e secondary cable neutral and the two paths through the transformer secondary and load. Note the polarity of t h e transformer windings with respect t o the direction of current flow. The bulk of t h e current flows in t h e secondary cable neutral, causing a large v o l t a g e d r o p across the length of the cable neutral conductor. This voltage is t h e determining factor in the amount of current that flows through t h e other two paths (Figure 13). The inductances limiting t h e currents in t h e s e two p a t h s a r e t h e secondary-to-secondary i n d u c t a n c e of t h e t r a n s f o r m e r , t h e secondary cable inductance, and the back emf due t o the mutual coupling of t h e phases of t h e secondary cable with t h e neutral. Note that if t h e secondary cable is short, or that i f t h e m u t u a l l y coupled voltage is nearly equal t o the v o l t a g e drop in the neutral, l i t t l e current will flow through the transformer and load.

QROUNO-END FIRST LAYER F L T A Q E


-MEMACE0 WlNDlNQ

I
VOLTS LOAD VOLTAQE CURRENT PATH WINDING CURRENT

1, $ t 7
I
I I

WINDING CURRENT

NoN-.NlERLAcED WINDING

I *

x1

x3

SECONDARY CURRENT INTERLACED WINDING

THROUGH TRANSFORMER AND LOAD

1
I Fig. 11. Comparison of computed voltage and current waveforms for interlaced and non-interlaced secondary windings for a 10-kVA transformer, 10-kA 8x20 surge; 65-ft. triplexed cable, 100 percent load.

LOA0

Fig. 13. Detail of t h e boxed area in Figure 12, showing how t h e secondary cable voltage drop determines t h e amount of current in t h e transformer and load.

64 1

r h e s e c o n d a r y - t o - s e c o n d a r y i n d u c t a n c e of t h e non-interlaced transformer is on t h e same order as t h e secondary c a b l e inductances, while t h e i n d u c t a n c e of t h e i n t e r l a c e d t r a n s f o r m e r is approximately an order of magnitude less. Therefore, t h e combination of non- interlaced transformer and cable offers a p p r o x i m a t e l y t w i c e t h e inductance t o t h e lightning s u r g e c u r r e n t a s does t h e combination of interlaced transformer and cable. When high c u r r e n t s can flow, due t o e i t h e r high load or secondary flashover, the interlaced connection will typically allow twice as m u c h c u r r e n t t o flow. T h e c u r r e n t s i n t h e s e c o n d a r y windings a r e not p e r f e c t l y balanced because of differences in impedances o f t h e secondary windings. The interlaced transformer has more nearly balanced secondary winding impedances than does t h e non-interlaced transformer. Differences in t h e ampere t u r n s between t h e secondary windings a r e made up in t h e primary winding.
As e x p e c t e d , t h e v o l t a g e s t r e s s a c r o s s t h e f i r s t l a y e r ( g r o u n d e d e n d ) of t h e primary winding of t h e i n t e r l a c e d t r a n s f o r m e r is much l o w e r t h a n i n t h e n o n - i n t e r l a c e d transformer. I t is proportional t o t h e primary a r r e s t e r discharge voltage by t h e r a t i o of t h e t u r n s i n t h e f i r s t l a y e r t o t h e t o t a l t u r n s on t h e winding. The voltage in t h e non-interlaced transformer o s c i l l a t e s a t t h e n a t u r a l frequency of t h e transformer and s e t t l e s out a t a value proportional t o t h e a r r e s t e r d i s c h a r g e v o l t a g e ( s e e F i g u r e 11). N o t e t h a t t h e a r r e s t e r discharge voltage adds t o t h e induced transient voltage a t t h e end of t h e winding shown, which is t h e grounded end. The transient layer-to-layer voltage a t t h e l i n e end of t h e winding is in t h e opposite polarity with respect t o t h e a r r e s t e r discharge voltage, w h i c h r e s u l t s i n somewhat lower voltages. This may help explain w h y failures due t o this s u r g e phenomenon a r e f o u n d m o r e f r e q u e n t l y a t t h e grounded end of t h e winding.

PRIMARY VOLTAGE ARRESTER

ARRESTERS BYPASS SURGE CURRENTS AROUND TRANSFORMER'S IMPEDANCE

I1
3
>OLE GROUND
+HOUSE

GROUND

Fig. 14. The e f f e c t of low-voltage a r r e s t e r s on t h e secondary d i s t r i b u t i o n system f o r low-voltage s i d e c u r r e n t surges. E f f e c t s of Load


As shown in Figure 2, t h e c u r r e n t t h a t flows through t h e transformer secondary must a l s o flow through t h e l o a d . Therefore, t h e load magnitude is a determining f a c t o r in t h e amount of c u r r e n t t h a t can flow. F i g u r e 1 5 shows how t h e peak c u r r e n t i n t h e transformer varies as a function of load. A 10-kVA transformer with a 65 f t ( 2 0 m ) triplexed secondary cable, consisting of a 50 f t span and 1 5 f t lead, was assumed f o r developing this curve. The load was assumed t o be 50 percent resistive and 5 0 percent motor load as described previously.

The v o l t a g e a c r o s s t h e load is proportional t o t h e s u r g e c u r r e n t flowing i n t o the load and is higher when t h e transformer has in t er 1 aced secondary w i n din gs. E f f e c t of Low-Voltage Arresters T h e simulations show t h a t placing a r r e s t e r s a c r o s s t h e s e c o n d a r y w i n d i n g is e f f e c t i v e i n l i m i t i n g t h e v o l t a g e within t h e transformer f o r this t y p e of surge. The voltages induced in t h e primary a r e approximately p r o p o r t i o n a l , by t h e t u r n s r a t i o , t o t h e low-voltage a r r e s t e r d r o p and would n o t be e x p e c t e d t o v a r y s i g n i f i c a n t l y for reasonable increases in s u r g e c u r r e n t m a g n i t u d e s . T h u s , i t would a p p e a r t h a t this is an e f f e c t i v e a l t e r n a t i v e t o i n t e r l a c i n g secondary windings where abnormally high secondary surge currents a r e expected. H o w e v e r , t h e l o w - v o l t a g e a r r e s t e r s o f f e r no p r o t e c t i o n f o r t h e load. In f a c t , a r r e s t e r s applied t o n o n - i n t e r l a c e d transformers will a c t u a l l y increase t h e voltage s t r e s s on t h e load t o approximately t h e same level as would an interlaced transformer. This may not be intuitively obvious because t h e a r r e s t e r s would seem t o be c l o s e enough t o t h e load t o p r o t e c t it as well as t h e transformer from t h e surges t h a t a r e c h a r a c t e r i s t i c of this problem. However, t h e a r r e s t e r is n o t connected a c r o s s t h e load, but is a c t u a l l y in series with t h e load and t h e s u r g e c u r r e n t path. This may be seen by adding a r r e s t e r s t o t h e redrawn c i r c u i t diagram as shown i n Figure 14. The a r r e s t e r s simply bypass t h e transformer, removing i t s impedance from t h e c u r r e n t path. Thus, more c u r r e n t can flow t h r o u g h t h e l o a d because o f the reduction in c i r c u i t impedance. In s h o r t , a d d i n g l o w - v o l t a g e a r r e s t e r s t o t h e c i r c u i t has approximately t h e same a f f e c t on t h e system as interlacing t h e transformer secondary windings.

0.0

0.2

0.3

0.4

0.8

1.0

L O A D , PER UNIT

Fig. 15. Variation of peak transformer secondary c u r r e n t as a function of load, 10-kVA transformer, 65-f t. t riplexed cable.

642
The peak voltage s t r e s s in t h e primary windings of t h e n o n - i n t e r l a c e d t r a n s f o r m e r will g e n e r a l l y be p r o p o r t i o n a l t o t h e c u r r e n t , although t h e r e is some d e c r e a s e in t h e proportion as t h e load increases due t o c h a n g i n g r a t e s - o f - r i s e i n t h e c i r c u i t . T h e primary w i n d i n g l a y e r s t r e s s in t h e interlaced transformer is relatively insensitive t o t h e change in current because t h e d i s c h a r g e voltage of t h e primary a r r e s t e r reflected t h r o u g h t h e winding is much larger than t h e voltage induced from t h e secondary s u r g e current ( a t least for lightning strokes in t h e 10-15 k A range). The limiting values of current in this c a s e occur when t h e load is short-circuited. They a r e 4.6 percent i n e a c h h a l f of t h e w i n d i n g for t h e non-interlaced transformer and 12 percent for t h e interlaced transformer. These limiting values a r e sensitive t o t h e secondary cable length, as described in t h e next section o f this paper. The peak voltage across t h e load corresponding t o t h e current variation of F i g u r e 1 5 is shown in F i g u r e 16. A s expected, t h e use of t h e interlaced transformer r e s u l t s i n s u b s t a n t i a l l y higher voltage s t r e s s across t h e l o a d f o r most of t h e load range. Both curves a p p r o a c h t h e same o p e n - c i r c u i t v a l u e as t h e l o a d approaches zero. The open-circuit ( n o load) voltage is t h e voltage drop a c r o s s t h e secondary cable neutral less t h e voltage induced i n t h e phase c o n d u c t o r s by mutual coupling. The induced voltage i n t h e secondary cable is substantial and t h e mutual coupling must be represented f o r a c c u r a t e simulations of this phenomenon. From t h e s e two c h a r t s , i t is observed that t h e transformer is more at risk when t h e load is high, o r when t h e secondary flashes over, and t h e load appliances a r e more at risk when t h e t o t a l resistive load is small. E f f e c t s of Secondary Cable Length The real c u l p r i t i n t h e system with respect t o this s u r g e problem is not t h e transformer, and i t s design, nor t h e load magnitude, but t h e secondary cable. The amount of current that flows through t h e transformer and load is dependent on t h e net voltage drop across the secondary cable.
1 . 4
3.0 9.0 LENGTH, FT

Of course, this drop across t h e secondary c a b l e is a f u n c t i o n of l e n g t h . To study t h e e f f e c t s of c a b l e length, a short-circuited load was assumed because this yields t h e worst case in terms of current magnitude i n the secondary. Of p a r t i c u l a r interest here a r e t h e c u r r e n t m a g n i t u d e , t o w h i c h t h e l o a d s t r e s s is proportional, a n d t h e voltage across t h e f i r s t layer o f the primary transformer winding. F i g u r e 17 and 18 show the results for t h e 1 0 - k V A non-interlaced transformer considered throughout this paper.

Fig. 17. Variation of maximum transformer secondary c u r r e n t as a f u n c t i o n of secondary c a b l e length; non-interlaced 10-kVA transformer.

I -

0.6

5 K 5
0
W

2.5

a
0

E
o

P
O.

ti
(3

2.0

ti
c)

L
f

z
I -

I -

1.5

I L

0.4

5
W

a *
K

1.0

L 2

1;
0.2
0.5

0.0 0.0 0.0 0.0 0.2 0.4


L O A D , PER UNIT

200.0

400.0

600.0

8 1 b . 0

1000.0

0.6

0.8

1.0

LENGTH. FT

Pig. 16. Variation of peak load voltages as a f u n c t i o n of load, 10-kVA transformer, 65-ft. triplexed cable.

Fig. 18. Variation of t h e peak v o l t a g e a c r o s s t h e f i r s t grounded-end layer of a 10-kVA transformer primary winding, as a f u n c t i o n of c a b l e length.

643

Tne current magnitude increases somewhat less than linearly for t h e first 400 f e e t , and then approaches a limiting value. The transformer voltage also increases, b u t is l e s s t h a n p r o p o r t i o n a l t o t h e i n c r e a s e i n c u r r e n t , and eventually begins t o d e c r e a s e as t h e cable g e t s longer. Although more current is being forced into the transformer secondary by t h e longer cable, t h e s l o p e o f t h e current is modified by t h e different inductanceto-resistance ratios of t h e t o t a l system at each c a b l e length, resulting in less induced voltage than might be ex pec t ed. Assuming equal ground resistance at t h e pole and a t t h e house, i t appears that this particular transformer sees no more than about 6.8 percent of t h e lightning s t r o k e current in each half of t h e secondary winding. In t h e same c i r c u i t , t h e interlaced connection of this transformer sees u p t o about 1 6 percent of t h e stroke c u r r e n t i n e a c h h a l f , which is approximately equal sharing of current between t h e t h r e e paths leading t o t h e house. E f f e c t s of Secondary Cable Type F i g u r e 19 secondary cable across t h e load) secondary surge 100 percent load shows t h e n e t v o l t a g e d r o p i n t h e c i r c u i t (i.e., t h e open-circuit voltage for t h r e e types of cable. The resultant current and primary winding voltage for a r e also shown.

w o u l d r e d u c e t h e t r a n s i e n t v o l t a g e s on b o t h t h e transformer and load simultaneously. Surrounding t h e two phase c o n d u c t o r s w i t h t h e ground c o n d u c t o r w i l l provide very c l o s e mutual coupling between t h e ground and phase conductors. Therefore, t h e induced voltage in t h e p h a s e c o n d u c t o r s is approximately equal t o t h e i n d u c t i v e d r o p i n t h e n e u t r a l . The net d i f f e r e n t i a l voltage forcing current through t h e transformer and load is t h e resistive voltage drop, which should be much lesy than with either triplexed o r open-wire cable. I n t e r e s t i n g l y , t h i s is o n e of t h e s o l u t i o n s t o t h e problem of secondary s u r g e c u r r e n t s proposed by Bryant and Newman in 1 9 4 2 [21. E f f e c t s of Multiple Customers The general e f f e c t of having more than one customer on t h e same transformer is t o reduce t h e s t r e s s on both the transformer and t h e loads. Consider t h e c a s e of two houses on one transformer and compare it t o t h e c a s e of one h o u s e per transformer. Assuming equal grounding r e s i s t an c e s t h r o u g h o u t t he ci r c u i t , ap p r ox i mat el y 3 3 p e r c e n t of t h e lightning s u r g e current w i l l travel in each of t h e t h r e e available paths: t h e pole ground and the two house grounds. In c o n t r a s t , 50 percent of t h e current flows in each path when t h e r e is only one house per transformer. Therefore, having multiple houses on a t r a n s f o r m e r r e s u l t s i n l e s s v o l t a g e drop a c r o s s t h e secondary cables because t h e s u r g e current is dispersed. T h i s , i n t u r n , r e s u l t s i n l e s s current being forced through t h e transformer secondary than for t h e c a s e with only one house per transformer. E f f e c t s of Grounding Resistance T h e f u n d a m e n t a l d i v i s i o n of t h e lightning s u r g e c u r r e n t s i n t h e c i r c u i t is determined by t h e ground r e s i s t a n c e at t h e various grounding points. There a r e some c h a n g e s i n waveform f o r d i f f e r e n t ground r e s i s t a n c e s , b u t , in general, the current i n the secondary cables can be a c c u r a t e l y determined by direct p r o p o r t i o n b a s e d on grounding resistance differences from a known case. The better t h e pole ground with respect t o t h e house ground, t h e more p r o t e c t e d a r e t h e transformer and load. Ideally, t h e pole ground should be much b e t t e r than t h e house ground and divert most of t h e s u r g e current away from t h e load. Of course, this is not always achievable and, in f a c t , nearly t h e opposite o c c u r s in some cases. In some areas it is d i f f i c u l t t o achieve pole grounds t h a t a r e s i g n i f i c a n t l y b e t t e r than house g r o u n d , especially when t h e house ground is also a t t a c h e d t o t h e municipal water system or a deep water well. I n such cases, much of t h e lightning current is diverted t o the house ground, imposing greater voltage stresses on t h e transformer and load. It would seem t h a t t h e homeowner would be well-advised t o i n s t a l l a low-voltage a r r e s t e r a t t h e service entrance. CON CLUSIONS

NET DIFFERENTIAL VOLTAGE IN SECONDARY CABLE TRIPLEX SHIELDED

SURGE CURRENT

IN SECONDARY
WINDINGS
( EACH HALF)

15
kV

VOLTAGE ACROSS FIRST LAYER OF PRIMARY WINDING

PRIMARY ARRESTER DISCHARGE VOLTAGE

(REFLECTED TO FIRST LAYER )

Fig. 19. Comparison of t h r e e secondary c a b l e types, 10-kVA transformer, 10-kA 8x20 surge, 65-ft. cable, 100 p e r c e n t load. The open-wire secondary cable develops more net voltage and allows more s u r g e current t o flow i n t h e t r a n s f o r m e r and load than the same length of triplexed cable. This is due t o t h e higher i n d u c t a n c e that results from greater spacing between c o n d u c t o r s and t h e lower mutual coupling between wires. The s h i e l d e d c a b l e develops a small net voltage, which is e s s e n t i a l l y t h e r e s i s t i v e d r o p a c r o s s t h e n e u t r a l c o n d u c t o r . Not only is t h e c u r r e n t that r e s u l t s from this voltage smaller than either of t h e other two t y p e s , b u t t h e r a t e - o f - r i s e is dramatically reduced, r e s u l t i n g i n a s i m i l a r r e d u c t i o n of induced voltage within t h e transformer. Thus, t h e use of shielded c a b l e

C o n c e r n i n g p r o p o s e d t e s t levels for low-voltage c u r r e n t s u r g e s i n distribution transformers, it would appear t o be i n a p p r o p r i a t e t o specify a single current m a g n i t u d e for a l l transformers. The amount of s u r g e c u r r e n t t h a t p a s s e s t h r o u g h t h e t r a n s f o r m e r varies s i g n i f i c a n t l y with t h e design and k V A rating of t h e transformer. The current t h a t flows is n o t a c o n s t a n t , b u t is d e t e r m i n e d , primarily, b y t h e net differential mode voltage drop a c r o s s t h e secondary c a b l e and, also, by t h e impedances of t h e transformer, cable, load, and ground. Sm a1 l e r - s i z ed t r a n s f o r m e r s w i t h i n t e r 1 a c e d secondaries o r low-voltage a r r e s t e r s will typically pass t w i c e as much s u r g e c u r r e n t as c o n v e n t i o n a l , non-interlaced designs of t h e same rating. The 10-kVA transformer described i n this paper would typically see

644

less than 7 percent o f t h e lightning s t r o k e c u r r e n t in each half of t h e secondary winding (14 percent t o t a l in t h e s e c o n d a r y ) when t h e secondary windings a r e n o t i n t e r l a c e d . I n t h e i n t e r l a c e d c o n n e c t i o n , t h e same transformer would be expected t o s e e nearly 16 percent of t h e lightning s t r o k e c u r r e n t in each half (33 percent t o t a l ) . T h e l a t t e r f i g u r e is t h e t h e o r e t i c a l limiting value assuming equal division of t h e s t r o k e c u r r e n t in t h e secondary cable and between pole and house grounds. 2 5- k V A and 50- k V A non- i n t e r 1aced transformers w i 11 a l l o w m o r e c u r r e n t t o f l o w t h a n t h e 10-kVA n o n - i n t e r l a c e d t r a n s f o r m e r b e c a u s e of t h e i r l o w e r i m p e d a n c e s . In f a c t , t h e i m p e d a n c e of a 50-kVA transformer is relatively insignificant when compared t o t h e c a b l e impedance and t h e s u r g e c u r r e n t in t h e t r a n s f o r m e r a p p r o a c h e s t h e t h e o r e t i c a l limit. This f a c t , in combination with different distributions of t h e t u r n s i n l a r g e r - k V A t r a n s f o r m e r s may e x p l a i n why r e s e a r c h e r s have reported t h a t transformers t h a t a r e 50-kVA, o r l a r g e r , a r e essentially immune t o f a i l u r e from low-voltage-side current s u r g e phenomena. Another f a c t o r might be t h a t these transformers a r e generally connected t o more than one secondary load, which our s t u d i e s i n d i c a t e would decrease t h e voltages induced within t h e transformers. T h e s y s t e m c o n s i s t i n g of t h e d i s t r i b u t i o n transformer and load has a complex response t o lightning s u r g e s a n d is v e r y s e n s i t i v e t o c h a n g e s i n t h e c h a r a c t e r i s t i c s of t h e components of t h e system. Changes t o t h e transformer cannot be considered without also considering t h e e f f e c t s of t h e changes on t h e rest of t h e s y s t e m . F o r example, p r o t e c t i n g t h e transformer insulation by interlacing t h e transformer secondaries o r by a p p l y i n g low-vol t a g e a r r e s t e r s will approximately double t h e voltage s t r e s s on t h e load f o r t h e smaller t r a n s f o r m e r s i z e s . The only solution found t h a t minimizes t h e problem in a l l areas of t h e system is t o use shielded secondary c a b l e t h a t has adequate mutual coupling between t h e n e u t r a l and phase conductors. This m i n i m i z e s t h e s u r g e c u r r e n t s by r e d u c i n g t h e n e t d i f f e r e n t i a l v o l t a g e i n d u c e d by t h e lightning s u r g e c u r r e n t s flowing in t h e c a b l e neutral. Other general trends found in this analysis include:
1.

REFERENCES
1.

K . D. Beardsley, W. A. McMorris, H. C. Stewart, "Voltage Stresses in Distribution Transformers Due t o Lightning C u r r e n t s in Low-Voltage Circuits," AIEE Transact ions, Volume 67, 1948, pp 1632-1635.
M. Bryant, V. Newman, "Abnormal C u r r e n t s in Distribution Transformers Due t o Lightning," AIEE Transactions, Volume 61, 1942, pp 564-8.

2.

J.

3.

C. J. VcMillen, C. W. Schoendube, D. W. Caverly, " S u s c e p t i b i l i t y of D i s t r i b u t i o n T r a n s f o r m e r s t o L o w - V o l t a g e S i d e Lightning S u r g e Failure," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-101, No. 9 , September 1982, pp 3457-3464. Surge C h a r a c t e r i s t i c s and Protection of Distribution Transformers, EPRI EL-3385, P r o j e c t 1532-1, Final R e p o r t , E l e c t r i c P o w e r R e s e a r c h I n s t i t u t e , Palo Alto, CA, January 1984.
S. N . Talukdar, "METAP - A Modular and Expandable

4.

5.

P r o g r a m f o r Simulating Power System Transients," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-95, Nov/Dec 1976, pp 1882-1891. 6. S t a g g and El-Abiad, Computer Methods in Power System Analysis, iVlcGraw-Hill Book Company, 1968, Chaps 3 and 4. E l e c t r i c a l Transmission and Distribution Reference Book, Central Station Engineers of t h e Westinghouse E l e c t r i c C o r p o r a t i o n , East Pittsburgh, P4, 1964, Chap 3. APPEND I X Transformer Model Data T h e transformer is modeled as a couple s e t of lossy i n d u c t o r s as shown in F i g u r e 6. T h e gamma matrix ( i n v e r s e of t h e i n d u c t a n c e m a t r i x ) f o r t h e 10 kva transformer in lower triangular form is: 185.0 -233.5 2.388 -.6256 .06838 -.04450 -.1358 -.1546 7.138 -2.555 517.8 -15.31 1.326 3.811 -.9121 -.6003 .1968 .2499 -.0531 .4417 -.0017 -.I3352 .0079 -.2519 .3003 -16.60 .7304 ( h y - l x 1000) 1.514 -.9124 1.455 .2118 -.8794 1.398 -.0457 .1996 -.a404 1.317 .0126 -.0358 .1752 -.7037 . a 6 2 2 -.2221 .0398 -.5565 1.589 -7.100 286.0 1.218 -158.8 -.1681 -.0057 .0970 -.3105

7.

T h e transformer is more susceptible t o f a i l u r e when t h e load is high, o r is s h o r t circuited, than a t low load. The load is more s u s c e p t i b l e t o f a i l u r e a t low load levels. To minimize problems t o b o t h load and transformer f o r low-voltage-side c u r r e n t surges, a. Secondary cables should be kept as s h o r t as p o s s i b l e ( a s h i e l d e d s e c o n d a r y c a b l e is desirable). Every a t t e m p t should be made t o achieve a b e t t e r pole ground than house ground. Having multiple houses fed from t h e same t r a n s f o r m e r helps by f u r t h e r diluting t h e 1 i g h t n i n g s u r g e c u r r e n t in t h e secondary c a b 1es. Where t h e s e conditions can not be p r a c t i c a l l y achieved, protection of t h e load w i t h l o w - v o l t a g e a r r e s t e r s would a l s o be advisable.

2.

3.

The sequence of windings represented by t h i s matrix goes from l e f t t o right in t h e diagram in Figure 6. I'o match t h e damping a t t h e natural frequency, p l a c e about 1000 ohms resistance in each primary winding group. The capacitances in t h e primary winding, beginning with t h e grounded end of t h e winding are: Capacitor
1

b. c.

pf
1372 1432 1496 1557 1619 1700

d.

2 3
4

5 6 Secondary C a b l e Data

T h e distribution transformer is only o n e element in a complex system consisting of transformers, a r r e s t e r s , s e c o n d a r y c a b l e s , p o l e g r o u n d s , house grounds, and l o a d s . Transformer t e s t standards must recognize t h i s fact.

T h e s e c o n d a r y c a b l e is r e p r e s e n t e d as a simple series impedance:


Z = R + jwL

645 The R and L matrices used f o r t h e d i f f e r e n t types of c a b l e follow. Each c a b l e was assumed t o consist of #2 A1 phase c o n d u c t o r s with a #4 A1 neutral. All matrices a r e symmetrical and t h e n e u t r a l c o n d u c t o r is represented by t h e middle row and column. The R matrix was assumed t o be t h e same f o r each t y p e of cable:

BIOGRAPHY
R o g e r D. D u g a n ( M 7 3 , S N I 8 3 ) w a s b o r n i n McConnelsville, ohia in 1950. He obtained t h e R.S.E.E. degree from Ohio University i n Athens, Ohio in 1972 and t h e M.E. d e g r e e in E l e c t r i c Power Engineering from Rensselaer Polytechnic I n s t i t u t e i n Troy, New York in 1973. He has worked f o r t h e Systems Engineering department of McGraw-Edison Power Systems s i n c e graduation. He has been primarily involved with t h e computer simulation o f power s y s t e m s , especially with respect t o harmonics, e 1 e c t r om a g n e t i c t r a n s i e n t s , t r a n s f o r m e r , a n d distribution engineering applications. He has also done work in computer s c i e n c e and computer graphics. S t e D h e n S m i t h is a s t a f f E n g i n e e r in t h e Dist ribufion Transformer Product Engineering Department a t t h e McGraw-Edison Zanesville, Ohio f a c i l i t y . He is a 1972 g r a d u a t e of Ohio S t a t e Universitv. holding both a BSEE -and MSEE in E l e c t r i c Power Engineering.- Before r e j o i n i n g McGraw-Edison i n A p r i l , 1984, he was a Professor of E l e c t r i c a l Technology and Physics at Muskingum Area Technical College, where h e had been s i n c e 1980. Steve has also gained considerable experience as an e l e c t r i c a l engineering consultant in the areas o f e l e c t r i c a l p o w e r d i s t r i b u t i o n a n d power transformer maintenance. He is a member of IEEE and a Registered Professional Engineer in t h e S t a t e of Ohio.

(ohms per 1000 f t ) Triplexed C a b l e (microhenries per 1000 f t )

12- inch Vert ically-spaced Open Wire (Neutral on t o p )


L = 525 525

785 483

772

(microhenries per 1000 f t )

Shielded C a b l e (microhenries per 1000 f t )


A l l R and L matrices a r e symmetrical.

Discussion D. R. Smith, T. J. Dionise, N. C. Abi-Samra, and J. Pun (Westinghouse Electric Corporation, Pittsburgh, PA and Athens, CA): The authors have done a thorough job in identifying those parameters which affect secondary winding currents and currents in the house load for lightning strokes to the primary system. They not only identify the parameters, but they show how parameter changes, such as service drop length and house load, affect winding currents. From their results, it is clear that secondary winding configuration, either interlaced or noninterlaced, can significantly affect winding currents and house load voltages for most conditions encountered in practice. We are conducting similar studies and our results are in good agreement with those reported in the paper. This is encouraging as their models for the transformer and service drop are different from ours. For example, they represent the secondary circuit with mutually coupled lumped inductors. In our studies the secondary is represented by positive and zero mode surge impedances and the propagation velocities for each mode. The authors studies used a IO-kA 8 x 20 ps lightning stroke current, whereas we are using 30 kA strokes with either 1.3 x 90 or 2.5 x 90 ps wave shapes. Our studies show that when the house load is shorted, stroke wave shape does not have a significant effect on surge current distribution, and in particular the current in terminal X2 of the transformer. It appears that any minor differences between the authors and our results are due to particulars of the systems studied and not the modeling techniques. There are several comments and questions that we would like to address to the authors. From Figs. 1-4 in the paper it appears that the authors did not represent the neutral conductor of the primary line. When the neutral conductor is represented, the surge impedance model for the line consists of three wye-connected resistors connected between transformer terminals H1, H2, and ground, rather than one 200-ohm resistor from HI to ground as shown in the paper. Would the authors comment on the effect of omitting the representation of the primary neutral conductor? From the open wire secondary, we have calculated the inductance values given in the Appendix to within two tenths of one percent for both the selfs and mutuals. To obtain the self-inductance values of 772 and 785 yH we had to use the conductor geometric mean radius rather than the conductor radius. It is, we believe, more accurate to use the conductor radius for these calculations as the current is on the conductor surface for lightning surges. Furthermore, to obtain the mutual inductance values given in the paper, we had to assume a spacing of 6, 6, and 12 in rather

than the 12-, 12-, and 24-in spacings given in Fig. 9 of the paper. For what spacings do these inductance values apply? In the section titled, Effects of Load, it is stated that with 65-ft drop, IO-kVA noninterlaced unit, and the load shorted, the current in each half of the secondary winding is 4.6 percent of the stroke current. However, in Fig. 17 which applies when the load is shorted, the maximum secondary current for the IO-kVA noninterlaced transformer with the 65-ft drop is about 3.2 percent. What is the reason for this difference? In summary, the paper shows that interlacing of secondary windings can have a significant effect on secondary winding currents and phaseto-neutral voltages at the house for lightning strokes to the primary. This is in agreement with the results of our studies. Manuscript received October 7 , 1986

C. J. McMillen (General Electric Company, Hickory, NC) and C. W. Schoendube (Retired, Hickory, NC): This paper contributes a valuable study identifying and quantifying the system parameters influencing surge currents entering the low-voltage side of the distribution system. It is noteworthy in that it complements our experimental work exploring the effects on the transformer of low-voltage-side surges originating at the customers premises [l], [2], [3]. However, there are many differences in the parameters described in this paper that account for differences from the results in our tests. The shape of the applied current wave used in this study was 8 x 20 as. This is a good wave for obtaining reproducible discharge voltages of lightning arresters, but does not correspond to typical lightning strokes. Lightning generally has higher maximum rates of rise and much longer duration tails [4]. A 1 x 50 wave is a more appropriate wave shape. The damaging oscillatory voltages generated in the high-voltage coil are sensitive in magnitude to the rate of rise duration of the front and the tail duration of the current wave [3]. The slowfront and short tail of the 8 x 20 wave used in this study will therefore reduce the voltage developed between high-voltage layers compared to our experimental tests. It is stated that other wave shapes were investigated. What were they and what were the results? The mutual inductance between each low-voltage coil and approximately half of the high-voltage coil is also significant. It appears from Fig.

646 8 that the mutual inductance of the IO-kVA transformers simulated was considerably lower than our transformers, thus yielding lower primary voltages. The shell-type transformer tested [l] in our scaled tests had 29 turns per 120-volt coil. The IO-kVA shell types used in our EPRI study [2], [3] had 39 turns. Figure 12 states 170pH for the noninterlaced secondary winding and 13 pH for interlaced. Our transformers with 39 turns had 226 pH and 59 p H for noninterlaced and interlaced, respectively, utilizing the same core design and turns for both transformers. The high ratio of inductances in your simulation suggests entirely different transformers were simulated. How many turns per 120 volts were simulated in your study? The voltage developed in the high-voltage winding is proportional to the mutual inductance of the high- and low-voltage coils and therefore proportional to the product of the high- and lowvoltage turns and a permeance term. We do not agree that the model very accurately duplicates the transient response of your transformer. The measured transformer response, Fig. 8A, has a natural period of 14.6 ps while the model Fig. 8B was 22.8 ps. The model peak voltage was 6 percent higher and the response was underdamped, i.e., the second peak is 77 percent of the first peak versus 65 percent of the transformer. The natural period is determined by the square root of the product of an inductance and capacitance. The difference in natural periods indicates an average error of 56 percent in calculation of each of those components. Figure 8[b] also shows a significant negative offset in the primary voltage trace which is not present in Fig. 8[a]. Figure 6, the schematic model of the transformer in this study indicates a high-voltage coil with layers connected in what we call a crossback type coil, rather than a conventional barrel type coil. The capacitor shown connected from the H2 terminal to the start of the second layer indicates that only the first layer has a crossback since capacitors are not shown across the corresponding layers at the bottom end of the high voltage layers. We also do not understand your capacitance matrix [Appendix]. How are only six capacitors connected to 10 primary groups? The paper stresses the magnitude of voltage impressed on the customers wiring and appliances, particularly in regard to the greater magnitude that interlaced transformers will allow. In reality, the meter protection [flashover at approximately 7 kV and outlet or wiring flashovers, approximately 2 to 5 kV] will limit the overvoltages impinging on susceptible appliances for currents allowed by both noninterlaced or interlaced windings. The best assurance for protection of the customers premises is the installation of home lightning protectors at the service entrance and spike protectors at more sensitive electronic appliances. We also believe the study may be flawed by not incorporating the effects of the service cable capacitance and the customers wiring. Our scaled tests included those capacitances which tend to reduce the slope of the current waves and voltage magnitudes calculated in your study. We do not understand the theoretical limiting value of current division in the second paragraph of your conclusions, nor the logic of the third sentence of the third paragraph. This fact referring to fact in the previous sentence does not logically follow as an explanation for the relative immunity to failure of 50-kVA noninterlaced transformers. The fact of having a lower mutual inductance between the low-voltage winding and half the high-voltage winding explains the induction of lower primary voltage as the kVA rating is increased, with a given magnitude of current. Your analysis of the effect of the home load on the system response, as well as line lengths, is a significant contribution toward understanding the complexity of the problem. The recommendation that shielded service cable is one solution that tends to protect both the transformer and the customer is also worthy of note. However, we doubt that it will be economically acceptable by either customers or utilities when compared to the application of home lightning protectors and interlaced windings. We do not agree with the implication that general adoption of interlaced windings in the lower kVA ratings of distribution transformers will lead to serious problems in the home. It is estimated that at least half of the distribution transformers, approximately 10 million transformers, have interlaced secondary windings, If there are serious problems, it seems that insurance companies would be advocating installation of home lightning protectors and spike protectors. Your work is a welcome contribution in that it thoroughly explores the effects of system parameters that help initiate the commonly occurring failures of shell-type transformers utilizing noninterlaced low-voltage windings. This work, combined with our work which defined the transformer parameters that influence susceptibility to failure, are both milestones towards establishing suitable low-side-surge withstand requirements for distribution transformers. The need for withstand requirements is amply demonstrated in [5], which documents the prevalence of distribution transformer failures attributable to this phenomenon. References
11 ] C . J. McMillen, D. W. Caverly, and C . W. Schoendube, Scaled

Low Voltage Side Surge Current Tests on a Model Distribution System, 86 T&D 552-4, IEEE/PES 1986 Transmission and Distribution Conference, Anaheim, CA, September 14-19, 1986. C. J. McMillen, C. W. Schoendube, and D. W. Caverly, Susceptibility of Distribution Transformers to Low Voltage Side Lightning Surge Failure, IEEE Transactions on Power Apparatus and Systems, Vol. PAS-101, No. 9, September 1982, pp. 3457-3470. Surge Characteristics and Protection of Distribution Transformers, Electric Power Research Institute Final Report EL-3385, January, 1985. H. M. Schneider and H. R. Stilwell, Measurement of Lightning Current Wave Shapes on Distribution Systems, paper #A79 526-5, presented at the IEEE-PES Summer Meeting, July 15-20, 1979. W. J. Ros, A Realistic Evaluation of Interlaced Secondary Designs, EEI Transmission and Distribution Committee, May 15, 1986. J. D. Borst, Susceptibility of Distribution Transformers to Low Voltage Surges, EEI Transmission and Distribution Committee, May 15, 1986.
Manuscript received October 10, 1986.

Roger C. Dugan and Stephen D. Smith: We appreciate the discussers taking the time to study our work carefully and making several good, constructive comments about this very interesting problem. First, we will address the comments of Messrs. Smith, Dionise, AbiSamra, and Puri. It is also encouraging to us to know that our studies can be duplicated with somewhat different models and methods. We are looking forward to publication of the discussers results. Our choice of lightning current waveshape was somewhat arbitrary, but did correspond to the waveshape used in the full-scale tests we conducted that preceded the analysis reported in the paper. It would, perhaps, be better to assume a steeper wavefront. We also have found that the results are not as sensitiveto the wavefront as one might think. We discuss this in greater detail below. As the discussers assumed, we did not represent the primary line neutral conductor. We believe that the general effect will be to bleed off a little of the lightning current and reduce the currents flowing in the distribution circuit. The amount of current flowing into the neutral will be dependent on the local ground resistances. For resistances in the 10-25 ohm range, the effect should be slight. For higher grounding resistances, the primary neutral will be more effective in limiting the surge currents. To model the primary neutral correctly, one should represent the coupling to the phase conductor and include several pole spans. We have not yet added this degree of complexity to our models. Regarding the inductances used in our model, we agree with the discussers that it would be more accurate to use the conductor radius instead of the GMR because of the higher frequencies involved. This will reduce the self-inductance relative to the mutual inductance and reduce the net differential voltage induced in the secondary cable loop. Thus our model as stated in the paper will predict higher than actual voltages. Our simulations to this point in time indicate that the difference is small. As the discussers have determined, the inductance matrix for the openwire cable given in the appendix is for 6-in spacing rather than 124x1spacing as we had believed. An open-wire secondary with 12-in spacing will generally yield somewhat higher surge voltages throughout the system than our simulatons would indicate. In the section titled Effects of Load there is indeed a typographical error. The actual maximum secondary current that we compute for a noninterlaced transformer and 65 ft of cable is 2.9 percent rather than the 4.6 percent printed. This is more apparent from Fig. 15 than from Fig. 17. In any case, it is much smaller than the proposed test values. The 4.6 percent figure applies to 165 ft of cable and was mistakenly transcribed from the wrong line of data. Now, we will address the comments of Messrs. McMillen and Schoendube. A steeper wavefront than the 8 x 2 0 ps wave is probably more typical of lightning surge currents as the discussers point out. We have looked at a variety of waveshapes including 1.2 x 50 ps wave. This causes

647
an increase in the computed first layer voltage from 10 kV to 12 kV. This is, perhaps, not as much as one might expect. Early in our research, we found that the voltages were not very sensitive to lightning current waveshape. It is the current through the secondary drop neutral that is the root cause of the voltages in the transformer. Due to typical values of inductance and resistance in the secondary circuit, the front of this current always seems to be significantly slower than the front of the lightning stroke current. Concerning the inductances of the transformer, we did use the same transformer model for both the interlaced and noninterlaced analysis and merely reconnected the secondaries. Perhaps, this transformer is built differently than the discussers,but at this time we do not think we have made any errors in the calculation of the inductances. We broke the transformer down into 10 winding groups and calculated the short-circuit impedances between each pair of groups. Then we constructed the inductance matrix (actually, its inverse) as described in the paper. The model checks out well with respect to high-to-low short circuit impedances and the transient response is well matched. It also is possible that we are not comparing apples-to-apples with respect to the discussers transformer. The inductance we refer to is the short-circuitinductance from one secondary winding to the other. The number of turns per 120 V in the transformers we tested and simulated were 29 turns for the 10 kVA, 19 turns for the 25 kVA, and 15 turns for the 50 kVa. These values are typical for our shell-type, loss-evaluated designs. The discussers point out that the simulated and measured results in Fig. 8 do not match very well. This is due to a misunderstanding with the artist who drew the figure from two plots on different scales. The time scale for Fig. 8a should be from 0 to 100 ps. The corrected figure is shown in Fig. A, included with this closure. Also, because the rendering makes it difficult to determine time scales, we have pointed out key time points in Fig. A. As can be seen, the dominant frequencies in the two oscillograms match very well, with the model exhibiting a slightly lower frequency. We apologize for difficulties caused by this error.
48 V O L ~ S
MIDPOINT OF PRIMARY WlNDlNQ

6.35 psec

13.8 VOLTS
0

25

Ir=

50

75

100

(a) Mensured transformer response.


DPOINT OF PRIMARY WINDING

1 4 . 7 VOLTS-

25

9s

50

I 75

(b) Computed response of the model. Fig. A. Correction of Figure 8.

It is true that the model has slightly less damping than the actual transformer. This will give voltages that are conservativelyhigh and we feel it is better for the model to exhibit slightly lower damping for this reason. We do not feel that the difference in damping pointed out by the discussers is significant. The negative offset questioned by the discussers is actually a transient swing due to the fact that the model has slightly less damping than the actual device. When the simulation is extended beyond 75 ps, this decays to zero. Figure 6 has no relationship to the actual physical design of the transformer, but is intended to show only the connections of the model. The transformer is indeed a conventional design with no crossback coils. The capacitors are not intended to depict a particular design, but only nodal connections in the model. Simply stated, the model consists of a capacitor in parallel with each primary winding section. Perhaps, this can be seen better by drawing the diagram differently. There are only six capacitors because there are only six primary groups. The transformer was broken into a total of 10 groups, four of which are secondary groups. The capacitances in the transformer secondaries were found to be of minor importance and were neglected. We agree that home lightning protectors should be installed at the service entrance and urge the power industry to develop methods to encourage their use. We have investigated the effects of service cable capacitance and the capacitance of the customers wiring, which is assumed by the discussers in their work to be 15 nF. As of this writing, we have not found any noticeable effect of this capacitance. Perhaps it is because we inject our current from the primary h e rather than the secondary. By theoretical limiting value we mean the maximum possible current that could flow in the transformer windings assuming the transformer winding impedance is zero and the load is shorted. If that were the case, the surge current would split equally between the three conductors of the secondary cable. Half of the surge current goes to the house ground in this example. Thus, 1/6 of the lightning current would flow in each wire of the cable. The fact that we were referring to with respect to the 50 kVA transformers is indeed the lower impedance. However, we have found that lower impedance is not the complete story. Lower impedance means that there w i l l be less voltage per ampere of surge current, but the 5O-kVA transformer allows more current to flow than does a IO-kVA transformer. Therefore, we often find that when the transformer is simulated in an actual circuit, the layer voltages are in the same ballpark as for IOIVA transformers. Of course, the voltage is always lower, but not by as substantial a factor as one might think. We have not completed our analysis of this, but it seems that other factors such as layer distribution, insulation thickness, wire size, and typical load application practices also play significant roles in the apparent immunity of SO-kVA transformers to low-voltage-sidecurrent surges. We are studying this to gain some insight into how to improve the withstand of the smaller sizes as well. We recognize that there are no data generally available from which we could correlate transformer winding connection with customer appliance failures. It very likely makes little difference in areas where it is possible to consistentlyachieve good pole grounds compared to house grounds; there will be few surge problems in any case. However, in other areas it is easily shown that the use of interlaced transformers will often inject two to three times more surge current into the customer load than will noninterlaced transformers. Thus, the potential for customer load failure is greater. These are the same areas that would be expected to experience higher-than-normal incidence of lightning failures of noninterlaced transformers. Thus, we should concentrate on these areas to try to gather data on customer equipment failure due to lightning. We will continue to work in this area to help the industry define rational and realistic surge withstand requirements. Manuscript received November 6, 1986.

You might also like