THE LONG ISLAND SOUND SUBMARINE CABLE INTERCONNECTION

P. Gazzana-Priaroggia, Senior Member, J. Piscioneri, Member,S. Margolin, Member, IEEE

ABSTRACT

This paper discusses the design and installation of a 300 MVA, 12mile, 138 kV, high pressure, oil-filled submarine cable.

The use of self-contained, oil-filled cable for this crossing representsa major contribution in the field of cable technology. Thislong-desired intertie became a reality with the development of newlow-viscosity impregnants capable of coping with severe hydraulictransients as well as improved design, manufacture and installationtechniques.

Also discussed are the associated terminal facilities and systemconsiderations.

INTRODUCTION

The Long Island Interconnection is the longest oil-filled submarinecable crossing in the world. This venture, undertaken jointly byThe Connecticut Light and Power Company, an affiliate of NortheastUtilities, and the Long Island Lighting Company, links for the firsttime the facilities of these two companies. Constructed at a totalcost exceeding $10,000,000, it will provide for economic interchangeof power as well as added reliability and system stability.

Each company has assumed title to that portion of thepothead-to-pothead system lying within its state boundaries. Thecost of terminal facilities, rights-of-way and civil work on each shorehave been borne separately.,

This 12 mile, 138 kV crossing provides a direct 300 MVA interchangebetween the LILCo. and CL&P station facilities at Northport andNorwalk respectively. To geographically isolated Long Island, thisattractive intertie, until very recently, was not economicallyjustifiable. From a technical point of view, the problem of crossingthe depths of the Sound was discouraging.

This project was undertaken as a result of joint studies by bothcompanies. What brought it into economic focus was the high costof alternatives in providing power supply requirements for eachsystem. Technically, the choice as to the nature of the system hadto be made. The high cost of ac/dc terminals precluded the use of d-c.Three types of cable installations were considered (1) self-containedoil-filled cable; (2) pipe-type cable; and (3) solid dielectric cable.Solid dielectric cable was rejected because of skepticism concerningits performance at 138 kV. Pipe-type cable was under seriousconsideration. However, it was ultimately rejected for the followingreasons:

1. With a single pipe, complete loss of capacity would result froma single contingency.

2. Repair time would be extensive.

3. Difficulty in positively preventing water ingress at a faultlocation.

A seven conductor system (six conductors plus a spare) of singleoil-filled cables was chosen. Optimum reliability results from its usebecause repair to a damaged cable can be made quickly and withoutloss of rated capacity. Comprehensive marine surveys indicated thatthe Sound bottom was amenable to the use of a multi-conductorsystem.

The cable was designed, manufactured and installed on a turnkey

basis by Pirelli SPA of Milan, Italy and Pirelli Construction Companyof Eastleigh, England. With modern, high capacity, massimpregnating facilities, the manufacturer was able to produce cablein splice-free lengths up to 40 miles. This capability, and thedevelopment of extra fluid alkylates as cable impregnants were twoimportant factors contributing to the feasibility of thisinterconnection. Using these low-viscosity oils, the manufacturer wasable to substantially reduce the core diameter and yet guarantee (1)positive oil feed to cable insulation under all conditions of load,and (2) prevent water ingress during contingencies. Accordingly, thisintertie represents a significant contribution to the state of the art.

THE GENERAL SYSTEM

This project was the culmination of extensive joint planning.Encouraged by recent technical progress in the development ofsubmarine cables, studies began some 5 years ago between NortheastUtilities and the Long Island Lighting Company. These studiesinvestigated the pooling benefits and nature of this newinterconnection, as well as the type and ratings on associatedequipment that would be required at each terminal.

Northport, Long Island and Norwalk, Connecticut were the logicalterminals as this was the shortest distance between two largegenerating stations on the two systems. Therefore, the tie wouldact as an immediate backup for both generating stations. In addition,substantial transmission capacity already existed at these locationsin both systems.

The choice of a 138 kV, 300 MVA tie took into account costs ofseveral alternatives plus certain limitations of the existing systems.345 kV cable was not practical because LILCo. did not have thistransmission voltage on its system and the existing NU 345 kV systemwas remote from Norwalk Harbor.

A nominal 300 MVA rating was chosen for the tie because of theexit capacity of the Northport and Norwalk Harbor generatingstations and any increase in rating would have resulted in aninordinate increase in cost. 138 kV was chosen over 115 kV becauseof its higher capacity.

Figures 1 and 2 show, in general, the plan and profile of the cablesystem from Northport to Norwalk, the details of which will bediscussed later. The choice of seven cables (six plus a spare) overa four cable system (three plus a spare) was adopted mainly to takeadvantage of flexibility during contingencies. Fifty percent of fullpower would be transmissible even with two or more cables faulted.

With the cable system parameters fixed, a transient network analyzerstudy (TNA) was initiated to establish equipment designs andoperating procedures for the submarine interconnection.(1) Studiesof many contingencies and configurations indicated a need for a 25degree phase angle regulator controlling circulating power in theConnecticut - Long Island - New York loop which the newinterconnection would complete. It also showed the need of a 10%load tap changer on the 138/115 kV autotransformer which joinsthe different transmission voltage levels of New England and NewYork.

The large capacitive energy - about 100 MVAR - that will be storedin the cable and the need for transformer terminations at both endswere unique enough to warrant the study of transient overvoltagesin various switching situations. The TNA study also established thenecessary insulation coordination and lightning arrester requirements

1863

Ur

Itit Federal Channel

tNorwalk Harbor Generating Station (C.L.&P.)%0' Spacing between cablesN

4

351 DEPTH

submarine cables I x 600 kcmil

between cables

cable #1 cable #7

35' DEPTH

)ff Shore Oil Platform

-r= 2500'

LONG ISLAND

Fig. 1. Plan View of 138 kV Cable System

1864

Transition

,_tc11 'ti .coti.

-RO,

Fig. 2. Profile of 138 kV Cable System

at the terminals. Table No. 1 summarizes the station equipment,ratings and operating restrictions determined as a result of the TNAstudy.

TERMINATING FACILITIES

The cable terminating facilities consist of individual potheads for theseven cables, each having 2000/5 amp multi-ratio external bushingtransformers. The cables connect to a 300 MVA 138/115 kVautotransformer through six manually removable cable links and twomanual group operated 3 PST disconnect switches with keyinterlocked grounding switches. The spare cable connects to a sparecable bus through fixed cable link grounding switches. The cablelinks are removable, permitting the spare bus to be connected inplace of any one of the cables while de-energized. The switchingof the spare cable CT's into the relaying circuit is accomplished bycurrent shorting type test switches.

Relaying for the interconnection consists of line primary, backupand transfer trip protection, transformer primary and backupprotection, 115 kV transformer bus protection and breaker failureprotection. Primary protection consists of a high speed solid staterelaying package containing a dual comparer phase-comparisonsystem with wideband frequency-shift audio tonetransmitter-receiver. Backup line protection consists of a high speedelectro-mechanical relaying system which is a combination ofstep-distance phase relays with self-contained overcurrent faultdetectors and an instantaneous ground directional overcurrent relayarranged in a permissive overreaching transfer trip scheme. The relaysact as the transferred tripping devices keying the transmitter andthe permissive devices for local tripping in the event of a receivedtransferred trip signal.

A 6,000 MHz space diversity microwave system using single side bandmultiplex is used to transmit all relay, telemeter and voiceintelligence. The microwave system has a capacity for 420 channels.The channel/signaling rack is wired for 36 channels of which thereare initially 9 channels in use. Two channels are allocated to voicecommunication, five are used for protective relaying, two are usedfor telemetering and two are spares.

SYSTEM DETAILS AND DESIGN

In the design of the cables, many factors peculiar to the physicalnature of the routes had to be taken into account. Northport andNorwalk are popular boating areas and to avoid damage from anchors,imbedment of the cables in water less than 35 ft deep was necessary.For the main crossing in deeper water, the cables lie directly onthe bottom. Because of the shallow water between Sheffield Islandand Norwalk, continuous embedment was required over an extendeddistance (about 8,000 ft). This fact, and the higher ambient watertemperature (100C average) in the inter-island area, posed a thermalproblem necessitating the use of larger conductors to avoid derating

system capacity. This was resolved by using 600 kcmil for the maineleven mile crossing and 700 kcmil cable in the inter-island area ofNorwalk and for the shore portions at both terminals. SheffieldIsland was a convenient location to construct the transition splicesbetween conductors. At Northport, the land/sea splices were madeat the shoreline. Shoreline splices were not necessary at Norwalkbecause of the use of 700 kcmil submarine cable from SheffieldIsland in to the terminals. The only requirement at the Norwalkshore was to strip the armor off the land portion to the potheadsand single-point bond the ends at the shoreline.

Figure 3 shows the cross sections of the 600 and 700 kcmil cables.Detailed specifications of the cable components are listed in theappendix. Shown in Table No. 2 are the physical, electrical andthermal data of the installation.

THE HYDRAULIC SYSTEM

The cable impregnant, a synthetic extra-fluid alkylate, is a benzenederivative with a C-9 hydrocarbon chain. It was introduced by thecable manufacturer for use with oil-filled submarine cable of longlengths. It behaves like a mineral oil, yet it has the advantage ofa lower viscosity and better electric stability. The fluid is colorlessand has a favorable evaporation rate so that leakage in controlledamounts due to failure will not constitute an environmental hazard.

In the cable, the oil is normally maintained between 170 and 270psig by a custom pressure system designed and fabricated by JeromeUTE, Inc. of Farmingdale, New York. A pump house is locatedat each terminal consisting of a 10,000 gallon storage tank, threeoil pressure pumps, three vacuum pumps, an annunciator andnecessary control equipment. Normally, at each end all seven cablesare served by pressure pumps, one of which starts automatically whenpressure drops below 190 psi. If pressure continues to drop, a secondpump will come in service at 170 psi and at 150 psi, low pressurealarm indication will occur. Action will then be taken to manuallystart a third emergency pump at a lower pressure after whichinvestigation will begin to determine the source of pressure loss.Normal pump operation is terminated after one hour. Pump reliefvalves are set to limit pressure at 250 psi.

Design of the pump units posed some difficult problems not normallyencountered with pipe-type units. The extremely low viscosity ofnonylbenzene oil makes it ineffective as a pump lubricant. Also,the oil has a powerful solvent action requiring a clean system toavoid contamination. The oil storage tanks must be maintained undera constant vacuUm seal to keep the oil in a degasified condition.Nitrogen, customarily used for pipe-type cable systems, could notbe used as an oil seal because it would readily diffuse into the oil.

The oil pumping units are designed merely to maintain a staticpressure and therefore, there is to be no intentional oil flow.However, drastic changes in load will produce pressure differentials

1865

COPPER CONDUCTOR

PAPER INSULATION

LEAD ALLOY SHEATH

POLYETHYLENE _SHEATH

TREATED JUTEBEDDING

Fig. 3. Cross Section of 138

along the cables. For example, a sudden switching off of full load(300 MVA) would result in an approximate 63 psig reduction inpressure at the center of the cable below terminal pressure. And,as might be expected, the switching on of full load at an ambienttemperature of 00C would cause an overpressure at the midpointof the cable system of approximately 70 psig. Figure 4 graphicallyportrays the pressure distribution of the oil in the cable for theseconditions of load variation.

These maximum and minimum pressures occur after an interval ofapproximately ten and twenty minutes respectively, from the timeof switching. The mechanical stress which ensues from thesetransients had to be taken into account in the design of cablereinforcement tapes, especially in the 600 kcmil crossing lengths.

Were it not for the oil feed problem during a complete severance,a lower pressure would have been adequate. The most severecontingency would be severance near one end in which a minimummanifold pressure of 190 psig would be required to prevent wateringress (see Fig. 5). A pumping unit was indicated for thisapplication, not only because of the high pressure levels required,but also to adequately cope with oil loss if the cable became cut.

A noteworthy advantage to the use of a small diameter oil ductis the surface tension phenomenon which occurs if the cable is cut.Very high surface tension on the "meniscus" separating water andoil favorably reduces water ingress tending to localize watermigration. Laboratory experiments have shown that for not toolarge diameters of the oil channel this surface tension very easilyprevents the mixing of the two liquids.

Reservoir capacity was designed to guarantee (1) automatic oil supply

kV Oil-Filled Submarine Cable

p i

250

200-

'50.

100

so

0

100

tOO

0 10e000 | |oboo soboo Ft.t0000 40000 60000

Fig. 4. Transient Pressure Distribution for Sudden Switching of NormalLoad (630 A)

1866

SCREEN

NOMINAL THICKNESS DIMTRNOMINAL

|

psi250-

200.

150.

100-

50S

00

100-

200-Ft. Ft.6 10000 30000 | 50000

20000 40000 60 00

Fig. 5. Transient Pressure Distribiution for Cable Severance at VariousPoints

at full pressure to a damaged cable during the interval of the thermaltransient caused by load interruption; (2) sufficient time to manuallyperform proper valving; and (3) oil supply at reduced pressure forlong periods of time during which damage can be evaluated and the"capping operation" can be mobilized and performed.

MANUFACTURE OF CABLE

The cable was manufactured at the Pirelli SPA Arco Felice Worksnear Naples, Italy. This relatively new facility has been designedand constructed with the capability of producing long, continuouslengths of high voltage submarine cable. Although the first submarinecable manufactured was paper insulated and impregnated with a-viscous compound (the Sardinia-Corsica-Italy 200 kV d-c cable), thefacilities were designed also for the production of oil-filled cables.

Essentially the machinery installed at Arco Felice is similar to, yetmuch larger than, conventional units used in the manufacture ofoil-filled cables by the mass impregnation process. The productionof long lengths of cable involves considerations not only of dimensionbut of technique imposed by extensive operation under stringentrequirements. Certain problems, occurring in various productionstages, had to be resolved prior to the successful manufacture in1966 of the 66 kV oil-filled cable linking Corfu Island with mainlandGreece and, in 1968, of the Long Island Cable.

Copper hardness must be selected to (1) guarantee adequate tensilestrength during laying and (2) allow for electrical welding duringstranding. Welded joints, where required, must be artificiallyhardened using special tooling. As can be seen in Figure 3, the singlelayer of 7 copper segments are interlocked for self-support duringmanufacture, shipping and installation.

To evacuate, dry and impregnate the cable in long lengths, a specialannular tank (15 meters in diameter, 200 cubic meters capacity) wasdesigned and constructed. The tank is able to rotate during coilingoperations so that the cable will not be subjected to torsion.The method of "syphon" lead sheathing has been long established

in the cable industry. However, in this case the exceptionaldimensions of the syphon and impregnating tank presentedsubstantial unknowns. Large amounts of degasified oil werenecessary (more than 8 cubic meters per hour) through the syphonin a direction opposite to the cable feed. This was to create anoil ambient around the cable having as low a gas saturation levelas possible. The volumes of oil needed and the lack of lubricityof the extremely fluid oil called for special circulating pump design.

The entire impregnation and lead extrusion for this project wasaccomplished in just two runs. The first consisted of 30 miles of600 kcmil cable and 14 miles of 700 kcmil cable. In the secondrun, a length of 40 miles of 600 kcmil cable was processed.Continuous lead extrusion for cables of such length and the use ofvery thin oil created thermal conditions much more critical than hadbeen encountered in the past. This required special developmentwork.

It is noteworthy that in manufacture, only three accidentalinterruptions occurred during the lead sheathing process. Eachdiscontinuity required the use of a factory repair flexible joint.

GROUNDING

The lead sheath as well as the aluminum armoring on the submarineportions are bonded together (but separately) at the two shore lines.(See Fig. 6) Therefore, in the lead sheath and in the armoring,circulating currents are induced. The sum of the induced currentsin the sheath, armor and the surrounding sea water (approximately10% of the total) is practically equal and opposite to the currentin the copper conductor. Each cable therefore, acts as a coaxialcircuit and the unusually low reactance results in an X/R ratio ofless than one.

To prevent excessive voltage gradient across the polyethylene sheath,additional high resistance grounding connections for the lead sheathwere made along each cable at two mile intervals.(2) This wasaccomplished by bringing out copper bonding wires through thepolyethylene jacket. Using a special helical wrapping procedure withself-amalgamating polyethylene tapes, the connection is madewatertight. The bonding wires are terminated with a tinned copperalloyed tape under the jute bedding, but there is no metal-to-metalcontact between this tape and the aluminum alloy armor wires. Thecontact is purely electro-chemical in nature using the salt water asan electrolyte.

As the cable system approaches the Northport shore, the armoringsare bonded at a single point using special bonding boxes. Bondingof the lead sheath is accomplished at the land/sea joints. The leadand aluminum have also been separately bonded at Sheffield Island.The reason for separate bonding is to avoid metal contact whichwould complete a metallic circuit tending to produce corrosion inthe cable grounding connections.

The cable grounding connections are designed to carry a current of6 kA for two seconds, which is sufficient -to cope with travelingor short circuit waves coming from sources external to the cablecircuit.

In the land section of Northport, the lead sheaths are bondedtogether only at the land/sea joints and grounded by connection tothe station ground bed. The lead sheaths are open circuited at thepothead bases to avoid circulating currents which would limit cablecurrent-carrying capacity.

At Norwalk, where there are no land/sea splices, special bondingboxes are employed at the shore and filled with a bituminouscompound. Here, separate bonding connections for the lead andfor the armor are provided. The lead sheaths are bonded at a singlepoint and connected directly into the station ground. The bondedarmor is not connected to the station ground, but relies on the

1867

NORWALK HARBOR GENERATING STATION C.L.P.CO.

LONG ISLAND SOUND

Fig. 6. Grounding Arrangement

presence of the salt water to effect a natural earthing.

In Norwalk as well as in Northport, the lead sheaths are opencircuited at the potheads, which means that electrically, the potheadbase is floating above ground. The induced voltage at the floatingconnection under full load conditions is approximately 70 volts atNorthport and 25 volts in Norwalk.

In the event of symmetrical short circuit of approximately 17.5 kA,the above values of voltage become 2,000 volts and 700 voltsrespectively. Both in Northport and in Norwalk, the bases of theinsulated potheads are connected to protective gap type surgediverters, which are tied into the station ground. The requirementsfor these surge diverters are as follows (1) to discharge any travelingsecondary waves induced in the lead sheath by primary lightningwaves traveling along the overhead lines in the cable conductors; (2)to discharge the currents induced in the lead sheath by primaryswitching surges traveling along the cable conductors. From thestudy of the oscillograms and from results obtained from the TNAstudy, these primary switching surges appear to be similar tosinusoidal damped waves having a natural frequency of about 250Hz and a peak value of about 220 kV with a duration of 30milliseconds. In this case, the induced currents on the lead sheathare evaluated at approximately 2.7 kA for 30 milliseconds; (3) toextinguish the discharge arc even when the permanent inducedvoltage of the base of the potheads reaches the maximum value of130 volts corresponding to the maximum emergency current of 1100amps; and (4) to limit the maximum discharge level which can betaken by the polyethylene jacket with no risk of puncture. Thislevel is specified as 3 kV and offers a very large margin of safetywith respect to the testing level of the jackets (30 kVa-c spark testduring manufacture).To limit the overvoltages which may be built up on the groundingpoints of the lead sheaths at both shores, these points are connectedto the main station grounds using bare copper conductors (500MCM). These ground conductors are transposed in between thecables of each circuit in order to minimize the voltage induced bythe currents of positive sequence flowing in the cable conductors.

The station grounds in both stations have a resistance of less than0.3 ohms and are cathodically protected by means of rectifierssupplying a minus 1.6 volt bias.

INSTALLATION OF CABLES

Prior to the cable installation, three separate marine surveys wereconducted. The first one was intended primarily for a pipe-type cable,but much of the information was -valid for use with any type ofinstallation. Four separate routes were investigated as to thefollowing:

1. Bottom contour and sub-bottom profile.

2. Current metering to determine magnitude and direction ofbottom current.

3. Measure of water temperature at different depths and atdifferent times of the year.

4. Inspection of the sea bottom by divers.

5. Core sampling.

When the decision was made to use seven single oil-filled cables,several factors dictated that the general area of their route systemwould be remote from the four survey routes previously studied.Thus, a second marine survey was undertaken to:

1. Specifically set the seven cable routes.

2. Measure the route distances precisely to determine the cableshipping lengths.

3. Inspect the sea sub-bottom to a depth of 5 feet in theembedment area. This with particular attention to bedrock andlarge boulders using seismic equipment and radio-navigationaldevices.

1868

NORT.HPORT POWER STATION L.I.L.CO.

Just prior to the start of the excavation, a third survey was madeto augment the previously accumulated bottom information. Thisfinal survey attempted to define along each route where boulders,rocks and ledge (if any) would be encountered. The routes werewalked by divers and hand probings, using water lances, were madeat 25 foot intervals. This data helped to establish the type ofequipment and method of excavation to be used. However, theinformation was general in nature and the precise amount of rockto be removed still remained unknown.

Three techniques were considered to embed the cable in areas lessthan 35 feet in depth:

1. Plowing and jetting simultaneously while laying.2. Laying the cable first and jetting and plowing in later.3. Pretrenching, laying and then backfilling.

Method 1 was rejected mainly because although acceptable for plasticand rubber type cable, it might have placed undue mechanical stresson an oil-filled lead sheathed cable. Method 2 would requirecomplete prior removal of all rocks without any chance of lateremedy after laying. Both methods would be slow and wouldincrease the exposure to sea action and boating hazards for extendedperiods (months). In spite of some anticipated minor drawbacks,the third method - pretrenching - was adopted. It was the onlypractical method to guarantee safe handling of the cable with aminimum of sea exposure. Also, a good portion of the requiredpretrenching would be in fact accomplished by the preliminary rockremoval phase (necessary in any case).Various modes of excavation were employed depending upon depthand nature of the bottom material. In areas of relatively soft,rock-free bottom and where there was sufficient flotation, a suctiondredge was used. Areas of rock and hardpan required use of a clamshell. In very shallow areas approaching the shore lines, the cablewas hand-jetted after laying.The theory of pretrenching called for the excavation of one trenchat a time. It was reasoned that time would be of the essence andthat each cable would have to be quickly installed before anypremature backfilling took place by sea action. With this in mind,it was planned to excavate the first trench using dredge and/or clamshell, side casting removed material where possible. Using a speciallybaffled discharge tube, the dredge would backfill the first trenchwhile proceeding to excavate the second trench. This would bepossible only where routes were parallel.

In some cases, the suction dredge in its subsequent route was unableto accomplish adequate backfill of the trench previously dredged.A combination of light silty material and bottom currents resultedin the disbursement of dredge discharge material over a wider areaso that in places, very little material actually fell into the trenches.Consequently, in those areas where backfilling was not properlyeffected by either dredge discharge or natural wave motion, obtaining3 feet of bottom cover over the cable posed a problem. In depthsless than 6 feet MLW, the Corps-of Engineers required transportationof excavated material away. Also, in very rocky areas, the excavatedmaterial was carried away to assigned dumping areas. Thus, afterthe cables were laid, there was a deficiency in available material insome areas with which to properly backfill the trenches. Thissituation was satisfactorily remedied as follows:

1. In silty and sandy areas, the wide slope trenches werebackfilled by means of a heavy steel beam towed along thebottom by a vessel.

2. In rocky areas and in those places where the beam methodwas not effective, the cable was covered with bags of portlandcement.

3. Hand jetting in those areas where it was deemed advisable bythe contractor.

Two procedures were used to lay the cable. In the inter-island area,the cables were payed from a barge, floated by sections from Norwalkto Sheffield Island, and then guided into the trenches by divers. Thebarge was kedged along with a system of four anchors using a lowintensity laser beam to keep on station.

This same procedure was used in the Sound crossings where the cablewas to be embedded (less than 35 feet deep). However, in deeperwater, the cable was laid directly on the bottom, using customdesigned tensioning machinery having a maximum braking capacityof 2 1/2 tons. The laying barge was propelled by a tug at abouttwo knots speed. Transverse control was maintained by two smallvessels secured at right angles to the cable barge responding to acontinuous electronic positioning system readout.

From the time the cable left the factory until the permanent oilunits were in service, constant oil pressure was maintained byportable oil tanks. These reservoirs were of the double chamber,pneumatic diaphragm type. During laying, however, portablepumping units were used capable of supplying .3 cubic meters perhour at a variable pressure. During the cooling transient occurringat the time of laying (assuming that the cable temperature wassignificantly higher than the water temperature) the feeding pressureof the pumping unit was continuously increased. This was inaccordance with calculations based on laying speed and temperaturedrop. This pressure increase had to match three requirements: 1- The relative oil pressure in the cable should never become negative;2 - The oil pressure in the cable while bending on the wheels ofthe laying machine should never exceed a safe limit; and 3 - Theflow meter installed aboard the laying barge should always indicatea very small but constant oil flow from shore to vessel. As anexample, during the laying across the Sound, the starting pressureof the pumping unit was 25 psi and the final pressure was 85 psi.

FACTORY TESTS

All of the 138 kV cable was constructed with a wall thicknessreduction of approximately 23% from the standard AEIC 0.505 in.Testing in accordance with AEIC specifications, as was done at thefactory, therefore imposed a degree of stringency greater thannormal. To prove the watertight integrity of the cables, other testswere performed as follows:

1. Internal Pressure Tests

Ten foot samples each of 600 and 700 kcmil submarine cablewere pressure tested internally for a 24 hour period at oilpressures of 600 and 400 psi, respectively. The pressure wasincreased 14 psi per minute. There was evidence of oil leakageat 1320 psi for the 700 kcmil specimen and at 1440 psi forthe 600 kcmil piece.

2. External Pressure Tests

One 15 foot sample of 600 kcmil submarine cable includingone grounding connection was pressurized externally for oneweek at ambient temperature to 135 psi in salt water. Nointernal pressure was applied to the cable. After the test, thegrounding connection was stripped off and the absence ofwater under the polyethylene jacket was ascertained by meansof a cobalt chloride test.

Another 15 foot sample was tested and inspected in a similarmanner except that the external salt water pressure was appliedaccording to the following schedule:

8 hours16 hours

at 118 psino pressure

1869

+ 12 Ft

Fig. 7. Flexible Repair Joint

8 hours at 294 psi16 hours no pressure8 hours at 118 psi

40 hours no pressure8 hours at 294 psi16 hours no pressure8 hours at 118 psi16 hours no pressure8 hours at 294 psi16 hours no pressure

FLEXIBLE REPAIR JOINTS

Not withstanding the assumption that "in principle" each of theseven cables from pothead-to-pothead should be splice-free, it wasrecognized that a repair joint must be available in case of emergencyeither during manufacture, laying or for any repair in future service.Such a repair joint should be flexible, as to allow the easiest andsafest means of handling the cable.

The flexible joint designed by the manufacturer for this crossing isshown in Figure 7. The ferrule is of the semi-flush compressiontype. The cable heads have a very long taper. The insulation isrestored tape by tape. The joint sleeve is welded to the cable sheathand makes a "tight fit." The reinforcement tapes are reapplied insuch a manner as to insure equal or better strength properties asthose of the original tape. The polyethylene jacket is also restoredby using self-amalgamating polyethylene tapes.

After restoring the jute bedding, the wire armor is reconstituted byusing specially designed screw connectors. The diameter of thefinished joint is practically identical to the original cable diameter.In order to insure that the highest dielectric quality is achieved inthe repair joint, the restored insulation is hand applied in a verylow humidity, air conditioned enclosure and then it is filled witha degasified oil under a vacuum seal.

DAMAGE AND REPAIR OF CABLES

Before the project's completion, the No. 7 cable (the most easterly)sustained damage from two separate accidents. The first occurredabout 2,000 feet north of Sheffield Island when a kedging anchorassociated with the dredging of another cable trench made contactwith the previously laid No. 7 cable. Although the conductor wasnot severed, the armor, polyethylene jacket, and lead sheath werebadly damaged. Oil leakage was kept under control by portablepumping units until the day after the accident when the damagedportion of the cable was cut away and the two ends were temporarilycapped.

The second accident occurred about 2 1/2 miles north of Northportin about 50 feet of water. The first indication of trouble was alow pressure alarm operation at the terminals. At the time, thecables were being temporarily fed by portable pressure tanks.

Emergency steps were taken immediately to insure an adequatesupply of oil at the two ends to avoid the risk of entrance of waterinto the cable. The magnitude of pressure drop (from 25 to 20psi) indicated that a cut in the cable had occurred.

The next day it was learned that an oil tanker, maneuvering withher anchors, had daught the No. 7 cable. The vessel was withouta tug due to a strike and, in rough seas, was approaching an oilfacility off Northport when the mishap occurred. Investigation bydivers revealed a cut in two places (only one severance). Withina period of 10 days, the damage was located and both ends wereraised, capped and resubmerged. The total rate of oil loss fromboth ends during this contingency period was approximately 120gallons per day which compared very favorably with calculatedvalues.

The length of damaged cable removed included the region of visualdamage plus an "insurance margin" in the event of brokenreinforcement tapes. The repair segment needed was in excess ofthe maximum available spare length (1800 ft), so two pieces werejoined together. It is felt that the disadvantage of this extra jointwas offset by the opportunity to replace a segment sufficiently long(3000 ft) to confidently insure that no incipient damage remained.All of the four required repair joints were of the flexible typepreviously discussed.

TABLE 1

I. Station FacilitiesAt Northport 300 MVA, 25 degree phase shifting

transformer.At Norwalk 300 MVA, 138/115 kV, autotransformer

with 10 percent tap changing under load.Oil Circuit Breakers Rated as Follows:At Norwalk 115 kV, 2000 amp, 10,000 MVA 550 BILAt Northport 138 kV, 2000 amp, 15,000 MVA 650 BIL

Lightning Arresters:At Northport 120 kV impulse with 168 kV switching

surge characteristics.At Norwalk 138 kV Bus 14.4 kV sparkover and

168 kV reseal.115 kV Bus 108 kV sparkover and

reseal value with IRcharacteristics of 120 kVarresters.

11. Operation Restrictions1. Phase-shifter, autotransformer and cable must be

energized together and only from Northport.2. Phase-shifter must be on 0 phase-shift when energizing.3. Planned de-energization accomplished from Northport.

1870

APPENDIX ACable Specifications

PhysicalDimensions

Single Core Cables RequiredSpare CableMax Laying DepthCable Separation in Deep WaterEmbedment DepthLand Cable Arrangement:Burial

ConfigurationSeparationBurial DepthProtection

II. ElectricalRated VoltageNominal PowerImpulse Insulation Level (BIL)3-Phase Circuit ConfigurationRated Current Per PhaseFrequency3-Phase Short CircuitSymmetric Currenta) From Norwalkb) From Northport

Refer to Fig. 3 and theAppendix61200 ft900 ft3 ft

,Direct (in thermal sandlayers)Flat40 in40 inConcrete Slabs

138 kV300 MVA650 kVGrounded Neutral630 A60 Hz

17.5 kA14.0 kA

Daily Loss FactorNormal 63%Emergency 50%

Voltage Level on Spare Cable 50 kV d-c(See Appendix B for Emergency Ampacities)

III. ThermalAmbient Temperatures (Deg. C)

Long Island SoundSheffield I. to NorwalkLandAir

Thermal Resistivity of SoilSound BottomOn Land (Design)

Min.0

105

-10

Max.

25352535

700C cm/w90C cm/w

61S

ConductorCopper, seven segment, hollow-coreInternal diameter-milsExternal diameter-milsConductor ShieldingInsulationPaper-milsInsulation ShieldingSimiconducting paper and fabriccopper tape, woven-milsI mpregnantNonylbenzeneViscosity range at 20C-centipoiseViscosity range at 85C-centipoiseLead Alloy Sheath-milsReinforcement3 layers copper alloy tapewith 1% cadmium-milsJacketExtruded polyethylene-milsBeddingTreated jute, thickness-milsArmorAluminum alloy-0.5%Silicon,0.5% Magnesium-wirediameter-milsDiameter over textilecover-milsOverall Cable Diameter-milsWeight of Completed CableApproximate lbs/ft

i00 Kcmil 700 Kcmilubmarine Submarine

6651030

6651080

Carbon black tapes

388

20

5-6

1.4-1.888.5

5.9

115

80

158

2052960

8

384

20

5-6

1.4-1.890.5

4.0

190

80

158

2053150

8.5Note: The Northport land cable is identical to 700 kcmil

submarine cable except for the lack of armoring.

ConditionNormal300 hour100 hour24 hour8 hour20 min.

APPENDIX BSystem Current Ratings in Amperes

Summer630660685700750

1,000

Winter630770785800850

1,jooThe above system emergency ratings are limited by the fourteen cablejoints. These joints can be replaced with lower temperature jointsto increase ratings if necessary.

ACKNOWLEDGEMENT

The authors wish to acknowledge the valuable work of the manypeople who took part in the research, development and design ofthis cable, as well as those responsible for the organization andinstallation of this difficult interconnection. They are grateful to theutilities and to the manufacturers for the permission to publish thispaper.

1.

REFERENCES

B. M. Fox, A. M. Madsen, Surge Study Sets Guides for LI.Sound Intertie, ELECTRICAL WORLD, May 12, 1969, PP.33-36.

2. G. Maschio, E. Occhini, Overvoltages on Anti-CorrosionSheaths of High-Voltage Cables with Particular Reference toLong Submarine Cables, CIGRE 1964, Report No. 224.

1871

TABLE 2Cable Data

DiscussionR. C. Waldron (The Okonite Company, Ramsey, N. J. 07446): Thecable supplied for this installation had aluminum alloy armor wireshaving 0.5% silicone and 0.5% magnesium. The individual wires on thiscable were individually covered with a textile wrap whereas our cableswere designed to avoid any contact with fibrous material. We have usedaluminum alloy 5052 containing 2.5% magnesium and 0.25% chromiumfor submarine cable armor. The cable described in our Paper (1) on thePuget Sound Cable Crossing is typical of our design.

At the time the Puget Sound cable was installed, there was con-siderable discussion as to the desirability of covering the armor wiresand, to study this point, a sample having four different coverings of thearmor wires was buried at one terminal of the crossing at a point be-tween low and high water.

These various coverings were on one continuous length of cable.This length was buried in 1962 and the armor wires connected to theground connection of the working cables. The cable ends were sealedto prevent any electrolytic action between the lead sheath and thearmor wires.

This sample was removed in 1970, after approximately eight years,and the various sections were examined for corrosion. The variouscoverings and the conditions found are described below with photo-graphs of typical sections of the armor wires.

1. Bare aluminum alloy wire (5052) with a separator of Polyestertape between the layer of wires and the neoprene covering over the leadsheath. This is the construction used for the cable crossing.

Some surface corrosion which seemed to be at points where mudor fine sand adhered to the surface in a small area. Average depth about5 mils with one area 14 mils in depth.

2. The same as "I", but slushed with asphalt.This was similar to the bare aluminum sample, as far as the area of

corrosion was concerned, but the depth of pitting was less, the maxi-mum measured being about 3 mils. The asphalt coating provided protec-tion.

3. Same as "2" with the addition of a layer of jute directly underthe armor.

General corrosion to 5 to 9 mils in depth. These seemed to belimited to the inner side of the armor in contact with the jute bedding.

4. Same as "3" but with an additional layer of jute over thearmor.

Jute Bedding and Covering Slushed. This corrosion was through-out resulting in a reduction in diameter in some areas to 127 mils from

* the original diameter of 203 mils.

Manuscript received February 12, 1971.

This study would indicate that slushing with an adhesive protec-tive coating prevents serious corrosion but the presence of fibrousmaterial even with slushing promotes corrosion. A more completedescription of the textile covering on the armor wires and the reason forthis covering and field experience with it would be desirable. We antici-pate greater use of aluminum alloy armoring of submarine cables andcorrosion of the armor could result in localized heating.

REFERENCE[I ] R. C. Waldron, "115 kV Submarine Cable Crossing of Puget

Sound" IEEE Trans. Power Apparatus and Systems Vol. 84,pp. 746-755, Sept. 1965.

C. T. Hatcher (1409 Wren Lane, West Chester, Pa. 19380): It isstated in the paper that the cable impregnant is a synthetic extra-fluidalkylate which is used by the manufacturer for long lengths of sub-marine cables and it is my understanding that this impregnant has onlybeen used in the Pirelli plant in Italy.

In the paper it is also stated that two of the submarine cables weredamaged by ships' anchors during the period of construction resultingin leakage of the impregnant. The leakage in one incident was approxi-mately 120 gallons per day. It is understood that at each terminal endthere is a storage. supply of 10,000 gallons for supplying all of thecables. If this amount was used to supply a leak of the above value itwould be sufficient for approximately 80 days. The damages occurredduring the period when construction forces were at the site of thecrossing and, therefore, it is reasonable to assume that the locations ofthe breaks were found quicker than if they had occurred at a time whenforces were not at the site. Having been associated with the installationand operation of the first 138KV low pressure oil-filled cable installedin conduit on the system of the Consolidated Edison Company ofNew York, I am familiar with the length of time required in some casesto locate an oil leak.

My question is, "assuming that a fast oil leak occurs and that it isnot readily found, is it possible to obtain the special impregnant in theUnited States or is it only available in Italy? If Italy is the only sourcecan it be supplied from that distance in such quantities as requiredto provide a sufficient flow of impregnant and how is it anticipatedthat it will be shipped?"

Manuscript received February 17, 1971.

Herman Halperin (275 Santa Margarita, Menlo Park, Calif.): A veryimpressive series of ideas and developments on cable and accessoriesare presented in this paper, which range from the fields of research,design, and manufacturing to installation and operation. My points andquestions are as follows:

1. After the initial installation of oil-filled, self-contained cablein 1926-7, a much thinner oil was adopted in this country in 1928,which had a pour point of -45C. What is the pour point for the newlydeveloped very thin oil described in the paper?

2. Regarding the authors' point that their cable insulation thick-ness of about 385 mils for 138 kv cable is materially lesslhan the AEICvalue of 505 mils, I indicated in my 1956 AIEE paper(l) that a thick-ness of 460 mils was feasible for low-pressure cable as made some 18years ago to operate normally at about three psig or more. The LongIsland Sound cable will normally operate at very high pressures, i.e., at170 psig or more. 'In the 1948 CIGRE paper(2) by M. Domenach, hepresented test data for oil-filled cable as then made, which indicated anincrease in unit dielectric strength of some 30% was obtained by in-creasing the internal pressure to the normal values employed by theauthors. I appreciate that one reason for the very high pressure is toavoid water migration into the cable in case of damage. However, itwould be of interest to learn the authors' ideas as to the increase in

Manuscript received February 22, 1971.1872

unit dielectric strength when the pressure goes from, say, 3 to 170psig.

3. Experiences with 138 kv low-pressure oil-filled lines in Chicagoeach up to 18 miles in length have indicated a significant change inamperes between the two ends of a line. Assuming that the 12-miledouble-circuit line is loaded to its normal limit of 300 mva, it would beof interest to learn (a) the change in amperes between the two terminalsand (b) the approximate load delivered in mw.

4. In connection with the effect of switching surges on creatingsurges across the polyethylene jacket on the cable, it is stated, in effect,in the paper that 115 mils of polyethylene can withstand the maximumexpected discharge level of only 3 kv. As the polyethylene can with-stand much higher transient voltages and as the surge level seems to bedetermined by the surge diverter, I wonder whether the polyethylenethickness was set by mechanical requirements.

5. Referring to the use of concrete slabs that were placed over theland portions of the cable, a brief description of the slabs and theirvertical spacing and information on the indicated beneficial value intheir use outside the United States would be of interest.

6. I know of three installations of high-voltage submarine cableof different types where failures occurred over the past 30 years due tooverheating of portions installed in mud in the bottom of the waters.The utilities involved were surprised that the material around the cablecould dry out enough to result in failures. Referring to the 69 kv pipe-type cable installed in the Pacific Ocean off Long Beach, Calif. a fewyears ago, thermocouples were installed on some of the pipe at myrecommendation. Were any thermocouples installed by the authors?

REFERENCES[1] Experimental 138-Kv Cable and Accessories, Herman Halperin.

AIEE Transactions, Vol. 75, Part III, 1956, 348-366.[2] Cables with Reduced Insulation Thickness of Insulant Under Pres-

sure, M. Domenach. CIGRE 1948 Session, paper 214.

P. Gazzana-Priaroggia, J. Piscioneri, and S. Margolin: In reply to Mr.Hatcher's discussion I should like to make the following observations:

The question of oil supply in case of damage was debated at lengthbut then it was concluded that in case of urgency the special oil couldbe easily transported by air from Italy.

Moreover the same oil can also be produced in the United Statesprovided the amount required is sufficient to justify special production.

Manuscript received April 12, 1971.

It must also be pointed out that if an oil leak is difficult to locatethis means that the leak is very small and consequently can be easilyfed without any urgency of repair.

In reply to Mr. Waldron's discussion I should like to make thefollowing observations:

The type of aluminum alloy wires used for the Long Island cableswas tested in laboratory in salt water for a long time (many months).During manufacture the wires were slushed with coal-tar and thenwrapped individually with coal-tar-impregnated jute textile-tape. Nochemical corrosion was found on the tested specimens.

The wires were not tested for electrolitic corrosion caused bycurrents of any type flowing from water to wires or viceversa; thiswould certainly be detrimental to aluminum alloy armouring. Alumi-num alloy armouring was used in several occasions in the past forunderwater cables and found to be entirely reliable.

In reply to Mr. Halperin's discussion, I should like to make thefollowing observations:

1. Pour point of the oil used for the cable is-750C.2. The cable operates normally at a pressure comprised between

190 and 250 psig and even during cooling transients never reachespressures lower than 130 psig.

Therefore its dielectric strength in normal service is always verymuch higher than that of a low pressure oil filled cable. The increase ofdielectric strength at 60 Hz from 3 psig to 170 psig can be evaluatedof the order of 50%.

However the insulation thickness was designed making referenceto the electrical gradients usually adopted in Europe for low pressurecables.

3. Assuming the 12 mile double circuit line loaded at its end atthe normal limit of 300 MVA and supposing cos 0 = 0.9, the change inamperes between the two terminals is approx. 60 A and the approxi-mate apparent power at the sending end of the circuit is 280 MVA.

4. We agree that the thickness of the polyethylene jacket wasmainly set by mechanical requirements, but one should not disregardthat the statistical dispersion of dielectric strength results on poly-ethylene jackets is such that the same stressing as accepted for cableP.E. insulation cannot be extended to P.E. jackets but should be muchlower.

5. The dimensions of the concrete slabs are 14"x28" and thevertical spacing over the cables was 8".

The protection provided by the use of these slabs is supported byat least 15 years excellent service experience in Europe.

6. Thermocouples were installed in the interisland section of theroute at the request of the Utilities where the cables were embeddedin sand.

OVERLOAD CLASSIFICATION OF SECONDARY NETWORK CABLES

MODIFIED TEST PROCEDURE

S. P. Lamberton A. L. McKean

Member - IEEECincinnati Gas& Electric Co.

Cincinnati, Ohio

Member IEEEGeneral CableCorporation

Bayonne, N. J.

Fellow IEEEPhelps Dodge

Cable & Wire Co.Yonkers, N. Y.

Member IEEEBoston EdisonCompany

Boston, Mass.

Abstract-This paper describes a new test procedure for determin-ing the damage threshold temperature for secondary network cables.Present network cables generally have a maximum damage thresholdtemperature of 260C and are protected by class L260 network

Paper 71 TP 33-PWR, recommended and approved by the Insulated Con-ductors Committee of the IEEE Power Engineering Society for presentation at theIEEE Winter Power Meeting, New York, N.Y., January 31-February 5, 1971.Manuscript submitted September 14, 1970; made available for printing November16, 1970.

limiters. The new procedure is considered suitable for evaluating cabledesigns for damage threshold temperatures up to 340C, and adaptablefor higher temperatures by increasing the rate of cooling of the conduit.

An earlier AIEE test procedure (1955), never adopted as standard,relied primarily on visual examination of the cable condition, aftercurrent overload in free air, as the principal means of determiningthreshold of damage temperature. The new procedure incorporates twomajor changes: (1) the cable is current-loaded in a metal conduit, and

1873

C. H. Hunter T. J. Zazulak