Surge Protection on Open Ended MV Cable 2

[KJvR]

I would appreciate any comments on the use of surge protection on open ended cables. I am doing a MV (11 kV) substation design connected to ring feeders that will be split during normal operating conditions. Long lengths of cables (up to 1500m) might be energised but not connected to a busbar/load. I should add that series reactors are used in the network (main incoming sub) to limit the fault level.

I have talked to an experienced engineers that know of cable failures on open ended cables. He reason that these failures are due to reflection of switching surges causing doubling of the amplitude of the surge and insulation failures. This make sense but I have never seen a MV cable installation with surge arrestors installed inside the metal clad switchgear panels (only motor feeders). I have the following querries:
(a) Is the installation of surge arrestors on a cable distribution system common practise (esp with XLPE cables)?
(b) Is long or short cable lengths worst?
(c) Where should you install the surge arrestors?
(d) Should I use normal Metal Oxide Surge arrestors?
Any other comments regarding this will be appreciated.

Thanks.

KJvR

 

[jghrist]

I doubt that switching surges are a problem at 11 kV. Lightning surges can double at the open ends of long cable runs, however. At 11 kV, there is normally enough protective margin (BIL over max surge voltage) to allow for doubling.

Long cable lengths are worse.

If installed, they should be at the terminations of the cable.

Normal metal-oxide surge arresters would be used.

 

[cuky2000]

(a) Is the installation of surge arrestors on a cable distribution system common practice (esp with XLPE cables)? Yes. There are four commonly used underground overvoltage protection schemes utilizing MOV arresters. They are, in order of increasing effectiveness and cost, (1) Riser Pole Protection, (2) Riser Pole and Open Point Protection, (3) Riser Pole, Open Point, and Equipment Protection, and (4) Riser Pole, Open Point, Equipment, and Scout Arrester Protection.

(b) Is long or short cable lengths worst? Long cable is worst

(c) Where should you install the surge arrestors? At the riser pole and end reflection point(s).

(d) Should I use normal Metal Oxide Surge arrestors? Yes
_{Any other comments regarding this will be appreciated. [blue]In recent years, new information has been revealed concerning lightning and its impact on underground power distribution systems. For instance, the IEEE Std 1299/C62.22.1-1996 states that the voltages developed on an unprotected line prior to flashover propagate along the line as traveling waves. At the surge impedance discontinuity of the overhead line to cable transition, part of the wave is reflected and not all of the surge propagates into the cable. A typical UG cable surge impedance is around 40 Ohms and only about 18% of the incoming surge voltage on the OH line would enter the cable because of the surge impedance discontinuity. However, because the voltage on the line prior to flashover could be 1000 kV or more, the voltage passing through to the cable can be excessive.
Therefore, improved protection is provided by the use of surge arresters at the terminal pole, mid-point of the cable run and at the open point.[/blue]}

 

[KJvR]

Thanks jghrist / cuky,

The 11kV system is an insulated cable (induvidualy screened cores) system only - no overhead lines that can be influenced by lightning. The 66 kV incomers are overhead lines feeding onto 66/11kV transformers. All lightning surges must past the 66kV surge arrestors and propogate through the transformers along a 1.5km cable. Is this possible?

jgchrist mentioned that switching surges will not be large enough to damage 11 kV cable with a BIL of 95kV at the open end. If that is true, is it still worth while installing surge arrestors on the open ended cables - or will it end up as dead capital hidden in the termination box of the switchgear?

KJvR

 

[cuky2000]

The system description provided is probably less severe than the one assumed from the original post. Switching surge may not be a major concern on 11 kV or even 69 kV systems. However, to determine the risk level of lightning effect propagating from the high voltage side passing throughout the transformer, a comprehensive insulation coordination study and transient analysis may be required.

The cost of those studies including the facility engineering time may be more than $100K.

Several options may be available to evaluate the risk:
a) Do nothing and accept the shortage in cable life or failure if happen.
b) Expend resources in study with open chance to recommend surge arrester.
c) Specify the arrester without detail studies.

NOTE: Cable insulation levels are 100%, 133% or occasionally 173%. BIL rating is typically used for equipment and power apparatus insulation level.

 

[jghrist]

Another option: Install arresters on the 11 kV side of the transformer. I think the risk level is low for a surge propogating through the tranformer with more voltage than the BIL, but as cuky says, it may take $100K to prove it. Lighting is a high frequency event, and the transformer impedance to high frequencies is high. At worst, there will be a surge but with a lower rise time. With a low voltage rise time, there will be less voltage buildup at open points.

 

[KJvR]

Thanks for your comments. I will investigate to see if there are any surge arrestors at the main 11 kV incomers. If there are surge arrestors, I will asume the risk to be low and will not install surge arrestors in the substation we are currently doing the designs for. If not, I will include surge arrestors.

KJvR

 

[mc5w]

You could also have lighting propogating into the 11 KV system because lighting elevates the potential of the grounds at the substation. That is, voltage drop in the grounding electrodes makes the ground mat and cable shields "hot" when lightning strikes. Surge arrestors at each end of the 11 KV cables will lessen the potential difference between the shields and the conductors. If your 11 KV system neutral is solidly bonded to the ground mats of the supply substation that is a plus and is the usual U.S. practice. However, solid grounding of the 11 KV transformer neutrals needs to be supplemented with surge arrestors.

 

[KJvR]

Hi mc5w,

Are you talking about direct strike on the 66kV incomer lines that will cause currents into the earth mat at the surge arrestors in the 66 kV substation OR indirect strikes close to the 11 kV substations causing voltage rise in the earth mat due to induction as normally happen on overhead lines?

One question: if the voltage spike is due to a direct lightning strike on the 66 kV lines, and increase the earth mat potential, will it not reduce the phase to screen (earth) voltage in the cable because the impulse also propogates through the transformer onto the the phases? I must admit I have never thought of what you just told me and run some ATP simulations.

Thanks

KJvR

 

[jghrist]

Good point, mc5w, that I hadn't thought of. The cable conductor will be at remote earth potential, but the local grid could be much higher if lightning struck a structure close to the cable terminations. Lightning will cause a large potential gradient in the ground grid, so the location of the equipment grounding point with respect to the point where the surge enters the ground grid is important.

I would not expect much of the surge to propogate through the transformer because of the high impedance of the transformer to surge current. Also, the surge may not be on the 66 kV conductor.

On the other hand, when doing direct stroke protection design for an outdoor substation, you don't consider that if the shielding successfully intercepts a stroke, the ground grid potential rise may flashover the bus insulation. At least not if you design in accordance IEEE Std 998-1996, IEEE Guide for Direct Lightning Stroke Shielding of Substations. Is this a problem with IEEE Std 998?

 

[amps21]

The transformer transferred surge, initially, is cosidered being transferred thru winding capacitance rather than inductance. How much gets transferred, is completely a transformer design parameter. but like someone suggested if properly corordinated arresters are used at the high side, surge thru transformer should be minimal. And mc5w has indeed a good point.