Question on a 6.6 kV Ring Main Distribution System

[krisys]

We have a large oil field having a length of about 75 kms and width of about 65 kms. The oil wells are randomly distributed in the field. The central oil gathering station is located somewhere in the center of the filed. The oil will be collected from the field at the gathering station, processed and exported.

As the oil field is remote, it has no utility power supply. The entire power supply is met with the in-house power generation. So far, the major power requirement is in the gathering station and the power generated at the gathering station is connected locally.

The individual well heads need only a very small amount of power (say less than 10 kW) to keep the instruments and controls operating. So far the well head power was catered by the locally installed solar power system at individual well head locations. Of late the maintenance of these solar systems became challenging. Hence the Company has decided supply the well head station from the central oil gathering station and discard the solar system completely.
As the power required for individual stations is very small (<10kW) and the well head stations are distributed along the field, it has been decided to distribute power at 6.6 kV through the Rings. Each well head station will have one Ring Main Unit (RMU). The RMU will feed the well head station load through a small step down transformer.

Four (4) rings, with each consisting about eight (8) RMUs will be installed. The average length of each ring is about 45 kms. A dedicated 6.6kV power system has been created at the central oil gathering station to supply the well head station power. Two nos. of 1000 kVA transformers will supply the 6.6kV switchboard. The 6.6 kV Switchboard has normally closed bus tie.

As the ring cable is too long, capacitance is high and as a result, a big voltage rise is expected at the central oil gathering station 6.6 kV bus. To overcome this, two nos. of shunt reactors, rated 350 kVA each (one reactor per bus section) will be installed.

The distribution system is as indicated in the below sketch.

<img src="

http://i61.tinypic.com/2gsplcj.jpg"

border="0" alt="Image and video hosting by TinyPic">

I would like the panel to discuss the pros and cons of this distribution with respect to the following:

1) The preferred ring operation philosophy (i.e. close ring or open ring)
2) The risk of switching transient. Whether the complete ring can be energized (i.e. say a cable length of about 45 kM) from the main switchboard without risking the transient over voltage?
4) Risk of parallel resonance in the system due to variable capacitance (depends on the no. of rings and the ring sections energized)
5) In my own view instead of having the shunt rectors, the installation of 1000 kVA transformers with very high impedance (say 35%) would regulate the voltage better. Also the possible parallel resonance would be eliminated. When the ring capacitive current increases, the inductive drop proportionately through the transformer.Thus resulting in a better voltage regulation.
6) The well head station load is mainly DC UPS (battery charger). Hence a good percentage of harmonic injection is also expected. I am afraid that some of the harmonics currents might excite the circuit to give rise to the over voltage/damage to the system.

 

[7anoter4]

I have no experience operating such a long cable. In my opinion has to be supplied at both ends but has to be interrupted in the middle. The transient voltage at "make or break" may be double- in my opinion-in the case of such long run. An average capacity-in my opinion-has to be 0.3-0.5 microF/km and the inductance-medium voltage cable 6/10 kV- about 0.31 miliH/km. You may calculate the resonance frequency.
The short-circuit current is low so for 1 sec. 35 sqr.mm could be o.k. But if you have to inject all 1 MVA one way the voltage drop will be huge. Even 3*300 sqr.mm copper will not be enough.
I think you may request a consult from Schneider or Siemens or other RMU manufacturer.

 

[7anoter4]

Sorry, I didn’t understand “40 kms” .Could be it 40 km length ? I considered 75+65=140 km
40 km. it is o.k. With 3*240 sqr.mm the voltage drop will be about 6% [for all 40 km and 1000 kVA].

 

[krisys]

7anoter4,
Thanks for your reply.

There are four (4) ring circuits. Each ring length is about 40 kilometers. The load of each ring is 150 kW. 85% of the load is non-linear, i.e; the loads which are generating harmonics.

Now regarding the resonance points:
As the load current in the ring is very small (for 150 kW @ 0.85 pf at 6.6kV the current is 15.4 Amps), the resonance I am talking is about the entire 6.6 kV system resonance. For that purpose, I consider each ring circuit as a capacitor. The source transformer will act as the series reactor. By the inherent property of the series reactor, there would be a voltage rise at 6.6 kV bus.

In order to bring down the bus voltage to nominal value, shunt reactors have been used. Please refer the Single Line Diagram (SLD) shown in the posting. Up to this point, it is fine.

Now, when we draw/run an impedance scan looking from the 6.6 kV bus, we may identify the resonance frequencies. If these points fall at the same frequency as the harmonic frequency of the load, I am afraid, that there could be some over voltages and equipment damages.

Hope with this clarifications, you may be able to give your view with more clarity.

 

[7anoter4]

I had not enough time- for the time being-to analyze the system, I am sorry. At a glance the voltage rise-Ferranti Effect- due to cable capacity is moderate-in my opinion.
Of course the transients may be high.It is a hard job to calculate it .You have to find a doctoral research as this one- for instance:

http://innovation.edf.com/fichiers/fckeditor/Commun/Innovation/theses/thesePagnetti.pdf

 

[krisys]

7anoter4,

Do you think there could be a possible voltage amplification at certain harmonic frequencies? We have adequate capacitive loads due to the energized cables of about 160 kms (4 rings of 40 kms length each). To counter this, there are two 350 kVAR shunt reactors. This would form a text book style resonance circuit (i.e. as some text books explain the parallel resonance circuit).

We will inject a good amount of current harmonics from the UPS, which is widely spread in the distribution system. Such non-linear load forms almost 85% of the total load.

In my view at least a system resonance study and preferably a harmonic analysis study should be done. It would identify, if the system is vulnerable for any resonance and voltage amplification.

Your valuable input is highly appreciated.