Touch & step potential 2

[halfweeg]

I have a project earthing philosophy that calls for calculations to confirm touch & step potentials are within acceptable limits as per IEEE-80.

I am seeking advice on how to apply IEEE-80 to the following situations:

- indoor substation
- outdoor equipment

I am not sure on how to calculate touch and step potentials when there is no grid system within the indoor substation or for equipment that is not part of the substation and is independently earthed to a separate plant earth.

In addition, the system is installed on an artificial island and it is intended to use the sheet steel pilings that encase the island as the earth electrodes. The island is made up of crushed limestone with an impervious polyethelene geomembrane. Earth rods are not an option as the membrane is design to prevent contamination from spreading beyond the island.

 

[halfweeg]

I forgot to mention that the system voltage levels and earthing method:

- 35 kV resistively earthed (4 x 200 A neutral earthing resistors, maximum 800 A)
- 6 kV resitively earthed (20 A)
- 690 V resistively earthed (200 A)
- 400 V solidly earthed

 

[jghrist]

You would need a sophisticate grounding program such as the SES CDEGS package to analyze your situation. See

http://www.sestech.com/Default.aspx?tabid=154

This would allow modelling of volumes of different resistivity soil and the sheet piling ground electrodes.

What is the source of power to the island? If all ground fault sources are on the island, and the ground grid extends over the whole island, then there will be no ground fault return through the earth and no step and touch potentials.

 

[halfweeg]

>>What is the source of power to the island?

The source of power is from barges moored adjacent to the island. Power is generated at 10 kV by 4 x 29.2 MW gas turbines, stepped up to 35 kV and transmitted to the islands via 35 kV cables. The barges' earth system is coupled directly to the islands earth system.

On satelite islands, the source is via submarine armoured cables fed from the central island's 6 kV system.

>>If all ground fault sources are on the island, and the ground grid extends over the whole island, then there will be no ground fault return through the earth and no step and touch potentials.

At the moment, there is not an earth grid planned, only a plant earth ring, with all equipment bonded to the plant earth. The entire plant area surface is concrete.

Would connecting the concrete pad rebar to the plant earth provide for the earth grid and provide an equipotential surface?

 

[jghrist]

The reinforced concrete pad will provide an equipotential surface, whether or not the rebar is connected to the plant earth.

 

[cuky2000]

One way to obtain a good grounding system is interconnecting a few of the steel piles in a close loop ring and also connected to the concrete rebar at several points.

To model approximately the piles, consider the perimeter divided by 2pi (6.2832) to obtain the equivalent solid round rod diameter electrodes. This approximation will be close to the actual value since the resistance of the electrode is mainly a function of the depth.

To calculate the step and touch potentials, consider apply the IEEE std 80. Faster and accurate results could be obtained using sophisticated grounding software but this option could be significantly expensive. I rather will contract a consulting firm to perform this task.

See the enclose reference for deep ground electrode resistance from "Computer Power & Consulting"

 

[jghrist]

I don't think there is a need for extended calculations. First of all, the ground current on the main island is limited to 800A by the resistance grounded system. Second, the source ground is coupled directly to the island ground, so practically all of the ground return current would flow through the ground connection, not through the earth/sea. The island is made of insulating crushed limestone with an impervious polyethylene membrane, ensuring that any return current that doesn't go through the cable ground will go through the perimeter piling instead of through the "soil". Everything sits on a reinforced concrete slab which will be essentially an equipotential surface. The satellite island ground current is limited to 20A.

 

[halfweeg]

Thanks jghrist, I think you've summed up the situation succinctly. The only real calculation needed is to size my main earth ring cable to handle local LV faults. GPR, and therefore step & touch potentials are not a problem.

 

[cuky2000]

I still believe that the engineer should conform with the project requirement in term how to prove by “…. calculations that the touch & step potentials are within acceptable limits as per IEEE-80”. Short cuts may be OK, however, the calculation extend of should be defined until fully satisfy the safety requirement of the project and accepted by other parties members.

Jaerist, let’s double check your statement: “Second, the source ground is coupled directly to the island ground, so practically all of the ground return current would flow through the ground connection, not through the earth/sea.” The earth/sea the impedance path is often lower than a direct ground cable because of the large area, moisture and salt presence.

Since concrete slab wet by salty water is very low (probably 5 to 30 Ohms-m or less) and the rest of the island is not fully isolated during the rain season or heavy storm, the expected resistivity of the upper layer will force to have an allowable step and touch potential low.
Therefore, from the practical point of view, this decision to obtain a thru ground is suggested to be reconsidered and provide additional scenarios to avoid hazardous situation to the personnel.

 

[jghrist]

If the earth/sea impedance path is low, then the GPR will be low. If the GPR is below the tolerable touch potential, then no further analysis is necessary (see IEEE Std 80, ¶16.4, Step 7).

If calculations are required, you can make a very conservative calculation of GPR by:

1. Calculate the ground fault current for a 35 kV fault, including the cable impedance. Assume grounding at both ends and a very low soil resistivity in calculating the zero-sequence impedance of the cable connection (sea water is around 0.2 ohm-m) to get a conservatively high fault current.
2. Calculate the GPR assuming the highest possible combined cable/earth parallel resistance. Assuming infinite ground resistance would be the worst case. In other words, all of the current flows through the cable ground. A lower ground resistance would reduce the GPR because this would be a parallel path and lower the total impedance for return current. Note that this is very conservative because you have already assumed a low earth resistance in your fault current calculation.

You'll run into trouble because you don't know the impedance of the cable return path, only the total cable impedance. You know the impedance of the return portion will be less than that of the total impedance, so assume as a worst case that essentially all of the impedance is in the return path.

This is extremely conservative. Essentially, you are assuming that the touch voltage on the island is the same as the voltage from the generator terminal to ground at the generator. I suspect that with a short cable connection, the GPR calculated this way will still be less than the tolerable touch potential if you have fast acting ground fault protection. Most of the voltage drop will be across generator impedance and the neutral grounding resistors.

The only way to calculate touch potentials with any degree of accuracy in your situation would be to model the system with a sophisticate grounding analysis package such as CDEGS by SES (see

http://www.sestech.com/Default.aspx?tabid=154).

If you're schedule and budget allows it, this would be a great learning experience.