Rebooting of Mark V controllers during grid earth fault & lightning

V

Thread Starter

V Rajendran

I have posted this issue early in the thread http://www.control.com/thread/1026250190#1244489188. The issue is all the computers, namely R, S & T reboots while lightning struck near by area or the grid trips due to earth fault. The problem could be resolved then by relocating the Mark V earth pits to a diagonally opposite direction. Since then there were no instance of Mark V computers re-booting while grid earth fault occurred or during lightning.

However recently the same phenomenon occurred in one of our machines when the lightning struck near by area. Machine tripped and caused production loss.

Here I am posting to understand the GE philosophy of installing the Mark V panel.

1. Is it required to gland all the field cables entering the Mark V panel? Presently at our site the cable are not glanded.

2. How the panel should be placed on the base frame? Presently the panel is placed on an insulator (rubber sheet).

3. What should be the earthing philosophy? Presently only one earthing connection is existing, going out from the ground bar from the panel to the earth pits.
 
1. If the cable is armoured, then the industry consensus is the cable should be glanded at both ends. (Personally, I've never understood this glanding business, but no one asked me. A gland at one end of a conductor (the armour) that grounds the armour and a gland at the other end that grounds the armour, creating a path for circulating currents to flow from one end of the cable to the other. Sounds brilliant! I must be missing something, because I've only ever seen insulating glands used haphazardly, and not on every cable at one end or the other. It just doesn't make sense. To me, anyway.)

2. I've never seen or heard of a Mark V panel mounted on a rubber sheet, or any other insulator for that matter, so congratulations! It was probably in anticipation of a potential problem with lightning strikes; omniscient designers, huh?

3. The Mark V was not designed to be segregated from the protective (safety; power) earth, however from the information in Question 2 it appears someone may have tried to do this. If they insulated the panel with a rubber sheet (that sure sounds funny, doesn't it?) then one has to hope they used non-conducting bolts to attach the panel to the base/support. And one has to hope they removed the ground strap between the earth bus bar and the panel/case before they attached the functional (instrument) earth strap.

And, the low-level cable shield drain wires would all have to be terminated on special bars, not the typical bars on the ends of the terminal board racks of the Mark V. Or some specially insulated, but earthed, ground bars. Are they?

But, if the lightning strikes were "blocked" on the other unit by "simply" relocating the earthing pit it would seem to suggest a problem with the original earthing pit, or some problem with the original earthing set-up, or just plain luck. Most likely some combination of the above (wouldn't the world be great if every problem could be traced to one, single problem?).

Trying to avoid electrical spikes from affecting a control system (<b>any</b> control system, not just a Speedtronic turbine control system) requires a <b>LOT</b> of planning and coordination and construction supervision, which almost <b>never</b> happens, especially the construction supervision required to ensure there are no field devices/instruments/shields grounded to protective earth and all shields are not grounded to protective earth.

Things can be done to reduce the likelihood of problems caused by lightning strikes but not to reduce every possibility of problems, particularly with regard to lightning strikes. They are unpredictable and downright nefarious. (I've never used 'nefarious' before. Score one for me.)

The problem(s) is(are) not the fault of the Mark V, but with the installation and the earthing pit(s). And Mother Nature, who doesn't like being anticipated.

And can usually outwit most engineers.

Rubber sheet or no.

;-)
 
Top