We have DP level transmitters connected to the bottom of the tanks. Few days ago it was raining, and there were lightning strikes in nearby areas. Immediately after the lightning all 4 tanks experienced increase in level reading (increase about 20% level). The peak was not immediate, and the rise due to lightning and return to normal value(time to de-energize) took about 1-2 hours for the transmitters. We experienced the same scenario several months ago and there was lightning on that day too.
The trends for 3 tanks had same characteristics: had relatively fast linear increase and gradual decrease - One line up, Peak, and line down with a period of app. 1 hour. All of these tanks had process fluid inside of them. The 4th tank with no fluid, had different behavior. While the 3 tanks had one peak, the empty tank had about 6-7 ups-and-downs for same period and then it stabilized.
To solve the issue I am thinking of connecting the transmitter ground port to the earth. I had few questions about this subject and would appreciate your assistance:
1) The tank has its own ground connection. Is it ok to connect the transmitter to the same ground? Or is it better to make a new one?
2) The tank is placed on concrete. It seems so that the only way for it to get energized is through the ground connection and through the pipes. If so, if we connect the transmitter to the ground, would the ground connection exert even more disturbance to the transmitter during next strike? I am little bit confused.
3) Did you have similar experience? How did you solve it?
I'm only marginally familiar with what is inside the surge protection devices the transmitter vendors sell - something that shunts a surge to ground; not something that 'absorbs' the energy.
I checked out the Crouse-Hinds MTL TP-48 surge suppressor spec sheet and found an interesting performance fact that sounds somewhat like what has been described.
"Any surge current appearing as a series-mode or common-mode transient is converted into a common-mode voltage - whereupon the transmitter electronics are temporarily raised to some higher voltage level before ‘floating’ down automatically (and without damage) to resume normal operation." (first page, 2nd column, 1st paragraph)
The datasheet does not describe the time frame for the input to "float" down. But I have to ask, are surge suppressors installed? Are they installed installed properly (that paragraph starts out "The all-important earthing connection")?
Maybe a call to tech support of whomever your site uses for surge suppression could enlighten you and you could enlighten us.
1) The tank has its own ground connection. Is it ok to
connect the transmitter to the same ground?
The tank must be grounded in several places and the grounding system tested. You need to get electrical involved.
The signal grounding separate from the piping/vessel grounding and typically tied to the control room ground where a central DCS is used. If wireless, then the transmitter power ground. The transmitter enclosure is grounded to the vessel/piping.
2) The tank is placed on concrete. It seems so that the only
way for it to get energized is through the ground connection
and through the pipes.
Looking at the original problem description, it certainly appears that you experienced an electrical surge due to the lightning strike. It also appears that the transmitters are totally isolated from ground.
Good wiring practice with 4-20mA instruments calls for shielded twisted-pair wiring from the transmitters to a controller I/O card. The shield must be grounded at the source of the 24-35 volt DC necessary to energize the field instruments via the 4-20mA signal wires.
Usually the negative side of the power supply is grounded, so the shield should be grounded at the I/O termination. It is not necessary to separately ground the transmitters, and if they are now grounded at the tanks, that ground should be removed. If the transmitters are grounded at the tanks, you will form a ground loop between the tanks and the instrument ground in the I/O termination area. Ground loops do strange things to low level instrument readings. I cannot state that what you saw was the effect of a ground loop, but it might have been. Just follow code wiring practice and avoid ground loops. Never ground a single circuit or shield at two (or more) different places.