The stator winding temperatures of 210 MW generator are measured by using the RTDs. These RTD carry odd harmonics when the set is synchronised with the grid or the set starts its generation. The DCS having the analogue input module which converts the RTDs directly into the Digital form doesn't convert this signal because of odd harmonics present in the signal. The DCS gives the above temperatures perfectly when the unit is not generating. There is no RTD to mA converter at present. If I had gone through this, I have to change the type of Analogue card in DCS which I don't want to. Is there any solution to filter out the harmonics present in RTDs? Any other user in power station have come across with this problem?

 

It is not clear what magnitude and frequency of harmonic voltage and current you are experiencing. It sounds like you suspect the source is subcycle currents due to differential winding pitches (alternator is paralleled with others). Are you sure that the voltage/current you are seeing is induced? Could it be insulation leakage? Are the RTD's 2 wire? 3 wire? 4 wire? Not knowing anything else, I'd say that 4 wire RTD's with seperate, shielded, twisted pairs for current source and potential leads would serve best, particularly if the existing DCS (PLC?) RTD interface module can support them. Do you know if the current is induced in the RTD's, or if it is induced in the leads? Are the leads between the RTD's and the DCS twisted/shielded pairs? If the RTD's are 2 wire, might either a low ESR shunt capacitor or a PI filter network at the DCS end of the RTD current loop serve? It would have the disadvantage of slowing RTD response time, but more important concerns might be whether the DCS's RTD current source will tolerate the capacitor inrush current, and also, depending its magnitude, whether the shunted AC circulating current will cause additional Joule heating (error) of the RTD. Last, but not least, are the RTD's dipped and baked in place? In my magnetics experience, physical placement of RTD's and their associated leads is critical. Assuming you can stay in the alternator hot-spot, you may be able to mitigate most of the noise by relocating the RTD's, and/or rotating the RTD's in one axis or another, and/or relocating their leads, thus altering how they cut across the flux.
[email protected]

 

As you're probably only gonna see very slow changes in temp, you can simply brute force filter the input with high value low leakage caps across the input and perhaps from the each input to analog ground. It's just a first pass, but that might knock the noise levels down to where the inherent CMRR can handle it. In any case it can't hurt. As a quick try you can use cheap high value monolithic ceramic caps which can give you 1 microfarad in a small package, but for long term use you might want to use more stable types. The caps from input to ground should be well matched to avoid creating a differential noise signal. Short leads will be most effective. All the normal stuff like twisted and shielded leads also applies.

Regards

cww

 

What's the bus voltage for the generators? I've seen something similar on load sharing lines. It turned out that the control wire cable tray was only three feet from the power cables' tray (11KV), and the loaded generators were inducing a signal into the control wiring. In this case, it was bad enough that it was damaging equipment. Are you certain that a different card will cure it?

If your RTD signal isn't clean, you might have a wiring problem. Changing a card might cure it, or might cure it for a while, but it might come back or show up in other parts of the controls. Make sure the shields are properly grounded on the signal wires, if you haven't already. You might have to move the signal wires farther from the load cables, or build barriers between them.

On something like this, it might be easier in the long run to find out why you're getting a dirty signal and cure that, instead of changing a card this week, a governor next week, etc, etc. Check and see if you're showing the harmonic on any other wiring.

Paul Baker