Need help with thermocouple measurements

[Pandiani]

In your first post, you said some of the steps you took (lifting drain wires) made the spikes worse. Please elaborate or post a plot. Is the digital filtering on the PLC making EMI noise events look like this? I'd be very interested to see how real noise is plotted.

Hello Keith,

I must admit that this was only my first impression and it is not for sure, since I instruct technicians to remove drain wire connection and then I waited not longer than 10 min. This is not sufficient time to make conclusion.

I also thought that these spikes are caused by opening spray water valves (in the closed loop control), but then I found in the archive how trends look like when the unit is off and all spray water control valves are closed.
I checked some measurments on a twin system and found smoothing time constants set to 5 sec!!! I removed then and observed spikes also.

I suspected EMI but definitely I'm not sure and don't know what to think after inspecting the behavior during the plant downtime.
This picture is given in attachment. EMI from energetics cable would be in miliseconds I agree.

I don't know how to explain this behavior.

Remark: some of the values are different color that on the first pics, because one of operators change the trend, but spikes are easily seen. This light red is smoothed measurement (5 sec time constant).

 

[danw]

1) The first page of your #2 analog input document shows paralleled T/C's for redundancy.

Does your configuration use parallelling to a 2nd AI card for redundancy?

2) The copper cable document that you attached
a) shows solid wire, not stranded
b) is translated from German, and says: "Stranding: cores twisted into pairs, 4 pairs layed up into sub-units (2-pairs cable stranded as one star-quad)"

so I've concluded that the cable contains twisted pairs, since the word 'stranding' in the translation (below) does not refer to 'stranded' wire.

Is my conclusion correct?

3) Could you describe what the process is that these T/C's are monitoring?

Is there a regulator valve involved?

 

[keithkyll]

I checked some measurements on a twin system and found smoothing time constants set to 5 sec!!! I removed them and observed spikes also. ... light red is smoothed measurement (5 sec time constant).

So, this is normal operation?
On a closed loop system, we need to isolate to discover what is causing the upset - Input, Control, or Output?
1) Are T/C's lying? I'm not convinced.
2) Is PLC dropping control, then recovering? Need a plot of control output. Intermittent 0-10V cable going to controller?
3) Is heater element opening or loosing AC power, brownout? System going to full output, then heater comes back, overshoots...

Some spikes are negative only, indicating a return of power, but not an upset with control loop.

 

[danw]

Although your experience previously indicated that temperature excursions of superheated steam do not show temperature spikes like these, look at the evidence for real event, as Keith is suggesting:

- a spike duration of multiple seconds could be long enough to be a change of several degrees.
- the period between spike incidents does not appear to entirely random, on some trends the spikes are quite regular, almost a 'timed' regularity
- the nature of the change does not appear to be random, it can be quite regular (yellow: negative, then positive)
- The AI cards are high end cards
- signals from this cable go to two different AI cards
-changing an AI card made no difference

I'm leaning towards an actual temperature changing events, rather than noise because I'm stumped as to what could possibly induce a negative change, then a positive change over a period of multiple seconds.

 

[Pandiani]

Ok guys, I made another test, but first to try to answer some of your questions.

We have no redundancy measurement setup.

First of all, these TCs are measuring steam temperatures on different sections of superheaters. Usually these locations are before and after watery spraying. Motorized control valves control amount of water that is injected into the steam. Mechanical construction is like shown here in Fig. 15.2.8.

In this "big" cable there are 40 wires and it is like 20x2x0.8 mm2, since each two wires are twisted pair inside. We used this type of signal cable a lot.

Today, I performed the following test:
1. we disconnected TC from the field
2. we connected new TC in the junction box to the terminal strip. With this setup, line from the junction box together with extension cable and TC (in steam pipeline) is removed.
TC is put in the junction box and connected directly to the terminal stripes.

Results of the experiment are shown in the attached pics.

Right now, I can be sure that TCs are not showing the exact temperature changes of the steam, since this time, it was measuring temperature inside the junction box. In addition to these temperatures, there are 4 signals (binaries) 24V as feedback signals (ON/OFF) from safety valves. These valves didn't operate for a long time (they open in case of trip or excess pressure in the boiler).

The only thing is to wait the overhaul and detail inspect the cable trace, and try to find if some power cable is nearby through which there is some ON/OFF control to some drive.

However, what is puzzling is the fact that these spikes are present during downtime, when most of high voltage equipment is off and control valves are closed.

If this is not EMI and if this is not induction, I don't know what is it since, real temperature fluctuations can by now excluded by this experiment.

Next test would be to try to connect TC directly to AI card in the control room, but this will have to wait, since I cannot take the risk during the plant operation. Connections are wire wrapped and not under screw in the cabinet.

Thank you all for the help.

Rehgards,
Pandiani

 

[keithkyll]

Great info! In the future, you could have done the same test by removing one of the thermocouples, and shorting the terminals. The extention wire connected to the terminal block becomes the primary T/C.

I'm at a loss to explain it. Something causes a few degrees shift. Sustained positive or negative shift of a few degrees, with about a 2 minute heartbeat. T/C extension wire has a magnetic wire. Maybe EMI, but not AC???

 

[Pandiani]

Thank you all for your contributions. Temporarily, I have set first order filtering with 4 sec time constant to get rid of these disturbance and it will stay that way until next year overhaul, when we'll try to exchange the cable.