1 RTD Temp Increase while 5 kept constant 7

[E. Richy]

From our siemens motor rated: 660V,900kw, 3Phase, 1492RPM, Ta: -20...+40 degree Celsius, I'm observing that one of the RTD (i.e., 5) Temperature is always increasing and
triggers alarm everyday while the 5 other RTDs gives nominal reading. At one occasion, the motor trips when RTD 5 exceed 130degrees celcius. The alarm limit is set at 120-degree Celsius while trip level is at 130 degrees celcius. Need your help, can anyone explain the cause and what's the possible solution I can do to save the motor in the future?
the motor is F Class (155 degrees Celsius).

Thanks.

 

[waross]

Are the cooling passages through the core clear?
Blocked cooling or a faulty RTD. (or RTD wiring)

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[E. Richy]

Hi Waross, yes the cooling passage is fine, we checked and nothing wrong with it.
I suspected to be faulty RTD but still unsure what really causing it.

 

[panter]

How high is measured temperature on other sensors .If is diference is big probably RTD is bad .

 

[E. Richy]

Hi Pante, its about 10 to 15degress Celsius difference in normal operating condition.

 

[che12345]

Mr. C. Richard (Electrical)(OP)6 Aug 23 03:52
"....From our motor .....I'm observing that one of the RTD (i.e., 5) Temperature is always increasing and triggers alarm everyday while the 5 other RTDs gives nominal reading...."
I have the following opinion for your consideration.

1. RTDs are ragged, it is either in good condition or open circuit. But the connection leads are of fine wires which are subject to breakage. The monitoring relays are of electronic device, subject to failures.
2. With the motor stopped > say 3h, measure with simple DMM, the ohmic value of all 6 RTD should be the same to say within 0.1 ohm. The exact ohmic value and at which ambient temperature is irreverent.
3. Transfer the "faulty" RTD to a proven healthy monitor relay and observe the reaction. If the RTD is NOT open circuited, the fault could the monitoring relay.
Che Kuan Yau (Singapore)

 

[Hoxton123]

When i worked in the rotating electrical machine industry,(motors generators) the designers favoured thermocouples as they were much more robust than RTDs. For RTDs we ALWAYS wound the stator with additions RTDs on a 1 for 1 basis. The minimum was 3 (1 per phase) this was doubled to 6. The more the wanted, we always added more spares.

The RTDs were put in the slot separator between the two layers of windings, and thus are inaccessible once the machine was wound.

The RTD leads were brought out to a terminal box for main and spare leads. Each terminal was marked with the slot number (i think that slot 1 was the one after top dead center, but at 75 my memory fails me!). This was not repeated on the machine wiring diagram.

i do not know if this was common industry practice, but it was certainly common sense!

So find the terminal box and see if there are spares and connect up the ones nearest the problem slot and see if the temperatures are are similar to the suspect slot. DO NOT LEAVE RTDS OPEN CIRCULT WHEN THE MACHINE IS RUNNING, as the rotor field will generate current like an open circuit CT....

You could always try looking at the machine through a thermal camera when it is running, if you can get access, alternatively glue a new RTD to the top of the suspect slot and also a satisfactory slot and see what the temperatures are.

Talk to your local rewinder and see if he can offer advice, they may have seen this on similar machines.

A hot spot in the RTD may not be detected on the surface. Thermocouples sense a small area, RTDs are usually a loop along the length of the slot and thus sense the average temperature along the slot. A stripdown wash and bake might cure the problem as restricted ventilation often results in temperature problems.

 

[waross]

Good advice Mr. Che. A star for you.

2. With the motor stopped > say 3h, measure with simple DMM, the ohmic value of all 6 RTD should be the same to say within 0.1 ohm. The exact ohmic value and at which ambient temperature is irreverent.

That would have been my next suggestion.

3. Transfer the "faulty" RTD to a proven healthy monitor relay and observe the reaction. If the RTD is NOT open circuited, the fault could the monitoring relay.

I would have remembered this tomorrow.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[electricpete]

We have a lot of erratic readings on motor stator RTD's which we attribute to termination conditions causing high resistance as I discussed here: thread238-502743

The terminations of interest are mounted either inside the motor frame or in an auxiliary terminal box attached to it, so they are all subject to vibration. For outdoor motors they are also subject to corrosion and in some cases the motor is in a salt spray environment. I'd say our motors in salt environment experience these problems 5-10x more than our other motors.

We use 10 ohm copper RTD's for stator temperature which I suspect are more susceptible than 100 ohm platinum (since a small change in connection resistance is a bigger fraction of the 10 ohm resistance).

When I look at the winding temperature trend on our plant computer, there are telltale signs that it is an indication problem (rather than actual temeprature problem)... specifically it rapidly might jump up and then back down (our motors operate at constant load). Sometimes it jumps below the reasonable temperature as a result of high resistance in the compensating lead connection of 3 wire RTD... I know there is no physical fault that makes the temperature lower than expected. And if only one RTD jumps but the others don't, I'm more inclined to suspect indication problem (sure there are a few scenarios that would icrease temperature of only 1/6 rtd's but those are less common). If it is a real condition, it will most often appear on all the RTD's, and it will most often increase gradually, and it will never decrease (in absence of other explanatory factors such as load decrease, ambient temperature decreaese, air filter change).

 

[E. Richy]

Thanks, Mr. Che, it's a good advice, will try do that when the motor is offline.

 

[E. Richy]

Hi Hoxton, thanks for your advice.

 

[electricpete]

It may or may not be obvious, but based on my comments, it will be very relevant WHERE You check the resistance.

You have option to check remotely where the leads enter the remote computer interface, or you can lift the leads at the motor.

You can in theory use successive measurements closer to the RTD to try to narrow down any high resistance in the connections.

You can lift the leads at the remote measurement and then check resistance at the local (motor terminal strip) without lifting leads at terminal strip (to capture as-found condition). (one RTD at a time). In the simplest terminal strip configuration you have a remote lead connected on one side and a local/RTD connected to the same terminal on the other side of the same point (connected by the terminal strip itself). Depending on where you put your probe you are including/excluding different parts of the resistance that may reside in that terminal strip. Bearing mind in theory the high resisstance could be at the crimp (have seen that... didn't strip the insulation correctly for the lug type... this type of problem tends to shows up immediately after a rewind), you rule out the most of the potential high resistance sources near the terminal strip by measuring the barrel of the lug of the lead going back to the RTD (rtd side, not monitoring side).

For our for our 3-wire RTD's we have to do 3 resistance measurements (based on the 3 pairs of leads). It should be one low and 2 high resistance measurements (where the high resistance measurements are very close). Subtract the low from the high, and that is the resistance that you will use with your appropriate RTD table to convert to temperature. You can compare that computed temperature to your plant indication. Yes there are RTD calibrators to do it for you, if you have those available that may be easier.

An important excercize is to consider what the temperature difference is that you suspect is attributable to indication, and then convert that to a resistance reading. That gives an idea how accurately you need to measure record the resistance. We use 10 ohm copper RTDs and we're looking for big resistance differences, we don't find a need to do a 4 wire resistance measurement (the only kelvin probes are the ones we use with our ductor DLRO... those are too big for this purpose anyway)... we just uses a digital multimeter. Again we have 10 ohm copper RTD's, I'm not sure if a more accurate measurement would be needed with 100ohm since smaller contact resistance have bigger effect there.

An important thing to realize in the process is that lifting the leads and relanding often FIXES the indication problem because for whatever reason (mechanical agitation or retightening) it resolves the high resistance at that termination.

In the end the simplest procedure we follow usually end up as follows:

[ol 1]
[li]The work is done with the motor running because we have no trips associated with winding temperature, only indication. If you have a temperature trip, then that deserves extra consideration for conditions of test and precautions to prevent trip during test[/li]
[li]Notify operators / control room that temperature indication will be affected by the work and alarms may be received[/li]
[li]Either lift remote leads at the remote instrumentatio, or skip this step Apparently it's not needed for us and our measurement system (*) .[/li]
[li]Measure the resistance of all 6 RTD's (3 resistance pairs per RTD) from the local terminal strip without lifting any leads[/li]
[li]Repeat measuremtn of the suspect RTD by lifting the lead.[/li]
[li]During previous 2 steps, inspect and clean lugs as applicable while they are lifted[/li]
[li]With only a single channel available for remote monitoring, reconnect so that the hottest of the RTD's is connected to that channel[/li]
[li]Check remote computer indication of the connected channel[/li]
[li]Figure out additional actions from there (most of the time the problem was resolved by lifting/cleaning/landing the problem lead and the indication now appears correct)[/li]
[/ol]

(*) I'm not sure how we get away with skipping lifting leads at the remote instrumentation since I'd think anything connected on that end would affect the measurement. I used to tell our guys to do that. They came back and told me they were getting the same resistance readings either way. Don't ask me to explain it.

We have had a few times when this procedure did not resolve the problem... all the connections were good and we still have higher indicated temeprature on one RTD. IN that case you have to look for other causes, but at least you have ruled out the cause which is by far most common at our plant (high resistance RTD terminations at the motor). In all the cases I recall with one or more indicating high temperature and several indicating normal temperature, we still ended up eventually attributing it to indication problem of some kind. Often the problematic resistance goes to infinity (open circuit) and there is presumably open circuit somewhere further into the motor. We have had rats chewing on the wire in one of those cases. Another case very fine gage RTD wire was used and ended up breaking.

For the last 23 years my work responsibilities have included health monitoring for 100+ large motors equipped with stator RTD's (all 10 ohm copper 3 wire RTD's). In all that time,
I've never encountered a problem where in the end we attributed it as a real temperature difference. So my opinion is that's a very unlikely outcome even though we have theoretical possibilities of blocked air passages (by foreign material getting past the filters, or by overly aggressive dip and bake that left extra resin... I don't ask for dip and bake but some of our shops have done it during refuribhsment anyway) or localized core hotspots. We have experienced both of those conditions affecting temperature on a motor wide basis but never on a single-rtd basis. What else could cause localized higher temperature?..

[ul]
[li]Severely unbalanced current could cause the localized higher temperature on one phase, but we have never seen it happen. At our plant, we have very stable / reliable power system for our motors so they are not subject to unbalanced voltage originating from the power system. We did have high resistance at the T-leadr lug enough to eventually burn it open and trip the motor, but we never saw any abnormal winding temperatures prior to that (perhaps we weren't monitoring the right rtd)[/li]
[li]There is one more scenario that is often brought up by our plant operators which is an electrical fault within the winding causing local high temperature. I tell them it's not possible because either entire motor would heat (if bizarre winding failure decreased torque capability) or else motor would trip in very short order. They remind me that our synchronous generator has temperature monitoring on individual coils precisely to find problems there like a strand to strand short. But there is a difference in coil construction, that generator uses Roebel bar with effectively single turn per coil. For them strand to strand short is a big deal, for multiturn diamond coil in large motor strand to strand short is insignificant. Multi turn diamond coil in large motor is also susceptible to turn to turn short, but that type of fault will inevitably escalate to trip very quickly, it is not something that will linger long enough to show on winding temeprature.[/li]
[/ul]

 

[Gr8blu]

E Richy : Plenty of useful information from previous posters. From my own experience: most likely culprit(s) are either a bad joint somewhere between device and data collection point, or a bad data collector. I would proceed as follows.
1) If the circuit resistance checks out, switch the "high" sensor for a "good" one at the data collection point. If the problem follows the switched connection, there's a problem with the circuit or device. If the problem does not follow the switch, the issue is the data collector.
2) Disconnect each detector circuit individually (in sequence) from data collector after warning operations that I was going to be messing with temperature measurements for subject machine. Measuring here captures all connections between collector and device, including device itself. Depending on size/length of lead extension to reach between motor termination point (typically auxiliary conduit box and terminal strip) and data collector, the measured circuit resistances may vary a bit. For perspective, a 15 C differential on a 100-ohm platinum detector at "operating" temperature would be roughly 6 ohms. For a 120-ohm nickel detector, the difference would be 7 ohms, and for a 10-ohm copper device it would be 1.4 ohms. If the measured value for an individual circuit differs from the AVERAGE of all circuits by more than 0.5% (roughly 0.5 ohm for a platinum detector), further investigation is warranted.
3) If there is enough difference to start checking more closely, go to the next "joint" closer to the motor itself and try again. If the difference persists, keep going closer. If it disappears, you've found the suspect section and/or termination.

If it's not the data collector, you'll be looking for loose connections, damaged lead wire, corroded connections/terminations, proximity of lead run to other high-power circuits (power cables, for example).

If still nothing - take a look at the motor itself. Is one side (which has the "bad" detector) close to some other heat source that might be raising the local temperature in that particular area? Examples: hot air exhaust from another part of the process, direct sunshine on only a portion of the motor frame, lack of a good exhaust ventilation path on one side of the motor (harder to get hot air out, so local temps rise accordingly over time).

Converting energy to motion for more than half a century

 

[edison123]

Just swap the troublesome RTD channel with another RTD channel (after disabling trip for both these RTD's). If the high temperature travels with the troublesome RTD, RTD is the problem and if it stays in the same channel as before, then winding (high temp) or RTD terminal connection (loose) is the issue. This is the quick and dry method.

Muthu

http://www.edison.co.in

 

[E. Richy]

Thanks Gr8blue, edison123 and electricpete for all your information and advice, all noted.

 

[onatirec]

Lots of good information above.

We design and manufacture electric machines at my place of work. In my experience, it's not uncommon to see up to 10-15C differences between thermal devices on our test motors that we dyno. We mostly use thermocouples, but have at least 1 RTD installed in every machine as well. The location at which they are installed, even with an experienced operator carefully inserting into the windings, can lead to reasonably large differences in readings. All of the literature I'm familiar with either explicitly or implicitly mentions variations in temperatures or hot spots - that is why even though Class F is rated for 155C, the permissible temperature rise is lower (40C ambient + 105C rise = 145C) to account for hotspots. So, it could be faulty sensor or connection, or could be a hotter spot in the machine. at 900kW, it's probably not feasible to put the whole thing in an oven and check the temperature sensors that way.

That said, 130C seems quite low to me for a tripping point, especially with embedded sensors. Class F can handle continuous operation at 155C for 20,000 hours and shorter spikes as high as 190C. Maybe you need to get 80,000+ hours out of the machine at 130C, in which case you could ignore this part. Most of our machines are inverter driven, so we may fold back the current for dynamic control when approaching 130C, but the tripping point would always be higher for Class F.

 

[E. Richy]

Thanks Onatirec, noted.