Are winding RTD's affected by VFD operation ? 3

[edison123]

My client has 2 nos. 645 V, 1.6 MW, 1760 Amps, 750 RPM, TEFV VFD driven motors for ID fan application in a thermal power station. Since commissioning these motors eight months back, the stator winding temperatures consistently show very high temperatures of over 140 deg C at load of around 1200 Amps at 600 RPM. The RTD resistance readings taken at the RTD terminal box match with the temperature readings in DCS. The motor body temperature does not exceed 50 deg C. The OEM simply increased the temp limit to 170 deg C to avoid temp trip. The motors were dismantled recently and the windings were inspected. There is not any sign of winding overheating and the original VPI winding looks healthy. No grey or brown marks on the overhangs indicating high temp operation. Is it possible that the winding RTD’s (3 wire PT 100) are affected by the VFD and they show erroneous readings during operation ?

 

[jbartos]

Suggestion: RTD and its module are supposed to be rated for the expected temperature range. Visit

http://search.netscape.com/ns/boomf...B+%22data%22&clickedItemURN=http%3A%2F%2F

http://www.wago.com%2Fwagoweb%2Fdoc...agoweb/documentation/750/eng_dat/d046100e.pdf

http://search.netscape.com/ns/boomf...B+%22data%22&clickedItemURN=http%3A%2F%2F

http://www.iprocessmart.com%2Fapt%2...processmart.com/apt/tr001_rtd_pt100_probe.htm

for:
RTD PT100 that has range at least -50 to 260degC

 

[electricpete]

You don't mention insulation class.

It may not be appropriate to rule out the possibility that the motor has actually seen these temperatures, based solely on visal inspection

The hot spot may perhaps be 15C higher than what you measure with RTD. 140C+15C=155C could be below the insulation system rating for class H system or at the insulation system rating for class F system. (although it looks like if the reading were valid you probably would exceed even the class H rating at full load).

Considering you may not exceeded the insulation rating and it has only been 8 months, visible signs of overheating may not be expected.

Other possibilities:

- High resistance anywhere in the circuit can intheory look like high temeprature. I have seen motor temperature indication erraticaly bouncing due apparently to loose connections. But doesn't seem likely to have a steady loose connection.
- I have also seen voltage indicated at the terminals of rtd (disconnected from the computer system) while the motor was running.... indicating there was some induced voltage (non-vfd). I never fully got to the bottom of it. I suspect maybe we were using a digital fluke with high input impedance and a lower impedance measuring device may not have had any problem.
- are you perhaps assuming an incorrect type of rtd and using wrong curve to convert for both dcs and local measurements.
- are you using the correct technicque to compensate for lead resistance in the local and dcs measurement? As you probably know for 3-wire the resistance between one pair (connected to same termination on the rtd) gets subtracted from the other pair (connected accross the rtd). For 4 wire you have one pair of current/source leads and one pair of voltage/sensing leads.

 

[vanstoja]

What's the basis for your belief that the motor winding temperature of 140C is "very high" when operating at 600 RPM with 1200 amps current? If the insulation system is Class B rated at 130C, then 140C is 10C too high and insulation life would halved (by rule of thumb) but evidence of damage would not be likely to show up in an 8 month period (5840 hours maximum). But if it's a Class F system rated at 155C then 140C winding temperatures are quite normal and and should show no evidence of any damage over an 8 month operating period. If the 1760 anps at 750 RPM is the rated motor operating condition, then 600 RPM operation is off-rated, and the motor's power factor will be degraded causing higher than speed-cubed-proportional running current. This is reflected in the 1200 amp current that is 33% higher than than the 901 amps that would result from the 80% rated speed operation assuming no power factor
change with speed.

 

[jbartos]

Suggestion: 170degC temp limit is supposed to fit Class 185degC insulation system. It leaves 15degC temp margin.
Please, notice that the Class H insulation system, if indicated, is only 180degC.
NEMA Class H permits 140degC temperature rise over 40degC ambient temperature.

 

[GordS]

Is it possible to check your theory of the VFD impacting your readings by noting if any immediate change occurs in the readings if you shut the motor off? If there is a VFD noise type of relationship, then the temperature should show a step change when this is done.

Whether there is a problem in your case depends on the insulation class of the motor. It would be expected that the motor would run hotter under VFD rather than sinusoid power. I have noted motor temperatures drop about 10C simply by changing the style of the VFD (from a 15 yr old thyristor based drive to a new IGBT based drive)

 

[edison123]

Here is some more info.
The machines are class F (and are not derated for class B temp rise) with closed cooling (TEFV). 3 wire RTD’s are correctly calibrated to cancel lead resistances. The VFD is a 4 section IGBT directly feeding two terminal boxes on the motor, where the motor is delta connected. The winding is 8 circuit delta. All the 12 RTD’s show temps higher than 140 deg C. When I ran the motors in open shaft at 750 RPM/50Hz thru’ the VFD, the winding temps stabilized at 75 deg C after two hours with a cold air temp of 28 deg C and the motor body temp at 33 deg C. The stator winding copper is liberally (or conservatively??) designed with 2.2 Amps per sq mm. As suggested by GordS, I had already taken the winding temp readings when VFD was shut down and found no step drop in the temp readings at DCS.

Now for the basis for my doubts about RTD’s getting affected (i.e. heated) by VFD:

1. As I mentioned in my posting, the motor body temp at the load of 1200 Amps at 600 RPM (when the RTD’s were showing over 140 deg C) was only 50 deg C. I would expect a much higher body temp with such high winding temps.

2. Both the stator winding condition during inspection (at over 140 deg C operation, in my experience of two decades, I expect to see some discoloration of the winding) and the very low current density (especially at 1200 Amps, it is only 1.5 amps per sq mm, which is normally used in copper bus bars where we don’t see such high temps notwithstanding an enclosed structure like a motor body) argue against such high winding temps indicated.

3. In open shaft condition, the winding temps are high at 75 deg C when the stator current was only about 400 Amps and the motor was running at the rated 750 RPM (thus ensuring adequate cold air flow).

4. Also, the rotors do not show any indication of overheat (the bare copper bars or the core do not show any discoloration).

vanstoja : you mentioned “motor's power factor will be degraded causing higher than speed-cubed-proportional running current. This is reflected in the 1200 amp current that is 33% higher than the 901 amps that would result from the 80% rated speed operation assuming no power factor”. Could you please expand on that ?

GordS : Yours was a good suggestion which I had already carried out. But, if the RTD’s were getting heated by the VFD, being located in an area where there is no cooling medium, will not the RTD’s temp drop slowly (as against a sudden drop) ?

 

[gsimson]

The RTD installed in a motor winding is a pure Platinum wire and it follows only the temperature curve. It is not affected by any way by voltage surges / harmonics that may be present in the voltage wave form due to VFD. So if the temperature shown by the RTD is 145 degrees, the actual temp. around RTD is 145 degrees.
The RTD leads may pickup some stray voltages but it will be very negligible.
Low power factor may contribute more current. The system reactive power may be more but the actual "wattloss" which generates heat in the winding may not be in direct proportion to the increase in current. - expert's views please?
If the net heat generated does not change then only the net heat taken away by the cooling system may have to be checked. The cooling at 600 rpm may not be adequate or the motor is designed to operate at 145 degree at 600 rpm for this application.
One of the methods to verify the actual temperature is to compare the resistance of the motor winding at ambient temp. and resistance measured immediately after stopping the motor.

 

[vanstoja]

I said..."assuming no power factor change with speed".
Power factor is a function of running speed and degrades with speed reduction from the presumably maximum value at design rated speed. Motor electrical input power for motor-driven fans, blowers and centrifugal pumps is a cubic function of speed because pumped fluid power is proportional to the product of flow (proportional to speed) and head/pressure rise (proportional to speed squared). From the electrical power equation, P=EI *cos theta, at constant volts, E, power and current maintain the same proportions if power factor stays constant (which it doesn't do if the speed changes). Using the speed cubed ratio on rated current gives you the estimated current for constant power factor and the ratio to actual current at the reduced speed gives a rough estimate of how much the power factor has actually changed due to the speed change when you can't or don't measure the reduced speed power factor. This simple-minded approach may be less reliable for variable speed power supplies because hte volts/Hz ratio is generally kept constant whatever speed setting.
Your assessment of high winding temperatures at 600 RPM appears to be based solely on the differential between winding RTD temperatures and motor body temperatures and may, as suggested by someone else above, be a motor cooling system effect rather than a motor electrical issue.
Your expected finding of motor insulation discoloration after operating for 8 months at not-so-high winding temperatures around 140C (for a 155C rated Class F insulation system)kind of mystifies me. For the typically multi-shaded brown and orange colors of newly wound and impregnated insulation systems, how does one judge discoloration effects as a qualitative indicator of winding overheating?

 

[jbartos]

Question to edison123 (Electrical) Dec 6, 2003 marked ///\\The OEM simply increased the temp limit to 170 deg C to avoid temp trip.
///Please, how was the increased temperature limit to 170degC justified?\\\

 

[gsimson]

Increasing the temperature setting without proper
investigation that too at 170 degrees is not a proper engineering solution.

With VFD with Vector controls, there will not be reduction in power factor. The existing drive may be checked.

As mentioned earlier, probably the motor was designed to run at 145 degrees at 600 rpm for this application. Why not check the actual temperature of the winding by measuring the winding temperature?

 

[edison123]

Ok folks, I modified the original winding connections in both the motors and both are now running well with a temp of about 105 deg c with 1200 Amps at 600 RPM. Both the client and OEM are happy with our work.

Thanks for all your inputs. However, I still retain my doubts about RTD's getting heated by VFD's. I would be happy to know if anybody has done some practical research on that.

gsimson: I will take up your good suggestion of measuring winding DC resistance to know the actual temp when my client could manage a next shutdown. I have taken the winding resistance at the ambient.

Earlier, OEM raised the temp limit to 170 deg C since the unit (210 MW) could not be shut down immediately and the client was willing to take the risk involved. Now, the temp limit has been reset to 120 deg C.

 

[jraef]

I doubt that the PWM firing pattern would heat up the RTD any more than it would cause increased heating in the motor windings. If so, the thousands of Multilin relays currently applied to VFD driven motors would be showing problems and GE would have done a white paper by now. They have not, I checked.

It is an established fact that motor windings can experience increased heating when being powered by a PWM signal so I would suspect that the RTDs were showing real data. The fact that you changed the winding connections and the temperature dropped would tend to support that. You probably achieved better attenuation of harmonics present in the motor windings, decreasing the heating effect.

Just a guess though, for what its worth.

Quando Omni Flunkus Moritati

 

[edison123]

jraef,

Thx for your view.

"I doubt that the PWM firing pattern would heat up the RTD any more than it would cause increased heating in the motor windings." Did you mean that RTD's also get heated by VFD just like the windings ? If so, then the temps shown by them are not real, are they ? I still have this doubt because even with my modified winding, the open shaft temp at rated 750 RPM with 400 Amps was high at 75 deg C. Do you think that VFD's perform better (visa vis harmonics) when they are loaded as against with small loads ?

 

[gsimson]

Edision123,
What is the modification done? From Star to Delta? The load current is same at 1200 Amp. What was the original design?

VFD can generate harmonics. And harmonic can not induce any voltage in a platinum wire. It can sense only temperature.

With the "marginal concept" the % increase in temp. at no load will certainly more than the % increase at load.

75 degrees at open shaft condition with 400 Amp. is before modification or after modification?

This is an interesting case study. Some more details please!

 

[tmahan]

my initial thought was that a RTD should not be affected by the use of VFD as a power source for the motor. Then I began to wonder if an induced voltage could cause a mis read by the actual device sensing the voltage drop across the RTD. Or could the RTD wires act as an antena for for the RF/EMI and that impact the RTD reader.

does anyone have any thoughts on this???

 

[jbartos]

Suggestion: If there is a twisted shielded cable installed, the RF/EMI may be shielded to high frequencies, e.g. MHz. Also, EMI interference is possible to measure.