Chiller motor, will not perform

[Mendit]

We recently rebuilt a chiller motor (600hp 3600rpm)It had a die cast rotor. The motor failed from broken rotor bars.
We removed all the old die cast alum, manufactured new rotor bars and shorting rings from alum, all sizes were dulicated within 1-2%. The rotor bars are welded to the shorting rings. The motor has been installed. it starts and runs up to speed in approx 3-4 seconds, inrush current is normal, no load current is (normal 250 amps), however when the IGVs are opened to allow the freon to flow the motor current increases to 320 amps and does not compress the freon. The chiller company (Carrier) say the chiller clearances are fine and all is working well, there is plenty of freon in the system. Puzzled ?

 

[rbulsara]

Sounds like a mechanical problem!

What was the reason for the original motor failure to begin with? Was compressor checked at that time? Rotor failure usually occure due to locked rotor conditions (overheating).

If the motor is turning and taking the load, something else is wrong. Plus motor does not compress the freon, the compressor does, is the compressor turning and compressing?

May be motor is not turning sufficiently due to mechanical issues?

 

[Mendit]

Motor is turning at normal speed, motor size is not like any other motor (much smaller hermetic) compressed cooled freon flows over the windings and rotor. If the rotor were turning too slow current would be high.

 

[testtech]

I suggest both power and vibration testing for this machine. A power analyzer should be used that is capable of measuring torque and conducting a rotor analysis. This will help confirm that the rotor is operating properly and indicate whether the consumed power is acutually doing work.

THe vibration test will indicate motor speed and compressor speed and any irregular mechanical conditions. I would be most concerned with compressor impellor condition. The vibration test can indicate deficiencies in the impellor as well as irregular gas flows in the machine.

If you have a similar machine, obtain data from both for comparisons.

 

[electricpete]

I always suspect mechanical problem first before electrical. But since that is the other guys problem and this is motors forum:

Consider the possibility that you have significantly altered the rotor resistance and therefore the torque speed curve. Maybe when you load the motor it is exceeding breakdown torque situation. Have you measured speed under load?

 

[RAMConsult]

The only puzzle is the statement "...inrush current is normal, no load current is (normal 250 amps)..." I'm assuming it should read "inrush current is normal (normal 250 amps), no load current is normal...", then the real key to the puzzle was how the rotor was remanufactured.

A few questions:

1. What was the LRA of the original motor?
2. What was the rotor speed when the IGV opened?
3. What type (turbine, screw, vane, axial, etc.)of compressor?
4. Was there any flow and what were the readings from the suction and outlet side?

Without seeing the answers, my guess is that either the motor stalled or ran significantly slower under load than the original. Unfortunately, despite that the fact that the physical dimensions were copied, your rebuild may not have used the same type of aluminum as the original diecast material, similarly for the shorting rings. This would have a definite effect on the resistivity of the material which in turn would affect the current flowing in the rotor. The result is a motor that works fine under no-load conditions but has higher than expected slip and an altered speed-torque curve under load.

It will be obvious if the motor stalls under load, but if it runs too slow (has high slip) then the type of compressor becomes important since the characteristics vary by type. Some compresors roughly follow the cube law; therefore a 15% reduction in shaft speed would reduce its output by nearly 40%. You might try modulating the freon flow if possible, but now I'm outside my area of expertise.

 

[Mendit]

A note about chiller motors, this rotor is approx the size of a 200HP 2pole motor, but due to the fantastic cooling afforded by the product flowing over the windings and rotor bars, the motor is around 600HP Full load current is 580amp at 575V
Normal inrush current is approx 5x or 2800 amps.
Because the motor is fully sealed it is not possible to put on a tach to measure RPM, we are going to try and monitor it through vibration, but because we balanced it so well (4W/N) the motor is very smooth, you can not feel by hand any vibrations. The compressor is centrifugal, it has 2 impellers approx 32" diameter one behind the other on the end of the motor shaft. Rotors commonly fail on these machines due to the high current at start up followed by rapid heating and rapid cooling, which stresses the di-cast aluminum, to the point of fracture at the junction of the shorting ring. The cast aluminum and the extruded aluminum have very similar electrical qualities (conductivity 64% vs 62%) The extruded material being 64%, however we reduce the section by 2% to make the bars fit easier into the slot.

 

[edison123]

As against the rated load current of 580 A, you have only 320 A indicating the motor is not fully loaded. This could be due to lower RPM (due to increased rotor resistance or lower supply frequency) which you are unable to measure. Or the compressor is not working properly at the rated RPM (maloperation of the valves ?).

 

[imok2]

Your motor may be fine the problem may be in the loading vanes not going full open, if this was the original problem(running unloaded) this could have caused the original motor to fail.

 

[jbartos]

Suggestion: I agree with the previous posting. It would be prudent to have a root-cause-analysis performed. They usually pay for themselves handsomely as you can see now by this situation.

 

[electricpete]

I didn't realize first time around that the current was so much less than fla. If it were some kind of stall problem, I would think the current would go higher.

Measuring speed using vibration should be no problem, regardless of the balance. The lowest bearing housing vibration I have ever seen at 1x was on the order of 0.001 ips. We could still pick it up. Put the spectrum on a log scale and it should show even the slightest peak at running speed above the noise floor. Modern 12-bit or more data collector would do a better job than a dinosaur 8-bit data collector.

If the motor is maintaining speed at/above nameplate speed, then the motor is doing its job and the problem is mechanical.