Narrative of a Monitoring System of Gird Tied Induction Motor Generators

[Phil Timmons]

Narrative of a Monitoring System of Gird Tied Induction Motor Generators

Abstract:

This diary entry is a continuation of a prior entry regarding monitoring of motor speed for induction motor generators.  

There is a LOT more to this than just over-speeding an induction motor and slamming it onto the grid.  At best, this could cause "hunting" while the motor tries to match the cycle point of the grid, and at worse, could cause a short if it were to engage while significantly out of phase.

To avoid that, if we bring the motor up to speed by allowing the grid to start spinning it, when we increase the speed towards generation, it will already be in phase.

While there are many good ways to implement this - hardware and software-wise, (including PLCs, PICs, Embedded Controllers, PC, on and on) I am trying to start this all hardware independent so that anyone interested can look over the concept details before we try to get married to any particular implementation.

=============


[DRAFT] Functional Narrative Outline:


1. Observe if grid is available and acting normally - voltage within set range [Note 1], freq = set freq [Note 2]
   
2. Read induction motor/generator for normal winding and bearing temperatures at rest
   
3. Engage induction motor generator to grid and check direction of rotation.
   
4. Motor speed will rise from rest to somewhere between Full Load Motor and Nominal RPM.  (see table, below) Observe motor speed, current, voltage, temperatures.
   
5. After all observed data are within normal range, begin to supply mechanical over speed power from turbine to the motor shaft.
   
6. Raise motor speed to nominal motor speed.[Note 3]
   
7. Observe current lower to near zero, and all other data stay normal
   
8. Raise motor speed towards generator mode (over speed).
   
9. Observe temperatures, voltage, freq.  Current should reverse phase. (output / generator mode)  [Note 4]
   
10. The generator output current is controlled by varying motor shaft speed from turbine.  Increased shaft speed will increase the current being sent up line.  
    
11.  Voltage may rise during generation, but only slightly and still stay within acceptable range.  Outgoing current should never exceed rated full load current, and shaft speed should not exceed full loaded generator speed in table.
    
12. If frequency varies or voltage drifts outside of range, it may indicate a loss-of-grid fault, and system should disconnect from the grid immediately.  [Note 5]
    
13. Other fault conditions should be monitored as listed below in Fault Conditions, and the system should be taken off line
    
    

Notes

1.  Voltages in table below are nominal.  Depending on local loads, time of day, and individual location on the grid, these may vary by almost +/- 10%
   
2.  Frequency is exact across the grid.  Any deviation indicates an error condition -- either at the grid level, or within local generation.  Some fault conditions, such an open grid, will show up as higher frequency on this system due to the over speed of the local induction motor / generator with no load on it.  Such condition must disconnect from the grid immediately.
   
3. Nominal motor speed is the speed the motor would operate at if there were truly no load and no friction, and were being swept along with the field.  At this point the only power the motor is consuming from the grid is the power to energize the field windings.
   
4. While current direction changes within every cycle of an Alternating Current system, when the current is coming down the gird (being consumed) it is in phase with the voltage falling and rising each cycle - as seen across the leads of a current transformer.  When the current is being generated up onto the grid, it is 180 degrees out of phase with the voltage - again as seen across the leads of a current transformer.  
   
5.  The loss of stable grid-maintained frequency is the standard for disconnection from an open or faulty grid per UL 1741
   
   

Fault Conditions, Monitoring and Response

Temperature -

I am pondering epoxy attached RTDs on the external sides and bearings of the motors being used.  As these will not give actual winding temps, but should be able to indicate some level of overheating, I am not sure what the threshold settings should be for alarm and automatic shutoff, and will require some experimenting.  

Internal thermal switch type cut-offs may also be present depending on the manufacture of the motor(s), but these generally operate as failure mode, shutting the system down without prior alarm.    While Motor Starter style "heaters" could also be placed in the control circuit, they actually operate based on motor overload current and not motor winding temperatures.  As this system will operate over-speed, with dynamic loads from a turbine on the shaft, it may be wise for the bearings to also be monitored.  

Current -

Current can be read directly from Current Transformers placed on the motor leads.  This should also allow detection of whether the system is producing or consuming power from the grid, as the phase should shift 180 degrees between incoming and outgoing current.  Alarm and cut-off settings can be determined from the maximum current rating on the motor nameplate or from data tables.  Further protection can be placed in the circuit as conventional circuit breakers (preferred over fuses for multi-phase systems), and "heaters," as described above under the Temperature topic heading.

Voltage -

As mentioned in the Notes section of the Functional Narrative, local voltages can vary, even at the same location by time of day, so there is no point on measuring for an exact value.  What we do have to look for is a minimum value, below which we can assume there is a grid or local distribution problem, and the same with a maximum value.  .  

A method of measuring voltage in this type application would be to use small instrumentation transformers on the grid side of the disconnect relay to monitor grid conditions when not connected and overall voltage while connected.

Frequency -

This is a very critical factor to monitor, as life safety issues are involved.  In the event the grid were to fail or be open, there is the possibility that any generator such as the induction motor generator discussed here could continue to pump power onto the grid, placing line workers and others at risk.  

The case is often made that an induction motor generator should tend to shut down on its own, because of the loss of power to the field windings, when faced with the loss of the grid.  However, there is a condition known as "islanding" where with a load closely balanced to the induction motor generator there could be sufficient load to hold the output voltage steady allowing the field windings to stay energized and the generator still producing power.

In this condition the local motor generator would be the source of the local base frequency.  As the local motor generator would be operating above its nominal speed, without the grid to hold the frequency down, the local frequency being produced would be around 61 hertz.   This condition is what is to be monitored for under UL 1741.  When an over-frequency is detected, the motor generator must  automatically and immediately disconnect from the grid.

Motor Speed -

Motor Speed sensing was covered in a prior Diary entry

http://www.fieldlines.com/story/2007/6/28/211842/419

Direction of Rotation -

While a fixed stationary system is not likely to ever encounter issues with the direction of rotation other than first time set-up, a mobile system is subject to direction error every time it is reconnected at a new site.

After the motor is started by grid power and before the turbine begins to put power to the motor/generator, the direction needs to be confirmed to prevent any mishap from the turbine fighting the motor.  

Direction of rotation can be observed by use of two pickup devices on speed monitor, whether magnetic hall effect or optical.  The method is called quadrature, and pickup sensors are so aligned that the pulse signal from one sensor leads the other depending on the direction of rotation.  By observing which sensor gets a signal first, the direction of the rotation can be determined.

Vibration -

For out of balance turbines and failing bearings, the monitoring of vibrations in a system like this can prevent serious mechanical failure.   Another issue that vibration monitoring can help with is harmonics - like an out-of-balance tire, the system may vibrate at certain speeds, and monitoring can help detect and cure this.  What clever ideas do you all have for vibration sensors?  Do any of the windmill folks here use vibration sensors on your equipment?

[Stonebrain]

Hi Phil,

I read your 'narrative' with interest.

I appreciate your systematic approach,and your effort to be complete.

For research it's usefull to monitor about anything you can.

But for practical application there's need for simplification.

I would simplify step 1 to 8 with:

When the grid is on start motor and after some seconds,apply the overspeedpower.

Of course,the safetyfactor is crucial.

I understand an electronic device that shutdown the generator in case of gridfailure is essential.

I think there are guys on the board that can make this(I surely cannot).

Personally,I would prefer to find a box on the market(approved by authorities)that can do the job.

Thanks for presenting your work.

I will be following any contributions on the subject with interest.

cheers,

stonebrain

[tdmack]

Phill

"overspeeding and slamming it onto the grid"      

I wonder what is happening when an induction motor is "slammed" onto the grid from  a  standstill?  This is a severe speed and phase mismatch (not to mention a short) and it's done millions of time each day with no monitoring system at all.

Induction generators, just like the motors, "want" to be locked to the grid.  It is the torque applied to the rotor (trying to break this lock) that makes them generate.  This lock "sync"  is so great that the motor/generator will overheat and be  destroyed if too much torque is applied (and remain "synced" to the smoky end).

Momentary series load switching and phase angle triac firing are two techniques that have been used to bring the generators on-line a little more humanly.  But "slamming" as you refer has been used (and still is) for around 30 years.  However, it can be monitored.

Tim