Transient supression of a PT signal

[TurbineGen]

Can anyone comment about ways to control transients in a potential transformer circuit without causing a phase shift?

Here's my scenario. We have a 30 MW generator that we synchronize online by using a signal from a PT to both determine it's speed and phase angle. This unit is a peaking unit which starts and stops daily. The problem is that at approimately 15 hz (electrical frequency) a few breakers close and cause false readings by our control system, causing speed reading in excess of 100% synchronous speed. Obviously this is a false reading, and it tends to last for three to five seconds (not 3-5 cycles). One issue is that at 95% speed we close the exciter breaker and start adjusting our controls to prepare for synchronization. Another problem is that as the transient falls off, the control reads that the generator is rapidly slowing down, which it isn't, and will trip.

Since the PT is used to obtain the phase angle of the generator for synchronization, a phase shift in this signal could be catastrophic to the generator. How can I control this transient without a phase shift on the PT signal? The signal must usable and accurate from 5hz up to 65 hz. Any assistance with this problem is greatly appreciated.

Thanks in advance.

Quote: If it is broken, fix it. If it isn't broken, I'll soon fix that.

 

[GTstartup]

It seems strange to me that you are using a PT and (presumably) a frequency transducer to measure speed. Why don't you use speed pickups? Then your PT would be used for what it was designed for (synchronizing).

Anyway, the transient signal you need to address is the 4-20mA from the transducer, this won't affect synchronizing.

I have to admit I am making a few assumptions here but I really don't expect your control system is taking a PT directly.

 

[davidbeach]

What is your VT? Is it truly a wound transformer or are you using a CCVT? What kinds of relays are you using? Most numeric relays will ignore the higher frequency components.

 

[TurbineGen]

GT,

It is true that our system doesn't rely soley on the PT signal. There is an optical speed sensor on the unit which is the primary for speed. Unfortunately that has proven unreliable, so the control monitors the PT frequency as well.

The PT signal is stepped down to 24 volts. I'm not sure whether that is done by a transformer or some other device. The PT is a wound transformer, not a CCVT. As for the relays we're using, they are mechanical relays unfortunately. The signals come from our control and are stepped up to the voltage needed for the mechanical relays.

This was an old system that was just updated and we are working the bugs out of it.

Quote: If it is broken, fix it. If it isn't broken, I'll soon fix that.

 

[GTstartup]

OK so the signal that needs suppressed is the stepped down voltage, no worries about phase shift on the PT. How does it get to the control system? Transducer?

Anyway, how does this PT have any usable voltage at 15Hz? The excitation isn't on so all you have is field residual magnetism and that might get you a few hundred volts on the primary at most. Seems like there are a few pieces of the puzzle missing. Is it possible that there is a logic error and this frequency should not be used until at full speed with excitation on?

Don't you love it when you inherit a scheme where the backup is more unreliable than the primary system? Now you have two systems that don't work. Maybe you should consider fixing the original scheme

 

[ScottyUK]

Abandon the optical sensor and use an inductive pickup. Very reliable. Don't you have any vibration monitoring on this unit from (eg) Bently Nevada or Emerson-CSI? The probes these systems use are induction based. If you have this type of equipment already, see if you can add another probe. In Bently's world the probe is a 'Proximitor'.


----------------------------------

Sometimes I only open my mouth to swap feet...

 

[GTstartup]

To elaborate on my previous post. With the information you gave us on the turbine we can assume that the generator voltage is 13.8kV. Problem occurs at 15Hz so excitation is not on. Lets assume a primary voltage of 500V due to residual magnetism. That results in a PT secondary voltage of 4-5VAC (assumes 13.8/115V PT). Then you state it is stepped down to 24VAC Now we have about 1VAC to the control system until excitation is on.

Am I missing something?

 

[TurbineGen]

GT,

This is one of the reasons why the problem is so perplexing. The residual magnetism at 15 hz gives about 1.5 kv from the generator according to our monitoring controls. The unit is a 13.8kv machine. Voltage on the secondary of the PT is obviously less than 24VAC, but is still read by the control, which the manufacturer states should be no problem at the lower voltage.

We checked the PT low side connections and found them loose. After tightening, we still have the same problem. The more I think about it, the more I think this could be solved with a little creative programming. And yes I really dislike it when the back up signal is less reliable than the primary... especially when the primary's reliablity is poor to begin with.

Scotty UK, I'll look into the inductive pick up. The optical we are using has got to go.

Quote: If it is broken, fix it. If it isn't broken, I'll soon fix that.

 

[stevenal]

"The signal must usable and accurate from 5hz up to 65 hz."

VTs made to IEEE standards have a rated frequency of 60Hz. Have you expressed this need to a VT manufacturer?

 

[sreid]

If it is high frequency noise that's messing things up, a low pass filter at 6500 Hz (100x 65 Hz) would have a phase shift of about 0.5 degrees.

 

[NorthSlope]

TurbineGen:

Let me go back to the original post. My first question would be - which breakers are closing and WHY?

Checking actual voltage should be easy with a PT and a Fluke. Then you'd know what the voltage actually is.

NorthSlope

 

[GTstartup]

Firstly, no way you will have 1.5kV at 15 Hz just with residual voltage on a 30MW machine. What you are reading, most likely is starting device voltage. I am assuming it's a GT since you said it was peaker. According to what you told us about the situation (no excitation, PT secondary voltage 13800/115/24V) you will be reading maybe 3V AC. How does this get to the control system? Measured directly or transducer?

Anyway bottom line, I agree with Scotty, this is no way to read speed on a GT. Bently is the way to go,