Jump to content


Photo

Supply Resonance


  • Please log in to reply
13 replies to this topic

#1 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,601 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 27 April 2002 - 08:47 PM

I am looking for information on supply resonance problems, and especially fixes for same. We see lots of problems, especially in rural areas where power factor correction is used and there are long overhead lines supplying the region.

#2 wareagle

wareagle

  • Members
  • Pip
  • 1 posts

Posted 09 January 2005 - 04:09 AM

Marke
Just what do you mean by "supply resonance"? Are you referring to voltage
rise?

#3 RalphChristie

RalphChristie

    Intermediate Member

  • Members
  • PipPipPip
  • 30 posts
  • Location:Northern Cape, South Africa

Posted 10 January 2005 - 01:46 PM

Ressonance - frequency point where inductive (transformers, cables, etc) and capacitive (capacitors etc ) reactances are equal.

Mark:
As far as I know it becomes a problem only if a source of harmonics exists at the frequency where the impedances are the same. You can estimate this frequency with this formula:

h(f) = (MVAsc/MVAr)^0.5

where
h(f) = harmonic frequency
MVAsc = minimum 3-ph fault level in MVA (at capacitors)
MVAr = 3-ph rating in MVA of the capacitor bank

If you have any harmonic current flowing with a magnitude almost the same as the calculated value, you'll face some problems.
Ways to solve harmonic problems are:
1. Install de-tuned harmonic filters or reactors in series with the capacitors. I've read somewhere (I think it was in an old Strike technical paper - www.strike.co.za, a company in South Africa who manufactures protective relays like their RLC-relay for capacitor protection) that they tune the filter a little lower than the harmonic frequency. For instance at the 5th harmonic they would tune it to 4,7. But I do not know why they do it like this.
2. Change capacitor location.
3. Change capacitor size.
4. Ungrounded capacitor banks - floating wye point to prevent harmonic current to flow to ground. However, due to safety concerns utilities should consider this options only as the last resort.


ABB do have a section who specialize in HV capacitors/capacitor banks (although I'm not sure if they are also located in NZ) - maybe you can contact them for some assistance.

Hope it helps.

Regards
Ralph

#4 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,601 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 12 January 2005 - 10:24 AM

Hello wareagle

A resonant supply will "ring" at its resonant frequency if there is some form of stimulation.
One of the causes of problems with a resonant supply, is harmonics that are close to the resonant frequency of the supply. Another problem is that if there are transients on the supply, these can also cause the supply to resonate for a period of time. The result of resonance is that inaddition to the normal supply voltage and frequency, there can be bursts of another voltage and frequency supperimposed, and depending on the Q of the resonant circuit, ths voltages can be very high.

Best regards,

#5 Guest__*

Guest__*
  • Guests

Posted 15 January 2005 - 12:55 AM

Mark
It sounds like you are describing utility distribution lines with voltages in the 12 kv or 13.2 vk level. I am familiar with harmonic problems that you are describing at industrial plants with the sub of large transformer at the site. Never had this problem on distribution line in a rural area.

Bob

#6 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,601 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 22 January 2005 - 07:28 PM

Hello Bob

Where I have seen issues is certainly in areas with higher voltage distribution systems, typically 11KV over head lines where there are local distribution transformers for each pump station. These transformers may be in the order of 100KVA at 400volt output.

If you close a contactor with power factor correction onto the output of the transformer, you get a high voltage surge due to the resonance of the capacitance with the transformer leakage reactance and the line reactance on the 11KV circuit. Any switching transients on the supply can cause this same resonance to occur with high voltages developed, especially when the load is light. If the load is high, the resonances are damped and much smaller in magnitude.

This supply resonance can cause problems with soft starters that are installed without a line contactor. When the pump is not running, the supply voltage is blocked by the SCRs only and if there is a resonant surge, the voltage can cause the SCRs to be damaged. MOVs across the SCRs tend to explode.

Best regards,

#7 Guest__*

Guest__*
  • Guests

Posted 24 January 2005 - 10:42 PM

Mark,

As Ralph pointed out, it seems that the source of harmonics sees a high impedance load at the resonant or near nerosant frequency causing it to develop high voltage.

If such is the case, then you need to detune the network to a new frequency by adding passive filter. Of course, the tuning will depend on the extent of study you make to the network and the economics involved.

Assuming a harmonic source having a significant 5th harmonic,

1) If the purpose is power factor correction only then you can tune it between 4.0 and 4.4 harmonics. However, this tuning does not provide significant improvement to the harmonic distortion levels.

2) If the purpose is to improve the power factor and the harmonic distortion then you can tune the filters between 4.4 and 4.8.

3) If you are planning to significantly dumped all the harmonics to the filter then you can tune the filters to exactly the fifth harmonics. However, this scheme requires careful study since dumping the harmonic content can overload the filters.

At any rate, economics dictates that tuning closer to the fifth harmonics requires higher rating for the reactor and capacitor thus it is more expensive.

Hope this helps.


Glenn Bonita

#8 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,601 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 26 January 2005 - 06:59 PM

Hello Glenn

Thank you for your comments.
Yes, detuning reactors etc can help to shift resonance away from harmonic frequencyies, but this does not alleviate problems due to the transient response of the suppy resonance.
The issue that I am concerned about, is the reaction of the supply to switching and fault transients on the network.

Best regards,

#9 jOmega

jOmega

    Senior Member

  • Full Member
  • PipPipPipPipPip
  • 254 posts
  • Gender:Male
  • Location:Midwest, USA

Posted 11 September 2006 - 06:51 PM

QUOTE(RalphChristie @ Jan 10 2005, 08:46 AM) View Post

Ressonance - frequency point where inductive (transformers, cables, etc) and capacitive (capacitors etc ) reactances are equal.
Mark:
As far as I know it becomes a problem only if a source of harmonics exists at the frequency where the impedances are the same. You can estimate this frequency with this formula:
h(f) = (MVAsc/MVAr)^0.5
where
h(f) = harmonic frequency
MVAsc = minimum 3-ph fault level in MVA (at capacitors)
MVAr = 3-ph rating in MVA of the capacitor bank
X----------------------- s n i p --------------------------------------------X

Ralph


Hi Ralph .... I hope you're still out there and can respond...
Happened to be reading your post from back in Jan '06 .... and
something struck me as being not quite right ....

specifically, h(f) as you defined it therein, and above.

The equation you stated does not render a harmonic frequency; i.e., 57 hZ, or 32 hZ, etc...

Rather, it renders a Resonant Frequency, expressed as a number that is a multiple of the fundamental (supply) frequency ....

Ergo, one would obtain numbers such as 5.18 which would correspond to a resonant frequency of 310.8 hZ which would place it in very close proximity to the 5th harmonic in a 60 hZ supplied system.

Ref: IEEE Std 519-1992 para. 8.2.1

Kind regards,


#10 bodolman

bodolman

    Junior Member

  • Members
  • PipPip
  • 4 posts

Posted 02 March 2007 - 08:21 AM

How many kVAR are you switching on at any one time? I have noticed that out on rural networks there are more likely to be multiple small pole mounted capacitor banks (in the order of 500kVAR to 1000kVAR)instead of one big capacitor bank.

I am not certain but could this method reduce the effects of resonance?



#11 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,601 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 02 March 2007 - 09:01 AM

Hello bodolman

Typical installations are irrigation pumps where the pump is operated from it's own transformer. There is no other load on the transformer.
The transformer is fed from an 11KVor 22KV overhead line.
Static correction is typically applied to the supply once the motor is running.

Best regards,

#12 jerry

jerry

  • Members
  • Pip
  • 1 posts

Posted 24 August 2007 - 11:14 PM

You may want to try a device that has non switching capacitance, high and low band pass filtering, and tvss all in the same unit. Its called a Circuit Master. you can look it up on the web. seems to work quite well in rural as well as manufacturing where long distances distort. Jerry

#13 EMartist

EMartist

  • Members
  • Pip
  • 1 posts

Posted 29 January 2008 - 04:41 PM

29 January 2008

Hi,

I'm a newcomer, and I thought I'd offer a sort of "resonance 101" which may or may not help with solving the issue of over voltages and equipment failure on your supply line.

If you suspend a steel bar and strike it, it will ring. This is a mechanical resonance whose frequency and duration depend on the inertia (mass), compliance (resilience), and the damping (friction) of the system under analysis. In the case of steel, compliance is low (steel is stiff) and damping (internal molecular friction) is low, so the frequency will be high and the transient will be long. The system is said to be under damped.

An identical bar of wood will resound with a thud. The damping is much higher, and is right around the amount the amount just necessary to prevent any ringing. The system is said to be critically damped.

An identical bar of soft plastic won't make much of a sound at all. The damping is still higher, and the system is said to be over damped.

In the world of electrical systems, inertial forces are represented as inductance (a larger inductance corresponds to larger inertia), compliant forces are represented as capacitance (a smaller capacitance corresponds to higher compliance), and damping forces are represented by resistive dissipation (a larger resistance offers more dissipation for a given current).

The differential equations which model both of these simple mechanical and electrical systems are identical, as are their transient and steady state solutions, so the analogy is completely accurate, and therefore offers insight.

In the scenario under consideration, if I understand it correctly, there is a long transmission line terminated on one end by a distribution transformer and on the other end by a step down transformer connected to a large irrigation pump motor through a contactor, and possibly soft start electronics.

When the contactor closes, you are banging the resonator, which is formed by distributed inductance of the transmission line which is proportional to length, and the lumped inductance of the transformer windings, together with the distributed capacitance to ground of the transmission line, the capacitance between the primary and secondary of the transformer, the large capacitors following the rectifiers on the input of the soft start electronics, and possibly a power factor correction capacitor across the transformer input. The damping in the circuit is negligible.

The banging is caused by the inrush current filling the empty capacitors in the soft start, or by the lower impedance, and hence higher current demand, of a motor stator with a stalled (stationary) rotor, which is a normal circumstance at start up.

More precisely, a banging is a shock wave excitation applied to the resonator. Altering the associated capacitances or inductances will do nothing except shift the frequency of the transient. In this particular scenario, there is an additional issue in that some of the reactances are distributed. This makes the power feed line a transmission line in the RF engineer's sense, that is, it has a characteristic impedance, and will propagate a shock wave excitation, which will reflect at impedance discontinuities and unmatched terminations. The result will be transient standing waves on the line, which will superpose with the line voltage to cause repeated brief (perhaps a few hundred microseconds) over voltages. A MOV is a clamping device which deteriorates each time current flows in it. In this situation, the clamp will be activated hundreds of times during a single transient, and it is for this reason the MOV melts down. A more durable clamp can be constructed from a high speed power rectifier and a suitably sized power resistor.

An extreme example of this sort of problem was the north eastern US power grid failure of a few years back. Since an RF transmission line propagates, there is a characteristic delay. The rate of propagation is controlled by the velocity factor of the line which depends on the materials and geometry of the line, but generally results in a velocity slightly less than the speed of light (light is also electro-magnetic in nature). A simple calculation shows that a single wire wrapped three times around Lake Erie will delay single phase power by about the right amount so that the relative phase in the three wires at a single point on the line is about the same as that for three phase power. Stated another way, another simple calculation shows the inductance and capacitance of the infamous Lake Erie loop (as it is called), gives a resonance around 60 Hz. It is easy to see how a power plant suddenly going off line would bang the loop, and how this might cause a few small problems now and again, including possibly, a cascade failure.

The best solution is to avoid banging the resonator at all. If you gradually apply pressure to a suspended steel bar, and then gradually release the pressure, it will remain silent.

A soft start circuit is in essence a single purpose variable frequency drive, designed to gradually start large motors by applying a ramp in voltage and frequency to the motor during starting to keep the current at modest levels throughout the process. For small motors, the motor is then operated at speed through the soft start circuit, and for large motors, the motor is synchronously switched to the line at a zero crossing of the current with a contactor. The operative word here is synchronous, which is necessary to avoid banging the line resonance with a contactor induced current discontinuity.

The purpose of the soft start is therefore to avoid banging the line, in addition to treating the motor and load more gently. If the soft start does not switch synchronously, this purpose will be fulfilled only by chance every now and again.

Even with a properly designed soft start, the inrush current to the electronics will bang the line when the control is started cold. This problem is averted by an inrush current limiter which is a sort of electronic soft start for the soft start. On a modern control this should be a standard feature, but often is not.

It is also important to realize that you can bang the line when you turn the motor off or suddenly remove the working load, so soft stopping and load monitoring can be important in some circumstances.

So, I hope this has been helpful.

Regards,

Peter


#14 elmger

elmger

    Junior Member

  • Members
  • PipPip
  • 3 posts

Posted 13 July 2010 - 10:31 AM

Supply resonances often cause problems especially in the transient.

The key to damping the resonance that occurs on switching transients is to provide wide band damping or energy absorption.

One way to do this is to add some non linear components such as a TVS. Alternatively a diode rectifier with capacitors on the DC side. The effect of this is to connect energy absorption to the AC system during the transient and then remove it when the tranisent dissapperas. Diode rectifiers do this well. This has been succesful in the past.

Hamish

hamish.laird@elmgnz.com




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users