ferroresonance mitigating resistance on open-delta VTs...

To me, this is a nonsense post.

There is no statement of the problem if there is one. The terms are bandied about without any explanation. There is no clear question. It is like getting a bag of parts and being asked to put them together into a circuit without a schematic diagram.

Bill

Oh, I know what he's talking about. I've no idea of the answer. And the last time I drew some broken-delta transformers on a protection schematic I was told we don't do that any more.

A different Bill

Have a bit of pity on him. This is the first serious question on this group for weeks. He does not understand ferro-resonance, few people do so he is looking for some guidance.

To get ferro-resonance you need to have an inductance that saturates in series with a capacitor. On distribution systems in the UK the classic way to get it is to blow one or two of a set of a three phase set of fuses, and then end up with the remaining live phase(s) feeding through a transformer to the capacitance of the disconnected line(s).

I assume that moris has an overhead line (probably long) with this VT at the end of it. He does not say what method of system earthing is in use on the network, or whether the VT is only used to provide an open delta voltage for earth fault protection, or whether it also has other windings with voltmeters, or distance protection relays connected across them.

To give him any sort of sensible answer one would also want to know whether there is also a system transformer in parallel with the VT.

At 33 kV in the Uk I never came across conditions where a damping resistor was required.

Again in the UK the rating of the open delta winding would normally be 50 VA, and this does not match the example figures quoted.

Morris does need to give us more information, but ferro resonance is always an interesting topic.

John

I do have some pity. However, if he is seeking help, he needs to ask better questions and explain his situation.

I must admit to not understanding ferro-resonance well. I can understand that a nonlinear inductor with an iron core can move onto resonance or not depending upon the current level in the winding. The explanations I have seen have been less than compelling and lucid.

My primary experience has been with the use of ferro-resonant transformers for charging capacitor banks. Since then the technology for charging capacitors has greatly changed. Direct charging from the line, even with transformers and rectifiers between the line and the capacitor bank is not an option. The inrush current is just too great and must be limited. Using a resistor for the purpose greatly reduces the efficiency.

To get around that Wabash ferro-resonant transformers were used. The idea was that this transformer would limit inrush current, Later in the charging cycle, it would act as a constant voltage transformer limiting how high the charge voltage would go.

In fact, these transformer were lousy. They heated up very badly. The current waveform was terrible with current spiking. The true rms current was three or four times the fundamental rms current. We finally did away with these transformers by using a current limiting (high leakage reactance) transformer with capacitive power factor correction.

I still would like to understand how these things were supposed to work.

Bill

Hi again,

Firstly, I'm sorry if my question is not that clear. The bottom line of my post is how to calculate the power and resistance value of a ferro-resonance suppressing resistor connected parallely to a broken (open) delta VT. I wonder if there is a general calculation method appliable to various system combinations (eg. different system voltages, different grounding methods, intended protection schemes etc.).

Despite open-delta VTs for ground protection are national utility company's standard requirement here in Turkey, that ferro-resonance suppressing resistor is something I never came across before.

Actually, this question is being asked to me by one of my colleagues. He is working on a GIS project recently and their foreign design company wants to use the suppressing resistor. Therefore, all I know about the system is that it is a GIS (probably 380 kV with at least 80 MVA transformer) for the moment. But I'm gathering all other details needed for answering my question such as a long overhead transmission line (probably the GIS is at the end of a long HV overhead line), grounding method and, core number/connection type of the VT and their intended use.

I'll update my post with system details.

Thanks again,

Moris

BIG SNIP

The answer to understanding all of these problems with saturating lumps of iron, whether the problem is magnetising inrush current, open circuit current transformers, or ferroresonance is to get a simple model which will show what is happening. The way I look at it is to simplify the magnetic component to two fixed inductors in series, the unsaturated inductance of the core and winding, and the inductance of the coil with no core inside of it. Then the high inductance withe coil unsaturated has a switch across it which closes when the flux in the core reaches saturation value, and opens again when the flux drops below saturation value. The beauty of a model like this is that you can construct it using a network analysis programme like SPICE, and then add in the other components which make up the actual problem.

Although in some ways the model is quite crude the graphs that are produced show waveforms that are very close to those measured experimentally. At different times in the past I have used the technique to model all three of the problems mentioned above.

I am not surprised that your ferro resonant transformers heated up. The iron losses would be high, as would the resistive losses in the transformer when it was passing peaks of current in saturation.

John

Hello Morris

BIG SNIP

At 380 kV I doubt whether it will be a wound VT. The signal is more likely to be obtained from a set of 3 single phase Capacitor Voltage Transformers. This is out of my field of expertise. I only worked on systems up to 132 kV. There are however known ferro-resonace problems with CVTs.

Look forward to more information.

John

---- Hide quoted text -

Hi John,

I don't want to mislead you; yes, it is a GIS with 380 kV HV level but the measuring point in question is on the MV side. MV level could be

34.5, 31.5, 20 or 6,3 kV. Therefore the VT is a wound type in middle voltage measuring cubicle. I'll let you know about the voltage levels with all other details.

P.S. You are right, it is the standard application here, too, to use CVTs for such high voltage measurings.

Moris

It has been a long time since I seriously considered ferro-resonance. The problem is simpler these days because of computer power available to almost anyone. Nonlinear problems have always been an analytical problem. Ferro-resonance is even worse than the typical nonlinear process because there is retention. The way the device is shut down can affect the start-up transient.

As you point out, the key to the problem is a workable model for the iron including hysteresis and eddy currents. Then, typical solvers for ordinary differential equations (ODEs), and presumably programs like SPICE, can handle the nonlinearity.

Bill

Has EMTP or it's modern counterpart ATP been tried? It has models which are much better suited to power network systems than SPICE.

I am afraid that I am showing my age ! Electrical circuit laws are the same whether you are working with micro amperes or kilo amperes, and I started using computers to help understand difficult circuit problems using the IBM programme ECAP around 1970. I moved on from there to SPICE after a few years.

I have continued to use SPICE intermittently up to my retirement from active work a couple of years ago. I still have SPICE on my computer.

I am sure that the newer programmes that you mention will do a very able job.

The important thing about using computer analysis is having sufficient understanding about what is happening in the circuit to be able to assess whether the answers that you are getting make sense. I am afraid that GIGO (Garbage in garbage out) still happens.

I was lucky in that for part of my career I worked in a research centre where it was possible to do limited real measurements on non linear behaviour, and then use computer modelling to match the results and then extend them to other cases.

John

---- Hide quoted text -

EMTP originated in the 1960's and was used on mainframes. In 1984 it eveolved into ATP which could run pn PC's. I'd be surprised if it doesn't predate Spice and it's clones. AND it is ideally suited for ferroresonance type problems amongst others. For more info pls see :

hope this helps.