Trap ( notch ) filters with multiple frequencies

[homoly]

Hello,

I would like to ask you if somebody can suggest me what configuration of trap filters could handle to mitigate 5th and 11th harmonics on the mains input voltage. It seems to me that for paralleling 2 trap filters tuned to 5th and 11th harmonics of mains the 2 frequencies are not enough different.

Is there some configuration of trap filters that should do the job and it should be built preferably from passive components ?

Any hints, suggestion and links are welcome.

Gabriel.

 

[tem1234]

Hi,

There's a lot of possible configuration for a harmonic filter, and every system is different so it's hard to tell on a public forum. What's the voltage of the filter, what type of non linear load...for what reason you're doing this (you have problem or you must respect utility standard?)

You should modelise your system with a software and try different configuration, observe the impact of the filter and then choose the best for your system.

you must be aware of resonance, since a filter will amplify other harmonic.

A 5th and 11th notch filter is one possibility, maybe not the best. With this kind of setup, you can have 7th harmonic problem even if you haven't before, so it's important to look at this. High pass filter are very common too.

 

[dpc]

I'd suggest working a vendor who specializes in harmonic filters if this is an actual application.

 

[homoly]

Thank you for the answer. The filter is for 3phase 400V application. This trap filter supposed to be used as an input filter for 12 pulse rectifier which is through isolation transformer feeding the inverter. The load is linear resistive.

 

[ozmosis]

Do you need a trap filter? If you have a 12 pulse rectifier arrangement and phase shift transformer, then this should be canceling the 5th and 11th within the tranx.

 

[LionelHutz]

Yes, I'll second the comment that the 12-pulse arrangement should cancel those harmonics.

 

[homoly]

The problem with those harmonics is not critical, indeed but it has to be reached strict THD value in the concrete application, so that is the reason why normally the additional trap filter is used tuned on 11th harmonics by us. After connecting this filter the 5th and 7th appears generated by transformer , what we normally do not care about, but now it is an application when filtering out additionally 5th will be desirable.

Nevertheless, I would like to ask if for the computer analysis mentioned above the paralell connection of 2 serial resonant circuit tuned 11th and 5th should be used or there is some other topology I could prefer. I have quite a rich experience with low power active, passive and digital filters in low power electronics but none with trap filters used in power electronic so that is the reason that I am looking for suggestion to double tuned trap filter realization here.

Btw could somebody suggest me what method or empiric values should I use to determine the source impedance of 3 phase distribution networks in US or UK?

Gabriel

 

[ozmosis]

homoly
Sorry, I'm not clear on what you are saying now. Harmonic currents are produced by the rectification process in combination with the charging of the internal capacitors in your VFD, not the transformer.
The 5th and 11th harmonic currents will be in phase from a 12-pulse rectifier and therefore be cancelled out within the phase shift transformer, assuming a 30degree shift of phase. The 5th,7th,17th and 19th order harmonics of 12-pulse rectifier will typically be cancelled. Your 11th may be higher but the overall current level will be lower.
I would say a correctly designed 12-pulse configuration with the correct designed balanced phase shift transformer should eliminate 5th harmonic orders from your VFD. There should be no need for the trap filter in my view.

 

[sibeen]

I'm not sure I understand why some are saying that an 11th harmonic is cancelled out in a 12 pulse rectifier. In theory the harmonics caused by a rectifer of pulse number n is:

xn plus/minus 1, where x is an interger value (1, 2, 3..).

So, for a 12 pulse rectifier the major harmonics should be the 11th, 13th, 23rd, 25th etc.

In practice there is some 5th harmonic that is normally also generated, but of a fairly low level.

homoly, rather than trying to fit some tuned filters at the input of the rectifier you may consider using an in-line choke.

What value of THDi are you trying to achieve?

 

[homoly]

Sibeen you are right, the major harmonic is 11th.
there is already an inline choke involved, the desired THD should be 3.5%.

A short description of the filter: standard series LC trap filter tuned on 11th harmonics.Capacitors of the filter are also used for power factor correction. This filter cancels out effectively the 11th harmonics, but after connecting the filter the 5th ( major ) and 7th harmonics appears. My idea was if it would be possible to cancel out those harmonics appearing after filter connection.

I suppose the real problem is this power factor correction and filter function combination in one block. As said by osmosis without filter with 11th harmonics present the THD is lower ( around 2,3% but the power factor is out of requested limits) than with filter connected which is tuned to 11th of harmonics - in this case the THD created majorly by 5th and 7th is higher ( around 4% but PF is OK).

So as a result I see basically two solutions - use double tuned trap filter which together with 11th harmonics should cancel also 5th, or redesign the PF and filter stage to split up to 2 separate blocks.

 

[sibeen]

Whenever I have worked on this type of application I've always used an external PF correction facility. The PF capacitors would be switched depending upon the load by a small PLC.

3.5% is fairly low. Any particular reason that the THDi needs to be at such a low level. In the UPS field, which I'm familar with, we normally try to get down to about 5 - 6%.

 

[homoly]

Sibeen, in this application the current sensing relay is used and the capacitors are switched off in two stages at small loads in order to avoid overcompensating thus being a capacitive load for mains and especially generators.

The THD level is defined by a customer ( unfortunatelly the marketing was not able to interpret our opinion to the customer that 4% is fine ). So that is the reason for my effort to try to suppress the THD furthermore.

Nevertheless I am curious how the proper solution for double tuned trap filter should look like , and I would like to know the typical source impedancies US and UK or place when I can get this info. It would be nice to know this also in order to make the designed filter capable of working connected to difference load sufficiently.

 

[electricuwe]

Two recommendations to you:

-teach you sales guys not to accept voltage THD specifications when selling rectifiers. Most probably you suffer from existing harmonic voltages on the grid

- if you provide a trap for a higher order harmonic also trap provide traps for the lower order harmoncs typical for rectifiers (5th and 7th, if a trap for 11th is added). The high order filter always bears the risk that you get a high impedance for a lower order harmonic.

 

[sibeen]

homoly, you are talking about 3.5% current distortion here and not voltage, I assume.

What is the size of the application? Maybe you need to consider an active harmonic filter on the front end.

 

[homoly]

sibeen, yes it is current THD and the application is 800 KVA UPS. The catch is this is not directly the customer specification, they have contracted a consultation company. This company how I believe compiled some parameters for specification normal for UPS units for much lower power levels and our sales department unfortunatelly accepted this.

What you are saying about cancelling 5th and 7th harmonics is not valid in this case - the opposite is measured: after connecting filter ( with PFC function ) tuned to 11th harmonics the 5th and 7th are raised.

 

[jghrist]

Nevertheless I am curious how the proper solution for double tuned trap filter should look like , and I would like to know the typical source impedancies US and UK or place when I can get this info. It would be nice to know this also in order to make the designed filter capable of working connected to difference load sufficiently.

The 400 volt source impedance is mostly the MV-LV transformer impedance. The utility should be able to provide either the source impedance of the fault currents that will allow calculating the source impedance.

 

[homoly]

sibeen, yes it is current THD and the application is 800 KVA UPS. The catch is this is not directly the customer specification, they have contracted a consultation company. This company how I believe compiled some parameters for specification normal for UPS units for much lower power levels and our sales department unfortunatelly accepted this.

What you are saying about cancelling 5th and 7th harmonics is not valid in this case - the opposite is measured: after connecting filter ( with PFC function ) tuned to 11th harmonics the 5th and 7th are raised.

 

[homoly]

jghrist, thank you for the information.

 

[sibeen]

homoly, OK, it sounds like you're caught between a consultant who has no bloody idea and a salesman who is looking for a commission, unfortunately you're not the first who has ever run into this scenario.

I'd be getting a copy of IEEE 519-1992 and trying to arrange a mutually agreed compromise with the consultant and the client. If this doesn't work I'd look seriously at putting an active harmonic filter on the input of the UPS. These devices can be programmed to inject anit-phase currents at different levels for individual harmonics. I have used them on large UPS systems and they can get the THDi down to about the 3% level. They can also be programmed to adjust the input power factor.

I believe this solution would be cheaper that designing, building, commissiong and testing of passive filters that may still not reach the required specifications.

At the end of the process I'd make sure that senior management within your organisation are aware of the problems and the problems caused by the sales department, and strongly suggest that no sales commissions be paid on this project.

Active harmonic filters can be obtained from Chloride, ABB, Merlin Gerin and others.

 

[homoly]

sibeen, thank you for your practical and very useful advice!