Which would be better choice to suppress harmonics generated by variable speed drives (VSD): phase shift transformers, or traditional 5% line reactors? It looks that two phase shift transformers (one at phase angle 0 and the other at -30) feeding two VSD’s will eliminate the higher order harmonics; however both drives have to be running at the same time. What if only one drive is running? Thanks
Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations cowski on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
Harmonic Suppression
- Thread starter timm33333
- Start date
timm, why would you bother having a 0 degrees phase shift transformer?
The reactor would certainly be the better way to go. As you've already surmised, using a phase shift will only work if both drives are operating, but also both drives would have to contain equal input currents for them to operate ay their best. Also, even at best case, it won't be eliminating the higher order harmonics.
What a phase shift transformer will do is make the two drives appear, to the upstream source, as a single twelve pulse VFD. Under perfect conditions you would be then only having harmonic currrents of the order n*12 +-1; so that your predomninant harmonics will be the 11th, 13th, 23rd, 25 etc. So it has gotten rid of your lower end harmonics, the 5th, 7th etc.
The reactor would certainly be the better way to go. As you've already surmised, using a phase shift will only work if both drives are operating, but also both drives would have to contain equal input currents for them to operate ay their best. Also, even at best case, it won't be eliminating the higher order harmonics.
What a phase shift transformer will do is make the two drives appear, to the upstream source, as a single twelve pulse VFD. Under perfect conditions you would be then only having harmonic currrents of the order n*12 +-1; so that your predomninant harmonics will be the 11th, 13th, 23rd, 25 etc. So it has gotten rid of your lower end harmonics, the 5th, 7th etc.
- Thread starter
- #3
Actually some people think that for 18 pulse VSD's, phase shift transformer is a better choice. Is it possible to re-arrange the secondary of a phase shift transformer to suppress the harmonics generated by a single variable speed drive (even though the second drive is not running)?
Far too many variables to say one approach is better than the other. Neither approach "will eliminate higher order harmonics", in fact the phase-shift approach just lowers lower order harmonics (5th and 7th) and introduces higher order harmonics instead. In many cases that is beneficial.
You can run the two phase shifted transformers into the same VSD by using two different rectifier front ends. That way you get the benefits of 12-pulse operation with having to load balance the two VSDs. For 18-pulse operation you need 3 phase shifts and 3 rectifiers.
You can run the two phase shifted transformers into the same VSD by using two different rectifier front ends. That way you get the benefits of 12-pulse operation with having to load balance the two VSDs. For 18-pulse operation you need 3 phase shifts and 3 rectifiers.
Is it possible to re-arrange the secondary of a phase shift transformer to suppress the harmonics generated by a single variable speed drive (even though the second drive is not running)?
timm, if one drive is off then the only thing a phase shifting transformer can do to the harmoic content of the other drive is to phase shift it. It won't effect the amount of harmonics, nor their order.
timm, if one drive is off then the only thing a phase shifting transformer can do to the harmoic content of the other drive is to phase shift it. It won't effect the amount of harmonics, nor their order.
The grossly over simplified version of why the phase shift works is that in a 3 phase non-linear device like a VFD, even order harmonics cancel each other out, and 3rd order (triplen) harmonics cancel each other, but odd order non-triplen harmonics are additive. By shifting 1/2 of the load through a transformer by 30 degrees, much of those non-triplen harmonics become even with relation the other half at that same time point, which makes them cancel. The reason you have the zero shift transformer is just to match impedance, otherwise all of the current would try to go through the lower impedance half.
If you add another transformer, much of the remainder become divisible by 3. So an 18 pulse drive has 2 (more) transformers to create the two phase shifts, and it has 3 separate bridge rectifiers, but all three are feeding the exact same load, the inverter, so all of the bridges therefore share the load equally.
If you do not have the load being shared EXACTLY in either of these configurations, it doesn't work as well, so doing it with multiple separate drives with transformers can only work if the drives are sharing the load evenly. Doing it without the other drive does nothing at all.
A simple 5% line reactor does absorb some harmonics, but can almost never be enough to mitigate a single drive for meeting IEEE 519 limits. But if the drive is small and the system is large, it may be fine in terms of the effect that drive has on the total harmonics at the PCC. So we cannot tell you if it is better or not, each and every installation must be evaluated on its own merit. The main reason that you see engineers require 18 pulse drives however is because if all of your individual devices within a facility will mitigate themselves to acceptable levels, the cumulative effect is that the levels at the PCC will carry through to being acceptable. So it's kind of a "cheat code" way of engineering a system that you don't know anything about in advance. That cannot be said of just slapping a line reactor ahead of a drive. Might be ok, might not.
"Will work for (the memory of) salami"
If you add another transformer, much of the remainder become divisible by 3. So an 18 pulse drive has 2 (more) transformers to create the two phase shifts, and it has 3 separate bridge rectifiers, but all three are feeding the exact same load, the inverter, so all of the bridges therefore share the load equally.
If you do not have the load being shared EXACTLY in either of these configurations, it doesn't work as well, so doing it with multiple separate drives with transformers can only work if the drives are sharing the load evenly. Doing it without the other drive does nothing at all.
A simple 5% line reactor does absorb some harmonics, but can almost never be enough to mitigate a single drive for meeting IEEE 519 limits. But if the drive is small and the system is large, it may be fine in terms of the effect that drive has on the total harmonics at the PCC. So we cannot tell you if it is better or not, each and every installation must be evaluated on its own merit. The main reason that you see engineers require 18 pulse drives however is because if all of your individual devices within a facility will mitigate themselves to acceptable levels, the cumulative effect is that the levels at the PCC will carry through to being acceptable. So it's kind of a "cheat code" way of engineering a system that you don't know anything about in advance. That cannot be said of just slapping a line reactor ahead of a drive. Might be ok, might not.
"Will work for (the memory of) salami"
An 18 pulse drive already has phase-shifting. It is essentially three 6 pulse drives phase shifted by 20°. Phase shifting by 30° will eliminate 5th and 7th harmonics in a 6 pulse drive if the transformers have identical ratios and impedances, and if the drives have the same delay angles and load current.
You won't have any reduction if only one of two drives are operating, but maybe the distortion of a single drive is within acceptable limits and you only need suppression if you have two drives running.
You won't have any reduction if only one of two drives are operating, but maybe the distortion of a single drive is within acceptable limits and you only need suppression if you have two drives running.
- Thread starter
- #8
How long does it take for the harmonics to overheat the incoming cables? I mean to say that if we have 2 drives which are fed by phase shift transformers, and the individual harmonic distortion of each of them is not acceptable. So when we start the first drive, there will be harmonics in the system as the phase shift transformers will not start suppressing harmonics yet. Let's say after 15 seconds we start the second drive, now phase shift transformers will start suppressing harmonics so the harmonics in the system will come under acceptable limit and system will be Ok. Will this 15 seconds be enough to overheat/damage the upcoming cables?
timmm, I really don't know where you're going with this. If you're looking at the heating effects of harmonics over such a short period of time then your cables are woefully undersized. In fact I've never even heard of someone trying to use harmonic mitigation because of heating effects.
If you have two separate drives, and use a phase shifting transformer upstream of one of them, to shift the voltage by 30 degrees, it will not effect the current at the input of either of the drives one jot. What it will do is mitigate the harmonic content at the point of common coupling of both the drives; ie at the upstream point.
jraef
The reason you have the zero shift transformer is just to match impedance, otherwise all of the current would try to go through the lower impedance half.
On a single drive with a 12 pulse bridge, sure; but with two individual drives I wouldn't be putting a 0 phase shift transformer in. timmm was talking about the two individual drive scenario, so that's what I responded to.
If you have two separate drives, and use a phase shifting transformer upstream of one of them, to shift the voltage by 30 degrees, it will not effect the current at the input of either of the drives one jot. What it will do is mitigate the harmonic content at the point of common coupling of both the drives; ie at the upstream point.
jraef
The reason you have the zero shift transformer is just to match impedance, otherwise all of the current would try to go through the lower impedance half.
On a single drive with a 12 pulse bridge, sure; but with two individual drives I wouldn't be putting a 0 phase shift transformer in. timmm was talking about the two individual drive scenario, so that's what I responded to.
- Thread starter
- #10
LionelHutz
Electrical
What level of harmonic reduction do you want? What is the site? What percentage of the total system load is VFD's?
Generally speaking, I find 5% reactors get you down to about 30% current THD and most plants find this acceptable.
At about 20x the cost of a reactor, the MTE Matrix AP filter will get you to 5% current THD. For similar costs, the Mirus Lineator will get you to about 8% current THD. My testing found the TCI H7G filter didn't perform any better than a line reactor.
Generally speaking, I find 5% reactors get you down to about 30% current THD and most plants find this acceptable.
At about 20x the cost of a reactor, the MTE Matrix AP filter will get you to 5% current THD. For similar costs, the Mirus Lineator will get you to about 8% current THD. My testing found the TCI H7G filter didn't perform any better than a line reactor.
- Thread starter
- #12
Acceptable limit would be 5% voltage distortion and 15% current distortion. There is no transformer on the supply side of PCC because the supply is direct from 4.16 kV utility. About 75% load is VFD (2 large VFD's). Due to the high proportion of VFD loads, perhaps reactors will not work here and the only practical option would be phase shift transformer.
LionelHutz
Electrical
What VFD? Most of the 5kV VFD's have 12-pulse or 18-pulse or some other form of harmonic mitigation built in.
- Thread starter
- #14
- Thread starter
- #17
LionelHutz
Electrical
This is why you need to provide some real details with your question. I believe the Harmony is a 18-pulse design. At any rate, it will easily meet your harmonic requirements as supplied.
- Status
Similar threads
- Locked
- Question
- Locked
- Question
- Locked
- Question