Transcript
Introduction
Hello, everyone, and welcome to today’s white paper webinar. Today, we’re here with Wes, and we’re talking about voltage imbalance and its effect on variable frequency drives. The VFD is the modern way to power a motor. In these days, most motors don’t really see the AC voltage right from the utility. They’re shielded from that, but the VFD itself does see that.
The VFD has some very interesting and different sensitivities of voltage imbalance than a traditional induction motor. Wes is going to review some of the effects of voltage imbalance on a VFD here in this paper.
Voltage Imbalance and Its Effects on VFDs
Voltage imbalance is a small signal imperfection with large signal consequences in VFDs. Unequal phase magnitudes and angles drive disproportionate diode current, accelerating diode capacitor heating, and amplify DC bus ripple. This stresses the inverse switchers and motor insulation, causing premature failure of equipment.
Here’s the typical VFD power path. You have the rectifier, the DC link, inverter, motor, shown here in figure one, page one. In a balanced system, each line-to-line voltage surpasses the DC link by the same amount, so six diodes conduct equal time pulses.
But just 1% unbalance, those pulse sites diverge. At 5%, the pulse site can carry twice as much of the current as the shortest. The more current moving through the diode, the hotter they run. Hotter diodes have a decreased forward voltage drop, which invites even more current and more heat. This is a classic example of thermal runaway.
DC Link Capacitor Effects
Meanwhile, in the DC link capacitor sized for a ripple of three hundred and sixty hertz, which is six times the line frequency, now sees higher ripple and higher RMS ripple current. As the capacitor heats and ages, the ripple grows further, and the drive can trip on DC bus overvoltage or even blow the link fuse.
An example of this is on a seven point five horsepower drive. The DC bus ripple climbs from three and a half when balanced to 18.6% at 5% unbalance. Over the same sweep, the mean bus voltage rises from six hundred and fourteen volts to six hundred and twenty-seven volts. Those changes push more RMS ripple current through the capacitors, sharply raising the capacitor heating factor and shortening the capacitor’s life.
Heightened ripple also boosts instantaneous DC bus peaks by as much as a hundred volts on a four hundred and eighty volt system at 5% unbalance. This shrinks the inverter device margin on each switching event. Figures two and three on page one visualizes negligible unbalance versus 5% unbalance and the resulting ripple current.
Motor and Insulation Impact
When those higher, faster DC bus peaks reach the motor terminals as sharper surges with long cable runs, they can erode insulation and induce partial discharge. Unequal diode conduction also injects ripple and harmonics into the DC link. The inverter reflux that as torque pulsation and acoustic noise.
A well-documented result of unbalance is extra motor heating. 2% unbalance at full load can raise winding temperatures by thirty degrees Celsius and half the insulation life. Figure four on page two shows the triple and harmonic signature recorded on a VFD under voltage imbalance, which was captured by a PMI Revolution.
Monitoring and Diagnosis Workflow
Because diode and current pulses are narrow, approximately sixty electrical degrees, you need a high-resolution waveform capture to really see them. The workflow outlined in this paper is to record RMS voltage and current, THD for both, voltage and current unbalance, and harmonic magnitude from three to fifteen to spotlight the triplens.
Figure five on page two is a waveform capture. The voltage looks uniform, but the currents are asymmetric. Classic unbalance symptom.
To separate supply problems from failing bridge or drive, use the ANSI C84 style overlay, the voltage unbalance and current unbalance as seen in figure six. Rule of thumb, if the current unbalance is ten to twenty times the measured voltage unbalance, the source is probably fine and the drive or front end is the culprit. If both rise together, the supply is likely the issue.
Figure six on page two shows the overlay approach with the voltage unbalance, one to one and a half percent, and the current unbalance greater than 35%. An extreme mismatch, the overlay points to the source likely being unbalanced.
Fixing Voltage Imbalance
To fix unbalance, you want to start with the least costly fix, which is the source. Correct open delta secondaries, single-phase transformer banks, unequal single-phase loading, and blown capacitor bank fuses. Where you can’t make it perfect, add series impedance. DC link chokes cut capacitor RMS ripple current into the inverter.
For important installations, apply prudent derating. NEMA MG-1 guidance suggests about ninety-five percent load at 2% unbalance, eighty-eight percent at 3%, eighty-two percent at 4%, and get the VFD maker’s curve for your model.
For tougher cases, active front end or AFE drives can equalize currents and tolerate modest asymmetry, but the actual performance is vendor algorithm dependent.
Bottom Line Summary
The bottom line, small unbalance, big consequences inside six-pulse rectifiers. It skews diode current, amplifies DC bus ripple, heats capacitors, squeezes inverter margin, and shortens motor life. The practical playbook is early high resolution monitoring, source side correction first, and impedance derating and electronics based solutions where needed.
Discussion: Why Voltage Imbalance Is a Silent VFD Killer
Thanks, Wes. The bottom line here is that voltage imbalance is kind of a silent degradation of VFD. Heat is a real killer of semiconductors. The average lifetime of something like a diode or transistor is cut in half with every ten-degree Celsius rise. So if some of these pairs of diodes are just running a bit hotter than the other pairs, even if it doesn’t go into thermal runaway, that uneven heating will cause a shortened life for the pairs of diodes that are suffering under that extra burden.
So if we see the waveforms down here, some of the diodes are conducting a higher current. If some pairs are, these pulses are higher than others, the ones that are higher are gonna make those diodes run hot, and the VFD is going to have a shorter life than if that heating were very balanced.
That results in a customer VFD failure earlier than it should have, and that can be tough to figure out why, because the voltage may be balanced at that time, or it may not be obvious that a one or two percent imbalance just caused a gradual shortening of the life. It’s not an immediate catastrophic failure.
Ripple Effects Through the System
Like Wes said, that kind of ripples through the entire chain here. This capacitor has higher ripple on this DC bus, and there’s a faster heating factor is something that VFD manufacturers worry about. That extra current in the capacitor makes it run hot.
And as Wes mentioned, if this DC bus has 100 volts of ripple on top of it, that’s a higher voltage on the transistors, on the motor winding, that rides on top of everything else. Everything downstream is generally either seeing a higher voltage or a higher current than it would normally see with voltage imbalance, and that can also cause issues on the motor itself.
Effects on the Utility: Triplen Harmonics
Now those are all disadvantages that the customer sees for the utility. The utility also has detrimental effects. Like Wes said, these triple harmonics. These are harmonics that are not present with a normal VFD. A typical six plus VFD with even pulse heights won’t have any third harmonic or no triple harmonics. It’ll just be the classic fifth and seventh, 11th and 13th or 17th and so on, multiples of six plus or minus one.
But those are still there when you have voltage imbalance, but you also have third harmonic and other triple harmonics pop up that weren’t supposed to be there. So that leads to a higher current distortion from the customer, which then translates into higher voltage distortion for the utility and higher heating on the transformers.
If the customer’s filtering harmonics, they’re likely not filtering the third harmonic because they don’t expect it to be there. So voltage imbalance on a feeder overall will lead to higher voltage distortion on that feeder because of all the VFDs that draw extra unexpected harmonic distortion because of that. So if you see a VFD creeping up on a feeder, you want to make sure your voltage is as balanced as possible on that feeder to try to tamp down that effect.
Applies to All Rectification Front Ends
This is not just VFDs. We’ve talked about VFDs here, but really any sort of three-phase load that has this rectification front end. This is more and more common for any sort of modern load where the voltage coming in is rectified to a DC bus and that powers whatever the load is. It could be a VFD, it could be anything else. Some industrial three-phase load using this AC to DC converter block is very common these days. Any of these rectification front ends will have the similar effect with voltage imbalance.
Active Front End as a Solution
Like Wes said, one solution for this on the customer side is this active front end. This is a whole new architecture that is more complex and more expensive, but a lot more efficient. With an active front end, those diodes are replaced by transistors that are controlled by the inverter. The VFD itself decides when to switch it. They’re not passive diodes that are forced to switch when they’re forward biased.
With that sort of architecture, it’s a lot smarter architecture. It can change its switching pattern based on the voltage imbalance to undo that effect. Of course, these are more expensive and not as common, but this is one way to address the issue.
Additional Resources
We also have a one-hour webinar on voltage imbalance if you’d like to learn, and also on VFDs. If you’d like to learn more about either one of those topics, feel free to go to our website and click on the training tab and for classes, and you’ll see free one-hour webinars on voltage imbalance in more detail and also on VFDs in particular.
If you have a question later, give us a call anytime at 1-800-296-4120 or send an email to support@powermonitors.com. Thanks for tuning in, everyone. Have a great day.
