VFD Filters

Status
Not open for further replies.
B

baumancl

Member
I am aware of the various issues to consider when powering a motor with a VFD like harmonics, voltage spikes, reflected waves at the motor terminals, cables lengths..etc. I am also have a decent understanding of when to apply filters.

I know that the dv/dt filter takes care of line voltage spikes and that sinewave filters smooth out the the voltage waveform to more closely match a pure sinusoidal power source.

My main question is what is the real difference between a dv/dt filter and a sine wave filter because it seems that some manufacturers "sine wave" filters claim to also address dv/dt problems. Which one is better?

Is there a clear cut way to determine in what situations to apply a dv/dt filter or a sinewave filters? Should they ever be applied together?
 
baumancl said:
I am aware of the various issues to consider when powering a motor with a VFD like harmonics, voltage spikes, reflected waves at the motor terminals, cables lengths..etc. I am also have a decent understanding of when to apply filters.

I know that the dv/dt filter takes care of line voltage spikes and that sinewave filters smooth out the the voltage waveform to more closely match a pure sinusoidal power source.

My main question is what is the real difference between a dv/dt filter and a sine wave filter because it seems that some manufacturers "sine wave" filters claim to also address dv/dt problems. Which one is better?

Is there a clear cut way to determine in what situations to apply a dv/dt filter or a sinewave filters? Should they ever be applied together?
Click to expand...
Do you want to clean up the line side or load side?
Typically, dv/dt filters are used on the load side to reduce the rate steep-fronted voltages typical of a PWM output voltage - we'd normally aim for less than 1600V/us on a modern motor. But they don't do a whole lot for line side voltage spikes.
Sinusoidal filters on the motor side are generally more complicated and a lot more expensive than simple dv/dt filters.
http://www.reo.co.uk/files/cnw933_2_115_2.pdf
For a VFD retrofit on an older motor, the sinusoidal filter might be the safe way to go.

Active front end (AFE) configurations can reduce the harmonics on the line side. At a cost. From recent experience, it may be anything from 1.5 to 1.8 times more expensive.

There is no "one size fits all" answer.
 
I guess it might be easier to give an application. On deep submersible well pumps with long cable lengths from the VFD we have applied sinewave filters to address surpressing peak voltages at the motor terminals due to reflected waves from harmonics. This comes as advice from design related problems as well as recommendations from manufacturers of the VFD's as well as the well pumps.

I am trying to determine how to most efficiently solve the problem in future applications to best economically serve the client. I never have included a dv/dt filter in a design as sinewave filters have always seemed to solve any potential problems and filter out harmonics which cause the standing voltages waves at the motor terminals. We have never come to the conclusion that peak voltages have caused problems in any of our designs involving VFD's powering motors (mostly pumps and fans). Do sinewave filters offer adequate protection against voltage spikes as well?

In summary, I am looking for a application where I might know going into design to consider a dv/dt filter when motor lead lengths are not significantly long and the requirments to justify a sinewave filter are not present. Are there certain loads that might cause more frequent or worse voltage spikes?
 
baumancl said:
I guess it might be easier to give an application. On deep submersible well pumps with long cable lengths from the VFD we have applied sinewave filters to address surpressing peak voltages at the motor terminals due to reflected waves from harmonics. This comes as advice from design related problems as well as recommendations from manufacturers of the VFD's as well as the well pumps.

I am trying to determine how to most efficiently solve the problem in future applications to best economically serve the client. I never have included a dv/dt filter in a design as sinewave filters have always seemed to solve any potential problems and filter out harmonics which cause the standing voltages waves at the motor terminals. We have never come to the conclusion that peak voltages have caused problems in any of our designs involving VFD's powering motors (mostly pumps and fans). Do sinewave filters offer adequate protection against voltage spikes as well?

In summary, I am looking for a application where I might know going into design to consider a dv/dt filter when motor lead lengths are not significantly long and the requirments to justify a sinewave filter are not present. Are there certain loads that might cause more frequent or worse voltage spikes?
Click to expand...
Interesting that you should ask about this application.
Here is an excerpt from a report I did:
Background
At Moorhall PS there are two pump houses, each with two borehole pumps. The pumps are driven by submersible motors each rated at 59 kW, which are powered by Mitsubishi PWM inverters. There have been repeated insulation failures on two of the motors, one from each pump house.

Measurements
Voltage waveforms were recorded at 40Hz, 45Hz, 50Hz, and 53Hz for all four motors at the boreholes.

Summary of principle measurements from the waveforms
The voltage waveforms measured at the pump houses are typical for PWM Inverter fed motors at the end of a significant cable run. It should be noted that the waveforms were measured at the well head, and not the motor terminals.
Voltage Overshoot
At the points it was measured the voltage overshoot is within the limits of Technical Specification IEC TS60034-17 third edition. The limit given for the switching times measured is about 1200V. Using figures from the Gambica Technical Report No1, (second edition) the peak voltages recorded give a very low probability for partial discharge and even if it did cause PD, the number of pulses to breakdown would exceed 1013. Bearing in mind that the switching frequency is below 1 kHz, the motor will never see that number of pulses.
Rate of change of voltage (dv/dt)
Rate of change of voltage was very fast. In all cases it was either outside the limits of the IEC TS60034-17 technical specification, or marginal. For the rise times measured the maximum dv/dt should not be greater than approximately 1600V/us. The average dv/dt measured at the well head was 2500V/us.
Insulation resistance
The insulation resistance was checked and found to be very low even on the motors which had (not yet) failed.
Conclusion
The insulation failures have the appearance of partial discharge failures such as would be created by the very high electric field intensity caused by the very fast dv/dt. The recorded dv/dt is as high as any we have measured at or close to the motor terminals.
In summary, the installation of dv/dt filters to reduce the rate of change of voltage applied to the motor is highly recommended.

Well, dv/dt filters were fitted and I was invited back to repeat some of the measurements.
Before:
P1before.jpg


After:
P1after.jpg


Note the horizontal time scales differ by an order of magnitude. Before is 1us per major division, after is 10us per major division. Peak voltage and dv/dt were both significantly reduced and, most importantly, there have been no motor failures since the installation of the dv/dt filters - three years ago.

I had no particular commercial interest in this installation. It wasn't our kit. I was just paid to come in as an independent consultant.
 
Status
Not open for further replies.
Top