VFD Load Reactor / Ferrite Core Placement

[emoltzan]

We are currently working on solutions to reduce bearing currents in our motors, that lead to raceway fluting / EDM. With our VFDs we use load reactors and now are considering ferrite cores on the VFD load side. My main question is; after reading several references/recommendations I understand you should place the reactors/filters as close to the drive as possible. None of these sources/references dove into the reason/science as to why it should be placed near the switching device. Can someone please enlighten me with the science and not just the blanket statement?

 

[WWWJD]

Peak voltages at the motor are due to the fact that you're sending a high frequency PWM signal along the motor leads. The motor leads have an associated inductance and capacitance value with them, depending on conductor type, length, insulation, etc. You have to think of motor leads in these installations more like transmission lines in RF systems. The impedance of the transmission line (motor leads) and the load (motor) are quite different. At the point of mismatch is where the damage is done. When you place the reactor at the source end of the transmission line, you're lengthening the voltage rise time, adding inductance to the system, thus lowering the peak voltage the overall system and ultimately the motor, sees.

If you added a reactor to the load end, you'd likely experience failure in the reactor due to voltage overshoot as opposed to the motor. There are load end filters called wave traps, that are some kind of RLC network, which dumps excess energy to a bank of resistors and earth(?), but they're quite large and much more costly than a simple reactor. There are also sine wave filters, which again, are a RLC network but placed at the drive end. Again, they're quite large and more expensive, but they do a good job of making the voltage signal down the line much more sinusoidal. I think if I were in your shoes, I would first make sure that my motors are rated premium efficient; these motors have tougher insulation and a higher corona inception voltage usually. Second, I would definitely use a load reactor at the VFD end of the system. Thirdly, I'd investigate options for adding a grounding ring to the motor rotor shaft, or using ceramic bearings (non-conductive); these break the current circulation path across the bearing which is what's causing your fluting to begin with.

My description probably lacks some polish; not an engineer but I do occasionally sleep at a Holiday Inn.

 

[EdStainless]

Since the HF harmonics are being produced there you want to suppress them as close as possible. Doing it further away just increases the risk of sympathetic excitation.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[FacEngrPE]

The attached Eaton paper on dV/dt filters also applies to load reactors, except that load reactors are less expensive and less effective (but often good enough).

 

[edison123]

If you are looking at reducing bearing currents, insulating one/both side bearing housing and properly isolating the bearings from ground will be a cheaper and permanent solution.

Muthu

http://www.edison.co.in