Picking the Wrong Harmonic Solution for an Application…. Resonance!

Earlier this year, I was asked to test a drive/filter combination in West Texas to try to figure out why the customer was experiencing nuisance drive trips and drive failures… the above graphic was the voltage wave trace I got when looking at the circuit between the filter and the drive, i.e. load side of the passive filter and line side of the ASD. Not pretty! Below is the individual phase to phase traces for a clearer understanding. A-B, B-C, C-A sequencially.

I highlighted a few of the issues that immediately jump out about the installation… resonance, notching, and transients. This is an ugly waveform and indicative of an issue associated with the mismatching of a filter with a drive, but what issue? Where would you start to look for a reason for the abnormalities associated with the voltage waveform. 

We have two issues presenting here, not just a single challenge. If you said that the transients/notching and the resonance were the two issues you would be correct. The resonance conditions is indicative of the over-excitation of the circuit associated with too much capacitance reactance ahead of the drive. The harmonic filter in this case featured a high capacitance reactance to kW ratio, i.e. prox.40% kVAR to kW. Remembering ELEC 101,

So, the greater the value of the capacitance, the more you shift the resonance point to the left, i.e. lowering the frequency at which resonance begins to develop. In this case, the system harmonic is in fact the harmonic being created by the drive, being injected back into the filter. Harmonic Passive Filters do feature a capacitive reactive element, as do Active Harmonic Filters with their EMI passive filter element. The Mirus International graphic highlights what can happen when this resonance exists…

Knowing that the amount of line side filter capacitance is creating the resonance, what then would account for the transient event? How about damage created by the presence of the resonance to begin with? If the drive had been exposed to the voltage resonance condition for an extended period of time, then logic would indicate that the drive may have been compromised by the resonance?

Yes… the resonance present would impact the front end diode bridge and have accelerated the aging of the DC bus caps. So, the selection of a high capacitance reactance passive filter created the resonance condition that ultimately damaged the drive itself, resulting in voltage transients and notching. Once the drive was rebuilt with a new diode bridge and new DC bus cap, the transients and notching disappeared. But, the resonance was still present. Capacitor Isolation at no load and low load will not prevent a resonant condition.

The only way to minimize the potential for this scenario occurring within your harmonic mitigation scheme is to specify a low capacitance reactance to power ratio for the filter construction. Below is a guide specification I use when reviewing a harmonic mitigation project…

• To ensure compatibility with source, distribution and load equipment, the harmonic mitigation equipment must never introduce a capacitive reactive power (kVAR) which is greater than 15% of its kW rating for sizes ≥ 100HP and 20% for sizes ≤ 75HP, at any given load level. All filters rated 75 HP and above must have fusing protection for the capacitive element of the circuit. An internal fault disconnect device within the capacitors are required for all size filter designs, but cannot be used in lieu of fusing at 75 HP and above. Capacitor contactor isolation schemes to isolate capacitance during low load or no load conditions is not acceptable.
• The harmonic mitigation equipment shall not resonate with system impedances, drive load equipment or attract harmonic currents from other harmonic sources, and verified via field testing.
• Harmonic compliance shall be verified with onsite field measurements of both the voltage and current harmonic distortion at the input terminals of the harmonic mitigating equipment with the equipment operating at full load. A background voltage distortion measurement must be taken with the drive not running and any downstream capacitance isolated, to ascertain the true state of the Utility source. In addition, a wave trace of the harmonic mitigation secondary, at nominal load, must be taken to assure that the installation of the harmonic mitigation equipment has not created any signs of system resonance created by its placement within the circuit. If harmonic compliance with IEEE 519 standards cannot be meet, and/or, a resonance condition is witnessed between the harmonic mitigation equipment and the drive, the harmonic mitigation equipment supplier shall remedy the condition at no charge to the end user.
• At Full Load and Nominal Load, True Power Factor (pf including harmonic reactive component) and Displacement Power Factor (dpf), Reactive Power with harmonic contribution (kVAR), Apparent Power (kVA), Real Power (kW), Current Draw Imbalance at full load, and Voltage Imbalance must be documented for system compatibility verification and to ascertain proper drive operation.

A word format document is available with my suggestions for your harmonic mitigation and drive systems specificational review. This suggested specification has been developed as a result of years of diagnostic testing and harmonic performance verifications within the field. Real world specifications and applicational testing requirements are the only way to ascertain a soundly engineered and properly deployed harmonic mitigation strategy for your clients and customers…

For more information, or to recieve a copy of the NSOEM, Inc. Guide harmonic specification, please contact me directly:

Michael A McGraw, NSOEM, Inc.

Phone: (713) 208-8534, E-Mail: [email protected]