Triplen harmonics - A devil in your network software tool?
This blog post addresses potential “issues” with triplen harmonics when performing a network harmonic study.
Positive, negative and zero sequence
The theory of symmetrical components tells that a periodic signal can be decomposed into harmonics of positive, negative and zero sequence.
Positive sequence: Phase B lags behind A, C lags behind B
- Phase A: 0°
- Phase B: 120°
- Phase C: 240°
Negative sequence: opposite to positive sequence
- Phase A: 0°
- Phase B: 240°
- Phase C: 120°
Zero sequence: no phase shift between the three phases
- Phase A: 0°
- Phase B: 0°
- Phase C: 0°
What are triplen harmonics?
Triplen harmonics (odd multiple of 3) represent so called zero sequence. In a 3-phase system the triplen harmonic has the same phase angle in each of the 3 phases (no relative phase displacement). These harmonics are not characteristic harmonics and an ideal VFD does not produce them. Neither does the theory derive them. In a real world, such harmonics can occur. The reasons can be various – starting from an inherent voltage ripple in the dc link up to control effects in case of AFE. Some VFD topologies have little bit more triplen harmonics than others.
Luckily, the magnitudes of those harmonics are small (typically fraction of %) and do not present any issue for the grid. They shall fit the limits of the commonly used standards for harmonic control such as IEEE 519 or IEC 61000-2-4 and IEC 61000-3-6 or marginally exceed them (since the limits for triplen harmonics are often quite strict).
Triplen harmonics in a network harmonic study
A problem can occur when performing a network harmonic study where the triplen harmonics are considered as well. The user may get shocked by seeing significant
(up to huge) voltage harmonic distortion. Before going into panic mode, let’s look at the results a bit closer…
In case of a diode rectifier, the VFD is considered as a current harmonic source towards the grid (note: this has nothing to do with the classification of voltage
source inverter and current source inverter). Thus, also the network model normally includes a current source representing the harmonics injected by the VFD.
The only way to have a high distortion of voltage is either high grid impedance (indicating a grid resonance at certain frequency) or very high current harmonic
component. The grid resonance is a “local thing” from frequency point of view.
When you see a high THD of grid voltage, check the individual harmonics.
- Are few adjacent harmonics dominant? Then it is probably a network resonance issue.
- Are multiple harmonics across the frequency spectrum abnormally high? You shall be alerted – this is suspicious.
- Are only the triplen harmonics high while the rest of the spectrum is within acceptable range (see Figure 2)? Then something in the numerical calculation is likely not correct.
We highly recommend a simple sanity check.
Which one? Purchase our premium subscription and learn more about this and plenty of other VFD related topics.
