DriveLab
Account

Power quality: the harmonics a drive draws

Why a six-pulse drive draws 6k ± 1 harmonics (eq. 6.3), what they do to the supply and to transformers, where IEEE 519 applies, and how to choose between a reactor, a 12-pulse front end, a filter and an active front end.

30 min

Show me

Each case study so far was a drive-selection problem, and in each the drive was the interesting part. In practice, a correctly chosen drive still fails in service if the system around it is inconsistent with it. The next three lessons cover the issues that most often turn a sound paper design into a troubled installation, and that a capstone project is expected to address rather than assume away. This lesson is course section 6.2.1.

Why a drive distorts its supply current

A conventional VFD draws current through a six-pulse diode rectifier feeding a DC-link capacitor. The capacitor charges only while the line voltage is above the link voltage, so the current comes in short pulses twice per cycle instead of as a sine wave. The result is a supply current rich in harmonics of order

h=6k±1,k=1,2,3,…(6.3)h = 6k \pm 1, \qquad k = 1, 2, 3, \dots \tag{6.3}

that is the 5th, 7th, 11th, 13th and so on. Triplen harmonics are absent by symmetry, the amplitudes fall roughly as 1/h1/h, and the 5th and 7th dominate. The total harmonic distortion of the current of an uncompensated six-pulse drive typically reaches 80 to 120 %.

Why it matters

Harmonic currents flow through the source impedance and distort the voltage at the point of common coupling (PCC), where they affect every other user. They also cause extra heating in transformers and cables that a simple RMS calculation misses, because eddy-current and skin-effect losses rise with the square of frequency. A transformer feeding many drives must be derated, conventionally through a K-factor rating.

IEEE 519 sets its limits at the point of common coupling, not at the drive terminals, a distinction students routinely miss. The allowance depends on the supply's stiffness, the ratio of short-circuit current to maximum demand current Isc/ILI_{sc}/I_L: a stiff supply tolerates more harmonic current from a given load than a weak one, because the same current produces less voltage distortion. For Isc/IL<20I_{sc}/I_L < 20 the total demand distortion (TDD) limit is 5 %, rising through 8, 12 and 15 % to 20 % for very stiff supplies above 1000.

Mitigation, in grades

The right level depends on how much of the plant load is non-linear:

  • AC line reactors or DC-link chokes, typically 3 to 5 % impedance. Cheap, always worth fitting: they roughly halve the current THD, from about 100 % to the 40 to 50 % band, and protect the rectifier from supply transients.
  • Multi-pulse rectifiers, 12- or 18-pulse, using a phase-shifting transformer to cancel the lower orders. A 12-pulse front end eliminates the 5th and 7th, leaving the 11th and 13th as the lowest present.
  • Passive tuned filters, effective and economical for a fixed, well-characterised load, but they interact with the supply impedance and can resonate with power-factor-correction capacitors elsewhere in the plant.
  • Active front ends, an IGBT rectifier drawing near-sinusoidal current at unity (or commanded) power factor: current THD below 5 % and, as a by-product, full regeneration. The most expensive option, and the only one that solves harmonics and regeneration together, which is why it is the default on the reversing mills and hoists of the earlier chapters, where regeneration is needed anyway.

The choice cannot be made from the drive alone: on a stiff supply a line reactor is enough, while the same drive on a weak supply may need a 12-pulse front end to meet the same standard.

Try it: harmonics at the point of common coupling
15807601115139176195233253Harmonic order h (bar height: % of the fundamental)

Short-circuit current at the point of common coupling over the maximum demand current. A weak supply is below 20.

Linear loads (heaters, DOL motors) add fundamental current and dilute the distortion.

Over the IEEE 519 limit: this plant needs more mitigation on this supply.

Current THD of the drive
102 %
TDD at the PCC
61.1 %
IEEE 519 limit
8 %

h = 6k ± 1 ; THD = √(Σ I_h²) / I_1 ; TDD ≈ THD × (drive share of the load)

Predict first

The supply has I_sc / I_L = 30, and drives make up 60 % of the plant load. Which is the cheapest front end that meets IEEE 519?

FoundationStart here if this is new to you

Imagine a water main shared by a street. Most houses draw water smoothly. One house fills a big tank in short, violent gulps twice a second. The pressure in the whole street jumps with each gulp, and the neighbours notice. A six-pulse drive is that house; the reactor makes its gulps gentler, and the active front end makes it draw as smoothly as everyone else.

ExplorerGo deeper: derivations and open questions

Why no triplens? In a balanced three-phase bridge, what do the 3rd harmonics of the three line currents look like relative to each other, and where would they have to flow?

Transformer derating. Using the widget's bare six-pulse spectrum, compute ∑(Ih/I1)2h2\sum (I_h/I_1)^2 h^2 for the orders shown. Why does this sum, rather than the RMS current, set the extra heating in a transformer?