Retrofitting a 30 kW Pump with a 6-Pulse VFD: Overheating and Harmonics
A 30 kW, 415 V, 4-pole, 50 Hz IE3 pump motor (nameplate 55 A) is retrofitted with a 6-pulse PWM voltage-source drive fed from a 415 V board with a 1 MVA, 5% transformer. The board's short-circuit current is 18 kA and the drive's fundamental input current at full load is 48 A. The motor cable is 130 m of unshielded 4-core 35 mm². The drive has no output reactor, a 100 ns IGBT rise time and a 4 kHz carrier. After commissioning, at 30 Hz the pump delivers about 60% of rated power but the motor draws 58 A and its frame runs at 85 °C against 70 °C at 50 Hz. The line-side current THD measured at the drive terminals is 42%. The utility limits voltage THD at the point of common coupling to 5% and applies IEEE 519 current limits.
1. For constant V/f operation, compute the output voltage and modulation index required at 30 Hz, and state whether the drive is voltage-limited at that operating point.
Read the worked answer
Base V/f = 415/50 = 8.3 V/Hz, so at 30 Hz the drive must produce 8.3 × 30 = 249 V line-to-line.
Diode-rectifier d.c. bus ≈ 1.35 × 415 = 560 V at full load (about 587 V at no load).
Required modulation index: m = (V_LL × √2)/V_dc = (249 × 1.414)/560 = 352/560 = 0.63
That is comfortably inside the linear range of sine-triangle PWM (0.785 without third-harmonic injection, 0.907 with it). The drive is therefore not voltage-limited at 30 Hz, and the high current is not caused by a shortage of bus voltage or by over-modulation. That conclusion matters because it redirects the investigation to the motor parameters, the meter and the actual pump operating point rather than to the d.c. bus.
2. The 58 A reading at 30 Hz is offered as evidence that the drive is 'not suited to the motor'. Test that claim and give a structured diagnostic.
Read the worked answer
Check the arithmetic first. At 60% of rated power the fundamental current should be roughly 0.6 × 55 ≈ 33 A. Six-pulse PWM output typically adds 10–15% current THD at the motor terminals, so a true-RMS reading incl harmonics would be about 35–38 A. 58 A is around 70% higher than that, so something real is happening - but it may not be the drive's fault.
Diagnostic sequence:
- The meter. A general-purpose true-RMS meter has limited bandwidth and reads garbage on a PWM output. Use a drive-rated power analyser, or simply read the drive's own fundamental current display.
- The operating point. The affinity laws assume no static head. A pump with a large static head can absorb 70–80% of rated torque at 30 Hz. Check the measured head and flow against the pump curve before blaming the drive.
- Drive parameters. Wrong motor nameplate voltage, base frequency or a manual torque boost setting will over-flux the machine and inflate magnetising current at part speed.
- Installation. The 130 m unshielded cable with a 100 ns rise and no output reactor raises high-frequency losses and can cause circulating bearing currents; a proper shielded cable and a dv/dt reactor are the fixes.
- Cooling. A shaft-mounted fan delivers roughly speed-proportional airflow, so 30 Hz gives about 60% of rated cooling. If the pump genuinely needs 80% torque at that speed, fit an independent blower - do not fit a larger motor, which would simply run at lower load and worse power factor.
3. Evaluate the harmonic situation against IEEE 519 and recommend the mitigation you would actually specify, separating the current-side and voltage-side issues.
Read the worked answer
Current side (the binding constraint). Isc/IL = 18 000/48 = 375, so the IEEE 519-2014 Table 2 limits for the >100 to 1000 band apply at the PCC: 12.0% for h < 11, 5.5% for 11 ≤ h < 17, and TDD 15%. The measured 42% THD is far outside this. The 5th alone at roughly 30% of fundamental is 14.4 A against a limit of 12% × 48 = 5.8 A.
Voltage side. Transformer reactance referred to 415 V is X = 0.05 × 415²/1×10⁶ = 8.6 mΩ. The 5th-harmonic voltage is 14.4 × 5 × 0.0086 = 0.62 V, about 0.15% of 415 V, and the 7th adds roughly 0.10%. Total voltage THD at the board stays under 1%, well inside the 5% utility limit - the board is electrically stiff, so voltage distortion is not the problem.
Mitigation, in order of cost:
- A 4% AC line reactor or 3–4% DC-link choke - cheapest, protects the drive, but only brings THD to about 35–40%, which is still non-compliant on its own.
- A 12-pulse rectifier with a phase-shifting transformer - about 10–12% THD, usually compliant.
- An active front end - below 5% THD and regenerative, justified only if other loads on the board are also distorted or if regeneration is valuable.
Also confirm with a full board harmonic study: IEEE 519 limits apply to the whole facility at the PCC, so other linear load on the board dilutes the TDD - but do not rely on that dilution when this drive dominates the board's load.