A UK harmonic study usually exists to answer a Distribution Network Operator. Either you are applying to connect non-linear load or generation and have to show what you will inject, or the DNO has measured distortion at your boundary and asked you to account for it. Both need the same evidence: what your site emits, at the point of common coupling, across the real duty cycle.
BS 7671 · IET Wiring Regulations · HSR25 · IEEE 1584 · IEC 60909
The deliverable is an argument the DNO can check, not a distortion figure on its own.
Harmonic assessment is made at the boundary your connection agreement names, not at the terminals of the equipment causing the distortion. Where that boundary sits determines the impedance, and therefore the result.
Distortion depends on the network you are connected to. Fault level and impedance data come from your DNO for its licence area, and assuming them is the fastest way to produce a study the operator will not accept.
Power factor correction capacitors and cable capacitance can form a resonant circuit with network inductance, amplifying a harmonic order that was previously harmless. This is why adding correction equipment sometimes makes distortion worse.
Variable speed drives, UPS systems and any G99-connected generation all contribute. Their phase relationships matter, so summing individual equipment datasheets overstates or understates the combined result depending on the mix.
Confirm whether this supports a connection application or answers a DNO measurement, and identify the point of common coupling that applies.
Request fault level and harmonic impedance data from the DNO for your licence area in writing, since the network determines how your emission appears at the boundary.
Survey drives, UPS, rectifiers and any connected generation, and establish the duty cycle the site genuinely runs rather than its rated condition.
Calculate distortion by order at the assessment point, and test for resonance against existing power factor correction and cable capacitance.
A submission-ready report with inputs, sources and mitigation options carrying predicted post-mitigation figures.
Most rejected harmonic submissions are rejected on inputs rather than arithmetic: the network data was assumed, or the duty cycle used was not the one the site actually runs.
Usually at one of two moments. Either you are applying to connect significant non-linear load or generation and the Distribution Network Operator wants to see the predicted emission, or the DNO has measured distortion at your boundary and asked you to account for it. Adding large drive or UPS capacity to an existing site frequently triggers the first.
Because harmonic voltage distortion at your boundary is the product of your current emission and the network impedance at each order. The same equipment on two different connections produces different results. Assuming impedance rather than obtaining it is the most common reason a harmonic submission is returned for rework.
Yes, and it is a common surprise. Correction capacitors form a resonant circuit with network inductance, and if the resonant frequency falls near an order the site already produces, that order is amplified rather than absorbed. This is why capacitor banks should not be added to a site with significant drive load without a resonance check.
No. Individual harmonic currents from different equipment do not add arithmetically, because their phase relationships differ. Summing datasheet figures typically overstates the total on a mixed site and can understate it where sources happen to align. The assessment has to model the combination at the point of common coupling.