AI Skill Report Card

Conducting Harmonics & Power Quality Studies

A89·Sep 26, 2026·Source: Web
Markdown
--- name: conducting-harmonics-studies description: Performs full electrical supply harmonics and power quality studies, evaluates compliance against UK (ENA G5/5, BS EN 50160, BS EN 61000 series) and international standards (IEC 61000, IEEE 519), and recommends mitigation/filtering measures. Use when assessing harmonic emissions from non-linear loads (VSDs, UPS, EV chargers, rectifiers, LED drivers), investigating power quality complaints, sizing harmonic filters, or performing supply connection compliance assessments. --- # Conducting Electrical Harmonics & Power Quality Studies
14 / 15

Given a site with non-linear loads and a connection agreement, run the study in this order:

  1. Characterize loads → build harmonic current spectrum per load (from datasheets, IEC 61000-3-2/-12 emission limits, or measurement).
  2. Model the system → source impedance at PCC, transformer/cable impedance, existing PF correction capacitors (resonance risk).
  3. Calculate/measure THD & individual harmonics at the Point of Common Coupling (PCC) and Point of Evaluation (PoE).
  4. Compare against limits → ENA G5/5 (UK planning levels), BS EN 50160 (supply voltage characteristics), IEEE 519 (if no UK limit applies).
  5. If exceeded → recommend mitigation (passive filter, active filter, line reactors, phase shifting, isolation transformer) and re-verify compliance.

Output a compliance table: Harmonic order | Measured/Calculated % | Planning Level | Compliant Y/N.

Recommendation▾
Fix the truncated final pitfall ('Don't overlook interharmonics') to complete the sentence with guidance
15 / 15
Progress:
- [ ] Step 1: Gather load data and site single-line diagram
- [ ] Step 2: Determine applicable standards and limits
- [ ] Step 3: Build/collect harmonic spectra for each load
- [ ] Step 4: Model network impedance and identify resonance points
- [ ] Step 5: Calculate voltage/current harmonic distortion at PCC
- [ ] Step 6: Compare results against planning/compatibility levels
- [ ] Step 7: If non-compliant, design mitigation measure
- [ ] Step 8: Re-verify post-mitigation and document findings

Step 1: Gather Load Data

Collect for each non-linear load:

  • Rated power (kW/kVA), type (VSD, UPS, rectifier, EV charger, LED, arc furnace, etc.)
  • Manufacturer harmonic spectrum (current distortion by order, %THDi) — request if not supplied
  • Diversity/loading factor (harmonics scale with load level, not just rated current)
  • Connection point in the network (which busbar, which transformer)

If no manufacturer data exists, use typical spectra for the load class (6-pulse VSD, 12-pulse, switch-mode PSU) as a conservative estimate, flagged as an assumption.

Step 2: Determine Applicable Standards

Default hierarchy for UK work:

  1. ENA Engineering Recommendation G5/5 (or latest revision) — UK planning levels for harmonic voltage distortion at LV/HV/EHV, and the assessment methodology (Stage 1/2/3 assessment).
  2. BS EN 50160 — voltage characteristics of public distribution networks (compatibility levels).
  3. BS EN 61000-3-2 / -3-12 — emission limits for equipment ≤16A / >16A≤75A per phase.
  4. BS EN 61000-3-3 / -3-11 — voltage fluctuation/flicker limits, if flicker also in scope.
  5. DNO/DNO-specific requirements — check the relevant Distribution Network Operator's connection conditions; they may impose stricter G5/5-based limits.

Where UK standards don't cover the scenario (e.g., large industrial converters, HVDC, offshore, specific inter-harmonics, or MV/HV connection studies not fully detailed in G5/5):

  • IEEE 519 — recommended practice for harmonic control, current/voltage distortion limits by SCR (short circuit ratio).
  • IEC 61000-4-7 — measurement methods and interharmonics.
  • IEC 61000-3-6 / -3-7 — MV/HV emission limit assessment methodology (complements G5/5 Stage 2/3 approach).

State explicitly in the report which standard covers which part of the assessment, and flag any gap-filling with international standards.

Step 3: Build Harmonic Spectra

Express each load's harmonic current as % of fundamental, at each order (2nd–50th typically; consider higher/interharmonics for modern SMPS/EV chargers if relevant). Aggregate multiple loads using vector or arithmetic summation per G5/5 Appendix/IEC 61000-3-6 diversity guidance — don't just add arithmetically without applying diversity factors, this overstates the problem.

Step 4: Model Network Impedance

  • Source impedance at PCC: from DNO fault level data (Ohm's law: Z = V²/Sfault).
  • Include transformer impedance (%Z from nameplate), cable/line impedance for the relevant length.
  • Check for resonance: parallel resonance between supply inductance and any PF correction capacitance.
    • Resonant order: h_r = √(Ssc / Qc), where Ssc = fault level (kVA), Qc = capacitor bank rating (kVAr).
    • Flag if h_r lands near a dominant harmonic order (5th, 7th, 11th) — high risk of amplification.

Step 5: Calculate Distortion

  • Individual harmonic voltage: Vh = Ih × Zh (at that harmonic's frequency, Zh = h × Z1 approx, adjust for skin effect on cables at higher orders).
  • THDv = √(Σ Vh²) / V1 × 100%
  • THDi similarly for current.
  • If measured data available (power quality analyzer, per IEC 61000-4-7/-4-30 Class A), use recorded 10-minute aggregated values (95th percentile over the week per G5/5 methodology), not instantaneous peaks.

Step 6: Compare to Limits

Use G5/5 Stage 1 (simple check against planning levels by voltage level) first. If load contribution is significant relative to available headroom (typically >the load's allocated capacity per the "fair share" allocation method), escalate to Stage 2 (detailed assessment with actual network impedance) or Stage 3 (network operator involvement for large connections).

Typical LV planning levels (indicative — always confirm current G5/5 revision):

HarmonicOdd non-triplen (5th,7th...)Triplen (3rd,9th)Even
THDv5% (LV)——
Individual~4-6% depending on order~1.5-3%~1-2%

Step 7: Mitigation Design (if non-compliant)

Select based on dominant harmonic order, load type, and cost/space constraints:

MitigationBest forNotes
Line/AC reactors (3-5% impedance)6-pulse VSDs, general THDi reductionCheapest, reduces THDi ~30-40%
DC link chokesVSDsSimilar effect to AC reactor, smaller footprint
Passive harmonic filter (tuned)Dominant single harmonic (5th/7th)Must check resonance with rest of network; also provides PF correction
12-pulse/18-pulse rectifiersLarge VSD/UPS installationsCancels 5th/7th (12-pulse) or more (18-pulse) inherently
Active harmonic filter (AHF)Multiple harmonics, variable load profilesMost flexible, higher cost, real-time correction
Isolation/phase-shifting transformerMultiple 6-pulse loadsCancels harmonics between groups via phase shift
Hybrid filterLarge installations needing PF + harmonic controlCombines passive + active

Always re-run Step 5 calculation with the mitigation's expected attenuation applied, to confirm compliance margin — don't just assert the fix works.

Step 8: Document

Report structure:

  1. Executive summary (compliant/non-compliant, key risk)
  2. Site description and single-line diagram
  3. Load harmonic data and assumptions
  4. Standards applied and rationale
  5. Calculation methodology and results (table by harmonic order)
  6. Resonance analysis
  7. Mitigation recommendation with predicted post-mitigation performance
  8. Conclusions and any conditions (e.g., recommend re-measurement after commissioning)
Recommendation▾
Add a third example covering a non-compliant case requiring active filtering to show more mitigation diversity
17 / 20

Example 1: Input: 250kW 6-pulse VSD connecting to a 500kVA LV transformer supply, DNO fault level at PCC 8MVA, no existing PFC capacitors. Output:

  • Load THDi (typical 6-pulse, no reactor): ~35%, dominant 5th (~30%) and 7th (~10%)
  • Zsource at PCC calculated from 8MVA fault level; THDv estimate ~6.2% → exceeds 5% G5/5 LV planning level
  • Resonance check: no PFC capacitors present, no parallel resonance risk
  • Recommendation: add 3% AC line reactor → reduces THDi to ~18%, recalculated THDv ~3.8% → compliant. Note as cheaper alternative to passive filter given single dominant load.

Example 2: Input: Industrial site with 400kVAr PFC capacitor bank and multiple VSDs, fault level 6MVA, complaints of nuisance tripping. Output:

  • Resonant harmonic order h_r = √(6000/400) ≈ 3.87 → close to 5th harmonic risk zone but more critically near dominant low-order components
  • Flag high risk of resonance amplification; recommend converting capacitor bank to detuned (tuned to ~4.2x, i.e., below 5th) filter rather than plain PFC
  • Re-assessment after retrofit shows THDv reduced from 9.1% to 4.3% → compliant with G5/5
Recommendation▾
Include a brief note on typical measurement duration/equipment cost tradeoffs when full IEC 61000-4-30 Class A instrumentation isn't available
  • Always state which harmonic standard/limit applies at which point (PCC vs equipment terminals vs supply network) — they are not interchangeable.
  • Apply diversity factors when summing harmonics from multiple loads — arithmetic addition without diversity overstates worst case significantly.
  • Always check for resonance whenever capacitors (PFC banks) exist anywhere in the network, even if the harmonic-injecting load is elsewhere.
  • Prefer measured data (IEC 61000-4-30 Class A instrument) over datasheet estimates when the study is for a compliance dispute or connection application.
  • When citing standards, note the specific edition/revision considered, as limits and methodology get updated (e.g., G5/5 revisions).
  • Clearly separate "voltage distortion" (network-level compliance, G5/5/EN50160) from "current distortion" (equipment emission limits, EN 61000-3-2/-3-12) — they're assessed differently.
  • Don't assume manufacturer nameplate THDi applies at all load levels — request or estimate the spectrum at actual operating load, since harmonic % varies with loading.
  • Don't add a PFC capacitor to "fix" power factor without checking harmonic resonance risk first — this is the most common cause of nuisance tripping/capacitor failure.
  • Don't apply IEEE 519 limits directly to a UK LV connection without justification — use it only to fill genuine gaps not addressed by G5/5/BS EN standards, and say so explicitly.
  • Don't rely on instantaneous/peak readings for compliance — G5/5 methodology uses 95th percentile of 10-minute averages over a representative period (typically one week).
  • Don't overlook interharmonics
0
Grade AAI Skill Framework
Scorecard
Criteria Breakdown
Quick Start
14/15
Workflow
15/15
Examples
17/20
Completeness
19/20
Format
15/15
Conciseness
13/15