![]() |
| Distorted current waveform formed by adding 5th, 7th, 11th and 13th harmonics to the fundamental |
Harmonic Distortion in Industrial Power Supply: Engineering the Problem Out of the Plant
A plant with a healthy power factor, correctly sized cables and a modern switchboard can still overheat transformers, blow capacitor fuses and trip drives for no obvious reason. When the load current and the voltage no longer look like sine waves, the cause is usually harmonic distortion. Variable frequency drives, rectifiers, UPS units, LED drivers and induction furnaces all draw current in pulses rather than smoothly, and the supply system pays for it in losses, heat and misoperation.
This article covers how harmonics are generated, how to quantify them (THD versus TDD), what current standards require at the point of common coupling (PCC), three worked calculations (TDD compliance, capacitor-bank resonance, and transformer K-factor), a mitigation selection guide, and the safety practice around filters and capacitor banks. It is written for electrical engineers working on LV/MV industrial distribution, including the factory, EPZ and re-rolling mill installations typical of Bangladesh.
What Harmonic Distortion Actually Is
A harmonic is a component of a periodic waveform whose frequency is an integer multiple of the fundamental. On Bangladesh's 50 Hz system, the 5th harmonic is 250 Hz, the 7th is 350 Hz, and the 3rd is 150 Hz. A distorted current is the sum of the fundamental and a series of these components.
![]() |
| Distorted current waveform formed by adding 5th, 7th, 11th and 13th harmonics to the fundamental |
The waveform above is the classic "double-hump" input current of a six-pulse diode-bridge drive without a line reactor. The distortion is not caused by the supply; the load draws current only when the supply voltage exceeds the DC-link voltage, and the harmonic currents then flow through the source impedance, creating voltage distortion for every other consumer on the same bus.
Two rules explain most field behaviour:
- Characteristic harmonics of a p-pulse converter appear at h = kp ± 1. A six-pulse bridge produces 5th, 7th, 11th, 13th, 17th, 19th and so on; a 12-pulse bridge cancels the 5th and 7th and leaves 11th, 13th, 23rd, 25th.
- Voltage distortion is current distortion times system impedance. V_h = I_h × Z_h, with Z_h rising roughly in proportion to h for an inductive source. A weak source (a generator, a long feeder, a lightly rated transformer) turns the same harmonic current into much more voltage distortion.
Fig. 2 shows why pulse number matters: a 12-pulse front end removes most 5th and 7th, and an 18-pulse front end removes most 5th through 13th, leaving the higher orders.
Where Harmonics Come From in an Industrial Plant
Table 1: Common Non-Linear Loads and Their Dominant Harmonics
| Load | Dominant Harmonic Orders | Typical Plant Context |
|---|---|---|
| Six-pulse VFD / DC drive | 5, 7, 11, 13 | Pumps, fans, compressors, textile and RMG lines |
| Single-phase SMPS, LED drivers, UPS (small) | 3, 5, 7, 9 (triplen-rich) | Office floors, lighting circuits, server rooms |
| Large UPS/battery chargers (six-pulse) | 5, 7, 11, 13 | Data rooms, telecom, hospitals |
| Induction furnace (medium-frequency) | 5, 7, 11, 13 plus interharmonics | Steel re-rolling and melting shops |
| Arc and ladle furnace | Broadband, 2–7 plus interharmonics, flicker | Steel and ship-recycling operations |
| Welding sets | 3, 5, 7 plus unbalance | Fabrication yards, shipbuilding |
| Saturated transformers (overvoltage, light load) | 2, 3, 5 | Lightly loaded feeders at night |
In Bangladesh specifically, the rapid growth of VFD-driven HVAC chillers, LED lighting retrofits and rooftop-solar inverters has changed the harmonic profile of many factories whose switchgear was specified for linear loads. Captive generator sets also matter: their sub-transient reactance is typically far higher than a utility transformer's, so a plant that passes harmonic limits on grid supply can show severe voltage distortion when transferred to the generator.
THD vs TDD: The Metric That Actually Sets Compliance
Total Harmonic Distortion (current):
THD_I = sqrt( I2² + I3² + ... + In² ) / I1 × 100 %
Total Demand Distortion:
TDD = sqrt( I2² + I3² + ... + In² ) / I_L × 100 %
where I_L is the maximum demand load current at the PCC (average of the maximum demand over the previous twelve months in the standard's definition, with provisions for new facilities without a year of data).
The distinction matters. A lightly loaded drive can show 80% THD_I while contributing almost nothing to system distortion, because the reference is its own tiny fundamental. TDD normalises to the facility's maximum demand, so it is the enforceable number. At full demand, the two are equal.
Standards Framework: What Applies and Which Edition
Table 2: Harmonic-Related Standards and Their Role (editions verified at time of writing)
| Standard | Role | Current Edition Note |
|---|---|---|
| IEEE Std 519-2022 | Harmonic control limits at the PCC for current (TDD) and voltage (THD_V) | 2022 edition supersedes 2014; voltage and TDD limit values are unchanged, with revised treatment of even harmonics, DER-containing sites, and measurement approach |
| IEC 61000-3-2 | Harmonic current limits for equipment up to 16 A per phase | Consolidated Ed. 5.2 (2018 + A1:2020 + A2:2024) |
| IEC 61000-3-12 | Limits for equipment above 16 A and up to 75 A per phase on public LV | 2011 edition; confirm whether a newer edition exists before citing |
| IEC TR 61000-3-6 | Assessment of emission limits for distorting installations on MV, HV and EHV systems | 2008 (Ed. 2); confirm current status |
| IEC 61000-2-4 | Compatibility levels in industrial plants | Ed. 3.0, 2024: new classes 2a, 2b, 2L, revised class 3 levels, and a 2–150 kHz range |
| IEC 61000-4-30 | Power quality measurement methods (Class A / S) | Ed. 4.0, published October 2025, replacing the 2015 third edition |
| IEC 61000-4-7 | Harmonic and interharmonic measurement guide | 2002 + A1:2008 is the widely referenced version; verify whether it has been revised |
| IEC 60364-5-52 | Cable selection, including harmonic reduction factors for four- and five-core cables | Verify the edition in force in your project specification |
Verification flags: the Bangladesh Grid Code and utility connection agreements (PGCB, BPDB, DESCO, DPDC, REB) may impose their own supply-quality limits. I could not confirm specific harmonic limit values from the Grid Code during research, so check the connection agreement and tender specification for the governing figures rather than assuming IEEE 519 applies by default. BNBC 2020 clause references have not been cited here for the same reason.
IEEE 519-2022 Voltage Limits at the PCC
Table 3: Voltage Distortion Limits (percent of nominal fundamental)
| Bus Voltage at PCC | Individual Harmonic | THD_V |
|---|---|---|
| ≤ 1.0 kV | 5.0% | 8.0% |
| 1 kV < V ≤ 69 kV | 3.0% | 5.0% |
| 69 kV < V ≤ 161 kV | 1.5% | 2.5% |
| > 161 kV | 1.0% | 1.5% |
IEEE 519-2022 Current Limits (120 V – 69 kV)
Current limits depend on the short-circuit ratio Isc/I_L. A stronger source (higher ratio) is allowed to carry more distorted load current, because the same current produces less voltage distortion.
Table 4: TDD Limits and Individual Odd-Harmonic Limits, 2 ≤ h < 11
| Isc / I_L | Individual limit, h < 11 (%) | TDD limit (%) |
|---|---|---|
| < 20 | 4.0 | 5 |
| 20 to < 50 | 7.0 | 8 |
| 50 to < 100 | 10.0 | 12 |
| 100 to < 1000 | 12.0 | 15 |
| ≥ 1000 | 15.0 | 20 |
The TDD column is confirmed against published summaries of the 2022 table. The individual-harmonic column shows only the h < 11 band; the higher bands (11 ≤ h < 17, 17 ≤ h < 23, 23 ≤ h < 35, 35 ≤ h ≤ 50) are lower and should be read from the standard itself. Even harmonics are limited to 25% of the odd-harmonic limit in the 2014 edition; the 2022 edition changed the even-harmonic approach, so check the clause before using it in a compliance report.
Two points that trip up designers:
- The limits apply at the PCC for the installation as a whole, not to each drive. A 90 kW drive is not "non-compliant" in isolation.
- Compliance is assessed statistically over a measurement period using the percentile approach in the standard, not against a single instantaneous reading.
Effects of Harmonics on Plant Equipment
Table 5: Harmonic Effects and Typical Symptoms
| Equipment | Mechanism | Field Symptom |
|---|---|---|
| Transformers | Eddy-current loss rises roughly with h²; extra winding and structural heating | Oil/winding temperature alarms below nameplate load, shortened insulation life |
| Cables and busbars | Skin and proximity effect raise AC resistance; added neutral current | Hot terminations, discoloured insulation, neutral overheating |
| Neutral conductor | Triplen currents (3rd, 9th, 15th) add arithmetically in the neutral | Neutral current exceeds phase current in 4-wire systems with heavy SMPS/LED load |
| Power factor capacitors | Capacitive reactance falls with frequency, absorbing harmonic current; resonance with source inductance | Blown fuses, bulging cans, repeated bank failures |
| Induction motors | Negative-sequence 5th and positive-sequence 7th cause extra heating and torque ripple | Hot motors, vibration, shortened bearing life |
| Protection and metering | Peak-sensing or non-true-RMS devices misread distorted current | Nuisance trips, measurement error |
| RCD / ELCB | High-frequency leakage through filter capacitors | Unexplained earth-leakage trips |
| Generators | High source impedance converts harmonic current into large voltage distortion; rotor heating | Voltage waveform distortion on genset, AVR instability |
Several of these overlap with problems covered in the WAZIPOINT circuit breaker article. A breaker that trips under a load that clamp-on measurements say is "within rating" may be responding to a high peak or a true-RMS current that a non-true-RMS meter understated.
Worked Example 1: TDD Compliance at the PCC
A factory takes supply from a 1000 kVA, 11/0.415 kV transformer with 5.75% impedance. The PCC is the 415 V transformer secondary. A 200 A (fundamental) six-pulse drive load with no reactor is added.
Step 1: Short-circuit current at the PCC (infinite-bus source assumed):
Isc = S / (√3 × V × Z_pu) = 1000 / (1.732 × 0.415 × 0.0575) ≈ 24.2 kA
Step 2: Maximum demand current. Taking I_L = 800 A for the whole facility:
Isc / I_L = 24,200 / 800 ≈ 30.3 → row "20 to < 50" → TDD limit 8%, individual h < 11 limit 7%
Step 3: Drive harmonic currents at 200 A fundamental, using the typical six-pulse spectrum (35%, 25%, 9%, 6% for the 5th, 7th, 11th, 13th):
I5 = 70 A, I7 = 50 A, I11 = 18 A, I13 = 12 A
I_h(rms) = sqrt(70² + 50² + 18² + 12²) = sqrt(7,868) ≈ 88.7 A
Step 4: TDD and individual check:
TDD = 88.7 / 800 × 100 ≈ 11.1 % > 8 % → FAIL
5th individual = 70 / 800 = 8.75 % > 7 % → FAIL
7th individual = 50 / 800 = 6.25 % < 7 % → pass
Step 5: Options. Using the indicative values in Fig. 4, a 5% line reactor (THD_I about 28–38%) reduces the drive's harmonic current to roughly 56–76 A, giving TDD of about 7–9.5%: marginal. A 12-pulse arrangement or an active filter (THD_I roughly 3–15%) brings TDD to about 1–4% and passes with margin.
This is a simplified screening calculation. It ignores the existing background distortion, other harmonic sources in the facility, harmonic phase-angle summation, and the source impedance's frequency dependence. It does not replace a measurement-based assessment or a harmonic study.
Worked Example 2: Capacitor Bank Resonance
Power factor correction capacitors are the most common way a harmonic problem becomes a failure. The capacitor and the source inductance form a parallel resonant circuit; if its resonant frequency lands near a harmonic the plant produces, current and voltage amplify.
Parallel resonance order:
h_r = sqrt( S_sc / Q_c )
Using the transformer above (S_sc = 1000 / 0.0575 ≈ 17.4 MVA) with a 300 kvar APFC bank:
h_r = sqrt( 17,391 / 300 ) = sqrt(58.0) ≈ 7.6
A resonance near h = 7.6 sits directly between the strong 7th and the 8th, so the 7th-harmonic current (350 Hz) is amplified and the bank fails prematurely. This is a textbook cause of repeat capacitor failures in factories that added VFDs to an existing APFC panel.
Detuned reactor solution. Series reactors of p % of the capacitor reactance move the series-tuned frequency below the lowest significant harmonic (5th):
f_tuned = f_1 / sqrt(p) (p as a fraction)
p = 7 % → 50 / 0.2646 ≈ 189 Hz (h ≈ 3.8)
p = 5.67 % → 50 / 0.2381 ≈ 210 Hz (h ≈ 4.2)
Both place the bank's resonant point below 250 Hz, so the combination looks inductive at 5th and above and cannot resonate with the source. Note that detuned banks have a rated voltage higher than system voltage (the series reactor raises the capacitor terminal voltage), so the capacitor voltage rating must be selected for the reactor percentage used.
Worked Example 3: Transformer K-Factor and Neutral Sizing
K-factor quantifies the heating a transformer experiences from a given harmonic spectrum:
K = Σ ( I_h / I_rms )² × h²
For the six-pulse spectrum used above (I1 = 100, I5 = 35, I7 = 25, I11 = 9, I13 = 6, in percent):
THD_I = sqrt(35² + 25² + 9² + 6²) / 100 ≈ 44.4 %
I_rms = sqrt(100² + 35² + 25² + 9² + 6²) ≈ 109.4
K ≈ 7.3
A K-factor of about 7.3 points to a K-9 rated transformer (the next standard rating up) or a standard-design transformer derated per IEEE C57.110 methodology. Confirm the current edition of C57.110 and the manufacturer's loss data before specifying; the 2018 edition is the version I would expect to be current, but I did not verify it here.
Neutral conductor sizing for triplen harmonics. In a balanced three-phase four-wire system, third-harmonic currents of the three phases are in phase and add in the neutral:
I_N = 3 × I_3
For a 100 A phase current carrying 40% third harmonic, I_N = 3 × 40 = 120 A, which exceeds the phase current. IEC 60364-5-52 reduction factors for four- and five-core cables, based on third-harmonic content of phase current:
Table 6: Cable Reduction Factors for Harmonic Currents (four- and five-core cables)
| Third Harmonic in Phase Current | Size Based On | Factor |
|---|---|---|
| 0–15% | Phase current | 1.0 |
| 15–33% | Phase current | 0.86 |
| 33–45% | Neutral current | 0.86 |
| > 45% | Neutral current | 1.0 |
At 40% third harmonic, size the cable for 120 A / 0.86 ≈ 140 A. Never assume the neutral can be halved or omitted in systems with heavy LED, SMPS or UPS load.
Mitigation Technology Selection
Table 7: Mitigation Options Compared
| Technique | Indicative THD_I | Strengths | Limitations |
|---|---|---|---|
| AC line reactor (3–5%) or DC choke | 28–45% | Cheap, simple, also protects the drive front end | Only partial reduction; may not meet TDD at weak PCC |
| Passive tuned filter (5th/7th) | 8–12% | Good at rated load, also supplies reactive power | Needs detailed study; can resonate with the source; poor at light load |
| 12-pulse rectifier | 10–15% | Robust, no active electronics | Needs phase-shifting transformer, sensitive to unbalance |
| 18-pulse rectifier | 4–8% | Very low THD_I | Larger, costlier, harder to retrofit |
| Active front end (AFE) | 3–5% | Regenerative, near-unity PF | Higher cost, switching-frequency EMI |
| Active harmonic filter (AHF) | 3–5% | Retrofits at the PCC, adapts to load changes | Current rating limits compensation, cost |
| Detuned APFC reactor (7% / 5.67%) | n/a (prevents resonance) | Makes PF correction safe in harmonic environments | Does not reduce load harmonics |
The THD_I figures are indicative ranges at full load drawn from general industry practice; request manufacturer test data for the specific converter and DC-link inductance.
A practical selection sequence:
- Measure first (see below). Without baseline data, mitigation sizing is guesswork.
- If the problem is a capacitor bank, fix the resonance before adding filters.
- If a handful of large drives dominate, treat them at source (reactor, 12/18-pulse, or AFE).
- If the harmonic load is diffuse (many small SMPS and LED loads), consider an AHF at the main board or PCC.
- Re-measure after commissioning and document the result against the contractual limit.
Measurement and Diagnostic Practice
A harmonic assessment is only defensible if the instrument and method are. Use a Class A analyser to IEC 61000-4-30 (Ed. 4.0, 2025) where a compliance or contractual dispute is possible; Class S is adequate for surveys. The IEEE 519-2022 assessment is statistical over a measurement period, using short-time and very-short-time values; read the percentile multipliers directly from the standard's measurement clause before writing a report.
Field survey checklist:
- Record V and I harmonics to at least the 50th order, plus THD_V, THD_I and TDD, at the PCC and at the main suspect loads.
- Log neutral current, not only phase currents, on 4-wire distribution boards.
- Log across at least one full production cycle (a week for continuous plants) including the generator transfer condition.
- Note capacitor bank status (steps on/off) at each recorded interval; resonance often depends on bank steps.
- Record transformer winding temperature and ambient/enclosure temperature alongside the electrical data. In Dhaka-area switchrooms, enclosure ambients of 40–50 °C compound harmonic heating.
- Use true-RMS clamp meters and flexible CT probes rated for the frequency range under test.
Risk and Safety Instructions
Harmonic work routinely involves capacitor banks, filter panels and live measurement on busbars. Treat each as a distinct hazard.
- Capacitor stored energy. Isolate and wait for the discharge time marked on the capacitor or required by the panel specification, then verify de-energisation with a voltage tester before touching terminals. Capacitors with failed discharge resistors can retain a dangerous charge.
- Do not energise or switch capacitor banks in a plant with known harmonic problems until resonance has been calculated. A bank switched into a resonant condition can fail violently.
- Arc-flash risk assessment before live measurement. Apply the incident-energy method used by the facility (IEEE 1584 methodology or the utility's equivalent procedure) and select arc-rated PPE to IEC 61482 series or the applicable local standard. Do not default to generic PPE.
- LOTO for filter and capacitor panel work. Filters can be energised from the load side through back-feed or from parallel banks.
- Never remove, undersize or fuse the neutral to "stop" overheating. A broken neutral in a harmonic-rich 4-wire system causes severe overvoltage on single-phase loads. Increase conductor size or treat the source of the triplens.
- Do not upsize breakers to eliminate nuisance trips caused by harmonic heating. Resize or treat the cause; an oversized breaker that no longer protects the cable is a fire risk.
- Verify filter rating against measured harmonic spectrum and fault level before purchase. A passive filter tuned incorrectly can amplify a harmonic instead of absorbing it.
- Document every study, measurement and change. A harmonic compliance file is the evidence in a dispute with the utility.
Quick-Reference Decision Matrix
| Symptom | Most Likely Cause | First Check | Typical Fix |
|---|---|---|---|
| Capacitor bank fuses or cans failing repeatedly | Parallel resonance with 5th/7th | Calculate h_r = √(S_sc / Q_c), check bank step status | Detuned reactors (7% / 5.67%) or passive filter |
| Transformer hot below nameplate load | Harmonic eddy-current loss | Spectrum, K-factor, winding temperature | Derate, replace with K-rated unit, treat sources |
| Neutral conductor overheating | Triplen (3rd, 9th) currents | Neutral current vs phase current | Resize neutral (IEC 60364-5-52 factors), treat SMPS/LED loads |
| Drives tripping on DC overvoltage / undervoltage | Voltage distortion or sags at DC link | THD_V at drive terminals | Line reactor, upstream filter, review source impedance |
| TDD exceeds contractual limit at PCC | Drive harmonics at weak PCC | TDD calculation vs Table 4 | AHF at PCC, 12/18-pulse, AFE |
| Voltage distortion on generator supply | High source impedance | THD_V on genset vs grid | Oversize genset, add filter, limit non-linear load fraction |
| Unexplained RCD trips | High-frequency leakage from filters | Leakage current spectrum | Use RCD type suited to the load, review filter earthing |
Conclusion
Harmonic distortion is a design and procurement problem before it is a maintenance problem. The four things that decide the outcome are the source impedance at the PCC, the type and quantity of non-linear load, the presence of power-factor correction, and whether anyone measured before specifying equipment. A short-circuit calculation, a TDD estimate and a resonance check cost a few hours; their absence costs transformers, capacitor banks and production time.
For Bangladesh plants adding VFDs, LED retrofits, rooftop-solar inverters or generator backup to older distribution, the practical routine is to establish a baseline survey, check Isc/I_L at the PCC, detune any capacitor bank, treat dominant sources at the drive, and verify against the limit named in the connection agreement. Because I could not confirm Grid Code-specific harmonic limits for Bangladesh, confirm the governing figure with the utility before finalising any compliance design.
Editorial verification flags before publishing: (1) the individual-harmonic column of Table 4, which should be checked against the purchased IEEE 519-2022 Table 2; (2) edition status of IEC 61000-3-12, IEC TR 61000-3-6, IEC 61000-4-7 and IEC 60364-5-52; (3) the IEEE C57.110 edition; (4) Bangladesh Grid Code harmonic limits; (5) the indicative THD_I and spectrum figures in Figs. 2 and 4, which are textbook-order ranges rather than measured data; (6) the worked examples are screening calculations, not substitutes for a harmonic study.
Have you traced a capacitor bank failure, an overheated transformer or a neutral-current problem to harmonics in a Bangladeshi plant? Share the case in the comments. WAZIPOINT is building out a technical reference series on power quality, switchgear coordination and protection for South Asian field conditions.




No comments:
Post a Comment
WAZIPOINT:
Thank you very much to visit and valuable comments on this blog post. Keep in touch for next and new article. Share your friends and well-wisher, share your idea to worldwide.