The practical question is never whether perfect power quality can be achieved, but what level is technically reasonable and economically justified. The operator manages quality cost-effectively; the regulator ensures customers receive an appropriate level without unnecessary cost. Because systems develop under different technical, climatic, commercial and regulatory conditions, no single rigid international objective suits every network — the right approach is a shared framework, locally adapted:
\[ \text{common framework} + \text{local adaptation} \]
Key idea
- Objectives target a reasonable, economic level — not perfection — under normal operating conditions.
- Steady-state objectives use weekly 95% indices; dips and interruptions get no universal limit.
- Typical unbalance: 2% at MV, 1.5% at HV, 1% at EHV; flicker needs the transfer coefficient.
- Planning levels control emissions internally; voltage characteristics report supply quality externally.
Section 1
What power-quality objectives are
Unlike measurement indices, objectives are not purely technical. They balance the cost of improving quality against the value customers place on it, and they must suit the network where they apply. The framework should be common, but the numbers adapt to local conditions.
The question is not “can perfect power quality be achieved?” but “what level is technically reasonable and economically justified?”
Section 2
Application and normal operating conditions
Objectives normally apply under normal operating conditions unless a contract states otherwise — the usual range of generation and load variation, reactive-compensation states, planned outages, maintenance, and the contingencies the system is designed for. They are not intended to apply during exceptional events:
Table 1 — Conditions normally excluded from quality objectives.
| Excluded Condition | Reason |
|---|
| Faults beyond normal security standards | Not part of normal operation |
| Multiple abnormal contingencies | Outside the normal planning basis |
| Force majeure / exceptional weather | Outside reasonable operator control |
| Acts by public authorities | External cause |
| Industrial action | External operational constraint |
| User non-compliance with emission limits | Disturbance caused by another party |
| Temporary supply arrangements | Used to avoid interruption during works |
A system should be planned and operated to meet objectives in normal conditions — but not every exceptional event should be treated as non-compliance.
Section 3
Monitoring limitations and responsibility
It is neither realistic nor economic to monitor every supply point permanently; monitoring covers selected sites, problem locations or representative categories. And crucially, a monitor shows what is happening at a point — not who is responsible. Harmonic distortion at a PCC, for instance, may come from customer emissions, background distortion, neighbouring users, network resonance or upstream sources:
\[ \text{measured disturbance} \;\neq\; \text{automatic responsibility of one party} \]
Determining responsibility usually needs coordinated voltage and current monitoring at several locations, plus network studies — expensive, and normally practical only for limited-duration investigations.
Section 4
Minimum monitoring periods
Steady-state disturbances need about a week to capture weekday/weekend variation; event-based disturbances need far longer, because they are driven by faults and weather:
\[ T_{\text{monitoring}} \geq 1\ \text{week (steady-state)} \qquad\qquad T_{\text{monitoring}} \geq 1\ \text{year (events)} \]
Table 2 — Typical minimum monitoring periods.
| Disturbance | Typical Minimum Period |
|---|
| Harmonics | One week |
| Flicker | One week |
| Unbalance | One week |
| Voltage dips | One year or more |
| Long interruptions | One year or more |
Section 5
Harmonic objectives
Retain the established planning-level philosophy for MV, HV and EHV. Planning levels should not be raised just because some measured values already exceed them — doing so would signal that rising distortion is acceptable. The internal check is:
\[ U_{h,\text{index}} \;\le\; U_{h,\text{planning level}} \]
where the index may be a 3-second or 10-minute statistical value. For external reporting, the voltage-characteristic site index is the weekly 95% value of the 10-minute harmonic voltage and THD:
\[ U_{h,sh95,weekly} \;\le\; U_{h,\text{voltage characteristic}} \qquad \mathrm{THD}_{V,sh95,weekly} \;\le\; \mathrm{THD}_{V,\text{voltage characteristic}} \]
MV characteristic values usually align with existing supply-voltage practice; HV and EHV values should reflect measured performance and local planning. The lower HV range may need interpolation between the MV and HV/EHV objectives, since HV spans a wide range of voltages.
Section 6
Flicker objectives and transfer
Flicker objectives need particular care because perception depends on lighting and human vision. Planning is based on \(P_{st}\), checked as \(P_{st95,weekly}\le P_{st,\text{planning level}}\). Transfer between levels is central: a fluctuating load at HV/EHV produces flicker that transfers to MV/LV, where lamps are connected. The transfer coefficient between an upstream point A and downstream point B is:
\[ T_{Pst,AB}=\frac{P_{st}(B)}{P_{st}(A)} \]
- \(P_{st}(A)\)
- flicker at the upstream / source location
- \(P_{st}(B)\)
- flicker at the downstream location
so \(T_{Pst,AB} < 1\) means attenuation and \(T_{Pst,AB} > 1\) means amplification, and the allowable HV/EHV emission is found from the downstream value:
\[ P_{st,\text{downstream}} = T_{Pst}\,P_{st,\text{upstream}} \]
Motor loads and network characteristics often attenuate transfer downstream, but not always — it varies with impedance, transformers, local sources and operating state — so the coefficient should come from synchronous measurements or network analysis, not from ratios of statistical percentiles alone:
Table 3 — What governs flicker transfer between levels.
| Factor | Influence |
|---|
| Network impedance | Governs the voltage-fluctuation transfer |
| Motor load | Can attenuate flicker |
| Transformer characteristics | Affect transfer between levels |
| Local flicker sources | Can mask the upstream contribution |
| Operating condition | Transfer changes with load state |
Section 7
Unbalance objectives
Unbalance objectives use the negative-sequence factor, with the assessment \(U_{\text{neg},sh95,weekly}\le U_{\text{neg},\text{objective}}\):
\[ U_{\text{neg}}=\frac{U_2}{U_1}\times 100 \]
- \(U_2,U_1\)
- negative- and positive-sequence voltage
Table 4 — Typical recommended unbalance objectives.
| Voltage Level | Planning Level | Voltage Characteristic |
|---|
| MV | 2% | 2% |
| HV | 1.5% | 2% |
| EHV | 1% | 1.5% |
where there is heavy single-phase loading, MV unbalance up to \(3\%\) may occur. Unbalance matters most for its thermal effect on machines — the negative-sequence voltage drives a negative-sequence current that causes heating and derating:
\[ U_2\;\rightarrow\;I_2\;\rightarrow\;\text{machine heating} \]
Section 8
Voltage dip objectives
Universal dip objectives are not practical, because the number and severity of dips vary enormously with fault rate, clearing time, topology, earthing, construction, weather and system strength. The recommended approach is to calculate and report the basic characteristics rather than impose one limit:
\[ U_{\text{retained}} = \min\!\left(U_{\text{rms}(1/2)}\right) \qquad\qquad t_{\text{duration}} = t_{\text{end}} - t_{\text{start}} \]
- \(U_{\text{retained}}\)
- lowest half-cycle RMS voltage during the dip
- \(U_{\text{rms}(1/2)}\)
- half-cycle refreshed RMS voltage
- \(t_{\text{start}},t_{\text{end}}\)
- times the RMS crosses the start and recovery thresholds
Table 5 — Voltage dip reporting methods.
| Method | Use |
|---|
| Voltage-dip table | Events by retained voltage and duration |
| Contour chart | Performance on the magnitude-duration plane |
| SARFI indices | Events below retained-voltage thresholds |
| Curve-based count | Events below SEMI, ITIC or a customer curve |
| Custom process criterion | Customer-specific compatibility |
Reports should always declare the reference voltage, dip threshold, phase measurement, time aggregation and monitoring period. At HV/EHV, phase-to-phase measurement is often more representative of what downstream equipment experiences, especially in non-solidly-earthed systems.
Section 9
Long interruption objectives
Universal interruption objectives are also impractical, because performance depends so strongly on topology, structure (radial or meshed), generation location, environment, geography, voltage level and redundancy. Direct benchmarking between companies can mislead; the more reliable comparison is each system against its own history. MV distribution uses customer-based indices — \(\mathrm{SAIFI},\ \mathrm{MAIFI},\ \mathrm{SAIDI},\ \mathrm{CAIDI}\) — while HV/EHV reporting should follow one declared philosophy of three:
Table 6 — Transmission interruption reporting philosophies.
| Philosophy | Main Question |
|---|
| Connection point performance | Was the transmission connection point supplied? |
| End-customer performance | Did end customers lose supply? |
| System energy performance | How much energy was not supplied? |
The energy philosophy uses estimated energy not supplied, which best represents severity when a single transmission customer is very large — \(\mathrm{EENS}=\mathrm{PNS}\times T\), the power not supplied times the duration. A connection-point interruption is not the same as an end-customer interruption, so the chosen basis must be declared.
Section 10
Planning levels versus voltage characteristics
For steady-state disturbances, two index categories must be kept distinct — internal planning levels and external voltage characteristics:
\[ \text{planning indices}\;\rightarrow\;\text{internal control} \qquad\qquad \text{voltage-characteristic indices}\;\rightarrow\;\text{external reporting} \]
Planning levels are the operator’s tool to manage emissions and protect future performance; voltage characteristics describe the supply quality customers can expect under normal conditions. And each is reported at two levels — a site index for a specific point, and a system index for a network, voltage level or region.
Section 11
Conclusions for each phenomenon
Drawing the recommendations together:
- Harmonics — keep planning broadly at existing levels; report on \(U_{h,sh95,weekly}\) and \(\mathrm{THD}_{V,sh95,weekly}\); prefer 99% over maximum values, which transients inflate.
- Flicker — 10-minute \(P_{st}\) for planning; report on \(P_{lt95,weekly}\); at HV/EHV apply the transfer coefficient, since lamps connect at lower levels: \(P_{st,\text{LV}} = T_{Pst}\,P_{st,\text{HV/EHV}}\).
- Unbalance — treat like harmonics (a thermal effect); report on \(U_{\text{neg},sh95,weekly}\); recognise voltage level, network type and single-phase loads.
- Voltage dips — report \(U_{\text{retained}}\) and \(t_{\text{duration}}\) consistently; no universal limit yet; use the data for compatibility and network improvement.
- Interruptions — \(\mathrm{SAIFI},\ \mathrm{MAIFI},\ \mathrm{SAIDI},\ \mathrm{CAIDI}\) at distribution; a declared philosophy at transmission; benchmark against own history.
Section 12
Final engineering message
Power-quality objectives must be technically meaningful, economically reasonable and suited to their network. A robust framework defines the disturbance type, measurement method, index, statistical basis, site or system level, the normal operating conditions assumed, the objective, and the treatment of responsibility and exclusions. And the single rule that runs through all eight parts of this series:
Do not compare power-quality results unless the index, method and operating context are the same.
Key message
Harmonics, flicker and unbalance are assessed with weekly statistical indices; voltage dips and long interruptions need longer observation and event-based reporting. Planning levels are internal tools for controlling future disturbance, while voltage characteristics are external values for reporting supply quality — and planning levels must always sit below the characteristics they protect. Monitoring equipment should compute daily and weekly percentiles from 3-second and 10-minute values, flag events appropriately, and report indices without hoarding raw data. Only with consistent indices, clear objectives and transparent reporting definitions can power quality be assessed fairly and improved over time — the conclusion of the entire Power Quality Indices and Objectives series.
