NEC 2026 Load Calculations: Article 120, the 2 VA Rule, and What Changed for Dwelling Services

A technical guide to the relocated and revised load-calculation article — the standard and optional methods worked end to end, the new branch-circuit count rule, EV charging, existing dwellings, multifamily buildings, and the errors circulating in early 2026 study material

Executive Summary

The 2026 edition of NFPA 70, the National Electrical Code, moved the load-calculation rules that every electrical designer, contractor, and plan reviewer has known for decades as Article 220 to a new home: Article 120, Branch-Circuit, Feeder, and Service Load Calculations. The section suffixes were preserved, so 220.82 is now 120.82 and Table 220.55 is now Table 120.55, but the move is more than housekeeping. It puts the calculation rules in Chapter 1 alongside the general requirements of Article 110, immediately ahead of the new Article 130 on power and energy management, and it arrived with substantive changes to the numbers.

Three of those changes affect nearly every dwelling calculation. First, the general lighting and general-use receptacle unit load for dwelling units fell from 3 VA to 2 VA per square foot in 120.41, a one-third reduction that reflects the disappearance of incandescent lighting and field data on what modern homes actually draw. Second, a new Section 120.13 requires the minimum number of general-purpose branch circuits to still be determined at 3 VA per square foot, so the smaller service load does not quietly translate into fewer circuits in the panel. Third, the optional method's 100 percent tier in 120.82(B) shrank from 10 kVA to 8 kVA, which partially offsets the lighting reduction for homes that use the optional path.



Around those headline items sit several smaller but consequential revisions: electric vehicle supply equipment is now calculated under 120.57 at the larger of nameplate or 7,200 VA, at 100 percent with no demand factor; power control systems are formally recognized in 120.7 so that a controlled setpoint, rather than connected nameplate, can enter the calculation; the clothes-dryer demand table in 120.54 was revised for small dryer counts; and the general rules in 120.5 clarify that the 125 percent continuous-load multiplier belongs to conductor and overcurrent-device sizing, not to the load calculation itself.

This guide walks through Article 120 the way a calculation actually flows: general rules, the general lighting group and its demand factor, appliances and cooking, heating and cooling, motors, and the assembly of a standard-method and an optional-method service calculation for the same house under both the 2023 and 2026 numbers. It then covers existing dwellings, multifamily buildings, neutral sizing, conductor selection, EV charging and load management, and closes with a correction section addressing errors we have found in early 2026 study material, a note on state adoption, and a 25-question FAQ.

1. Article 220 Becomes Article 120

NFPA's reorganization of the Code moved the load-calculation article from Chapter 2 (Wiring and Protection) to Chapter 1 (General). The reasoning is structural: load calculations are not a wiring method or a protection rule, they are a general requirement that every subsequent article consumes. Article 110 sets the general requirements for installations, Article 120 sets how loads are calculated, and the new Article 130 sets how power and energy management systems may control those loads. The three now read in sequence.



For practitioners the most important fact is that the section suffix survived the move. Anyone who can find 220.14(I) in the 2023 book can find 120.14(I) in the 2026 book. The table below maps the sections referenced most often in dwelling and light-commercial work.

2. The General Rules of Article 120

Part I of Article 120 sets the rules every downstream calculation depends on. Most are carried over from the 2023 text, but two additions matter.

2.1 Nominal voltages, rounding, and floor area (120.5)

Calculations are performed at nominal system voltages: 120, 120/240, 208Y/120, 240, 347, 480Y/277, 480, 600Y/347, and 600 volts. Where a calculation produces a fraction of an ampere, a fraction of 0.5 or larger is rounded up and a fraction below 0.5 may be dropped. Floor area is measured from the outside dimensions of the building, apartment, or other area involved; for dwelling units the area does not include open porches, garages, or unused or unfinished spaces that are not adaptable for future use. A basement that could become living space is counted; a crawl space is not.

The 2026 text adds a clarification practitioners have argued about for years: the 125 percent multiplier for continuous loads is not part of the load calculation. Continuous loads still drive conductor ampacity and overcurrent-device ratings through 210.19, 210.20, 215.2, 215.3, and 230.42, all of which remain in Chapter 2. But the calculated load that establishes the service size under Article 120 is the load itself. Keeping those two ideas apart prevents the double-counting that occurs when a designer applies 125 percent inside the calculation and then again when selecting conductors.

2.2 Noncoincident loads (120.6) and power control systems (120.7)

Where it is unlikely that two or more noncoincident loads will operate at the same time, only the largest load contributing to the total need be used. This is the general permission behind the heating-versus-cooling rule in 120.60 and the heat-pump provisions in the optional method.



New Section 120.7 formally recognizes power control systems, listed equipment that monitors current and actively limits it to a setpoint. Where such a system controls a load or group of loads, the setpoint may be used in the calculation in place of the connected nameplate, and the setpoint is treated as the calculated load for that group. This is the code hook that allows a 200 A service to accept two 48 A EV chargers, a heat pump, and an induction range without an upgrade, provided the control system is listed for the purpose and installed per Article 130 and, for EV charging, the energy-management provisions of Article 625.

3. The 2 VA Rule and the New Branch-Circuit Count

3.1 General lighting at 2 VA per square foot (120.41)

For dwelling units, the general lighting and general-use receptacle load for feeder and service calculations is now 2 VA per square foot of floor area. The 2023 value was 3 VA, a figure that had stood essentially unchanged since the era of incandescent lamps. Two things drove the change. Federal efficiency standards effectively removed general-service incandescent lamps from the market, moving residential lighting to LED sources that draw a fraction of the power. And measured data from a large sample of occupied U.S. homes, collected by the national laboratory community, showed median general lighting and receptacle density in the vicinity of 2.3 W per square foot. The Code-making panel adopted 2 VA as the new unit load for the service and feeder calculation.



In a dwelling the general-use receptacles are included in this unit load; they are not counted separately at 180 VA each. The 180 VA per receptacle rule in 120.14(I) applies to non-dwelling occupancies. The 2 VA figure flows into every dwelling method: the standard method in Part III, the optional one-family method in 120.82, the existing-dwelling method in 120.83, and the multifamily optional method in 120.84.

The reductions in the right-hand columns are connected-load reductions before the Table 120.42 demand factor. Because everything above the first 3,000 VA of the general lighting group is taken at 35 percent in the standard method, the effect on the final service current is smaller than the table suggests: roughly 0.35 × 2,400 VA ÷ 240 V, or about 3.5 A, for the 2,400 ft² house. Under the optional method, where the remainder above 8 kVA is taken at 40 percent, the effect is about 4 A for the same house.

3.2 Section 120.13: the circuit count stays at 3 VA per square foot



The panel recognized that if the same 2 VA figure were used to determine how many general-purpose branch circuits a dwelling needs, the circuit count would fall by a third and receptacle circuits would be more heavily loaded in practice. New Section 120.13 therefore requires the minimum number of 15 A or 20 A general-purpose branch circuits to be determined at 3 VA per square foot, independent of the 2 VA service calculation. The two figures answer two different questions for the same house.

4. Small-Appliance, Laundry, and the Lighting Demand Factor

Section 120.52 assigns 1,500 VA to each two-wire 20 A small-appliance branch circuit required by 210.11(C)(1), with a minimum of two, and 1,500 VA to the laundry branch circuit required by 210.11(C)(2). Where a dwelling has more than two small-appliance circuits, each is counted. Where a laundry receptacle is not required (for example, in a multifamily unit served by a common laundry that meets the conditions of 210.52(F)), the laundry load may be omitted.



These loads are added to the general lighting load and the sum is taken through Table 120.42. For dwelling units the demand factors are unchanged from the 2023 edition: the first 3,000 VA at 100 percent, from 3,001 to 120,000 VA at 35 percent, and the remainder above 120,000 VA at 25 percent. The demand factor recognizes that lighting, receptacle, kitchen countertop, and laundry loads do not peak simultaneously. It applies to this group only. Ranges, dryers, fixed appliances, heating, cooling, and EV charging each have their own rules and never pass through Table 120.42.

Two points deserve emphasis. The demand factor applies before the appliance, range, dryer, and HVAC loads are added, never to the whole-house total. And a demand factor never increases a load; where a reader sees a multiplier above 100 percent, it is the continuous-load rule for conductor sizing, which is a different concept living in a different article.

5. Fixed Appliances, Dryers, and Cooking Equipment

5.1 Fastened-in-place appliances (120.53)

Where four or more appliances fastened in place, other than ranges, clothes dryers, space-heating equipment, and air-conditioning equipment, are served by the same feeder or service in a one-family, two-family, or multifamily dwelling, a demand factor of 75 percent may be applied to their combined nameplate rating. Typical members of this group are the dishwasher, waste disposer, built-in microwave, water heater, well pump, sump pump, attic fan, and garage-door opener. With three or fewer, each is counted at 100 percent.

The group is defined by what is excluded. A range hood is a fastened-in-place appliance; a countertop microwave on a small-appliance circuit is not counted at all because it is already inside the 1,500 VA small-appliance allowance. Electric water heaters belong in this group even though they are often the largest single member.

5.2 Household clothes dryers (120.54)

Each household electric clothes dryer is counted at 5,000 VA or the nameplate rating, whichever is larger. Where a feeder or service supplies several dryers, as in a multifamily building or a laundry room, the demand factors of Table 120.54 apply. The 2026 edition revised the table for small dryer counts: the 2023 table held 100 percent through four dryers and stepped to 85 percent at five, whereas the 2026 table applies a reduced factor from the third dryer onward, reflecting the same diversity data that drove the lighting change. The table continues to decline with the number of dryers to a floor in the mid-20-percent range for very large counts. Because the small-count rows changed, designers should read the adopted 2026 table directly rather than relying on 2023 memory.



Where two or more single-phase dryers are supplied by a three-phase, four-wire feeder or service, the load is calculated on the basis of twice the maximum number connected between any two phases, exactly as for ranges.

5.3 Household cooking equipment (120.55 and Table 120.55)

Household electric ranges, wall-mounted ovens, counter-mounted cooking units, and other household cooking appliances individually rated over 1¾ kW are calculated using Table 120.55 and its notes. Appliances at 1¾ kW or below are treated as ordinary appliance loads at nameplate. The table has three columns: Column A and Column B give percentage demand factors by number of appliances for units rated under 3½ kW and from 3½ to 8¾ kW respectively, and Column C gives a maximum demand in kW for appliances rated over 8¾ kW up to 12 kW. For a single range not over 12 kW, Column C yields 8 kW.

5.4 Range and dryer neutrals (120.61(B))

Because the heating elements of ranges and dryers operate line-to-line at 240 V, only the 120 V oven lights, controls, timers, and motors return on the neutral. Section 120.61(B) permits the feeder or service neutral for household ranges, wall ovens, cooktops, and dryers to be taken at 70 percent of the demand calculated above. That reduction is one of the two available for the neutral; the other, in 120.61(C), takes the portion of the unbalanced load above 200 A at 70 percent for loads other than nonlinear loads and three-wire single-phase or two-phase systems.

6. Heating, Air Conditioning, and Motor Loads

Fixed electric space heating enters the standard method at 100 percent of the total connected load under 120.51, with no demand factor unless the authority having jurisdiction grants one for a system with demonstrated diversity or duty cycle. Air-conditioning equipment enters at its rated load; for hermetic refrigerant motor-compressors the branch-circuit selection current or rated-load current from the nameplate governs per Article 440.

Section 120.60 then permits the smaller of two noncoincident loads to be omitted, and heating versus cooling is the textbook case. In the standard method the designer includes the larger of the space-heating load or the air-conditioning load and drops the other. For a heat pump with electric resistance supplemental heat, both the compressor and the strip heat are counted unless a control prevents simultaneous operation, in which case the smaller may be omitted.



Section 120.50 adds 25 percent of the largest motor load, per 430.24, on top of the other loads. In a dwelling the largest motor is usually the air-conditioning compressor. Where the compressor has been omitted under 120.60 because heating governs, the common practice, and the one Keentel follows, is to apply the 25 percent to the largest motor that remains in the calculation, typically the waste disposer, a well pump, or a pool pump. The adder is small, but omitting it entirely is a plan-review comment waiting to happen.

7. The Standard Method, End to End

Part III of Article 120 is the standard method for feeders and services. It applies to any occupancy, and for dwellings it is the method against which the optional methods are compared. The example below is a single-family dwelling with a full set of electric appliances and a Level 2 EV charger, worked under the 2026 numbers with the 2023 result alongside.

The trade running through the table is between fault current and voltage stress. Reduce the earth fault current and you reduce equipment damage, arc flash energy on earth faults and ground potential rise — but you raise the voltage on the healthy phases, make detection harder, and lose effective grounding. Raise it and you get the opposite. There is no method that is best on every axis, which is why the decision has to be made against the specific installation and stated in the design basis.

8. You Cannot Ground a Delta

An assumption sits underneath the whole discussion and it is worth surfacing: that there is a neutral available to ground. Frequently there is not.



A delta winding has no star point. There is nothing to connect to earth, and a system fed from a delta winding is ungrounded unless something else grounds it. That is not a defect in the transformer; it is a property of the connection, and it is extremely common — the delta side of a delta / grounded-wye transformer, the tertiary of a three-winding unit, and the collector side of many generating plant arrangements.

Where a grounding reference is wanted on such a system, it has to be created deliberately, with a grounding transformer. The two usual arrangements are a zigzag transformer, which presents a low zero-sequence impedance path by construction, and a wye-delta grounding bank, where the delta provides the ampere-turn balance. Either way the device is a designed component with its own ratings, and the ratings are not obvious.

• Zero-sequence impedance, which sets the earth fault current and therefore whether the system is effectively grounded.
• Continuous rating, for the unbalance current it will carry in normal operation.
• Short-time thermal rating, for the fault current it must survive for the clearing time — typically a ten-second or one-minute rating rather than a continuous one, which is why grounding banks look small for their duty.
• Its own protection, because a grounding bank that fails open silently removes the grounding reference from the system, which is the failure mode in Section 12.

The related point is that a transformer’s zero-sequence behaviour is not readable from its winding connection alone. For a grounded-wye / grounded-wye transformer without a delta, the zero-sequence impedance depends on core construction — a three-legged core-form unit behaves quite differently from a shell-form or five-legged design with the same nameplate. Any effective grounding calculation needs the tested zero-sequence impedance, not an assumed one.

9. Generators Are a Different Problem

The circulating material treats transformer neutrals and generator neutrals as the same subject. They share a name and very little else, and the difference is worth understanding because it explains a practice that otherwise looks inconsistent.



On a transformer, the grounding method is chosen to control system behaviour. On a generator, the dominant consideration is the machine itself. A stator earth fault passes current through the stator core iron, and iron damage from that current is expensive and slow to repair — potentially requiring a rewind or core restack. The damage is a strong function of fault current magnitude, so the design objective is to limit stator earth fault current to a very low value.

The usual arrangement is high-resistance grounding through a distribution transformer with a resistor on its secondary. The transformer steps the neutral voltage down so that a modestly sized secondary resistor reflects as a very high effective primary resistance, limiting the earth fault current to a few amperes. That is enough for protection to detect the fault and far too little to damage the core.

Two consequences follow.

• The generator sits in its own grounding zone. It is normally connected to the system through a delta winding on the step-up transformer, which isolates the generator zero-sequence network from the system entirely. The generator can be high-resistance grounded while the system beyond the transformer is solidly grounded, and neither affects the other. That is by design and it is a good illustration of Section 8 in practice.
• Detecting faults near the neutral requires special methods. A conventional neutral overvoltage element sees very little for a fault close to the star point, because the voltage there is small. Achieving coverage of the full winding requires third-harmonic-based or injection-based schemes, which is why hundred-percent stator earth fault protection is a distinct and non-trivial function.

So the answer to "why is the generator neutral grounded through a resistor when the transformer next to it is solidly grounded?" is that they are solving different problems, in electrically separate zones.

10. Where the Two Systems Interact

Having separated the two functions, the paper should now put them back together, because the most consequential engineering point is the place they meet.



Ground potential rise is the product of the current returning through the grid and the grid’s resistance to remote earth. Step and touch potentials — the criteria the grid design has to satisfy — scale with that rise. So the earth fault current, which is set by the system grounding decision, is a direct input to the grid design.

The interaction runs in both directions and is worth stating plainly.

• Choose solid grounding and the earth fault current is high, which is good for protection and detection, and makes step and touch potential harder to satisfy. The grid needs more conductor, more rods, possibly surface treatment, and in poor soil that can become a substantial cost.
• Choose resistance grounding and the earth fault current is low, ground potential rise falls with it, and the grid design gets easier — but detection is harder, effective grounding may be lost, and the healthy phases see more voltage.
• Change the system grounding later — add a source, change a transformer connection, add a grounding bank — and the grid analysis is invalidated. A grid designed against one fault current is not automatically adequate for a higher one.

Both calculations also depend on the protection clearing time, which is a third discipline’s decision. The permissible touch voltage falls as duration rises, and the conductor thermal rating depends on the same duration. So system grounding, grid design and protection settings are one coupled problem, routinely solved by three different people at three different times.

11. One Bond, Not Several

There is a rule that follows from the distinction between the two earthing functions and is broken constantly in the field.



The neutral is bonded to earth at one point per separately derived system. Not at the transformer and again at the switchboard; not at every panel; once.

The reason is that a neutral conductor carries current in normal operation on a system serving unbalanced or single-phase load. If the neutral is bonded to earth at more than one point, that normal neutral current divides between the neutral conductor and every parallel path through the earthing system — the grid, the building steel, cable armour, conduit, piping. The consequences are real and cumulative.

• Current flows continuously through structural steel and equipment enclosures that were never intended to carry it.
• Ground fault protection sees a standing residual current that has nothing to do with a fault, which either causes nuisance operation or forces the sensitivity to be relaxed until the protection no longer does its job.
• Potential differences appear between points of metalwork, producing stray voltage and measurable interference in instrumentation and communications.
• Corrosion accelerates where the current enters and leaves buried metalwork.

The diagnostic is straightforward when someone thinks to do it: with the system energised and carrying normal load, measure current on the bonding conductors. A protective bonding conductor should carry essentially nothing under healthy conditions. If it is carrying steady current proportional to load, there is a second neutral-to-earth bond somewhere, and finding it is a survey rather than a calculation.

12. The Neutral Earthing Resistor Is a Component That Fails

A resistance-grounded system depends on a resistor. Resistors fail, and the failure mode that matters is the silent one.



If a neutral earthing resistor fails open — or its connection fails, or an isolating link is left open after maintenance — the system is no longer resistance grounded. It is ungrounded. Nothing about normal operation changes: the load is unaffected, voltages look normal, and nothing alarms unless something was specifically arranged to alarm.

The system then sits in the condition described in Section 5, exposed to arcing ground overvoltage, with its earth fault protection unable to see a first fault because there is no return path for the current the protection is looking for. The consequences appear when the first earth fault occurs, which may be months later, and the failure is frequently attributed to whatever equipment happened to fail rather than to the missing resistor.

The engineering responses are inexpensive relative to the exposure.

• Continuity monitoring on the neutral earthing path, alarmed to somewhere a person will act on it. This is the single most valuable measure and it is frequently omitted.
• Periodic testing of the resistor value and the integrity of the connection, as a scheduled maintenance activity with a recorded result rather than a visual inspection.
• Isolating links treated as a controlled item, with the position verified and recorded after any work on the neutral connection.
• Neutral displacement monitoring, which detects the neutral shift characteristic of an earth fault whether or not the resistor is intact, and which provides an independent indication.

The same reasoning applies to a grounding transformer. A grounding bank that has been isolated, has failed, or has had its protection operate without anyone noticing leaves the system ungrounded in exactly the same way.

13. Case Studies

The following scenarios are composite and illustrative. They are constructed from patterns that recur across industrial and utility systems to show how these failures develop and how they are found. They do not describe any specific client, site, project, manufacturer, or utility.

13.1 Case A — The System That Had Quietly Become Ungrounded

Situation.  A medium-voltage distribution system at an industrial facility, designed as low-resistance grounded with earth fault protection set to detect and clear faults on the resistor-limited current. The system had operated without incident for a number of years.

What happened

A phase-to-earth fault developed on a cable. It did not clear. Over a period, insulation failures appeared at several unrelated locations on the same galvanically connected system, including equipment remote from the original fault and equipment that had no history of problems.

What the investigation found

The neutral earthing resistor connection had been opened during maintenance work some time earlier and not restored. The system had been operating ungrounded. The first earth fault therefore drew only capacitive current, well below the earth fault protection setting, and was never detected. Because the fault was intermittent rather than solid, the system had been subjected to repeated arcing ground transients, stressing insulation across the whole network. The failures that eventually appeared were the weakest points in that population, not the fault location.

Why it was hard to see

Nothing in normal operation indicated a problem. Load was unaffected, phase voltages measured normally under healthy conditions, and no alarm existed for loss of the neutral earthing path. The condition was invisible until it caused damage, and the damage appeared to be unrelated equipment failures.

Remedy

The resistor connection was restored and tested. Continuity monitoring of the neutral earthing path was added with an alarm to the control room. Neutral displacement monitoring was installed to provide independent indication of an earth fault. Isolating links in the neutral path were added to the controlled-item register with position verification required after any work.

Lesson

A resistance-grounded system depends on a component, and that component can fail in a way that changes the entire system’s behaviour without changing anything observable. If there is no monitoring on the neutral earthing path, the design has an undetectable single point of failure at its foundation.

13.2 Case B — Arresters Selected on a Grounding Assumption That Was Not True

Situation.  A substation serving a generating installation, with surge arresters selected on the basis that the system was effectively grounded — permitting a lower arrester rating and a correspondingly better protective margin against the equipment insulation level.

What happened

Arrester failures occurred over a period, at more than one location, without an obvious lightning or switching cause coinciding with them.

What the investigation found

The effective grounding assessment had been performed for the normal system configuration, in which a particular transformer provided the grounding reference. In several credible operating configurations — including a routine outage of that transformer — the point was no longer effectively grounded. The impedance ratios exceeded the criterion, the temporary overvoltage during an earth fault exceeded what the arresters were rated to withstand, and the arresters absorbed energy they were never selected for.

A contributing factor

The zero-sequence impedance used in the original assessment had been taken from a typical value rather than from the transformer test report. The actual tested value differed enough to matter, which is a recurring issue for wye-wye transformers where core construction rather than the connection diagram governs the zero-sequence behaviour.

Remedy

The effective grounding assessment was repeated across the credible operating configurations using tested zero-sequence data, arrester ratings were revised for the worst case rather than the normal case, and operating instructions were revised so that configurations that removed the grounding source were identified and managed.

Lesson

Effective grounding is a property of the system at a point in a configuration, not a property of an installation. Assessing it only in the normal configuration, with assumed impedance data, produces an arrester selection that is correct most of the time and inadequate exactly when it is needed.

13.3 Case C — The Ground Fault Relay That Would Not Stay Set

Situation.  A low-voltage distribution system in a facility with substantial single-phase load. Ground fault protection had been progressively desensitised over several years because it operated on load changes that were not faults, until eventually it was set so high that it provided little useful protection.

What the investigation found

The neutral was bonded to earth at more than one point. In addition to the correct bond at the source, a second bond existed at a downstream distribution board, created during an earlier modification. Normal neutral current was therefore dividing between the neutral conductor and parallel paths through the building steel and the earthing system, and the residual quantity the ground fault protection measured included a standing component proportional to load rather than only fault current.

What else it was causing

Measurable potential differences between points of building metalwork, intermittent noise on instrumentation and communications circuits sharing routes with the affected steel, and evidence of accelerated corrosion where the stray current entered and left buried metalwork. None of these had previously been connected to the protection problem, and each had been investigated separately.

Remedy

A bonding survey located and removed the additional bond. Neutral-to-earth continuity was verified as a single point per separately derived system. Ground fault protection was reset to appropriate sensitivity and verified against measured residual current under load. The single-point bonding requirement was written into the facility’s modification procedure so that future work would not recreate it.

Lesson



A ground fault relay that keeps operating without a fault is reporting something real. Desensitising it removes the symptom and the protection at the same time. The underlying condition also produces stray voltage, interference and corrosion, which are usually investigated as unrelated problems by different people.

14. Reading the Source Correctly

The core distinction is right and well put

Body earthing protects people and metalwork; neutral earthing determines how the system behaves during an earth fault. Both may terminate on the same grid and their purposes are different. That is the correct framing and it is the answer to the question as asked.

The arithmetic is right, and it is not the dangerous number

Eleven divided by the square root of three is about 6.35 kilovolts, and healthy phases approaching eleven kilovolts on an ungrounded system is a factor of 1.73. Correct — and that sustained condition is manageable. The mechanism that actually matters is the intermittent arcing earth fault and the much higher transient overvoltages it can produce.

Effective grounding has a numerical definition and a cost

Zero-sequence to positive-sequence reactance ratio not exceeding three, and zero-sequence resistance to positive-sequence reactance not exceeding one. Meeting it limits the earth fault factor to about 1.4 and permits lower-rated arresters. Failing it makes insulation coordination more expensive. It is calculated, it is configuration-dependent, and it needs tested transformer data.

Not every winding has a neutral to ground

A delta winding has no star point. Systems fed from one are ungrounded unless a grounding transformer is provided, and that device has zero-sequence impedance, continuous, short-time and protection requirements of its own.

Generators are a separate problem

Generator neutral grounding is driven by stator core damage rather than system behaviour, which is why high-resistance grounding limiting the fault to a few amperes is the normal answer, and why the generator sits in its own grounding zone behind the step-up transformer delta.

The two earths meet at ground potential rise



The system grounding decision sets the earth fault current, which with the grid resistance sets ground potential rise, which is what the step and touch potential criteria are evaluated against. Heavier grounding makes protection easier and the grid harder. They are one design problem, and the clearing time is an input to both.

15. Keentel Grounding and System Earthing Services

Grounding and system earthing sit at the intersection of protection, insulation coordination and personnel safety, which is why they are so often designed in pieces. Keentel Engineering treats them as one problem.

15.1 System Grounding Design

• System grounding method selection against the specific installation — solid, low-resistance, high-resistance, reactance or resonant — with the trade between fault current, overvoltage, detectability and continuity of operation documented in the design basis.
• Effective grounding assessment using tested transformer zero-sequence data, evaluated across the credible operating configurations rather than the normal one only.
• Neutral earthing resistor and reactor specification including resistance value, continuous and short-time ratings, and monitoring requirements.
• Grounding transformer specification and sizing — zigzag or wye-delta — with zero-sequence impedance, continuous unbalance rating, short-time thermal rating and protection.
• Generator neutral grounding design, including distribution transformer and secondary resistor sizing and coordination with stator earth fault protection.

15.2 Ground Grid and Safety



• Substation grounding grid design from measured soil resistivity, with ground potential rise, step and touch potential analysis against the actual earth fault current and protection clearing time.
• Conductor and connection sizing for thermal duty, and evaluation of existing grids against revised fault levels where the system has changed.
• Protective earthing and bonding design, single-point bonding verification, and stray current and circulating current investigation.
• Lightning protection and shielding design, and coordination of the lightning, protective, system and functional earthing arrangements onto one grid.

15.3 Insulation Coordination and Protection

• Insulation coordination studies, surge arrester selection against continuous operating voltage and temporary overvoltage for the worst credible grounding configuration, and protective margin verification.
• Earth fault and zero-sequence protection design and settings, including sensitive earth fault, neutral displacement, restricted earth fault and hundred-percent stator earth fault schemes.
• Short-circuit, protective coordination and arc-flash studies, with the earth fault case evaluated rather than assumed to be bounded by the phase fault case.
• Current transformer sizing and saturation assessment for earth fault applications, where the quantities are small and accuracy matters most.

15.4 Investigation, Assessment, and Owner’s Engineer Support

• Investigation of arrester failures, repeated insulation failures, nuisance ground fault operation, stray voltage and interference problems.
• Condition assessment of existing grounding installations, including verification that the system is grounded the way the drawings say it is.
• Design review of third-party grounding studies and packages, and QA/QC of soil resistivity data and grid models.
• Commissioning specification and test procedures including grid integrity, resistance measurement, neutral earthing path continuity and protection verification.

Keentel Engineering holds a Florida Certificate of Authorization and maintains offices in Tampa, Austin, Sacramento, and Baltimore, supporting projects across the interconnections.

References and Further Reading

The following are referenced by subject in the body of this document. The current published edition of each standard governs its own requirements, and the edition adopted by the authority having jurisdiction governs code requirements.

System Grounding

• IEEE Std 142, Recommended Practice for Grounding of Industrial and Commercial Power Systems — the Green Book, covering system grounding method selection, equipment grounding and the grounding of sensitive electronic equipment  —  IEEE Standards Association
  https://standards.ieee.org/
• IEEE Std C62.92 series, Guide for the Application of Neutral Grounding in Electrical Utility Systems — including the effective grounding criteria and the treatment of generator and distribution system grounding  —  IEEE Standards Association
  https://standards.ieee.org/
• IEEE Std C37.101 for generator ground protection and IEEE Std C37.102 for generator protection, covering stator earth fault schemes including hundred-percent coverage methods  —  IEEE Standards Association
  https://standards.ieee.org/
• IEEE Std C57.32 for neutral grounding devices, covering grounding transformers, neutral earthing resistors and reactors and their rating requirements  —  IEEE Standards Association
  https://standards.ieee.org/

Ground Grid, Safety, and Bonding

• IEEE Std 80, Guide for Safety in AC Substation Grounding — ground potential rise, step and touch potential criteria, and grid design  —  IEEE Standards Association
  https://standards.ieee.org/
• IEEE Std 81, Guide for Measuring Earth Resistivity, Ground Impedance, and Earth Surface Potentials of a Grounding System, and IEEE Std 837 for permanent connections  —  IEEE Standards Association
  https://standards.ieee.org/
• NFPA 70, National Electrical Code — including the article covering grounding and bonding, separately derived systems and the single-point bonding requirement  —  National Fire Protection Association
  https://www.nfpa.org/
• ANSI C2, National Electrical Safety Code, for grounding requirements applicable to electric supply stations and lines  —  Institute of Electrical and Electronics Engineers
  https://standards.ieee.org/products-programs/nesc/



Insulation Coordination and Overvoltage

• IEEE Std 1313.1 and IEEE Std 1313.2, Standard and Application Guide for Insulation Coordination, covering temporary overvoltage and the relationship to system grounding  —  IEEE Standards Association
  https://standards.ieee.org/
• IEEE Std C62.11 for metal-oxide surge arresters and IEEE Std C62.22 for their application, including selection against continuous operating voltage and temporary overvoltage  —  IEEE Standards Association
  https://standards.ieee.org/
• IEC 60071 series, Insulation co-ordination, including the earth fault factor and its relationship to the neutral earthing arrangement  —  International Electrotechnical Commission
  https://webstore.iec.ch/
  

Frequently Asked Questions

Notice and Disclaimer

This document is original technical content prepared by Keentel Engineering LLC for general professional information. It is not project-specific engineering advice and does not constitute a design, a study, an equipment specification, or a safety determination for any installation. System grounding method selection, ground grid design, insulation coordination and protection settings must be established by project-specific analysis using measured soil data, tested equipment impedance data, and the applicable code and utility requirements.

Values, ratios and behavioural descriptions in this paper are general engineering discussion. Actual overvoltage magnitudes, fault currents, ratings and grounding requirements vary with system configuration, equipment characteristics and operating condition, and must be established from calculation and from the applicable standards rather than inferred from typical figures.

The case studies in Section 13 are composite and illustrative. They are constructed from patterns that recur across the industry to demonstrate how these failures develop and how they are found. They do not describe any specific client, site, project, manufacturer, or utility, and no inference should be drawn about any actual installation or party.

Work on or near neutral earthing equipment carries specific hazards, including the possibility that a neutral conductor is at elevated potential during a fault. All such work must be carried out under an appropriate safe system of work by qualified persons.



Keentel Engineering LLC is an independent engineering consultancy. Reference to any standard, code, industry organisation, regulator, or equipment category in this document does not imply affiliation with, endorsement by, or sponsorship from any such organisation or manufacturer.