Floating neutral in three-phase systems: Thermal collapse and destruction of single-phase loads due to voltage displacement

Electromechanical Fundamentals and Topology of Unbalanced Three-Phase Systems

The rigorous analysis of three-phase alternating current electrical power distribution systems requires a deep comprehension of the interactions among phase voltages, line currents, and the neutral conductor impedance. In an ideal, symmetrical, and balanced three-phase network feeding purely linear loads, the voltage at the neutral point coincides with the source ground potential, and the circulating current through the neutral conductor is strictly zero. Nevertheless, within modern industrial, commercial, and residential installations, the predominant presence of stochastically distributed single-phase loads generates a permanent unbalance in the phase currents.

To analytically model this phenomenon, consider a four-wire three-phase system featuring perfectly sinusoidal and balanced source phase voltages:

Where $V_p$ represents the root mean square (RMS) value of the line-to-neutral voltage (phase voltage). The individual loads connected between each phase and the neutral exhibit individual complex impedances denoted as \vec{Z}_A, \vec{Z}_B, and \vec{Z}_C. Applying the generalized Ohm's law for AC networks, the line currents flowing through the phase conductors are defined as:

By direct application of Kirchhoff's First Law (Kirchhoff's Current Law - KCL) at the common load node (the load neutral point), the return current flowing through the neutral conductor back to the source, denoted as \vec{I}_N, is the vector sum of the phase currents:

Under normal operating conditions with unbalance, the neutral conductor carries this resulting current, whose magnitude depends directly on the degree of asymmetry of the load impedances. Nonetheless, the system maintains a stable potential reference point thanks to the inherent impedance of the neutral conductor, which ideally approaches zero (\vec{Z}_n \approx 0).

Physics of Neutral Interruption and Neutral Point Displacement

The critical scenario analyzed in this paper occurs when a physical discontinuity, high contact resistance, or total open circuit arises in the neutral conductor upstream of the connection point for single-phase loads (for instance, in the main distribution switchboard, at the service transformer terminal, or due to the single-pole operation of a circuit breaker or fuse in the neutral, a practice strictly prohibited by international standards such as IEC 60364 and the NEC). When the neutral is interrupted, the return path impedance transitions from a value near zero to infinity (\vec{Z}_n \to \infty).

This discontinuity instantly transforms the circuit into a network of impedances connected in series across the line-to-line voltages. The single-phase loads connected between the phases and the floating neutral are subjected to a massive resistive-inductive voltage divider energized by the line-to-line voltages ($VLL$) instead of the phase voltages ($VLN$).

To analytically determine the resulting voltages across each individual single-phase load following the loss of the neutral, Millman's Theorem (or the nodal voltage method) is applied. The potential of the floating neutral node with respect to the source, denoted as \vec{V}_{N'N}, is calculated via the following expression:

Once the floating neutral potential \vec{V}_{N'N} is determined, the actual voltages applied across the terminals of the single-phase loads connected to each phase (\vec{V}'_{AN}, \vec{V}'_{BN}, \vec{V}'_{CN}) shift drastically, expressed as:

The physical impact of this neutral point displacement is catastrophic and asymmetrical. Voltages across the loads are no longer restricted to the nominal phase-to-neutral voltage (e.g., 230V in European systems), but are instead redistributed as a function of the apparent powers connected to each phase. Specifically, the phase possessing the lowest connected load (highest impedance) will absorb the greatest share of the line voltage, driving its RMS value dangerously close to the system line-to-line voltage (VLL = \sqrt{3} \times VLN, i.e., up to 400V in 230V/400V systems).

Forensic Failure Analysis and Dynamic Behavior of Single-Phase Loads

Forensic investigation of damage sustained by electronic equipment, household appliances, and single-phase machinery following a floating neutral event reveals highly predictable dielectric and thermal destruction patterns. Loads connected to an electrical system can be classified according to their dynamic response under sustained overvoltages:

• Purely Resistive Loads (Heating, incandescent lighting): The power dissipated by a pure resistance is given by P = \frac{V^2}{R}. If the applied voltage increases from 230V to 380V due to neutral displacement, the dissipated power rises by a factor of \left(\frac{380}{230}\right)^2 \approx 2.73. This represents a 173% increase in thermal dissipation, causing the immediate fusion of heating elements and the destruction of surrounding organic insulations.
• Capacitive Loads and Switch Mode Power Supplies (SMPS): Computing equipment, televisions, industrial controllers, and chargers contain input rectification stages followed by electrolytic filter capacitors typically rated to withstand nominal voltages of 275V to 300V AC (for 350V DC peaks in 230V networks). When the phase-to-neutral voltage climbs to 350V-400V AC, the electrolytic capacitors suffer violent dielectric breakdown, puncturing the aluminum oxide layer, generating internal overpressure, pressure-relief vent rupture, leakage of corrosive electrolyte, and ignition of adjacent components.
• Inductive Loads (Single-phase motors, control transformers, electromagnetic ballasts): The increment in applied voltage beyond the knee point of the ferromagnetic core's hysteresis curve provokes severe magnetic saturation. The excitation current ceases to be sinusoidal and exhibits elevated harmonic peaks. The effective inductive impedance drops drastically, increasing no-load current and causing thermal destruction of the copper winding's insulating enamel, culminating in an interturn short circuit.

Electrical / Operational Parameter	Nominal Condition (Balanced System)	Fault Condition (Interrupted Neutral / Extreme Asymmetry)	Dielectric and Thermal Consequence
Phase A Phase-to-Neutral Voltage ($VAN$)	230 V RMS (\pm 5\%)	385 V RMS (Peak > 540 V)	Dielectric puncture in filter capacitors; destruction of metal oxide varistors (MOVs).
Phase B Phase-to-Neutral Voltage ($VBN$)	230 V RMS (\pm 5\%)	110 V RMS (Undervoltage)	Operational failure; low-voltage disconnection in relays and motors (elevated slip).
Phase C Phase-to-Neutral Voltage ($VCN$)	230 V RMS (\pm 5\%)	65 V RMS (Severe undervoltage)	Equipment inoperability; low-voltage overheating in compressors.
Neutral Conductor Current ($I_N$)	Minimal or zero (< 10\% of $I_f$)	Infinite / Open Circuit (\vec{Z}_n = \infty)	Electric arc at the interruption point; melting of terminals and lugs.
Dissipated Thermal Power ($P$)	1.0 \times Pnom	2.73 \times Pnom (on overvoltage phase)	Carbonization of insulation, conductor melting, risk of electrical fire.

Advanced Design, Mitigation, and International Standards Strategies

To absolutely mitigate the risks associated with a floating neutral and sustained temporary overvoltages (TOV), international standards such as IEC 60364, IEEE Std 141 (Red Book), and the NEC (National Electrical Code) establish rigorous prescriptive and performance guidelines.

Oversizing and Mechanical Robustness of the Neutral Conductor

Historically, in networks featuring heavily non-linear loads generating high zero-sequence harmonic components (specifically the third harmonic and its multiples, 150 Hz and 250 Hz in 50 Hz networks), harmonic currents add algebraically in the neutral conductor. In these scenarios, the neutral current can exceed the nominal phase current, reaching up to 173\% thereof in fully unbalanced three-phase systems equipped with single-phase rectifiers. Consequently, design under IEC 60287 mandates that the cross-sectional area of the neutral conductor equals that of the phases (full sizing) when supplying significant non-linear loads, and strictly prohibits neutral cross-section reduction formerly permitted in purely linear networks.

Temporary Overvoltage (TOV) Protection via POP Devices

The most advanced active countermeasure to prevent catastrophic damage from neutral interruption in low-voltage installations is the implementation of Power Frequency Overvoltage Protection Devices (known in European standard UNE-EN 50550 as POP devices). These devices continuously monitor the phase-to-neutral voltage. If they detect a phase-to-neutral voltage rise above critical thresholds (for example, exceeding 275V AC sustained for more than a few cycles), the device emits a trip signal that activates a main circuit breaker or a shunt release coil coupled to an emission current release (MX), disconnecting the entire installation all-pole (including the phase and, optionally, the neutral depending on the TN-S or TT earthing scheme configuration) before loads suffer irreversible damage.

Practical Application and Analysis via Vexten Suite

To illustrate the application of the developed theoretical concepts, a case study solved using the calculation engine of the Vexten Suite platform is presented, applying normative methodologies from IEC 60909 and IEEE 141 to an industrial low-voltage three-phase distribution system.

Definition of the Industrial Network Scenario

A low-voltage main distribution switchboard (LVMS) fed by a 630 kVA, 10 kV / 400V-230V Dyn11 distribution step-down transformer with a transformer short-circuit impedance u_k = 4.5\% is analyzed. The earthing system is type TT with a service earth electrode resistance R_A = 2\,\Omega and exposed conductive parts earth resistance R_B = 5\,\Omega.

The switchboard feeds a set of unbalanced single-phase loads distributed across three sub-switchboards:

• Phase A Sub-switchboard ($ST_A$): Resistive and IT load equivalent to $S_A = 35 kVA$, with a power factor \cos(\varphi_A) = 0.95 inductive.
• Phase B Sub-switchboard ($ST_B$): Mixed load equivalent to $S_B = 15 kVA$, with a power factor \cos(\varphi_B) = 0.85 inductive.
• Phase C Sub-switchboard ($ST_C$): Light load equivalent to $S_C = 8 kVA$, with a power factor \cos(\varphi_C) = 0.90 inductive.

Execution of the Floating Neutral Calculation Algorithm (Vexten PowerFlow & Fault Analysis Module)

The Vexten Suite engine processes the complex equivalent load impedances referred to the source phase voltages:

Upon simulating the total interruption of the main neutral conductor at the LVMS output, Vexten Suite's node resolution algorithm computes the floating neutral potential displacement \vec{V}_{N'N} through the iterative application of Millman's Theorem:

The computational result returned by the Vexten Suite engine yields the following displacement vector:

From this value, the actual effective voltages applied across the single-phase loads in each sub-switchboard are calculated and reported in the Vexten Suite engineering report as follows:

As observed in the results of the Vexten Suite analysis, Phase C, possessing the lowest connected load ($8 kVA$), experiences a voltage of $368.2 V RMS$, exceeding the nominal permissible voltage of connected single-phase equipment by 60\%. This value confirms the predictive diagnosis of the tool: massive and immediate failure of switch mode power supplies, puncture of surge protective device varistors (Type 3 SPDs), and thermal destruction of equipment connected to Sub-switchboard C.

Thermal Sizing of the Neutral and Harmonic Derating (IEC 60287 / NEC 310 Standard)

To prevent thermal fatigue failures in the neutral conductor due to harmonic content, Vexten Suite's cable sizing module processes the total harmonic distortion of current ($THD_i$) measured in the installation, which reaches a value of 32\%, heavily dominated by the third harmonic (I_3 = 28\%).

Applying the harmonic content correction factors stipulated by IEC 60287 and IEEE Std 519, the permissible current-carrying capacity of the neutral conductor ($IzN$) is adjusted via the reduction factor $Kharm$:

For the case study with THD_i = 32\%, Vexten Suite applies the factor $Kharm = 0.86$, determining that the $95 mm ^2$ copper phase conductor section with XLPE insulation mandatorily requires a neutral conductor with an equivalent section of $95 mm ^2$ (rather than a reduced section of $50 mm ^2$), complemented by the installation of an active harmonic filter (AHF) at the LVMS to cancel zero-sequence homopolar currents before they saturate the distribution system and neutralize the thermal risk on the neutral.

Advanced Technical Conclusions

The interruption of the neutral conductor in four-wire three-phase systems represents one of the most destructive failure modes for connected single-phase loads. Mathematical and physical analysis demonstrates that the system ceases to operate under stable phase voltages and instead converts into an asymmetrical voltage divider circuit energized by line voltages. Loads connected to phases with lower power demand absorb severe overvoltages that vastly exceed the dielectric limits of electronic and electromechanical components. The strict implementation of design standards (full neutral cross-section sizing per IEC 60287), the mandatory use of temporary overvoltage protection devices (POP per EN 50550), and engineering validation via advanced calculation platforms such as Vexten Suite guarantee operational resilience, personnel safety, and asset integrity in modern industrial and commercial electrical installations.