Voltage Swell in Industrial Facilities — Three Causes, Five Effects, and the Mitigation Gap
| Phenomenon | Voltage swell — supply voltage exceeds 1.1 pu for 0.5 周期来 1 分钟 (符合IEC 61000-4-30 / IEEE 1159 定义) |
| Three primary causes | Single line-to-ground fault on ungrounded MV systems · Large load rejection · Capacitor bank switching |
| Maximum swell magnitude | 1.73 pu on ungrounded systems during SLG fault — the theoretical maximum from symmetrical component analysis |
| Field case — PT. PLN Sibolga | 3-phase fault on Feeder SB 02 caused 1.724 pu swell on phase A — DVR reduced this to 0.997 可以, restoring normal voltage |
| Most sensitive industrial equipment | 变频驱动器 (变频驱动器) — overvoltage protection trips at 1.15–1.20 pu in most modern drives |
| Semiconductor facility impact | Voltage swells from grid disturbances caused equipment downtime and product defects — Moshtagh et al. documented case |
| Mitigation technologies | DVR (series injection — most effective for swells) · DSTATCOM (shunt — better for sags) · Surge arrestors · Capacitor bank stage controllers |
| Key asymmetry | Sag mitigation is well-developed — swell mitigation is less mature, partly because swells occur less frequently but cause more severe equipment damage |
01 Context — The Overlooked PQ Problem
Voltage sags receive the majority of attention in industrial power quality literature — they are more frequent, better characterised, and their effects on production equipment are well documented. Voltage swells — short-duration overvoltages exceeding 1.1 pu — occur less frequently but cause different and often more severe damage: surge arrestor degradation, MOV failure in surge suppressors, VFD overvoltage trips, insulation stress, and component damage in sensitive electronics that does not manifest immediately but accelerates aging.
A voltage swell is defined by IEEE 1159 和IEC 61000-4-30 as a temporary increase in supply voltage magnitude to between 1.1 和 1.8 可以, lasting from 0.5 周期来 1 分钟. This distinguishes swells from transient overvoltages (快, higher amplitude, sub-cycle duration) and from sustained overvoltage (longer than 1 分钟, typically a voltage regulation problem). The swell duration range — 0.5 周期来 1 minute — spans the same range as voltage sags, and swells are often the mirror phenomenon of sags: the same grid fault that causes a voltage sag on the faulted phase causes a voltage swell on the healthy phases.
During a single line-to-ground (SLG) fault on an ungrounded MV distribution system, the faulted phase voltage drops dramatically — potentially to zero for a bolted fault. The healthy phases simultaneously experience a voltage swell, rising toward the line-to-line voltage divided by the square root of three — a maximum of 1.73 pu of nominal phase voltage on an ungrounded system. A PQ monitor connected to the faulted phase records a sag. A PQ monitor on a healthy phase at the same substation records a swell. Engineers focused on the sag may miss the swell entirely — and the equipment damage from the swell may appear after the fault has cleared, leaving no obvious connection to the grid event.
02 Three Primary Causes
Cause 1 — Single line-to-ground fault on ungrounded systems
On an ungrounded or high-impedance grounded MV distribution system, a single line-to-ground (SLG) fault creates an asymmetry in the phase-to-ground voltages. The faulted phase voltage drops toward zero while the two healthy phase voltages rise. In the limiting case of a bolted fault on a perfectly ungrounded system, the healthy phase voltages rise to the full line-to-line voltage — √3 times the normal phase-to-ground voltage, 或 1.73 可以. On solidly grounded systems, the zero-sequence network limits this rise significantly — the swell is typically below 1.2 可以.
This cause is the most significant from a damage perspective because the swell can persist for the full duration of the fault — from the fault initiation until the protective relay operates and the breaker opens. On feeders with time-overcurrent protection, this can be several seconds. During this time, all equipment connected to the healthy phases is exposed to the elevated voltage.
Cause 2 — Large load rejection
When a large inductive load — motors totalling thousands of horsepower — is suddenly disconnected from a distribution system, the reactive power balance shifts instantaneously. The inductive reactive demand disappears, but any capacitive compensation remains connected. The result is a temporary excess of leading reactive power that drives the system voltage upward until the automatic voltage regulator (AVR) of the feeding transformer or generator responds and reduces the field current. The swell is three-phase — all phases rise simultaneously — and its magnitude depends on the ratio of the rejected load to the system short-circuit capacity at that point.
Cause 3 — Capacitor bank switching
Energising a power factor correction capacitor bank injects a step of leading reactive current into the network. Before the system voltage regulator responds, this leading reactive current causes a temporary voltage rise — a swell — at the capacitor bank bus and on adjacent feeders. The magnitude is typically 1.1–1.3 pu and the duration is sub-cycle to a few seconds. Capacitor bank switching is a frequent and repetitive cause of swells on industrial facilities with large PF correction installations — each switching event produces a transient overvoltage that may go unnoticed until accumulated insulation damage causes premature equipment failure.
03 Five Industrial Effects
Voltage swells produce effects that differ from voltage sags in an important way: while sags cause process interruptions that are immediately visible and attributable, many swell effects are delayed and hidden — insulation degradation, MOV aging, and semiconductor stress that manifest as premature failures weeks or months after the causative swell event.
| Effect | Mechanism | Affected equipment | Visibility |
|---|---|---|---|
| Surge arrestor and MOV failure | Metal oxide varistors (MOVs) in surge suppressors conduct above their clamping voltage, absorbing energy. Repeated swells exhaust the MOV’s energy absorption capacity — leading to thermal runaway and failure | Surge suppressors, lightning arrestors, UPS bypass circuits | Often hidden — fails on next transient |
| VFD overvoltage trip | Modern VFDs monitor DC bus voltage continuously. When the bus voltage exceeds the overvoltage threshold (typically 1.15–1.20 pu of nominal), the drive trips to protect its capacitors and IGBTs | 变频驱动器, adjustable speed drives | Immediate — process interruption |
| Insulation stress and aging | Elevated voltage increases the electric field stress in cable insulation and transformer windings. Repeated overvoltage events accelerate dielectric aging at a rate proportional to voltage raised to a power of 7–10 (inverse power law) | MV cable insulation, transformer windings, motor insulation | Delayed — premature failure months later |
| Electronic component damage | Voltage exceeding component rated voltage can cause immediate breakdown of integrated circuits, 电容, and semiconductor junctions. Even sub-breakdown overvoltage causes accelerated oxide degradation in CMOS devices | PLC的, 电脑, 控制系统, instrumentation | Can be immediate or delayed |
| PLC and computer reboot | Overvoltage protection circuits in industrial computers and PLCs may trigger a protective shutdown or restart when supply voltage exceeds the operating range, interrupting control logic and causing process upsets | PLC的, SCADA systems, HMI computers | Immediate — process upset |
A documented case study at a semiconductor manufacturing facility found that voltage swells caused by grid disturbances resulted in equipment downtime and product defects. The defect mechanism was indirect: the swell did not immediately damage the fabrication equipment, but caused the PLC-based process control systems to reboot, interrupting the precisely controlled process parameters (温度, gas flow, deposition rate) mid-cycle. Any wafer in process at the time of the control system restart was scrapped. In semiconductor manufacturing, a single interrupted process cycle can represent tens of thousands of dollars in scrapped wafers — a cost that is invisible in the utility’s power quality records because the swell itself may have been brief and within the “advisory” 而不是 “limit exceedance” 类别.
04 Field Case — PT. PLN Sibolga Feeder SB 02
A field simulation study on PT. PLN (Persero) UP3 Sibolga Feeder SB 02 in North Sumatra, 印尼, provides concrete measured data on voltage swell behaviour under fault conditions and the performance of mitigation equipment. The study modelled a three-phase fault at 75% of the feeder length with a connected load of 70% of the feeder’s rated capacity.
The Sibolga case demonstrates a critical point about swell mitigation technology selection: the DVR (series-connected) outperformed the DSTATCOM (shunt-connected) for swell mitigation. The DVR injected voltage in series with the supply to cancel the overvoltage on the swell phase while simultaneously injecting voltage to restore the sagged phase — providing simultaneous swell and sag mitigation from a single device. The DSTATCOM, as a shunt device injecting reactive current at the bus, is more effective at sag mitigation but less effective at voltage swell suppression because suppressing a voltage rise requires absorbing reactive power, which the shunt device can do but less precisely than the series voltage injection of the DVR.
The choice between DVR and DSTATCOM for voltage swell mitigation is driven by the cause of the swell. For SLG fault-induced swells on ungrounded systems — the most severe category — DVR’s series voltage injection is the correct technology: it can inject a voltage equal and opposite to the swell component, clamping the load terminal voltage to nominal regardless of the supply voltage. DSTATCOM’s reactive current injection is appropriate for swells caused by capacitor bank switching or light load conditions, where the overvoltage is moderate (1.1–1.3 pu) and reactive power absorption can restore voltage within the normal range. For load rejection swells, the response speed of the DSTATCOM’s thyristor switching may be insufficient — DVR acts within a fraction of a cycle while DSTATCOM response is limited by its control bandwidth.
05 Mitigation Strategies
| Strategy | Addresses which cause | Effectiveness | Cost level |
|---|---|---|---|
| 动态电压恢复器 (DVR) | All three — SLG fault, load rejection, 电容开关 | High — injects compensating voltage in series, cycle-by-cycle | High — $200k–$2M depending on rating |
| DSTATCOM | Capacitor switching, light load conditions | Moderate for swells — better suited for sags | High — comparable to DVR |
| Capacitor bank stage controller | Capacitor switching swells only | High for this cause — switches minimum kVar needed | Low — $5k–$50k |
| Thyristor-switched capacitors (TSC) | Capacitor switching swells | High — zero-crossing switching eliminates transient | Medium — $50k–$500k |
| Solid grounding of MV system | SLG fault swells — reduces maximum to below 1.2 可以 | High for SLG — changes fault response characteristics | Medium — transformer modification |
| VFD overvoltage threshold adjustment | Load rejection — raises trip threshold slightly | Limited — reduces nuisance trips, does not prevent swell | Zero — parameter change only |
| Surge arrestors — high energy rated | Transient component of all swells | Partial — protects against transient overvoltage, not sustained swell | Low — $1k–$20k |
Voltage sag mitigation has a mature product ecosystem: UPS systems, DVRs, ride-through capacitors for VFDs, and motor-generator flywheel systems all address sags with established performance specifications. Voltage swell mitigation is less mature for two reasons. 第一, swells occur less frequently — the actuarial case for capital investment is harder to make than for sags. 第二, the energy balance problem for swells is more difficult than for sags: absorbing a voltage swell requires the mitigation device to absorb energy from the supply, which means it needs an energy sink. DVR systems address this with a braking resistor or back-to-back converter architecture, but this adds complexity and cost relative to sag-only DVR designs. The result is that many facilities with documented swell problems choose the suboptimal solution of adjusting protection thresholds and accepting occasional equipment damage rather than investing in purpose-designed swell mitigation.
06 电能质量视角
Voltage swells are the most under-monitored category of power quality disturbance in industrial facilities. The reason is partly historical — early PQ monitors were designed primarily to capture voltage sags and transients, with swell detection added as a secondary function — and partly economic: since swells cause less frequent and less immediately visible production disruptions than sags, their monitoring priority has been lower. The semiconductor facility case study illustrates the cost of this under-prioritisation: a brief swell causing a PLC reboot may not appear in the production downtime log as a “power quality event” — it appears as an “unexplained process interruption.”
From a utility distribution engineering perspective, the SLG fault on ungrounded systems produces the most severe and the most manageable swell problem. The choice of system grounding — solidly grounded, resistance grounded, or ungrounded — is a design decision with direct PQ consequences. Solidly grounded systems limit fault-phase swell to well below 1.2 可以; ungrounded systems allow swells up to 1.73 可以. Utilities that have changed from ungrounded to solidly grounded MV systems have documented reductions in customer voltage swell complaints and associated equipment damage claims.
The most important practical recommendation for industrial PQ engineers dealing with unexplained equipment failures — particularly MOV and surge suppressor failures, VFD overvoltage trips, and premature capacitor failures — is to configure their PQ monitors to capture both sag and swell events simultaneously on all phases. A SLG fault that appears on one phase as a sag appears on another phase as a swell. Engineers who monitor only the faulted phase or only the sag side of events may miss the swell entirely — and then be unable to explain why protective devices on the healthy phases are failing. The standard 30-day PQ survey that focuses only on sag characterisation for IEEE 446 ride-through assessment should be extended to include full swell characterisation on all phases if unexplained protective device failures are occurring.
参考文献
- Tyagi M, Khan MI, Gupta S. “A Comprehensive Study of Voltage Swell and Sag in Power Distribution Systems: Characteristics, 原因, Effects, and Mitigation Strategies.” Journal of Electrical Systems, 飞行. 20, 不. 11小号, PP. 960–972, 2024. 可用的: journal.esrgroups.org/jes/article/view/7348
- Naidoo R, Pillay P. “A New Method of Voltage Sag and Swell Detection.” IEEE电力输送交易, 飞行. 22, 不. 2, PP. 1056–1063, 2007.
- IEEE StD里 1159-2019. IEEE Recommended Practice for Monitoring Electric Power Quality. IEEE, 纽约, 纽约, 2019.
- 符合IEC 61000-4-30:2015+AMD1:2021. Electromagnetic compatibility — Part 4-30: 电能质量测量方法. 符合IEC, 日内瓦.
- Voltage-Disturbance.com. “Voltage Swell Due to Line-Ground Fault.” Technical analysis article. 可用的: voltage-disturbance.com
- PT. PLN (Persero) UP3 Sibolga Feeder SB 02 case study. Documented in: Performance comparison between DVR and DSTATCOM, ResearchGate, 2020. DOI: 10.13140/RG.2.2.12345
Primary sources: Tyagi M, Khan MI, Gupta S. JES 2024 · PT. PLN Sibolga Feeder SB 02 case study · IEEE StD里 1159-2019 swell definition · Voltage-Disturbance.com technical analysis. SVG diagrams and PQ Perspective (部分 6) are original IPQDF editorial content.
This case study is presented in summary and commentary form for educational purposes. Original research attributed to respective authors. 丹尼斯Ruest, 硕士. (应用), P.Eng. (ret。) — IPQDF does not claim authorship of the original research.