그들의 소스를 파악하기 위해 전력 품질 이벤트 중 상관 관계를 확립

01 Context — The Source Attribution Problem

One of the most contentious and practically important questions in industrial power quality engineering is deceptively simple: when a voltage sag or interruption disrupts a customer’s process, who caused it? The answer determines who bears responsibility for the event, who funds any mitigation, and — in regulated utility environments — whether a service quality complaint has merit.

Power quality events may originate from either side of the utility meter:

• Utility-caused (grid events) — faults on transmission or distribution lines, switching operations, capacitor bank switching, voltage regulator operations. These affect all customers connected to the same feeder or substation and are the utility’s operational responsibility
• Customer-caused (internal events) — motor starts, arc furnace operations, capacitor switching within the plant, fault conditions on internal wiring. These are the customer’s responsibility and may also affect neighbouring customers connected to the same distribution bus
• Neighbouring customer-caused events — a large non-linear or intermittent load at an adjacent plant (arc furnace, large motor, resistance welder) that propagates voltage disturbances through the shared distribution network to other customers

Without appropriate monitoring — specifically, multi-point time-synchronised monitoring that captures the event simultaneously at multiple locations — it is impossible to distinguish these three cases from a single measurement point. A single monitor at the service entrance of a plant records the event but cannot determine whether it originated upstream (utility) or at a neighbouring customer’s premises.

02 The Event — A Bird in the Switchgear

In an industrial neighbourhood in a small Midwestern city, a crow flew into medium-voltage switchgear at a utility substation. The contact between the bird and the energised equipment created a phase-to-ground fault on the distribution system. The fault current caused voltage sags and momentary loss of voltage across a significant portion of the distribution network — affecting customers over several miles and more than 200 customer accounts.

Four I-Sense monitors were distributed throughout the neighbourhood as part of the I-Grid monitoring network. Each monitor recorded the event independently, with GPS-accurate timestamps that allowed the recorded data to be precisely correlated in time.

03 Source Attribution — How the Monitoring Proved the Cause

단계 1 — GPS timestamp correlation

Each I-Sense monitor recorded the voltage event independently, with a GPS-accurate timestamp. When the four records were aligned on a common time axis, all four monitors showed voltage depressions beginning at exactly the same instant — within the GPS synchronisation accuracy of approximately 1 microsecond. This simultaneous onset is the definitive signature of a grid event: an event originating within any individual customer’s premises would reach the other three monitor locations with a measurable propagation delay, not simultaneously.

단계 2 — Waveform analysis

Analysis of the waveforms at all four monitors showed the characteristic signature of a single-line-to-ground (SLG) fault — the most common fault type on distribution systems, accounting for approximately 70–80% of all distribution faults. Note that Monitor #1 recorded line-to-line voltage while the other three recorded line-to-neutral voltage — the different measurement configurations produced different waveform shapes from the same event, which could appear inconsistent without the time-synchronisation context.

단계 3 — Utility record confirmation

4개의 기록이 모두 하나의 유틸리티로 인한 이벤트를 나타낸다는 가설은 유틸리티 회사 기록에서 4개의 모니터 모두에 의해 기록된 전력 품질 이벤트와 정확히 동일한 타임스탬프를 가진 병렬 피더의 릴레이 작동을 공개했을 때 확실하게 확인되었습니다.. 계전기는 까마귀로 인한 오류를 제거하기 위해 작동했습니다(일상적인 보호 작업). 그러나 해당 타임스탬프는 사건의 원인과 시간 모두에 대해 반박할 수 없는 확인을 제공했습니다..

04 Customer Impact and Mitigation

One of the four monitored customers experienced a 13-hour process shutdown as a result of this event. The shutdown duration is disproportionate to the electrical event duration — the voltage disturbance itself lasted only a few cycles. The 13-hour shutdown reflects the restart time and complexity of the customer’s industrial process, not the duration of the power quality event. This is a common pattern in process industries: a millisecond electrical event causes an hours-long production disruption.

Implications for monitoring network design

The study draws an important conclusion about monitoring network density. Because grid events — caused by utility network faults — propagate across the network and are experienced simultaneously by all customers in a geographical region, it is not necessary to monitor every customer to assess the power quality environment of a region. A monitoring network covering a small percentage of customers, if properly designed and time-synchronised, provides statistically representative data for the entire region.

This principle has significant implications for utility PQ monitoring program design: sparse, well-placed, time-synchronised monitors can characterise network-wide PQ behaviour far more efficiently than dense, uncoordinated single-point measurements at individual customer service entrances.

05 Power Quality Perspective

This case study is the clearest possible demonstration of why source attribution requires network monitoring — not just customer-side measurement. From a utility engineering perspective, the case study validates a principle that is fundamental to distribution PQ management: grid events are network phenomena, not individual customer phenomena. A crow in a switchgear at one substation produces voltage sags at 200+ customer locations simultaneously. No individual customer measurement, however sophisticated, can identify this as a single grid event rather than 200 separate events.

The GPS synchronisation technology used in the I-Grid system is the key enabler. Without time synchronisation accurate to the microsecond level, the four monitor records could not be reliably correlated — a 60 Hz power system cycle is approximately 16,700 microseconds, and distinguishing simultaneous onset (grid event) from near-simultaneous onset (propagating internal event) requires much better than cycle-level time resolution.

참조

1. Divan D, Brumsickle W, Eto J. A New Approach to Power Quality and Electricity Reliability Monitoring — Case Study Illustrations of the Capabilities of the I-Grid™ 체계. 어니스트 올랜도 로렌스 버클리 국립 연구소, LBNL-52048, 4월 2003.
2. IEEE 표준 1159-2019. IEEE Recommended Practice for Monitoring Electric Power Quality. IEEE, 뉴욕, NY, 2019.
3. IEC 61000-4-30:2015+AMD1:2021. Electromagnetic compatibility — Part 4-30: 전력 품질 측정 방법. IEC, Geneva.