descarga parcial do transformador

• Descarga parcial (DP) é uma ruptura localizada do isolamento que, deixado despercebido, degrada progressivamente o isolamento do transformador e pode causar falha catastrófica. Monitoramento PD on-line detecta esses defeitos no estágio inicial.
• Cinco técnicas de detecção complementares – elétrica, acústico, UHF, TEV, e químico (DGA) - cada um captura uma manifestação física diferente de descarga parcial, e nenhum método sozinho fornece cobertura diagnóstica completa.
• UM fusão multissensor arquitetura combinando sensores ultrassônicos (20 kHz–200 kHz), sensores de corrente de alta frequência (100 kHz–50 MHz), e Sensores UHF (300 MHz–3 GHz) elimina falsos positivos, permite localização de origem, e oferece a mais alta confiabilidade de detecção.
• Avançado PRPD (Descarga parcial resolvida por fase) análise de padrões tridimensionais e PRPS (Sequência de pulso resolvida por fase) a visualização permite que os engenheiros identifiquem o tipo específico de descarga - corona, descarga superficial, vazio interno, ou potencial flutuante - e priorize a manutenção de acordo.
• Moderno Sistemas de monitoramento de PD integrar com SCADA e plataformas de gerenciamento de ativos empresariais via Modbus, CEI 61850, e DNP3, incorporação de dados de integridade do isolamento no fluxo de trabalho mais amplo de manutenção baseado em condições da concessionária.

Índice

1. O que é descarga parcial em transformadores e por que deve ser monitorada?
2. Quatro tipos comuns de descarga parcial dentro de transformadores de potência
3. Cinco técnicas de detecção de descarga parcial comparadas - elétrica, Acústico, UHF, TEV, e Métodos Químicos
4. Por que a fusão multissensor supera a detecção de método único
5. Quais são os componentes de um sistema online de monitoramento de descarga parcial?
6. Instalação do sensor, Largura de banda, e Função - Ultrassônico, TCFC, e UHF
7. Principais especificações técnicas da unidade hospedeira de monitoramento PD
8. Como os padrões 3D PRPD e as sequências de pulso PRPS identificam os tipos de descarga?
9. Software de monitoramento de back-end – recursos e capacidades de diagnóstico
10. How Does a PD Monitoring System Integrate with SCADA and Asset Management Platforms?
11. Which Transformers Benefit Most from Online Partial Discharge Monitoring?
12. How to Select the Right Partial Discharge Monitoring Equipment — A Buyer’s Guide
13. Applicable International Standards for Partial Discharge Testing and Monitoring
14. Perguntas frequentes (Perguntas frequentes)

1. O que é descarga parcial em transformadores e por que deve ser monitorada?

Descarga parcial is a localised electrical breakdown that only partially bridges the insulation between conductors inside a transformer. Unlike a full flashover, a partial discharge event does not create a complete conductive path, but it does release energy — in the form of electromagnetic radiation, ondas acústicas, aquecer, and chemical by-products — that gradually erodes the surrounding insulation material. Ao longo do tempo, repeated partial discharge activity enlarges the original defect, accelerates insulation ageing, and can ultimately trigger a complete insulation failure, leading to catastrophic transformer damage, interrupções não planejadas, and significant financial loss.

The challenge is that partial discharge activity is invisible during normal operation. External symptoms such as dissolved gas accumulation in the oil or elevated winding temperatures often appear only after the defect has already progressed to an advanced stage. É por isso monitoramento on-line de descarga parcial has become an essential component of modern monitoramento da condição do transformador programmes. By detecting the electrical, acústico, and electromagnetic signatures of PD events in real time, an online system provides the earliest possible warning of insulation degradation — weeks, meses, or even years before the fault would be detected by conventional periodic testing.

2. Quatro tipos comuns de descarga parcial dentro de transformadores de potência

Not all partial discharge is the same. The physical mechanism, localização, and severity of the discharge depend on the nature of the insulation defect. Understanding the four most common PD types helps engineers interpret monitoring data and plan appropriate maintenance responses.

Descarga Corona

Corona discharge occurs at sharp metallic protrusions or poorly shaped electrodes where the localised electric field intensity exceeds the breakdown strength of the surrounding medium — typically transformer oil or gas. The discharge appears as a faint glow and produces predominantly hydrogen gas. While corona is often considered the least severe form of PD, persistent corona activity degrades oil quality over time and can initiate more damaging discharge types.

Descarga Superficial

Surface discharge develops along the interface between solid insulation (pressboard or crepe paper) and the surrounding oil or gas. It is frequently caused by contamination, entrada de umidade, or excessive tangential electric field stress at the insulation surface. Surface discharge can quickly escalate in severity because the carbonised tracking path it creates along the insulation surface progressively shortens the effective insulation distance.

Internal Void Discharge

Gas-filled voids or cavities trapped within solid insulation — typically caused by manufacturing defects, estresse mecânico, or thermal ageing — create regions where the dielectric strength is significantly lower than the surrounding material. When the applied voltage exceeds the breakdown threshold of the void, a partial discharge ignites inside the cavity. A descarga de vazios internos é particularmente insidiosa porque está totalmente encerrada no isolamento e não pode ser detectada por inspeção visual.

Descarga de Potencial Flutuante

Quando um componente metálico dentro do transformador - como uma blindagem, um suporte estrutural, ou uma conexão solta – não está conectado corretamente a um potencial elétrico definido, adquire uma tensão flutuante através de acoplamento capacitivo. Este potencial flutuante pode gerar descargas repetitivas entre o componente e estruturas adjacentes aterradas ou energizadas. A descarga de potencial flutuante normalmente tem alta energia e produz fortes assinaturas UHF e acústicas, tornando-o relativamente mais fácil de detectar, mas também mais prejudicial ao isolamento próximo.

3. Cinco técnicas de detecção de descarga parcial comparadas - elétrica, Acústico, UHF, TEV, e Métodos Químicos

Each detection technique captures a different physical phenomenon produced by partial discharge events. The table below provides a side-by-side comparison of the five most widely used methods, summarising their measurement principles, typical sensitivity, principais vantagens, and primary limitations.

Método de detecção	Physical Quantity Measured	Typical Sensor	Sensitivity Metric	Principais vantagens	Principais Limitações
Elétrica (CEI 60270)	Apparent charge (computador / nC)	Coupling capacitor, bushing tap	Down to ~1 pC	Standardised, quantitative, excellent for factory testing	Susceptible to EMI in field; primarily offline
Acústico / Ultrassônico	Emissão acústica (dB / mV)	Piezoelectric sensor (20–200 kHz)	Moderado	Imune ao EMI; enables PD source localisation via triangulation	Signal attenuated by tank structure and oil path
UHF (Frequência ultra-alta)	Electromagnetic signal (300 MHz–3 GHz)	UHF antenna (conical, spiral, Vivaldi)	Down to a few pC equivalent	Excellent noise rejection; em tempo real; suitable for online use	Sensitivity depends on sensor position; requires installation port
TEV (Tensão transitória da terra)	Surface voltage pulse (mV)	Capacitive plate sensor	Moderado a alto	Não intrusivo; no outage required; instalação simples	Limited to metallic-enclosure equipment; external PD only
Químico (DGA)	Dissolved gas concentration (ppm)	Online DGA monitor / lab chromatography	Indirect indicator	Detects cumulative insulation degradation; established standard	Resposta lenta; cannot pinpoint PD location or type

As the table illustrates, no single technique covers all aspects of partial discharge detection. Electrical methods provide the most accurate charge quantification but struggle with on-site noise. Acoustic and UHF methods excel at online monitoring and source localisation. TEV is ideal for quick non-intrusive screening. DGA reveals cumulative insulation damage but provides no real-time pulse-level information. This complementarity is what drives the industry toward multi-sensor fusion architectures.

4. Por que a fusão multissensor supera a detecção de método único

A single-sensor PD monitor — regardless of how sensitive it is — faces two fundamental challenges: false positives caused by external noise sources and diagnostic ambiguity when only one type of signal is available. Multi-sensor fusion technology addresses both problems by cross-correlating data from sensors operating in entirely different frequency domains and physical measurement principles.

Consider a practical example. An ultrasonic sensor mounted on the transformer tank detects an acoustic emission event. In isolation, the operator cannot be certain whether the signal is genuine PD or a mechanical vibration from a nearby cooling fan. No entanto, if a UHF sensor simultaneously detects a corresponding electromagnetic pulse, and a high-frequency current sensor at the grounding cable records a coincident current spike, the probability that the event is a true partial discharge rises to near certainty. The time-of-arrival difference between the acoustic and electromagnetic signals can further be used to estimate the spatial location of the discharge source inside the transformer.

This fusion approach dramatically reduces false alarm rates, improves diagnostic confidence, and enables the operator to not only confirm that PD is occurring but also determine where it is occurring and how severe it is — all from a single integrated monitoring platform. It is the reason why leading sistemas de monitoramento de descarga parcial de transformadores now incorporate three sensor types as standard, rather than relying on any one method alone.

5. What Are the Components of an Online Partial Discharge Monitoring System?

Um completo online PD monitoring system consists of three functional layers that work together to convert raw discharge signals into actionable diagnostic intelligence.

Sensores de campo

Three types of sensors are deployed on the transformer to capture different physical manifestations of partial discharge. Ultrasonic sensors detect acoustic emissions from PD activity within the windings and oil. High-frequency current (TCFC) sensors clamp onto the core grounding cable to measure pulse currents generated by discharge events. UHF sensors are installed at oil valve ports to capture ultra-high-frequency electromagnetic radiation propagating through the transformer oil. Each sensor is designed for harsh outdoor environments with an IP68 protection rating.

Unidade hospedeira de monitoramento PD

The monitoring host is the central processing hub of the system. It receives analogue signals from all connected sensors, realiza condicionamento de sinal (amplification, filtragem, and impedance matching), and digitises the waveforms at high speed using a multi-channel acquisition architecture. The host calculates key PD parameters — including maximum discharge amplitude, average discharge quantity, and discharge frequency — and applies intelligent algorithms for pattern recognition and fault classification. It is typically rack-mounted in a 2U enclosure inside a convergence cabinet or control panel near the transformer.

Backend Monitoring Software

Installed on a control room computer or server, the software platform provides real-time visualisation, tendências históricas, gerenciamento de alarme, and diagnostic analysis. Its core analytical capabilities include 3D PRPD pattern display, PRPS pulse sequence mapping, discharge amplitude statistics, and comparison against an expert pattern database for automated PD type identification. The software communicates with the monitoring host via Ethernet or RS-485.

6. Instalação do sensor, Largura de banda, e Função - Ultrassônico, TCFC, e UHF

The effectiveness of a sistema de monitoramento de descarga parcial depends heavily on correct sensor selection and placement. The table below details the three sensor types used in a full-spectrum multi-sensor architecture, including their monitoring bandwidth, método de instalação, mounting location, and primary diagnostic function.

Tipo de sensor	Monitorando largura de banda	Método de instalação	Mounting Location	Função Primária
Sensor ultrassônico	20 kHz – 200 kHz	Montagem magnética	Transformer tank surface	Detects acoustic emission signals generated by internal PD activity in windings and insulation structures
High-Frequency Current (TCFC) Sensor	100 kHz – 50 MHz	Fixação	Core grounding point	Captures high-frequency pulse currents flowing through the grounding cable as a result of discharge events
Sensor UHF	300 MHz – 3 000 MHz	Plug-in type	Oil drain valve port	Monitors ultra-high-frequency electromagnetic signals propagating through transformer oil, indicating internal insulation discharge

Installation Notes

Ultrasonic sensors attach to the tank wall using a magnetic holder, which allows flexible repositioning without drilling or welding. For optimal acoustic coupling, a thin layer of couplant gel is applied between the sensor face and the tank surface. The HFCT sensor is a split-core clamp that can be installed around the grounding cable without disconnecting it — meaning no transformer outage is required. The UHF sensor inserts into an existing oil drain valve or dedicated dielectric window port, placing the antenna element inside the oil space for maximum sensitivity to internal electromagnetic signals. All three sensor types are rated IP68, ensuring reliable operation in rain, pó, umidade, and temperature extremes from -20 °C a +125 °C.

7. Principais especificações técnicas da unidade hospedeira de monitoramento PD

The monitoring host is the heart of the system, responsible for high-speed signal acquisition, real-time processing, and data communication. The table below presents the core technical parameters of a representative industrial-grade PD monitoring host designed for substation deployment.

Parâmetro	Especificação
Signal Reception	Ultrassônico, corrente de alta frequência (TCFC), and UHF sensor inputs
Dynamic Range	-80 para -20 dBm
Taxa de amostragem	200 MS/s (200 million samples per second)
Configuração do canal	4 ou 6 canais (configurável pelo usuário)
Channel Consistency	≤ 0.5 dBm
Faixa de monitoramento	≤ 20 000 computador
Transmission Impedance	≥ 12 mV/mA
Interfaces de comunicação	Ethernet RJ45, RS-485
Protocolos Suportados	Modbus RTU/TCP, CEI 61850, DNP3
Fonte de energia	AC 90–240 V, 50/60 Hz
Gabinete	2Montagem em rack U (483 milímetros × 89 milímetros × 300 milímetros)
Método de instalação	Convergence cabinet or control panel mount
Sensor Protection Rating	IP68
Temperatura operacional	-20 °C a +125 °C (sensor); host per cabinet environment
Diagnostic Outputs	Magnitude de descarga (P), discharge phase (Ø), 3D PRPD patterns, PRPS pulse sequences, maximum amplitude, average quantity, frequência de descarga

Por que 200 MS/s Sampling Rate Matters

Partial discharge pulses are extremely fast transient events, often lasting only nanoseconds. A sampling rate of 200 MS/s — equivalent to a 5-nanosecond sampling interval — ensures that the host captures the full waveform of each discharge pulse without aliasing or distortion. This waveform fidelity is essential for accurate PRPD pattern construction and for distinguishing genuine PD pulses from noise artefacts. Lower sampling rates may miss critical waveform features, leading to misclassification or missed detections.

8. Como os padrões 3D PRPD e as sequências de pulso PRPS identificam os tipos de descarga?

Raw PD data — pulse counts, amplitudes, and timestamps — becomes truly diagnostic when it is visualised through Descarga parcial resolvida por fase (PRPD) patterns and Sequência de pulso resolvida por fase (PRPS) exibições.

PRPD — The Fingerprint of Discharge

A PRPD pattern plots discharge magnitude (vertical axis) against the phase angle of the power-frequency cycle (horizontal axis), accumulated over many cycles to build a three-dimensional density map. Different PD types produce distinctly different PRPD shapes. Corona discharge typically appears as clusters concentrated near the voltage peaks on one polarity. Internal void discharge produces symmetrical patterns on both positive and negative half-cycles, with the discharge magnitude remaining relatively constant. Surface discharge shows asymmetric, spreading patterns that increase in magnitude with applied voltage. Floating-potential discharge creates dense, high-amplitude clusters that shift in phase as the floating voltage changes.

By comparing a measured PRPD pattern against an expert database of known discharge signatures, o software de monitoramento pode classificar automaticamente o tipo de DP e avaliar sua gravidade – transformando um fenômeno eletromagnético complexo em uma recomendação de manutenção acionável.

PRPS — Acompanhamento da evolução da descarga ao longo do tempo

Embora o PRPD forneça um instantâneo cumulativo, PRPS exibe pulsos individuais em sequência, preservando a relação de tempo entre eventos de descarga consecutivos. Isto é particularmente valioso para detectar atividade intermitente de DP, observando como os padrões de descarga evoluem sob condições variáveis ​​de carga ou temperatura, e distinguir entre múltiplas fontes simultâneas de PD. PRPS data also supports advanced statistical analysis — such as pulse interval distributions and clustering algorithms — that can reveal degradation trends before they are visible in the PRPD pattern alone.

9. Software de monitoramento de back-end – recursos e capacidades de diagnóstico

The backend software platform transforms the monitoring host’s raw output into a decision-support tool for operators and asset managers. Installed on a control room workstation or accessible via a web interface, it provides four core functional modules.

Real-Time Monitoring and Visualisation

The system continuously acquires and displays live PD data, including 3D PRPD spectrum maps, PRPS pulse sequences, discharge amplitude bar charts, and trend lines for key parameters such as maximum discharge magnitude, average discharge quantity, and discharge repetition rate. Operators can view individual channel data or an aggregated system-level summary.

Consulta histórica e tendências

Todos os dados de medição são armazenados com carimbos de data/hora, permitindo que os engenheiros consultem registros históricos por intervalo de datas, canal, ou evento de alarme. Ferramentas estatísticas de tendências revelam trajetórias de degradação do isolamento a longo prazo, variações sazonais, e comportamento PD correlacionado com carga. Algoritmos de previsão de tendências suportam agendamento de manutenção preditiva.

Gerenciamento de Alarmes

Limites de alarme multinível — normalmente informativos, aviso, e crítico — pode ser configurado para cada parâmetro monitorado. Quando um limite é excedido, o sistema gera alertas visuais no painel e transmite notificações por e-mail, SMS, ou saída de relé. Os eventos de alarme são registrados com contexto completo (carimbo de data/hora, canal, valor do parâmetro, Instantâneo PRPD) para análise pós-evento.

Diagnóstico Inteligente

O software inclui um banco de dados de padrões especializados integrado que mapeia assinaturas PRPD e PRPS para tipos de descarga conhecidos. When new data matches a stored pattern, the system suggests the most probable PD type and recommended action. This reduces dependence on manual expert interpretation and accelerates the decision-making process, particularly for utilities managing large transformer fleets.

10. How Does a PD Monitoring System Integrate with SCADA and Asset Management Platforms?

Partial discharge data delivers maximum value when it is embedded in the utility’s wider operational data ecosystem rather than confined to a standalone display. A well-designed PD monitoring system supports this integration through standard industrial communication interfaces and protocols.

At the substation level, the PD monitoring host connects to the station RTU (Unidade Terminal Remota) or bay controller via Ethernet RJ45 ou RS-485. Standard protocols — including Modbus RTU/TCP, CEI 61850, e DNP3 — ensure compatibility with virtually any substation automation architecture. Key data points transmitted to SCADA include real-time PD amplitude values, sinalizadores de status de alarme, and diagnostic summary codes. Dispatchers can configure high-priority alarms for critical PD events — such as sudden acetylene-type UHF signatures or rapidly increasing discharge rates — ensuring immediate visibility on the SCADA overview screen.

Correlation with Other Monitoring Parameters

The greatest diagnostic insight comes from correlating PD data with complementary transformer health parameters. When the PD monitoring system feeds data into an integrated plataforma de monitoramento de transformadores alongside dissolved gas analysis (DGA), fibre optic winding temperature, bushing capacitance and tan-delta, and on-load tap changer condition data, the platform can perform automated cross-parameter analysis. Por exemplo, a simultaneous increase in UHF PD activity and a rise in hydrogen concentration in the oil provides much stronger confirmation of an active internal insulation fault than either indicator alone. This multi-parameter correlation approach significantly reduces diagnostic uncertainty and supports more confident maintenance decision-making.

11. Which Transformers Benefit Most from Online Partial Discharge Monitoring?

While any oil-filled or dry-type transformer can experience partial discharge, the investment in continuous online monitoring is best directed at assets where the consequences of an undetected insulation fault are most severe.

Highest-Priority Applications

Transmission-voltage power transformers (≥110 kV) at utility substations are the primary candidates, já que sua falha causa interrupções generalizadas e os prazos de substituição podem exceder doze meses. Generator step-up (UGS) transformers at thermal, hidro, and nuclear power plants are equally critical because an unplanned trip directly removes generation capacity from the grid. Large industrial transformers serving petrochemical complexes, semiconductor fabrication plants, centros de dados, and steel mills also justify online PD monitoring due to the enormous cost of production downtime.

Growing Adoption Scenarios

A expansão das energias renováveis ​​criou nova procura. Coletores e transformadores de interligação em parques eólicos e fazendas solares experience highly variable loading profiles and are often in remote locations where periodic manual testing is expensive and infrequent. Transformadores de potência de tração para eletrificação ferroviária systems carry safety-critical loads. Ageing transformers operating beyond their original design life are another strong candidate — continuous PD trending supports evidence-based lifetime extension decisions. Alta tensão comutador, SIG (painel de distribuição isolado a gás), e power cable systems are also increasingly equipped with online PD monitoring, using the same sensor technologies adapted for their specific enclosure geometries.

12. How to Select the Right Partial Discharge Monitoring Equipment — A Buyer’s Guide

The market offers a range of PD monitoring products, from single-sensor screening devices to full multi-sensor diagnostic platforms. The following criteria will help buyers match the right equipment to their specific application requirements.

Sensor Coverage and Fusion Capability

For comprehensive diagnostics on critical transformers, specify a system that supports all three sensor types — ultrasonic, TCFC, and UHF — with true multi-channel data fusion. Single-sensor systems (por exemplo, UHF-only or acoustic-only) are suitable for basic screening but cannot provide the cross-verification and source localisation capabilities that multi-sensor fusion delivers.

Sampling Rate and Dynamic Range

A sampling rate of at least 200 MS/s ensures that fast PD transients are captured without loss of waveform detail. The dynamic range should be wide enough — at least -80 para -20 dBm — to handle both very small incipient discharges and large discharge events without saturation or signal clipping.

Channel Count and Scalability

Evaluate whether four channels suffice for the intended transformer or whether six channels are needed to accommodate additional sensor positions. Systems with configurable channel options provide flexibility for both initial deployment and future expansion.

Diagnostic Software Quality

The software should include 3D PRPD pattern display, PRPS visualisation, an expert pattern database for automated PD type classification, multi-level alarm management, and historical trend analysis with forecasting. Web-based or remote-access capability is increasingly expected for fleet-wide management.

Communication Protocol Compatibility

Ensure the monitoring host supports the communication protocol already in use at your substation — Modbus RTU, Modbus TCP, CEI 61850, ou DNP3. Native protocol support avoids the cost and complexity of adding external protocol converters.

Environmental Rating and Sensor Durability

Sensors must be rated IP68 for outdoor installation and specified for the full operating temperature range of the site. Sensor mounting methods — magnetic, clamp-on, and plug-in — should require no modification to the transformer structure and no outage for installation.

Vendor Support and Expert Database Updates

PD pattern recognition accuracy depends on the quality and breadth of the expert database. Choose a vendor that provides regular database updates incorporating new discharge patterns and diagnostic refinements as field experience accumulates across their installed base.

13. Applicable International Standards for Partial Discharge Testing and Monitoring

Several international standards govern partial discharge measurement, interpretação, e desempenho do equipamento. Understanding these references helps buyers write better procurement specifications and ensures that the selected monitoring system meets globally accepted benchmarks.

CEI 60270 (High-Voltage Test Techniques — Partial Discharge Measurements) is the foundational standard for electrical PD measurement. It defines the apparent charge method, calibration procedures, and test circuit configurations. While primarily intended for offline factory testing, its measurement principles underpin many online system designs.

CEI 62478 (High-Voltage Test Techniques — Measurement of Partial Discharges by Electromagnetic and Acoustic Methods) extends the standard framework to cover UHF and acoustic detection techniques, providing guidance on sensor specifications, processamento de sinal, and data presentation for non-conventional PD measurement methods used in online monitoring.

IEEE C57.127 (Guide for the Detection, Localização, and Interpretation of Sources of Acoustic Emissions from Electrical Discharges in Power Transformers and Reactors) focuses specifically on acoustic PD detection in transformers, covering sensor placement, signal interpretation, and source localisation techniques.

Referências adicionais incluem Folheto Técnico CIGRE 676 (Partial Discharges in Transformers) which provides comprehensive guidance on PD phenomena, measurement techniques, and interpretation strategies, e CEI 61850 which defines the communication standard for substation automation and governs how PD monitoring data is exchanged with SCADA and asset management systems.

14. Perguntas frequentes (Perguntas frequentes)

---

Isenção de responsabilidade: As informações fornecidas neste artigo são apenas para fins educacionais gerais e de referência. FJINNO (www.fjinno.net) não oferece garantias, expresso ou implícito, em relação à completude, precisão, ou aplicabilidade do conteúdo a qualquer projeto ou instalação específica. As especificações técnicas aqui mencionadas representam valores típicos e podem variar dependendo do tipo de transformador, colocação do sensor, e ambiente do local. Engineering decisions should always be based on site-specific assessments conducted by qualified professionals in accordance with applicable standards including IEC 60270, CEI 62478, IEEE C57.127, and local grid codes. Nomes de produtos de fabricantes terceiros são marcas registradas de seus respectivos proprietários e são mencionados apenas para referência informativa. A FJINNO não será responsável por qualquer perda ou dano decorrente do uso ou confiança nesta informação.

Sensor de temperatura de fibra óptica, Sistema de monitoramento inteligente, Fabricante distribuído de fibra óptica na China