Common Failure Modes and Root Cause Analysis of Metal-Oxide Surge Arresters

A surge arrester that fails silently during an overvoltage event leaves your entire substation exposed to catastrophic damage.

1. Recognizing the Warning Signs: How Arrester Failures Manifest

Metal-oxide surge arresters (MOSAs) are designed to operate reliably for 20–30 years, but when failure begins, the warning signs are often subtle and progressive rather than sudden. Understanding how failures manifest — and what early indicators to watch for — is the first line of defense against catastrophic substation damage.

The Silent Nature of Arrester Degradation

Unlike circuit breakers that trip visibly or fuses that blow, a degrading surge arrester often shows no immediate external signs. The ZnO varistor blocks inside the housing deteriorate gradually, with the resistive component of leakage current increasing exponentially over time. By the time visible symptoms appear — housing discoloration, cracked porcelain, or tracking marks on polymer surfaces — the arrester may already be in an advanced state of failure.

The resistive leakage current (I_r) is the most sensitive indicator of ZnO block health. In a new arrester, I_r is typically 0.5–2 mA. As the arrester ages due to electrical stress, thermal cycling, and moisture, this current increases. The degradation follows a characteristic four-phase curve:

Figure 1: Typical resistive leakage current degradation curve over the service life of a metal-oxide surge arrester, showing four distinct phases and monitoring thresholds.

During Phase 1 (years 0–2), the resistive current may actually decrease slightly as the ZnO blocks undergo initial stabilization. Phase 2 (years 3–15) shows a very slow, nearly imperceptible increase. In Phase 3 (years 15–25), the rate of increase becomes noticeable — this is where condition monitoring becomes critical. Phase 4 represents imminent failure, where I_r rises rapidly toward thermal runaway.

Physical and Electrical Symptoms by Failure Mode

Different failure modes produce different symptom patterns. Recognizing these patterns allows maintenance teams to identify the root cause before catastrophic failure occurs:

2. Root Causes: What Actually Drives Metal-Oxide Arrester Failure

Arrester failures are never caused by a single factor in isolation. The root causes interact across six categories: electrical stress, thermal stress, environmental factors, manufacturing defects, installation errors, and aging. The fishbone diagram below maps these cause categories to their specific failure mechanisms.

Figure 2: Fishbone (Ishikawa) diagram categorizing the root causes of metal-oxide surge arrester failures into six primary categories with detailed sub-causes.

Thermal Runaway — The Most Catastrophic Mode

Thermal runaway arises from the fundamental nonlinear V-I characteristics of ZnO varistor blocks. Under normal MCOV, the arrester exhibits very high resistance with only a small capacitive current (1–5 mA for distribution-class). However, the resistive component of leakage current increases exponentially with temperature. When sustained overvoltage, harmonic heating, or solar radiation raises the internal temperature, this increases the resistive current, which generates more heat. If heat generation exceeds heat dissipation, the temperature rises uncontrollably.

The mathematical condition for thermal stability is: dQ_gen/dT < dQ_diss/dT. Once violated, the arrester temperature can rise from 80°C to over 600°C within minutes, resulting in housing rupture, expulsion of hot gases, and potentially explosive failure.

Moisture Ingress — The Leading Long-Term Degradation Mechanism

Moisture ingress is the most common long-term degradation mechanism. The internal ZnO blocks and insulating spacers are highly sensitive to moisture. When the sealing system — O-rings, gaskets, sealing compounds, and hermetic terminal seals — deteriorates, water vapor penetrates the interior. This causes: (1) surface contamination of ZnO blocks causing non-uniform current distribution; (2) corrosion of internal metal components increasing contact resistance; (3) reduction of insulation resistance; and (4) in severe cases, conductive paths leading to internal flashover.

The primary entry points are degraded O-ring seals at terminal ends, cracks in porcelain housing (from vandalism or thermal shock), failed gaskets in the base assembly, and permeation through polymer housing materials exposed to long-term UV and pollution.

Manufacturing Defects — The Infant Mortality Driver

Manufacturing defects cause premature failure even when the arrester operates within specified ratings. Common defects include:

• ZnO Block Formulation Errors: Incorrect proportion of ZnO, Bi₂O₃, Sb₂O₃, CoO, MnO, and Cr₂O₃ leads to non-uniform grain boundary characteristics and localized overheating.
• Incomplete Sintering: Insufficient sintering temperature or time results in poor grain boundary formation, high leakage current, and accelerated aging.
• Internal Assembly Contamination: Dust, metal particles, or moisture introduced during assembly create partial discharge sites within the column stack.
• Improper Block Grading: When ZnO blocks in a column have significantly different V-I characteristics, blocks with lower breakdown voltage carry disproportionate current and fail prematurely.
• Sealing System Defects: Inadequate sealing compound application, improper O-ring compression, or improper curing compromises long-term seal integrity.

TOV and Energy Overstress

Temporary overvoltages (TOV) are sustained overvoltages lasting from cycles to seconds, caused by single-line-to-ground faults, load rejection, or ferroresonance. IEC 60099-4 requires that a 10 kA arrester with rated voltage of 90 kV withstand 1.45 pu for 10 seconds. Failure to coordinate TOV capability with system studies is a common cause of post-fault arrester failure.

Energy overstress occurs when switching surges or multiple lightning strikes exceed the arrester’s energy absorption capability (kJ/kV of rated voltage). This is particularly common in arresters protecting capacitor banks and in areas with high ground flash density. The ZnO blocks can crack, melt, or suffer puncture under such conditions.

3. Prevention and Mitigation: Proven Strategies to Stop Failures Before They Occur

Preventing arrester failures requires a multi-layered approach: proper specification and selection, stringent quality control at procurement, correct installation practices, and a robust condition monitoring program. Each layer addresses different root cause categories identified in the fishbone diagram.

Specification and Procurement Safeguards

The first defense against failure is selecting the correct arrester for the application. This means:

• TOV coordination: Conduct system temporary overvoltage studies before selecting the arrester rated voltage. A 10-second TOV rating is insufficient if the actual system TOV lasts 30 seconds.
• Energy class matching: Select the energy absorption class based on lightning flash density (for lines) and switching surge energy (for capacitor banks). Under-rated energy class leads to block cracking during high-energy surges.
• Factory acceptance testing (FAT): Require FAT for critical station-class arresters, including residual voltage measurement, reference voltage test, and partial discharge measurement. A batch of 72 kV arresters from a new supplier was found to have sodium and potassium contaminants in the ZnO blocks — only caught through laboratory FAT.
• Supplier quality audits: Audit the ZnO varistor production process — mixing, sintering, and block grading — to ensure consistent quality.

Installation Quality Control

Studies by major utilities have shown that up to 30% of surge arrester failures can be directly attributed to installation errors. Critical installation safeguards include:

• Maintaining lead length below 1.5 m for station-class and 0.5 m for distribution-class arresters to limit inductive voltage drop during surge discharge.
• Ensuring ground resistance below 10 Ω (below 5 Ω for substations with grounded neutral systems).
• Connecting the arrester to the same ground electrode as the protected equipment to avoid ground potential rise differences.
• Using calibrated torque wrenches for all connections — under-torquing causes overheating, over-torquing damages threads and contact surfaces.
• Performing pre-installation visual inspection to reject units with housing damage, cracked porcelain, or abraded polymer housings.

Condition Monitoring and Diagnostic Program

A robust condition monitoring program is the most effective way to detect incipient failure before it becomes catastrophic. The key elements are:

Real-World Case Lessons

A utility in North America reported a 10 kV distribution arrester failure rate of 12% over 3 years in a coastal service area. Forensic analysis revealed saturated internal desiccant and water-induced corrosion of ZnO blocks — the root cause was inadequate sealing at the terminal end. After the manufacturer improved the sealing system with a secondary O-ring and low-outgassing compound, the redesigned arresters showed zero failures over 5 subsequent years in the same environment.

In another case, a 132 kV polymer-housed arrester failed catastrophically after 8 years of service following a 1.6 pu TOV for 3 seconds — the arrester’s TOV capability was only 1.45 pu for 10 seconds. Maintenance logs showed the housing temperature had been rising for 3 months (42°C to 78°C) and resistive current had increased from 1.8 mA to 8.5 mA. Both warning signs were present but unmonitored.

4. Diagnostic Methods Compared: Thermovision vs. Leakage Current vs. Partial Discharge Testing

Three diagnostic methods dominate surge arrester condition assessment: infrared thermovision, leakage current measurement, and partial discharge testing. Each method detects different failure mechanisms at different stages, and a comprehensive monitoring program should employ at least two of these complementary techniques.

The optimal strategy is to use all three methods in combination: leakage current measurement for continuous online health tracking, thermovision for periodic confirmation of thermal behavior, and partial discharge testing for early detection of internal defects — especially valuable at commissioning and after major system events.

5. Summary

• Pain point: A surge arrester that fails silently during an overvoltage event leaves the entire substation exposed to catastrophic damage — and by the time visible symptoms appear, failure may already be advanced.
• Finding 1: Arrester degradation follows a predictable four-phase curve, with resistive leakage current as the most sensitive indicator. Monitoring I_r annually and acting on the 5 mA warning threshold can detect incipient failure years before thermal runaway.
• Finding 2: Six root cause categories — electrical, thermal, environmental, manufacturing, installation, and aging — interact to drive failures. Thermal runaway is the most catastrophic; moisture ingress is the most common long-term mechanism; manufacturing defects cause infant mortality.
• Finding 3: Prevention requires a multi-layered approach: TOV-coordinated specification, factory acceptance testing, correct installation (lead length, grounding, torque), and a structured condition monitoring program combining at least two diagnostic methods.
• Comparison conclusion: Leakage current measurement offers the best early-stage sensitivity and online capability; thermovision provides visual thermal confirmation; partial discharge testing detects internal defects earliest but requires specialized equipment. Using all three in combination provides the most comprehensive protection against unexpected failure.