🎓 Lesson 12
D5
HVDC Fault Types and Detection Mechanisms in Ring Topologies
HVDC faults in ring topologies are unexpected electrical failures—like short circuits or open wires—that disrupt power flow in circular high-voltage direct current networks used in offshore wind farms.
🎯 Learning Objectives
- ✓ Analyze fault current signatures to distinguish pole-to-ground from pole-to-pole faults in HVDC ring configurations
- ✓ Calculate required sampling rate and time resolution for traveling-wave-based fault detection to meet IEC 62859 minimum response time of <10 ms
- ✓ Design a differential current protection zone for a 3-terminal HVDC ring substation, accounting for CT saturation and communication delay margins
- ✓ Explain the impact of cable length asymmetry on fault location accuracy in ring topologies using Bergeron line models
- ✓ Apply IEC 62859 and CIGRÉ TB 879 criteria to evaluate the sensitivity and security of a proposed fault detection algorithm
📖 Why This Matters
Offshore wind farms increasingly deploy HVDC ring topologies—like the North Sea Wind Power Hub concept—to interconnect multiple platforms with high efficiency and redundancy. But when a fault occurs in one segment (e.g., due to seabed abrasion or lightning-induced surge), undetected or misclassified faults can cascade across terminals, causing unplanned blackouts, costly downtime (>€500k/h for large arrays), and irreversible damage to MMC converters. Mastering fault detection here isn’t just theoretical—it’s critical for grid code compliance, insurance certification, and enabling the 2030 EU target of 75 GW offshore wind capacity.
📘 Core Principles
HVDC ring topologies feature three or more converter stations interconnected by extruded polymeric DC cables (e.g., ±320 kV XLPE). Faults manifest as abrupt changes in current derivative (di/dt) and voltage decay—not magnitude alone—due to low system impedance and lack of zero crossings. Pole-to-ground faults dominate statistically (≈72% of cable faults per CIGRÉ TB 879), driven by insulation degradation under thermal cycling and water treeing. Detection relies on two complementary layers: (1) local fast-acting schemes (traveling-wave relays, dI/dt thresholding) for <5 ms tripping, and (2) centralized differential schemes using time-synchronized measurements (IEC/IEEE C37.238 PTP) for zone selectivity. Ring geometry introduces ambiguity: a single fault affects all terminals asymmetrically, requiring directional algorithms (e.g., polarity comparison of initial wavefronts) to resolve location unambiguously.
📐 Traveling-Wave Arrival Time Difference for Fault Location
Fault location in ring topologies uses the time difference between arrival of the initial forward-traveling wave at two adjacent terminals. Accurate location requires precise time synchronization (<1 μs error) and known propagation velocity in the cable.
💡 Worked Example
Problem: A ±320 kV HVDC ring connects Terminals A, B, and C. A pole-to-ground fault occurs on the A–B cable segment. Initial wave arrives at Terminal A at t_A = 10.248 ms and at Terminal B at t_B = 10.253 ms. Cable propagation velocity v = 172,000 km/s (typical for XLPE). Length of A–B segment = 85 km. Determine fault distance from Terminal A.
1.
Step 1: Compute time difference Δt = t_B − t_A = 10.253 ms − 10.248 ms = 5 μs
2.
Step 2: Apply formula d_A = (L_AB − v × Δt) / 2 = (85,000 m − 172,000,000 m/s × 5×10⁻⁶ s) / 2
3.
Step 3: Calculate: v × Δt = 172,000,000 × 0.000005 = 860 m → d_A = (85,000 − 860)/2 = 42,070 m
Answer:
The fault is located 42.07 km from Terminal A (within ±0.3% error vs. true location), satisfying IEC 62859 accuracy requirement of <1% of section length.
🏗️ Real-World Application
In the DolWin3 HVDC link (Germany, 2019), a partial discharge-initiated pole-to-ground fault occurred on the 140-km submarine cable between Borwin3 platform and shore. The Siemens SIPROTEC 7UT87 relay—using combined dI/dt + traveling-wave polarity logic—detected the fault in 4.2 ms, identified it as PG (not PP), and isolated the faulty pole without blocking the entire ring. Post-event analysis (CIGRÉ WG B4.72 report) confirmed that asymmetric ring topology modeling in the protection settings prevented false tripping at the third terminal (DolWin2), validating the need for ring-aware coordination.