Global PV capacity has maintained rapid year-on-year expansion, accompanied by surging safety risks posed by DC arc faults. Supported by mandatory regulatory rollouts worldwide, Arc Fault Circuit Interrupters (AFCI) for DC arc fault detection have evolved from an optional add-on to a compulsory standard component for all PV systems.

1.1 Global PV Installation Landscape

Global new PV capacity additions hit 597 GW in 2024, marking a 33% year-over-year growth. PV power accounts for over 75% of the world’s total renewable energy capacity. By the end of 2024, cumulative global PV installations exceeded 2 TW.

China added 329 GW of new PV capacity in 2024, representing 55% of global new installations and cementing its position as the world’s largest PV market. Distributed PV consistently makes up 25%–30% of China’s total installed capacity, with rooftop PV, Building-Integrated PV (BIPV), agrivoltaics and other on-site applications gaining rapid traction. These distributed systems are built adjacent to buildings and occupied spaces, creating elevated safety hazards for personnel.

Figure 1: PPT Global Photovoltaic Installed Capacity Data Chart

Three root causes amplify safety risks across PV fleets:

1. High DC Voltage & Dense Wiring Points: System DC voltages have risen to 600–1500 V. MW-scale power stations feature thousands of terminal connections, drastically increasing failure risks from loose contacts, cable aging and insulation degradation.

2. Power Upgrades Intensify Thermal Stress: PV module power ratings have jumped from 300 W to over 600 W. Per Joule’s Law , doubling current quadruples thermal generation.

3. Hard-to-Manage Distributed Assets: Residential and commercial rooftop PV systems are geographically dispersed, delaying fault troubleshooting and leaving hidden safety hazards unaddressed.

1.2 DC Arc Faults: The Primary Trigger of PV Fires

DC arcs generate extreme temperatures ranging from 3,000°C to 7,000°C, capable of igniting surrounding combustible materials within seconds. Industry statistics confirm 50%–80% of PV fire incidents originate from DC arc faults. German insurance industry data shows PV fire claims consistently rank first among all energy-related insurance payouts.

Figure 2: Schematic Diagram of PPT DC Arcing Hazards, Temperature, and Accident Proportion

1.3 Global Mandatory AFCI Regulatory Framework Takes Full Effect

Three dominant standard regimes have been established across North America, Europe and China, while Brazil, Australia, Japan and other regions accelerate corresponding legislation. Mandatory AFCI deployment is a unified global trend.

1. North America (United States)

NEC Article 690.11 mandates AFCI installation for all DC PV systems operating above 80 V, paired with UL 1699B and UL 1699C (AI-enhanced variant) certification requirements. The North American DC AFCI market reached an estimated USD 828 million in 2025, making it the world’s largest single regional market.

2. Europe

The IEC 63027 standard is enforced with unified regional specifications from CENELEC, alongside stringent TÜV and VDE certification protocols. AFCI devices are required to interlock with Rapid Shutdown Devices (RSD) to protect first responders, forming the world’s second-largest AFCI market. New PV capacity additions hit 82 GW in Europe in 2024.

3. China

China accounts for 55% of global PV installations. A draft notice on distributed PV safety was released in 2021, followed by a formal GB/T standard draft published in March 2025. The new regulation proposes mandatory AFCI fitment for all new distributed PV systems with capacity above 10 kW.

Figure 3: Comparison Chart of PPT Global North America / Europe / China AFCI Policy Standards

Overarching Trend: AFCIs are shifting from regional optional features to national legal mandates, with compliance coverage expanding from newly built plants to retrofits of existing fleets, driving massive demand for DC arc fault detection technology.

II. DC Arc Fault Technical Characteristics & Core Development Barriers: 5 Key Technical Challenges

DC arcs exhibit unique time-domain and frequency-domain signal signatures, compounded by severe electromagnetic interference in PV systems, creating multiple technical hurdles that differentiate competing product solutions.

2.1 Fundamental Arc Signal Characteristics

Frequency-Domain Traits

Typical arc current ranges from 2 A to 10 A, with spectral energy spanning 3 kHz up to several MHz, classified as pink noise following the power spectral density rule . The industry’s optimal detection frequency band sits at 30–100 kHz, requiring a minimum system sampling frequency of ≥200 kHz per the Nyquist sampling theorem.

Time-Domain Traits

DC arc current waveforms display two distinct markers: flat plateaus and abrupt sharp spikes, paired with shortened conduction angles and drastically elevated current standard deviation. Arc progression unfolds in three distinct phases: initial fluctuation, transitional combustion and steady-state burning. Lab testing conducted by Zhejiang University verifies three core time-domain metrics that reliably distinguish normal operating conditions from arc faults: current variance, peak-to-peak amplitude and high-frequency energy content.

Figure 4.1: The time-domain waveform of the arc exhibits characteristics of "flat shoulder + sharp peak", with a three-stage combustion process

Figure 4.2: PPT Normal/Arcing State Spectrum Data Table, Time Domain Waveform Feature Diagram

2.2 Arc Simulation Models & Fault Classification

Large-scale PV power stations (1–4 MW, 1500 V, kA-class) predominantly adopt the Cassie arc model. Distributed-parameter line models accurately replicate real-world field conditions with significant current fluctuation and random arc ignition behavior. Traditional lumped-parameter models output smoothed waveforms that discard critical high-frequency arc signatures, rendering them low-value for simulation validation.

DC arc faults are categorized into five types based on fault location: series busbar arcs, parallel busbar arcs, intra-string series arcs, intra-string parallel arcs and inter-string parallel arcs. Algorithms incorporating Empirical Wavelet Transform, multi-entropy feature fusion and GRU neural networks achieve fault identification accuracy of 93.6%–97.5% alongside precise fault location capability.

Figure 5: Comparison Table of PPT Lumped Parameter / Distributed Parameter Model Data

2.5 Five Core Industry Technical Pain Points

PV DC arc fault detection extends far beyond basic signal sampling, with universal industry bottlenecks stemming from complex interference sources:

1. Severe PWM Noise Interference (High Severity)

10–50 kHz inverter switching harmonics overlap heavily with arc frequency spectra; noise amplitude often exceeds arc signal magnitude, rendering conventional fixed-threshold algorithms completely ineffective.

2. Tradeoff Between Spectral Resolution & Processing Power

High sampling rates and multi-channel FFT analysis impose heavy MCU memory and computing load constraints, creating a conflict between high spectral resolution and real-time system responsiveness.

3. Weak Arc Detection Limitations (High Severity)

Low-power faint arcs produce minimal waveform distortion, with characteristic signatures easily masked by background system noise, resulting in poor detection performance from traditional algorithms.

4. Poor Cross-System Operating Compatibility

Arc frequency spectra vary drastically across different system voltage levels, current ratings and inverter topologies; fixed-parameter detection schemes fail to deliver consistent performance across diverse hardware platforms.

5. Long-Distance Signal Attenuation

DC cables exceeding 200 m in length suffer severe arc signal degradation and plummeting signal-to-noise ratio (SNR), imposing strict sensitivity requirements on sensing hardware.

Figure 6: Diagram Marking the Five Major Technical Difficulties and Interference Levels of PPT

III. In-Depth Comparison of Mainstream Detection Solutions: Advantages, Drawbacks & Application Scenarios of 3 Core Technical Routes

Three global mainstream technical architectures exist for PV DC arc fault detection, featuring fundamental differences in operating principle, cost, performance and suitable deployment environments:

3.1 TI Full-Spectrum Detection Solution

Technical Architecture: CT current transformer signal acquisition → op-amp signal amplification → 30–100 kHz bandpass filtering → 16-bit ADC sampling → FFT spectrum analysis executed on C2000 MCU

Core Components: TMS320F280049C, ADS8363, OPA322; next-generation F28P55x series integrates an on-chip NPU supporting edge-AI fusion detection

Advantages: Mature solution, robust development ecosystem, supports synchronous multi-channel detection, optimized for standardized mass production

Drawbacks: Fixed frequency band limits compatibility with all inverter architectures, weak performance detecting low-magnitude faint arcs

3.2 Magnetic Ring Sensing Solution (Dominant for Domestic Chinese Manufacturers)

Operating Principle: Non-intrusive detection via external magnetic ring coils coupled to high-frequency arc electromagnetic signals, requiring no modification to the primary power circuit

The domestic product lineup has evolved to Generation 3.0, integrating adaptive dynamic threshold algorithms and AI deep learning. Effective detection distance reaches 200 m with response time controlled under 500 ms.

Advantages: Simple mechanical structure, low BOM cost, mature local supply chain, straightforward field installation

Drawbacks: Inconsistent magnetic core material performance, significant temperature drift, reduced SNR after long-distance signal transmission

3.3 ST AI Cloud-Based Solution

Technical Architecture: Edge-side magnetic ring signal capture → local preprocessing via STM32 MCU → data upload to cloud for large model training → quantized model deployed back to edge hardware for inference, supporting remote OTA firmware upgrades. GAN technology generates synthetic arc datasets to resolve limited physical arc sample availability.

Advantages: Data-driven algorithm design with superior anti-interference capability, continuous model iteration, capable of identifying unknown arc spectral signatures

Drawbacks: Reliant on large volumes of labeled training data, edge MCU computing limits complex model deployment, recurring cloud platform operational costs

3.4 Comprehensive Comparative Summary of the Three Technical Routes

Figure 7: Comprehensive Comparison Chart of the Three Major Technical Routes of PPT

Industry Consensus: No single architecture delivers universal optimal performance. Hardware selection must align with product positioning, cost targets and circuit topology. The long-term industry direction leans toward hybrid detection schemes that combine time-domain, frequency-domain and AI algorithms to leverage complementary strengths.

IV. Industry Technology Trends & Unresolved Technical Challenges

4.1 Core Development Direction: Integrated Arc Fault & Current Sensing Modules

Current AFCI arc detection and current measurement rely on two independent hardware subsystems, driving up BOM costs and occupying excessive PCB footprint. The next-generation industry focus is integrated all-in-one sensing solutions that share sensors, signal chains and ADC hardware to deliver three core functions from a single sampling circuit: arc fault detection, current measurement and fault diagnostics.

Time-division multiplexing is the primary implementation architecture: high-frequency burst sampling handles arc fault monitoring, while low-rate steady-state sampling undertakes current metering. Paired with high-end MCUs featuring integrated NPUs for single-chip algorithm fusion, this design cuts BOM costs by 20%–30% while boosting overall system reliability.

The integrated solution is not merely physical hardware consolidation, but deep algorithmic integration: one set of sampled data simultaneously powers arc fault detection, current measurement and fault diagnosis to achieve maximum hardware integration efficiency.

Figure 8: Integrated Solution Architecture Evolution Diagram

4.2 Four Critical Unresolved Industry Challenges

1. Lack of Standardized Testing Protocols: Global standards only specify detection rate and false alarm rate benchmarks, with no unified arc generator hardware or standardized test workflows. Cross-manufacturer performance comparison is not feasible.

2. Insufficient Weak Arc Detection Accuracy: Detection performance degrades drastically under low-current, high-voltage and long cable length operating conditions.

3. Difficult False Alarm Mitigation: Normal operating events including cloud shading fluctuations, load transients and inverter startup/shutdown frequently trigger false alarms; the industry target false alarm rate is ≤0.1%.

4. Multi-Channel Parallel Computing Bottlenecks: String inverters feature multiple independent MPPT channels requiring simultaneous fault detection, imposing extreme parallel processing demands on MCUs and ADCs.

4.3 Long-Term Industry Outlook

Driven by global mandatory regulatory requirements, the AFCI market will maintain sustained growth, projected to exceed USD 15.1 billion by 2034. Three core technological development pillars will define the industry: hybrid multi-domain algorithms, hardware integration and edge artificial intelligence. DC arc fault detection will transition fundamentally from rule-based passive monitoring to data-driven intelligent safety protection.

V. MAGTRON Custom PV DC Arc Fault & Current Detection Solutions

Across the full PV DC arc detection industrial chain, product performance is constrained by four critical bottlenecks: magnetic core material stability, high-frequency weak signal acquisition, long-distance signal transmission and temperature drift suppression — core technical areas where MAGTRON has focused R&D investment for over a decade.

Drawing on decades of proprietary expertise in magnetic sensing, magnetic core material development, high-frequency electromagnetic coupling and precision signal conditioning, MAGTRON delivers fully customized end-to-end solutions addressing key industry pain points:

1. Optimized Magnetic Core Material Formulation: Eliminates widespread industry issues of inconsistent magnetic ring performance and severe temperature drift, ensuring stable operation across full high/low temperature operating ranges.

2. Enhanced Long-Distance Weak Signal Capture: Boosts signal coupling efficiency and SNR for DC cable runs exceeding 200 m, resolving long-distance detection limitations.

3. Universal Cross-System Compatibility Design: Sensing hardware optimized with combined time/frequency domain feature analysis to support all voltage classes and inverter topologies, with robust suppression of PWM switching noise interference.

4. Cost-Effective Hardware Architecture: Retains the magnetic ring solution’s key advantages — low BOM cost, non-intrusive installation and simple field fitting — upgraded with cutting-edge on-inverter AI algorithms and adaptive threshold logic.

We fully understand the R&D, mass production and compliance requirements of upstream and downstream manufacturers including PV inverter suppliers, combiner box OEMs, energy storage system integrators and distributed PV operators. PV safety is the fundamental foundation of sustainable industry growth, and DC arc fault detection stands as a critical component for regulatory compliance and reliable asset operation.

MAGTRON welcomes collaboration with all industry partners to jointly overcome technical barriers and deliver highly reliable, cost-competitive PV AFCI detection hardware for global markets.