Abstract
The integration of distributed energy resources like photovoltaics and wind power, alongside a growing proportion of DC loads, necessitates a shift in distribution network architecture. This paper explores the use of multi-port Power Electronic Transformers (PET) to facilitate efficient AC/DC hybrid power supply. We propose a hierarchical, multi-time-scale control framework and a reliability evaluation method based on sequential Monte Carlo simulation. Our findings demonstrate that PET clusters significantly enhance renewable energy utilization and system reliability.
Introduction
Modern distribution grids are evolving to accommodate a surge in distributed renewable energy and DC-based loads, such as electric vehicle charging stations and data centers. Traditional AC networks struggle with the efficient integration of these diverse sources and loads. AC/DC hybrid networks utilizing multi-port Power Electronic Transformers (PETs) offer a promising solution but introduce significant control complexity and reliability concerns.
Managing bidirectional power flows across multiple voltage levels in real-time is computationally demanding. Furthermore, the perceived fragility of power electronics compared to traditional transformers necessitates robust reliability modeling. This research addresses these gaps by introducing a structured management framework and a comprehensive reliability assessment for PET-based hybrid networks.
Multi-Port PET and Hierarchical Control
The core of the proposed system is the multi-port PET, which serves as the interface between various AC and DC ports.
To manage this complexity, a three-layer hierarchical control architecture is implemented:
- Device Level (First Layer): Handles millisecond-level local control, such as bus voltage regulation, ensuring immediate stability.
- PET Cluster Level (Second Layer): Coordinates multiple PETs within a cluster on a seconds-level timescale. This enables "hot standby" functionality, where healthy PETs can compensate for a failed unit.
- Energy Management Level (Third Layer): Performs system-wide optimization on a minutes-level timescale, scheduling power based on load forecasts and renewable generation availability.
Reliability Evaluation via Sequential Monte Carlo
Quantifying the reliability of such complex systems requires advanced simulation techniques. The authors employ a Sequential Monte Carlo (SMC) method, which is better suited for high-dimensional hybrid systems than traditional analytical methods. Internal PET components—including voltage source converters (VSC), dual active bridges (DAB), and high-frequency transformers (HFT)—are modeled in a series-parallel reliability block diagram. This allows for the calculation of system-wide failure rates and repair times over extended operational windows (e.g., 100,000 hours).
Key Insights and Results
The study validated its approach using a modified IEEE 33-node distribution network. Several critical insights emerged:
- The "Cluster" Advantage: Networking PETs into clusters provides "N-1" redundancy. This structural resilience drastically reduces the probability of total blackouts during component failures.
- Renewable Utilization: Implementing bidirectional power flow through PETs increased renewable energy utilization from 51.7% to 74.1% in the test case.
- Port Capacity Sensitivity: System reliability is highly sensitive to PET port capacity. Sizing ports too close to peak loads limits the system's ability to transfer loads during faults, negatively impacting indices like the System Average Interruption Duration Index (SAIDI).
- Reliability Indices: As PET port capacity increased from 50% to 100% of load requirements, the Average Service Availability Index (ASAI) improved from 0.984 to 0.990.
Conclusion and Future Directions
This research demonstrates that a hierarchical control framework combined with PET clusters can effectively manage the complexities of AC/DC hybrid distribution networks. While the technical benefits for reliability and renewable integration are clear, future work should address component aging through "physics-of-failure" models and explore the system's resilience against communication failures or cyber-attacks. Additionally, a detailed economic analysis is needed to weigh the capital costs of multi-port PETs against long-term operational gains.
Summary updated