In modern electrified railway systems, the increasing demand for energy efficiency and power quality has driven the development of advanced power conditioning devices. Traditional railway energy routers (RERs) based on back-to-back inverter structures require two sets of transformers and inverters, leading to high costs and limited practicality. We propose an innovative single phase inverter-based railway energy router (IBI-RER) that bridges the α and β phase traction buses while integrating photovoltaic (PV) and energy storage systems (ESS) on the DC side. This structure fundamentally reduces hardware requirements while maintaining full functionality in regenerative braking energy recovery, renewable energy utilization, and power quality improvement.
The core innovation lies in utilizing a single phase inverter to create a three-port system between the DC bus and two AC traction buses. Unlike conventional two-port inverters, this configuration enables multi-directional power flow among all three ports. However, this introduces unique challenges in power coupling and control, which we address through a hybrid decoupling approach and multi-layer optimization strategy.
System Configuration and Operational Principles
The proposed IBI-RER topology consists of multiple parallel single phase inverter units connected through a multi-winding step-down transformer to both traction buses. The DC side integrates PV generation and battery storage through DC/DC converters. Additional reactive power compensation devices (thyristor-switched capacitors and static var generators) are installed on each traction bus to handle reactive power demands.
The mathematical foundation of the system begins with the voltage relationships between the two traction buses. For a three-phase system with V/V transformer connection, the α and β phase voltages exhibit a 60° phase difference:
$$V_\alpha(t) = \sqrt{2}V_m\sin\left(2\pi ft – \frac{\pi}{6}\right)$$
$$V_\beta(t) = \sqrt{2}V_m\sin\left(2\pi ft – \frac{\pi}{2}\right)$$
The voltage between the two phases creates a stable sinusoidal voltage:
$$V_{\alpha\beta}(t) = \sqrt{2}V_m\sin\left(2\pi ft + \frac{\pi}{6}\right)$$
The current injected by the single phase inverter can be expressed as:
$$i_{\text{RER}}(t) = \sqrt{2}I_{\text{RER}}\sin\left(2\pi ft – \frac{\pi}{6} – \phi\right)$$
where $I_{\text{RER}}$ is the RMS current and $\phi$ is the phase angle difference between current and voltage. The power injection into each traction bus contains both active and reactive components:
$$P_{c\alpha} = V_m I_{\text{RER}} \cos\phi$$
$$P_{c\beta} = V_m I_{\text{RER}} \cos\left(\phi – \frac{\pi}{3}\right)$$
$$Q_{c\alpha} = V_m I_{\text{RER}} \sin\phi$$
$$Q_{c\beta} = V_m I_{\text{RER}} \sin\left(\phi – \frac{\pi}{3}\right)$$
These equations reveal that the single phase inverter inherently generates reactive power while transferring active power between ports, creating a fundamental coupling challenge.
Energy Transfer Modes and Power Flow Analysis
The three-port nature of the single phase inverter enables four distinct operational modes, each with specific power flow characteristics:
| Operation Mode | Power Condition | ESS Status | Power Flow Direction |
|---|---|---|---|
| Peak Shaving | $P_L – P_{PV} \geq P_H$ | Discharging ($P_{\text{ESS}} > 0$) | DC → α,β phases |
| Regenerative Braking | $P_L – P_{PV} < 0$ | Charging ($P_{\text{ESS}} < 0$) | α,β phases → DC |
| Power Transfer | $P_{\text{low}} \leq P_L – P_{PV} < P_H$ | Idle ($P_{\text{ESS}} \approx 0$) | α ↔ β phase transfer |
| Valley Filling | $0 \leq P_L – P_{PV} < P_{\text{low}}$ | Charging ($P_{\text{ESS}} < 0$) | Grid → ESS |
where $P_L = P_{L\alpha} + P_{L\beta}$ represents total traction load, $P_{PV}$ is PV output, $P_H$ and $P_{\text{low}}$ are preset peak and valley thresholds.
The single phase inverter operates in four quadrants, enabling flexible power transfer based on the phase angle $\phi$:
- When $\phi \in (-30^\circ, 90^\circ)$: Power flows from β phase and DC to α phase
- When $\phi \in (-90^\circ, -30^\circ)$: Power flows from DC to both α and β phases
- When $\phi \in (-180^\circ, -90^\circ) \cup (150^\circ, 180^\circ)$: Power flows from α phase and DC to β phase
- When $\phi \in (90^\circ, 150^\circ)$: Power flows from both α and β phases to DC
Multi-Layer Decoupling Control Strategy
The control system addresses two critical coupling issues: the three-port power coupling and the active-reactive power coupling inherent in the single phase inverter structure. Our approach combines hardware decoupling through reactive power compensation devices with software-based optimization.
Hardware Decoupling with Reactive Power Compensation
The total reactive power requiring compensation includes three components:
$$Q_{\alpha} = Q_{L\alpha} + Q_{V\alpha} – Q_{c\alpha}$$
$$Q_{\beta} = Q_{L\beta} + Q_{V\beta} – Q_{c\beta}$$
where $Q_{Lj}$ represents load reactive power, $Q_{Vj}$ accounts for V/V transformer-induced reactive power, and $Q_{cj}$ is the reactive power generated by the single phase inverter. The transformer-induced reactive power can be calculated as:
$$Q_{V\alpha} = \frac{P_L – P_{PV} – P_{\text{ESS}}}{2} \tan\left(\frac{\pi}{6}\right)$$
$$Q_{V\beta} = -\frac{P_L – P_{PV} – P_{\text{ESS}}}{2} \tan\left(\frac{\pi}{6}\right)$$
We employ a hybrid compensation system using thyristor-switched capacitors (TSC) for bulk reactive power compensation and static var generators (SVG) for precise adjustment. For L-level TSC configuration, the optimal capacity distribution follows:
$$Q_{\text{TSC}_{Lj}} = 2^{j-1}Q_{\text{TSC}_{1j}},\quad j=1,2,\ldots,M$$
$$Q_{\text{SVG}_j} = Q_j – (2^M – 1)Q_{\text{TSC}_{1j}}$$
This configuration minimizes the SVG capacity requirement, reducing overall system cost.
Multi-Layer Optimization Control
The control strategy employs three hierarchical layers:
| Control Layer | Function | Output Variables |
|---|---|---|
| Energy Management | Determine operational mode and ESS power reference | $P_{\text{ESS}}^{\text{ref}}$ |
| Capacity-Constrained Optimization | Optimize power references considering inverter limits | $P_{\text{RER}\alpha}, P_{\text{RER}\beta}, Q_{\text{RER}\alpha}, Q_{\text{RER}\beta}$ |
| Converter Coordination | Execute current tracking and power distribution | Modulation signals |
The energy management strategy follows these principles:
- Regenerative braking energy priority: First utilized by trains in traction mode on the same feeder, then transferred to opposite feeder, finally stored in ESS if available
- PV energy priority: Directly consumed by traction loads, excess energy stored in ESS
- Peak shaving and valley filling based on net load and ESS state of charge (SOC)
The power optimization model ensures the single phase inverter operates within its capacity limits:
$$\min f = |P_{\text{RER}\alpha} – P_{c\alpha}| + |P_{\text{RER}\beta} – P_{c\beta}|$$
$$\text{subject to: } P_{\text{RER}\alpha}^2 + Q_{\text{RER}\alpha}^2 \leq S_{\text{max}}^2$$
$$P_{\text{RER}\beta}^2 + Q_{\text{RER}\beta}^2 \leq S_{\text{max}}^2$$
We solve this nonlinear optimization problem using particle swarm optimization, obtaining feasible power references that respect the coupling relationships between the three ports.
Current Reference Generation and Tracking
The current reference for the single phase inverter is derived from the optimized power references:
$$I_{\text{RER}} = \frac{\sqrt{P_{c\alpha}^2 + P_{c\beta}^2 – P_{c\alpha}P_{c\beta}}}{V_m}$$
$$\phi = \arctan\left(\frac{\sqrt{3}P_{c\beta}}{2P_{c\alpha} – P_{c\beta}}\right)$$
For current tracking, we employ proportional-resonant (PR) controllers in the stationary reference frame, which provide zero steady-state error for sinusoidal signals. The ESS converter operates in voltage-controlled mode to maintain DC-link voltage stability, while PV converters implement maximum power point tracking (MPPT).
Performance Evaluation and Comparative Analysis
We validate the proposed system through comprehensive simulation and practical data analysis. The system parameters used for evaluation are:
| Component | Parameter | Value |
|---|---|---|
| Single Phase Inverter | Transformer ratio | 27.5/2 kV |
| Capacity | 6.4 MVA | |
| DC voltage | 4000 V | |
| Energy Storage | Technology | Lithium Titanate |
| Capacity | 500 kWh | |
| PV System | Capacity | 2 MW |
| Reactive Compensation | TSC (α/β phase) | 6.9/6 Mvar |
| SVG (α/β phase) | 1/1 Mvar |
Simulation Results under Typical Operating Conditions
The single phase inverter successfully manages power transfer in all operational modes. During peak shaving (both feeders in traction mode), the system reduces grid power consumption from 18 MW to 14.2 MW while fully utilizing PV generation. In regenerative braking mode, the single phase inverter recovers 3 MW of braking energy while transferring 2 MW between phases. The DC-link voltage maintains stability with less than 3% fluctuation during mode transitions.
Power quality improvements are significant. The three-phase voltage unbalance decreases from 2.52% to 1.58%, well below the 2% limit in power quality standards. The average power factor improves from 0.688 to 0.966, eliminating potential penalty charges.
The mathematical relationship between the transferred active power and the resulting reactive power in the single phase inverter follows:
$$Q_{c\alpha} = P_{c\alpha} \tan\phi$$
$$Q_{c\beta} = P_{c\beta} \tan\left(\phi – \frac{\pi}{3}\right)$$
This confirms the inherent active-reactive power coupling that necessitates the decoupling compensation system.
Economic Analysis Based on 24-Hour Field Data
Using actual traction load data from a typical day with 1 MW PV installation, the system demonstrates substantial economic benefits:
| Performance Metric | Before IBI-RER | After IBI-RER | Improvement |
|---|---|---|---|
| Daily Energy Consumption | 328 MWh | 296.75 MWh | 9.53% reduction |
| Regenerative Energy Recovery | 0% | 61.65% | 61.65% recovery |
| PV Utilization | N/A | 94.16% | 94.16% utilization |
| Voltage Unbalance (95%) | 2.52% | 1.58% | 37.3% reduction |
| Average Power Factor | 0.688 | 0.966 | 40.4% improvement |
The economic advantage of the single phase inverter-based structure becomes evident in equipment cost comparison:
| System Configuration | Inverter Capacity | Transformer Capacity | Total Cost |
|---|---|---|---|
| Conventional Back-to-Back RER | 26 MVA (2×13 MVA) | 26 MVA (2×13 MVA) | 9.1 million USD |
| Proposed Single Phase Inverter RER | 15 MVA | 15 MVA | 5.71 million USD |
| Reduction | 42.31% | 42.31% | 37.23% |
The capacity reduction in the single phase inverter system stems from the efficient three-port power transfer capability, eliminating the need for duplicate conversion stages. The reactive compensation system further optimizes the utilization of the single phase inverter by offloading reactive power burden to dedicated compensation devices.
Conclusion
We have presented a novel railway energy router based on a single phase inverter structure that bridges two traction phases while integrating PV and energy storage. The key innovation lies in transforming the conventional two-port single phase inverter into a three-port system, enabling direct power transfer between α phase, β phase, and DC bus. This approach reduces main equipment capacity by 42.31% and overall system cost by 37.23% while maintaining equivalent functionality in regenerative braking recovery, renewable energy utilization, and power quality improvement.
The multi-layer decoupling control strategy effectively addresses the inherent power coupling challenges through hybrid reactive compensation and optimized power reference generation. Practical validation confirms excellent performance in voltage unbalance mitigation (reduced to 1.58%), power factor correction (improved to 0.966), and energy efficiency (61.65% regenerative energy recovery, 94.16% PV utilization).
The single phase inverter-based architecture represents a significant advancement in railway power conditioning technology, offering a cost-effective solution for modern electrified railway systems pursuing sustainability and power quality objectives. Future work will focus on reliability optimization and practical implementation challenges in actual railway environments.