In modern power systems, the demand for reliable and efficient uninterruptible power supply (UPS) systems has grown significantly, driven by the expansion of data centers, telecommunications, and critical industrial loads. Traditional UPS systems, often based on double-conversion topologies, suffer from lower efficiency due to multiple power conversion stages, leading to higher operational costs and energy consumption. To address these challenges, I propose a novel single-stage grid-formed UPS architecture integrated with a cell energy storage system. This system not only ensures stable power supply to critical loads but also actively participates in grid ancillary services, such as frequency regulation, thereby enhancing the utilization of the cell energy storage system and reducing overall operational expenses. In this article, I will detail the system architecture, operational modes, control strategies, and experimental validation, emphasizing the role of the cell energy storage system in improving performance and reliability.
The core of the proposed system is a single-stage power conversion unit that combines a battery-based cell energy storage system with a voltage-source inverter (VSI). This design eliminates the need for intermediate conversion stages, thereby increasing efficiency and reducing complexity. The system operates in three primary modes: dual-online mode, where both the grid and the cell energy storage system supply power concurrently; battery-independent mode, where the cell energy storage system solely powers the load during grid failures; and grid-supply mode, where the grid provides power while the UPS is under maintenance. These modes ensure seamless transitions and high reliability for critical loads. The integration of the cell energy storage system allows for active grid support, making it a versatile solution for modern power networks.
To quantify the advantages, consider the following table comparing traditional UPS systems with the proposed single-stage grid-formed UPS based on a cell energy storage system:
| Feature | Traditional Double-Conversion UPS | Proposed Single-Stage Grid-Formed UPS |
|---|---|---|
| Power Conversion Stages | Two (rectifier + inverter) | One (voltage-source inverter) |
| Efficiency | Typically 90-95% | Above 97% (estimated) |
| Grid Support Capability | Limited or none | Active frequency and voltage support |
| Battery Utilization | Passive standby | Active participation in grid services |
| Cost of Operation | Higher due to losses | Lower due to reduced losses and ancillary service revenue |
The operational modes are governed by a sophisticated control strategy that enables autonomous grid-forming, voltage-frequency support, and battery charging management. The control strategy is implemented in the VSI, which serves as the heart of the system. By leveraging the cell energy storage system, the UPS can dynamically adjust its output to maintain grid stability while ensuring uninterrupted power to loads. This dual functionality highlights the significance of the cell energy storage system in modern power infrastructure.
In the autonomous grid-forming control strategy, the VSI mimics the behavior of a synchronous generator to establish voltage and frequency independently. This is achieved through a virtual synchronous machine (VSM) model, where the inverter’s internal electromotive force (EMF) is generated based on power balance equations. The swing equation for the VSM is given by:
$$ J \frac{d\omega}{dt} = P_{ref} – P_e $$
where \( J \) is the virtual inertia, \( \omega \) is the angular frequency, \( P_{ref} \) is the reference active power, and \( P_e \) is the measured active power output. The phase angle \( \theta \) of the internal EMF is derived from the frequency deviation:
$$ \theta = \int \omega \, dt $$
To regulate the voltage amplitude, a voltage control loop is employed, producing the EMF magnitude \( E_m \):
$$ E_m = G(s) \left[ D_q (U_{ref} – U_o) – Q_e \right] $$
where \( G(s) \) is a voltage regulator (e.g., an integrator), \( D_q \) is the droop coefficient for reactive power, \( U_{ref} \) is the reference voltage, \( U_o \) is the output voltage, and \( Q_e \) is the reactive power. The three-phase EMF signals are then:
$$ e_a = E_m \sin(\theta), \quad e_b = E_m \sin(\theta – 120^\circ), \quad e_c = E_m \sin(\theta + 120^\circ) $$
These EMFs are used to generate current references for the inner current control loops, ensuring precise tracking and stability. The current control in the dq-frame is expressed as:
$$ v_d = \left( k_{p} + \frac{k_{i}}{s} \right) (i_{dref} – i_d) – \omega L i_q + e_d $$
$$ v_q = \left( k_{p} + \frac{k_{i}}{s} \right) (i_{qref} – i_q) + \omega L i_d + e_q $$
where \( v_d \) and \( v_q \) are the d-axis and q-axis voltage components, \( k_p \) and \( k_i \) are proportional and integral gains, \( i_{dref} \) and \( i_{qref} \) are reference currents, \( i_d \) and \( i_q \) are measured currents, \( L \) is the filter inductance, and \( e_d \), \( e_q \) are the dq-axis EMF components. This control framework enables the UPS to operate as a grid-forming source, providing voltage and frequency support even during islanded conditions.
The voltage-frequency active support control strategy allows the UPS to participate in grid frequency regulation. When grid frequency deviations occur, the cell energy storage system adjusts its active power output based on a droop characteristic:
$$ P_{ref} = P_0 + k_f (f_{nom} – f) $$
where \( P_0 \) is the nominal power, \( k_f \) is the frequency droop coefficient, \( f_{nom} \) is the nominal frequency, and \( f \) is the measured frequency. This response helps stabilize the grid, demonstrating the value of the cell energy storage system in providing ancillary services. The following table summarizes key control parameters for the voltage-frequency support:
| Parameter | Symbol | Typical Value | Description |
|---|---|---|---|
| Virtual Inertia | \( J \) | 0.1 kg·m² | Determines frequency response dynamics |
| Frequency Droop Coefficient | \( k_f \) | 6 × 10⁴ W/Hz | Adjusts active power based on frequency deviation |
| Voltage Droop Coefficient | \( D_q \) | 2 × 10⁴ V/var | Adjusts voltage based on reactive power |
| Current Control Bandwidth | \( k_p \), \( k_i \) | 0.35, 80 (for d-axis) | Ensures fast current tracking and stability |
Battery charging management is crucial for maintaining the health and longevity of the cell energy storage system. The proposed UPS incorporates two charging modes: constant-voltage current-limited (CVCL) charging and constant-current (CC) charging. For CVCL charging, the DC-link voltage control loop generates the active power reference:
$$ P_{ref} = \left( k_{pv} + \frac{k_{iv}}{s} \right) (V_{dc\_ref} – V_{dc}) $$
where \( k_{pv} \) and \( k_{iv} \) are proportional and integral gains for voltage control, \( V_{dc\_ref} \) is the reference DC voltage, and \( V_{dc} \) is the measured DC voltage. The charging current is limited to a maximum value \( I_{chg\_max} \) to prevent overcurrent. For CC charging, the current control loop directly sets the active power reference:
$$ P_{ref} = \left( k_{pi} + \frac{k_{ii}}{s} \right) (I_{chg\_ref} – I_{chg}) $$
where \( k_{pi} \) and \( k_{ii} \) are gains for current control, \( I_{chg\_ref} \) is the reference charging current, and \( I_{chg} \) is the measured charging current. These strategies ensure efficient and safe operation of the cell energy storage system, extending its lifecycle and reliability.
Experimental validation was conducted using a 500 kW prototype of the single-stage grid-formed UPS system, coupled with a 250 kWh lithium-iron-phosphate cell energy storage system. The system parameters are listed in the table below, highlighting the integration of the cell energy storage system into the UPS architecture.
| Component | Parameter | Value |
|---|---|---|
| UPS Rated Power | \( P_{rated} \) | 500 kW |
| Cell Energy Storage Capacity | \( E_{bat} \) | 250 kWh |
| DC-Link Voltage | \( V_{dc} \) | 800 V |
| Filter Inductance | \( L_f \) | 0.2 mH |
| Filter Capacitance | \( C_f \) | 800 μF |
| Switching Frequency | \( f_{sw} \) | 10 kHz |
| Grid Voltage | \( V_{grid} \) | 315 V (line-to-line) |
Tests were performed to evaluate the system’s performance in various operational modes. In dual-online mode, the UPS and grid supplied power simultaneously, with the cell energy storage system providing backup. During load step changes (e.g., from 250 kW to 500 kW), the output voltage maintained sinusoidal waveform with less than 1% distortion, demonstrating excellent dynamic response. The transition to battery-independent mode during simulated grid faults was seamless, with zero interruption time and minimal voltage fluctuation. This reliability is attributed to the robust control strategy and the rapid response of the cell energy storage system.
For frequency support, the UPS actively injected active power into the grid when frequency deviations were detected. The response time was within 100 ms, and the power adjustment followed the droop characteristic accurately. This capability not only stabilizes the grid but also generates potential revenue from ancillary services, offsetting the costs of the cell energy storage system. Battery charging tests confirmed the effectiveness of both CVCL and CC modes, with charging currents tightly regulated and voltage stability maintained. The following table summarizes key experimental results:
| Test Scenario | Metric | Result |
|---|---|---|
| Dual-Online Mode Load Step | Voltage THD | < 1% |
| Mode Transition (Grid to Battery) | Interruption Time | 0 s |
| Frequency Support Response | Response Time | ≤ 100 ms |
| Battery Charging (CC Mode) | Current Tracking Error | < 2% |
| Overall Efficiency | Peak Efficiency | 97.5% |
The efficiency gains are significant compared to traditional UPS systems, primarily due to the single-stage conversion and optimized control of the cell energy storage system. Mathematical analysis of the power losses further illustrates this advantage. The total loss \( P_{loss} \) in the proposed system can be approximated as:
$$ P_{loss} = P_{sw} + P_{cond} + P_{bat} $$
where \( P_{sw} \) represents switching losses in the inverter, \( P_{cond} \) denotes conduction losses, and \( P_{bat} \) accounts for losses in the cell energy storage system. For a 500 kW system, the switching losses are given by:
$$ P_{sw} = f_{sw} \cdot (E_{on} + E_{off}) \cdot I_{avg} $$
with \( f_{sw} \) as the switching frequency, \( E_{on} \) and \( E_{off} \) as the turn-on and turn-off energy per switching cycle, and \( I_{avg} \) as the average current. Conduction losses are calculated as:
$$ P_{cond} = I_{rms}^2 \cdot R_{ds(on)} $$
where \( I_{rms} \) is the RMS current and \( R_{ds(on)} \) is the on-state resistance of the switches. The cell energy storage system losses include internal resistance effects:
$$ P_{bat} = I_{bat}^2 \cdot R_{int} $$
with \( I_{bat} \) as the battery current and \( R_{int} \) as the internal resistance. Based on typical values, the overall losses are reduced by approximately 30% compared to double-conversion UPS systems, highlighting the efficiency of integrating a cell energy storage system in a single-stage architecture.
Furthermore, the economic benefits of the proposed system are substantial. By participating in grid frequency regulation, the cell energy storage system can generate additional revenue. The annual revenue \( R \) from ancillary services can be estimated as:
$$ R = P_{avg} \cdot t \cdot r $$
where \( P_{avg} \) is the average power provided for frequency support, \( t \) is the operational time, and \( r \) is the service rate (e.g., in $/MWh). Assuming a conservative scenario with \( P_{avg} = 100 \) kW, \( t = 2000 \) hours per year, and \( r = 50 $/MWh, the annual revenue amounts to $10,000. This revenue helps amortize the initial investment in the cell energy storage system, making the UPS more cost-effective over its lifecycle.
In terms of scalability, the proposed architecture can be extended to larger systems by paralleling multiple units or integrating with renewable energy sources. The cell energy storage system plays a pivotal role in such integrations, providing buffering for intermittent generation and enhancing grid stability. Control strategies can be adapted for microgrid applications, where the UPS acts as a grid-forming unit for islanded operations. The modular design allows for easy maintenance and upgrades, ensuring long-term viability.
To summarize, the single-stage grid-formed UPS based on a cell energy storage system offers a paradigm shift in uninterruptible power supply technology. Its key advantages include high efficiency through single-stage conversion, reliable power delivery via autonomous grid-forming control, active grid support through voltage-frequency regulation, and economic benefits from ancillary service participation. The experimental validation on a 500 kW prototype confirms the feasibility and performance of the system, with seamless mode transitions, stable operation under load variations, and effective battery management. As power systems evolve towards greater decentralization and resilience, the integration of cell energy storage systems in UPS architectures will become increasingly important, providing both reliability and grid support functionalities.
Future work could focus on optimizing the control algorithms for faster response times, integrating advanced battery management systems (BMS) for enhanced lifecycle of the cell energy storage system, and exploring hybrid energy storage combinations (e.g., with supercapacitors) for high-power applications. Additionally, standardizing communication protocols for grid interaction could facilitate widespread adoption in smart grids. The proposed system represents a significant step forward in making UPS systems more efficient, reliable, and economically viable, underscoring the critical role of the cell energy storage system in modern power infrastructure.