In recent years, battery energy storage technology has shown significant advantages in maintaining grid stability, making it a crucial solution for large-scale energy storage. With decreasing costs and maturing integration techniques, energy storage batteries are becoming more prevalent in power systems. Among these, the all-vanadium redox flow battery (VRB) stands out due to its long cycle life, safety, and flexible power and capacity variations. To accurately simulate and analyze the performance of such systems in practical applications, it is essential to develop engineering-ready models that account for various operational constraints. In this study, I focus on modeling a VRB-based energy storage battery system with energy limitation modules, using the Power System Analysis Software Package (PSASP). The model incorporates factors such as converter capacity and state of charge (SOC), and through simulation analysis, I compare control strategies and validate the model’s applicability across different-sized energy storage batteries. This approach ensures that the energy storage battery can effectively regulate grid stability while adhering to physical and operational limits.
The energy storage battery system primarily consists of the battery stack and the grid-connected power conversion system (PCS) controller. The VRB stack is composed of metal electrodes, electrolytes containing vanadium ions in different oxidation states, an ion-exchange membrane, and electrolyte tanks. During charge and discharge, electrolytes circulate through the stack, where redox reactions occur at the electrodes. The discharge reactions are as follows:
Positive electrode: $$VO_2^+ + 2H^+ + e^- = VO^{2+} + H_2O$$
Negative electrode: $$V^{2+} – e^- = V^{3+}$$
I adopt an improved loss-inclusive equivalent model for the battery stack. The stack open-circuit voltage \(U_{stack}\) is represented by a controlled voltage source, while the terminal voltage \(U_b\) and currents \(I\) and \(I_b\) account for internal losses. The model equations are:
$$U_b = C_{elec} \frac{dU_{C_{elec}}}{dt} + I R_{Res} + U_{C_{elec}}$$
$$U_{C_{elec}} = U_{stack} + I R_{Reac}$$
$$C_{elec} \frac{dU_{C_{elec}}}{dt} = I_b – I – \frac{U_b}{R_{fix}}$$
Here, \(R_{Reac}\) and \(R_{Res}\) represent reaction and resistive losses, constituting 60% and 40% of total internal losses, respectively. \(R_{fix}\) accounts for pump and auxiliary losses, and \(C_{elec}\) is the inter-electrode capacitance with voltage \(U_{C_{elec}}\). The open-circuit voltage of a single cell is given by the Nernst equation:
$$U_{cell} = E_0 – \frac{RT}{nF} \ln \frac{[V^{3+}][V^{4+}]}{[V^{2+}][V^{5+}][H^+]}$$
where \(E_0\) is the standard electrode potential (1.259 V at 298 K), \(R\) is the gas constant, \(T\) is absolute temperature, \(n\) is the number of electrons transferred (n=1 for VRB), and \(F\) is Faraday’s constant. For a stack with \(N_{cell}\) cells, the total open-circuit voltage is:
$$U_{stack} = N_{cell} U_{cell}$$
The state of charge (SOC) is a critical parameter, defined as the ratio of remaining energy \(E_{stack}\) to rated capacity \(E_C\):
$$SOC = \frac{E_{stack}}{E_C}$$
Under balanced ion concentrations, SOC relates to vanadium ion ratios. At 298 K, the cell voltage simplifies to:
$$U_{cell} = E_0 – \frac{RT}{nF} \ln \left( \frac{1 – SOC}{SOC} [H^+] \right)^2$$
SOC evolves over time based on charge/discharge power:
$$SOC_{t+1} = SOC_t + \Delta SOC$$
$$\Delta SOC = \frac{\Delta E}{E_c} = \frac{P_s \Delta t}{E_c} = \frac{I U_{stack} \Delta t}{E_c}$$
where \(P_s\) is the battery power and \(\Delta t\) is the time step. To protect the energy storage battery, SOC limits \(SOC_{min}\) and \(SOC_{max}\) are enforced, and the maximum output power \(P_{max}\) and minimum \(P_{min}\) are fixed regardless of SOC variations.
The grid-connected converter PCS enables bidirectional power flow and four-quadrant operation. It uses a dual-loop decoupling control strategy: an outer power loop for frequency/active power and voltage/reactive power control, and an inner current loop for precise current regulation. The outer loop generates active and reactive power references \(P_{set}\) and \(Q_{set}\) using PI controllers:
$$P_{set} = \left( K_{wp} + \frac{K_{wi}}{s} \right) \Delta \omega$$
$$Q_{set} = \left( K_{vp} + \frac{K_{vi}}{s} \right) \Delta U$$
where \(\Delta \omega\) and \(\Delta U\) are frequency and voltage deviations, and \(s\) is the Laplace operator. The inner loop approximates to first-order lags:
$$P = \frac{1}{1 + T_p s} P_{set}$$
$$Q = \frac{1}{1 + T_q s} Q_{set}$$
where \(T_p\) and \(T_q\) are time constants. The converter has a limited capacity \(S\), typically 0-10% larger than the maximum active power \(P_{max}\) of the energy storage battery, necessitating energy limitation modules to prevent overloading.
To ensure safe and efficient operation, I design two key limitation modules for the energy storage battery system. First, the SOC limitation module monitors SOC and halts power transfer when SOC exceeds \(SOC_{min}\) or \(SOC_{max}\). The flowchart involves integrating power to update SOC, with separate paths for charge and discharge based on sign detection. The power \(P\) is the system output, and battery power \(P_s\) accounts for converter efficiencies \(\eta_{charge}\) and \(\eta_{discharge}\):
$$P_s = P \eta_{charge} \quad \text{(charging)}$$
$$P = P_s \eta_{discharge} \quad \text{(discharging)}$$
Second, the converter limitation module restricts reactive power \(Q\) based on converter capacity \(S\) and actual active power \(P\). The constraint is:
$$-P_{max} \leq P \leq P_{max}$$
$$-Q_{max} \leq Q \leq Q_{max}$$
$$Q_{max} = \sqrt{S^2 – P^2}$$
This ensures that the apparent power \( \sqrt{P^2 + Q^2} \) does not exceed \(S\). The implementation uses a limiter to adjust \(Q\) dynamically, allowing flexible reactive power support depending on \(P\). Compared to fixed reactive limits, this adaptive approach optimizes the energy storage battery’s regulatory capability.
I implement the model in PSASP using the User-Defined (UD) modeling function. Since UD models output nodal current, I convert power outputs to current injections. Let \(S = P + jQ\) be the complex power injected at the bus, with bus voltage \(U = U_R + jU_I\) and current \(I = I_R + jI_I\). Then:
$$S = U I^* = (U_R + jU_I)(I_R – jI_I)$$
Solving for real and imaginary parts:
$$P = U_R I_R + U_I I_I$$
$$Q = U_I I_R – U_R I_I$$
Thus, the current components are:
$$I_R = \frac{U_R P + U_I Q}{U_R^2 + U_I^2}$$
$$I_I = \frac{U_I P – U_R Q}{U_R^2 + U_I^2}$$
These equations form the interface in the UD model, integrating the control strategies and limitation modules. The overall UD structure includes blocks for SOC calculation, power limitation, and PI controllers, ensuring comprehensive representation of the energy storage battery dynamics.
For simulation, I use the CEPRI-7 benchmark network in PSASP. The system base is \(S_B = 100 \text{ MVA}\), with a time step of 0.01 s and total duration of 15.00 s. I set a single-phase-to-ground fault between buses B3 and B4 at 0.02 s, cleared at 0.15 s, and successful auto-reclosing at 0.70 s. The energy storage battery is connected to bus B1 with initial SOC \(SOC_{t0} = 0.5\) and deadband of 0.01. I evaluate two scenarios: comparison of reactive limitation strategies and applicability to different-sized energy storage batteries.
In the first scenario, I compare the adaptive reactive limitation strategy with a fixed reactive limit. For an energy storage battery with \(P_{max} = 1.5 \text{ pu}\) and converter capacity \(S = 1.65 \text{ pu}\), the fixed limit sets \(Q_{max} = \sqrt{1.65^2 – 1.5^2} = 0.69 \text{ pu}\). However, the adaptive strategy allows \(Q_{max}\) to vary with \(P\), providing higher reactive support when \(P\) is below maximum. Simulation results show that during voltage dips, the adaptive strategy better suppresses voltage fluctuations and reduces oscillations, demonstrating enhanced grid support from the energy storage battery.
In the second scenario, I test the model with different energy storage battery sizes: \(P_{max} = 0.1, 0.5, 1.5 \text{ pu}\), and corresponding \(S = 0.11, 0.55, 1.65 \text{ pu}\). The table below summarizes the parameters and key outcomes:
| \(P_{max}\) (pu) | \(S\) (pu) | Voltage Recovery | Oscillation Damping |
|---|---|---|---|
| 0.1 | 0.11 | Moderate | Limited |
| 0.5 | 0.55 | Improved | Noticeable |
| 1.5 | 1.65 | Fast | Significant |
The results indicate that larger energy storage batteries with appropriately sized converters offer better transient stability, as seen in generator power angle and bus voltage responses. The limitation modules effectively constrain power within safe limits, validating the model’s scalability. This underscores the importance of properly sizing both the energy storage battery and converter for optimal performance.
Furthermore, I analyze the impact of SOC limits on long-term operation. With \(SOC_{min} = 0.2\) and \(SOC_{max} = 0.8\), the energy storage battery avoids deep discharge or overcharge, extending lifespan. The SOC variation during simulation follows:
$$\Delta SOC = \int \frac{P_s}{E_c} dt$$
where \(P_s\) is derived from \(P\) considering efficiency. The energy storage battery seamlessly switches between charging and discharging modes based on grid needs, with the limitation modules ensuring safe SOC levels. This highlights the robustness of the control strategy in real-world applications.
The integration of energy limitation modules is crucial for practical deployment of energy storage batteries. By accounting for converter capacity and SOC, the model prevents overheating and degradation, enhancing reliability. In grid simulations, the energy storage battery provides frequency regulation through active power control and voltage support through reactive power control, with limitations ensuring no exceedance of rated parameters. The adaptive reactive limitation, in particular, maximizes the utility of the energy storage battery by dynamically adjusting \(Q_{max}\) based on \(P\), a feature absent in fixed-limit approaches.
In conclusion, I have developed a comprehensive model for an all-vanadium redox flow battery energy storage system with energy limitation modules in PSASP. The model incorporates battery dynamics, converter control, SOC constraints, and adaptive reactive power limits. Simulations using the CEPRI-7 network demonstrate that the model effectively regulates grid stability while adhering to physical limits. The adaptive reactive limitation strategy outperforms fixed limits, offering flexible support, and the model is applicable across various energy storage battery sizes. This work provides a practical tool for designing and analyzing energy storage battery systems in power grids, ensuring safe and efficient operation. Future extensions could include multi-objective optimization and integration with renewable sources to further enhance grid resilience.
The energy storage battery technology continues to evolve, and models like this are essential for leveraging its full potential. By considering real-world constraints, we can ensure that energy storage batteries contribute reliably to grid stability, paving the way for a sustainable energy future. The use of advanced modeling tools like PSASP, combined with robust control strategies, will drive the adoption of energy storage batteries in modern power systems.