Abstract
This article discusses an adaptive control strategy for energy storage batteries participating in the primary frequency regulation of thermal power units. With the increasing penetration of renewable energy sources such as wind and solar, the stability of power grids faces new challenges due to their inherent variability and intermittency. Energy storage batteries, known for their rapid response and high accuracy, have emerged as a viable solution to enhance grid frequency regulation. This study proposes an adaptive control method combining virtual negative inertia control, virtual droop control, and virtual inertia control based on the battery’s state of charge (SOC). The effectiveness of the proposed strategy is validated through simulations on a typical single-area power grid model.
1. Introduction
The rapid depletion of fossil fuels and the increasing awareness of climate change have driven the transition towards renewable energy sources. However, the integration of renewable energy, especially wind and solar, poses significant challenges to grid stability due to their stochastic nature. To mitigate these challenges, energy storage batteries have been increasingly used for grid frequency regulation. This paper focuses on the adaptive control of energy storage batteries participating in the primary frequency regulation of thermal power units.
2. Background and Motivation
2.1 Grid Frequency Regulation
Grid frequency regulation aims to maintain the system frequency within an acceptable range by compensating for power imbalances. Primary frequency regulation is the first line of defense against sudden changes in load or generation. Thermal power units, as the primary source of baseload power, are traditionally responsible for frequency regulation. However, their response time is relatively slow, making them less effective in coping with sudden frequency deviations.
2.2 Energy Storage Batteries for Frequency Regulation
Energy storage batteries offer several advantages for grid frequency regulation, including fast response time, high accuracy, and flexibility. By rapidly absorbing or injecting power into the grid, they can effectively compensate for frequency deviations. The integration of energy storage batteries into grid frequency regulation can reduce the burden on thermal power units, extend their lifespan, and improve overall grid stability.
3. Control Strategies for Energy Storage Batteries
3.1 Virtual Droop Control
Virtual droop control mimics the droop characteristics of synchronous generators. It adjusts the battery’s output power based on the frequency deviation (Δf) to minimize steady-state frequency errors. The power increment (ΔPE1) under virtual droop control is given by:
DeltaPE1=−KEΔf
where KE is the droop coefficient.
3.2 Virtual Inertia Control
Virtual inertia control mimics the inertia response of synchronous generators. It adjusts the battery’s output power based on the rate of change of frequency (Δf/Δt) to dampen frequency oscillations. The power increment (ΔPE2) under virtual inertia control is given by:
DeltaPE2=−MEdtd(Δf)
where ME is the virtual inertia coefficient.
3.3 Virtual Negative Inertia Control
To address the limitations of traditional virtual inertia control, which can hinder frequency recovery during the restoration phase, this study introduces virtual negative inertia control. This control strategy promotes frequency recovery by reversing the power direction during the frequency restoration phase. The power increment (ΔPE3) under virtual negative inertia control is given by:
DeltaPE3=ME1dtd(Δf)
where ME1 is the virtual negative inertia coefficient.
4. Adaptive Control Strategy
The proposed adaptive control strategy dynamically adjusts the contribution of each control method based on the system frequency and its rate of change. The strategy also considers the battery’s SOC to avoid overcharging or overdischarging.
4.1 Frequency Deadbands
To reduce unnecessary battery operations and extend its lifespan, deadbands are set for both the energy storage battery and the thermal power unit. The battery deadband (fe) is set slightly smaller than the thermal power unit deadband (fg) to allow the battery to respond earlier and mitigate frequency deviations.
| Deadband | Value (Hz) |
|---|---|
| Battery Deadband | fe |
| Thermal Deadband | fg |
4.2 Control Proportion Allocation
The control proportions for virtual droop, virtual inertia, and virtual negative inertia control are dynamically adjusted based on the frequency deviation and its rate of change. The allocation coefficients (k1 and k2) are calculated as follows:
k_1 = \begin{cases} e^{N\Delta f} & \text{if } \frac{d(\Delta f)}{dt} < 0 \\ 0.25 & \text{if } \frac{d(\Delta f)}{dt} > 0 end{cases}
k2=1−k1
where N is a constant (e.g., N = 10).
4.3 SOC-Based Power Constraints
To maintain the battery’s SOC within safe limits, the charging and discharging coefficients (Kc and Kd) are adjusted based on the current SOC value using a logistic function:
Kc=⎩⎨⎧Kmax,Kmax⋅Kmax+K0⋅e2n(XSOC,max−XSOC)−1K0⋅e2n(XSOC,max−XSOC),0,if XSOC<XSOC,minif XSOC,min≤XSOC≤XSOC,maxif XSOC>XSOC,max
Kd=⎩⎨⎧0,Kmax⋅Kmax+K0⋅e2n(XSOC−XSOC,min)−1K0⋅e2n(XSOC−XSOC,min),Kmax,if XSOC<XSOC,minif XSOC,min≤XSOC≤XSOC,maxif XSOC>XSOC,max
where K0 and n are constants, and XSOC,min and XSOC,max are the minimum and maximum SOC limits, respectively.
5. Simulation and Results
5.1 Simulation Setup
A typical single-area power grid model is used for simulations. The thermal power unit has a rated capacity of 660 MW, and the energy storage battery has a configured capacity of 6 MW/6 MWh. Simulations are conducted under both step and continuous load disturbances.
5.2 Step Load Disturbance
A step load disturbance of 0.01 p.u. is applied to the system. The frequency deviation and SOC change are shown in Figures 6 and 7, respectively. The proposed adaptive control strategy demonstrates the best performance in terms of maximum frequency deviation, steady-state frequency deviation, and frequency restoration time (Table 1).
| Method | Δfm (p.u.) | Δfs (p.u.) | Δtm (s) | Δts (s) |
|---|---|---|---|---|
| No Battery Regulation | -1.62 × 10^-3 | -1.02 × 10^-3 | 8.5 | 18.3 |
| Droop Control (Fixed K) | -0.81 × 10^-3 | -0.76 × 10^-3 | 7.6 | 17.7 |
| Adaptive Control | -0.74 × 10^-3 | -0.68 × 10^-3 | 7.9 | 15.6 |
Table 1: Evaluation Metrics under Step Load Disturbance
5.3 Continuous Load Disturbance
A continuous load disturbance of 0.01 p.u. for 200 s is applied. The frequency deviation and SOC change are shown in Figures 8 and 9, respectively. The adaptive control strategy continues to outperform the other methods in terms of frequency regulation and SOC maintenance (Table 2).
| Method | Rf (×10^-4) | RSOC (×10^-5) |
|---|---|---|
| No Battery Regulation | 4.7096 | – |
| Droop Control (Fixed K) | 3.3741 | 3.7602 |
| Adaptive Control | 1.6444 | 4.2533 |
Table 2: Evaluation Metrics under Continuous Load Disturbance
6. Conclusion
This study proposes an adaptive control strategy for energy storage batteries participating in the primary frequency regulation of thermal power units. By combining virtual droop, virtual inertia, and virtual negative inertia control, the strategy dynamically adjusts the control proportions based on system frequency and its rate of change. The strategy also considers the battery’s SOC to maintain it within safe limits. Simulations demonstrate that the proposed adaptive control strategy outperforms traditional methods in terms of frequency regulation and SOC maintenance.