1. Introduction
The integration of renewable energy sources, such as wind and solar power, into the grid faces significant challenges due to their inherent intermittency and variability. These fluctuations threaten grid stability and limit the scalability of renewable energy adoption. Battery energy storage system (BESS) have emerged as a critical solution to mitigate these issues by providing rapid power balancing, frequency regulation, and voltage support. This paper proposes a hierarchical control framework for battery energy storage system (BESS), addressing key challenges such as power allocation, state-of-charge (SOC) management, flexible power control, and zero-sequence circulating current suppression. The proposed strategies are validated through simulations and experiments, demonstrating their effectiveness in enhancing grid stability and prolonging battery lifespan.
2. Hierarchical Control Architecture
The hierarchical control structure of the battery energy storage system (BESS) is divided into four layers: Grid Control Layer, Energy Control Layer, Power Control Layer, and Current Control Layer. Each layer addresses specific technical requirements and coordinates with others to ensure optimal performance.
2.1 Grid Control Layer
This layer manages power dispatch commands from the grid and local control strategies. It calculates the total active and reactive power demands (PSETPSET, QSETQSET) based on grid scheduling or local measurements (e.g., PCC voltage VPCCVPCC and power PPCCPPCC). Key algorithms include:
- Active Power Demand Calculation:PLOCAL=PPCCτsτs+1,τ=12πfcPLOCAL=PPCCτs+1τs,τ=2πfc1where fcfc is the cutoff frequency for low-pass filtering.
- Reactive Power Demand Calculation:QLOCAL=ev(Kpv+Kivs),ev=ΔV (voltage deviation)QLOCAL=ev(Kpv+sKiv),ev=ΔV (voltage deviation)
2.2 Energy Control Layer
This layer regulates the SOC of battery units to maintain charge/discharge capabilities. A feedback-based strategy adjusts the power reference PrefPref by superimposing low-frequency charging/discharging components:Pbat=PSOC⋅τsτs+1,PSOC=ΔSOC⋅(Kpsoc+Kisocs)Pbat=PSOC⋅τs+1τs,PSOC=ΔSOC⋅(Kpsoc+sKisoc)
where ττ is a variable time constant dependent on SOC (Table 1).
Table 1: Variable Time Constant (ττ) Based on SOC
| SOC Range | ττ Calculation |
|---|---|
| [0, 0.2) | τ=τ0+k1⋅SOCτ=τ0+k1⋅SOC |
| [0.2, 0.4) | τ=τ0+a(k1−k2)+k2⋅SOCτ=τ0+a(k1−k2)+k2⋅SOC |
| [0.4, 0.6] | τ=τ0+ak1+(b−a)k2τ=τ0+ak1+(b−a)k2 |
| (0.6, 0.8] | τ=τ0+ak1+(1−a)k2−k2⋅SOCτ=τ0+ak1+(1−a)k2−k2⋅SOC |
| (0.8, 1.0] | τ=τ0+ak1−k1⋅SOCτ=τ0+ak1−k1⋅SOC |
2.3 Power Control Layer
This layer ensures smooth transitions between operational modes (rectification, inversion, reactive compensation) by dynamically adjusting current and voltage limits. The flexible power control strategy unifies these modes through dynamic threshold adjustments:{id_max∗=(2MmaxUdc/3)2−(ωLPref∗/Ug)2−UgωLiq_min∗=−S2−(Pref∗)2Ug⎩⎨⎧id_max∗=ωL(2MmaxUdc/3)2−(ωLPref∗/Ug)2−Ugiq_min∗=−UgS2−(Pref∗)2
where MmaxMmax is the maximum modulation index, and SS is the apparent power.
2.4 Current Control Layer
This layer suppresses zero-sequence circulating currents (i0i0) in parallel converters using dual-carrier modulation and proportional-resonant (PR) control:GPR(s)=Kp+Krss2+ω02GPR(s)=Kp+s2+ω02Krs
where ω0ω0 is the resonant frequency.
3. State Classification and Power Allocation
3.1 Active Power Allocation
Battery units are classified into six states (S1–S6) based on SOC and fault conditions (Table 2). Power allocation prioritizes units with higher charging/discharging capabilities.
Table 2: Battery Unit State Classification
| State | SOC Range | Power Capability | Dispatch Priority |
|---|---|---|---|
| S1 | [0, 0.2] | Limited charging | High (charging) |
| S2 | [0.2, 0.4] | Moderate charging | Medium (charging) |
| S3 | [0.4, 0.6] | Balanced | Low |
| S4 | [0.6, 0.8] | Moderate discharging | Medium (discharging) |
| S5 | [0.8, 1.0] | Limited discharging | High (discharging) |
| S6 | — | Fault | Not dispatched |
The power allocation algorithm dynamically adjusts references for each unit:Pn_ref(t+1)={PH(S1 units at max charging)PCMAX(S2–S4 units)PL(S5 units at max discharging)Pn_ref(t+1)=⎩⎨⎧PHPCMAXPL(S1 units at max charging)(S2–S4 units)(S5 units at max discharging)
3.2 Reactive Power Allocation
Reactive power allocation prioritizes “active units” (those delivering active power) over “idle units.” The maximum reactive output for idle units is:Qmax=αS,0≤α≤1Qmax=αS,0≤α≤1
4. Simulation and Experimental Validation
4.1 Active Power Control
Simulations compared the proposed state-classification-based strategy with an average allocation method. Results demonstrated superior tracking accuracy and reduced startup cycles (Figure 1).
Key Metrics:
- Tracking Error: <2% for proposed vs. >10% for average allocation.
- SOC Balancing: Units converged to 50% SOC, ensuring sustained power capability.
4.2 Energy Management
A 1500 kW wind farm with 400 kW battery energy storage system (BESS) was simulated under three SOC control strategies:
- Power Limitation: Restricted charging/discharging at SOC boundaries.
- Direct Power Superposition: Overlaid fixed power on PrefPref.
- Proposed Strategy: Variable ττ low-pass filtering.
Table 3: Performance Comparison
| Metric | Power Limitation | Direct Superposition | Proposed Strategy |
|---|---|---|---|
| FHC (0.01–1 Hz) | 9.33% | 4.26% | 3.19% |
| Equivalent Cycle Life | 0.64 | 0.66 | 0.59 |
4.3 Zero-Sequence Circulating Current Suppression
Experiments on parallel converters showed a 75% reduction in i0i0 using dual-carrier modulation and PR control (Figure 2).
5. Conclusion
This paper presents a comprehensive hierarchical control framework for battery energy storage system, addressing critical challenges in renewable energy integration. Key contributions include:
- A state-classification-based power allocation algorithm for improved accuracy.
- SOC feedback control with variable time constants to balance energy management and grid stability.
- Flexible power control enabling seamless mode transitions.
- Dual-carrier modulation and PR control for circulating current suppression.
Future work will focus on real-time optimization of control parameters and scalability for large-scale battery energy storage system (BESS) deployments.
Formulas and Tables Summary
- Active Power Demand:PLOCAL=PPCCτsτs+1PLOCAL=PPCCτs+1τs
- Reactive Power Demand:QLOCAL=ev(Kpv+Kivs)QLOCAL=ev(Kpv+sKiv)
- SOC Feedback Control:Pbat=PSOC⋅τsτs+1Pbat=PSOC⋅τs+1τs
- Zero-Sequence Suppression:GPR(s)=Kp+Krss2+ω02GPR(s)=Kp+s2+ω02Krs
Tables:
- Table 1: Variable ττ based on SOC.
- Table 2: Battery unit state classification.
- Table 3: Performance comparison of SOC strategies.
This hierarchical approach ensures the battery energy storage system operates efficiently, prolongs battery life, and enhances grid stability, making it a pivotal solution for future renewable-dominated power systems.