Energy storage batteries play a pivotal role in modern power systems, particularly in renewable energy integration and grid stabilization. To address the challenge of phase current imbalance caused by inconsistent parasitic parameters in three-phase interleaved bidirectional Buck-Boost converters, this study proposes a sliding mode-predictive control (SMC-MPC) strategy. This method ensures rapid current sharing while maintaining high robustness and dynamic performance.
Converter Topology and Operating Principles
The three-phase interleaved bidirectional Buck-Boost converter, as shown in the figure, operates in two modes: Boost mode (energy transfer from battery to DC bus) and Buck mode (reverse energy flow). The voltage conversion relationships are expressed as:
Boost mode:
$$U_o = \frac{U_{in}}{1 – D_{Boost}}$$
Buck mode:
$$U_{in} = D_{Buck} \cdot U_o$$
Where $D_{Boost}$ and $D_{Buck}$ represent duty cycles in respective modes. The current ripple characteristics under different duty cycles are summarized in Table 1.
| Operating Mode | Duty Cycle Range | Current Ripple Formula |
|---|---|---|
| Boost | $0 < D_{Boost} < \frac{1}{3}$ | $$\Delta i_L = \frac{U_o D_{Boost}(1 – 3D_{Boost})T_s}{L}$$ |
| Boost | $\frac{1}{3} \leq D_{Boost} \leq \frac{2}{3}$ | $$\Delta i_L = \frac{U_o T_s(3D_{Boost} – 1)(2 – 3D_{Boost})}{3L}$$ |
| Buck | $0 < D_{Buck} < \frac{1}{3}$ | $$\Delta i_L = \frac{U_o D_{Buck}(1 – 3D_{Buck})T_s}{L}$$ |
SMC-MPC Control Strategy
The proposed hierarchical control framework combines sliding mode control (SMC) for voltage regulation and model predictive control (MPC) for current sharing:
1. Voltage Outer Loop (SMC):
Define sliding surface using voltage error and integral terms:
$$S = \alpha_1(i_{ref} – i_L^{avg}) + \alpha_2(U_{ref} – U_o) + \alpha_3\int (U_{ref} – U_o)dt$$
where $\alpha_1$, $\alpha_2$, $\alpha_3$ are sliding coefficients.
2. Current Inner Loop (MPC):
Predictive model for $k+1$ sampling instant:
$$i_L(k+1) = i_L(k) + \frac{T_s}{L}[U_{in}(k) – (1 – D(k))U_o(k)]$$
Cost function minimization:
$$J = \lambda_1(i_L^{ref} – i_L(k+1))^2 + \lambda_2(U_o^{ref} – U_o(k+1))^2$$
Key Implementation Aspects
For energy storage battery applications, critical design parameters include:
| Parameter | Value | Impact on Performance |
|---|---|---|
| Switching Frequency | 20 kHz | Higher frequencies reduce current ripple but increase switching losses |
| Inductor Tolerance | <±5% | Directly affects current sharing accuracy |
| Predictive Horizon | 3 steps | Balances computation load and dynamic response |
The optimal duty cycle calculation for phase balancing:
$$D_{opt} = \frac{L[i_L^{ref} – i_L(k)] + T_sU_{in}(k)}{T_sU_o(k)} + \Delta D_{comp}$$
where $\Delta D_{comp}$ represents the compensation term for parameter mismatches.
Performance Comparison
Experimental verification using a 500W prototype demonstrates significant improvements over conventional PI-MPC control:
| Metric | SMC-MPC | PI-MPC |
|---|---|---|
| Current Sharing Error | 1.2% | 4.5% |
| Voltage Regulation Time | 2.1 ms | 5.8 ms |
| Overshoot (Load Step) | 3.8% | 12.4% |
| THD (Full Load) | 8.7% | 15.2% |
The enhanced performance stems from the combined advantages of SMC’s robustness and MPC’s fast dynamics. For energy storage battery systems requiring frequent charge/discharge cycles, this strategy effectively handles:
- Nonlinear battery impedance variations
- Parasitic parameter mismatches (up to 15% inductance tolerance)
- Wide voltage conversion range (1:4 ratio)
Conclusion
This paper presents a novel SMC-MPC strategy for energy storage battery converters that achieves:
- Improved current sharing accuracy (≤2% imbalance)
- Enhanced dynamic response (63% faster than PI-MPC)
- Reduced output voltage ripple (42% reduction)
The methodology demonstrates excellent potential for large-scale energy storage systems requiring high reliability and fast response. Future work will focus on extending this approach to multi-port converters for hybrid energy storage applications.