Abstract
The integration of renewable energy sources, such as wind and solar, into distributed power systems has highlighted the critical role of energy storage batteries in stabilizing DC microgrids. Bidirectional DC-DC converters serve as the interface between energy storage batteries and the DC bus, enabling efficient energy transfer. This paper addresses the challenge of current imbalance in three-phase interleaved bidirectional Buck-Boost converters caused by parasitic parameter mismatches. A novel sliding mode-predictive control (SMC-MPC) strategy is proposed to enhance current-sharing accuracy, robustness, and dynamic response. The voltage outer loop employs sliding mode control (SMC) to generate reference inductor currents, while the current inner loop utilizes model predictive control (MPC) to optimize duty cycles. Experimental results from a 500 W prototype demonstrate that the SMC-MPC strategy significantly outperforms traditional PI-MPC in reducing overshoot, improving transient response, and achieving precise current balance.
1. Introduction
Energy storage batteries are indispensable in modern DC microgrids, compensating for power fluctuations caused by intermittent renewable sources and load variations. Bidirectional DC-DC converters, particularly interleaved topologies, are favored for their low current ripple, high power density, and reduced component stress. However, parasitic mismatches in parallel phases lead to unequal current distribution, accelerating component degradation and compromising system reliability.
Existing current-sharing methods, such as PI-based controls and neural network approaches, face limitations in dynamic response and parameter sensitivity. This work introduces a hybrid SMC-MPC strategy to address these challenges, leveraging the robustness of sliding mode control and the precision of predictive algorithms.
2. Topology and Operating Principles
The three-phase interleaved bidirectional Buck-Boost converter (Fig. 1) operates in two modes:
- Boost Mode: Energy flows from the battery (low voltage) to the DC bus (high voltage).
- Buck Mode: Energy flows from the DC bus to the battery during regenerative braking or grid support.
Key parameters include:
- Input voltage: Uin=10−40V
- Output voltage: Uo=50V
- Inductance: L1=0.82mH,L2=0.80mH,L3=0.78mH
- Switching frequency: fs=20kHz
The converter’s state-space equations for Boost mode are:{iL1(k+1)=iL1(k)+L1(DBoost(k)−1)TUo(k)+L1TUin(k)Uo(k+1)=C2(1−DBoost(k))TiL1(k)+(1−RC2T)Uo(k)
where DBoost is the duty cycle.
3. Proposed SMC-MPC Strategy
3.1 Voltage Outer Loop: Sliding Mode Control
A sliding surface S is designed to regulate the output voltage and inductor currents:S=Δ1(iL,ref−m=1∑3iLm)+Δ2(Uo,ref−Uo)+Δ3∫(Uo,ref−Uo)dt
Solving S=0 yields the reference inductor current:iL,ref=m=1∑3iLm−Δ1Δ2(Uo,ref−Uo)−Δ1Δ3∫(Uo,ref−Uo)dt
3.2 Current Inner Loop: Model Predictive Control
The MPC cost function minimizes tracking errors:Q=[Uo(k+1)−Uo,ref]2+[iL1(k+1)−iL,ref]2
Optimal duty cycles for Boost mode are derived as:DBoost,m(k)=1+C22Uo2(k)TUo(k)iLm(k)(C2Lm2−C22Lm)+C22Uo2(k)TC22LmUo(k)iL,ref−C2Lm2iLm(k)Uo,ref
4. Simulation and Experimental Results
4.1 Steady-State Performance
Table 1 compares current-sharing errors between SMC-MPC and PI-MPC under Boost and Buck modes.
| Mode | Control Strategy | Current Imbalance (%) |
|---|---|---|
| Boost | PI-MPC | 3.56 |
| SMC-MPC | 1.98 | |
| Buck | PI-MPC | 2.69 |
| SMC-MPC | 1.12 |
4.2 Dynamic Response
Load step changes (0.4 Ω ↔ 0.8 Ω in Buck mode, 10 Ω ↔ 20 Ω in Boost mode) validate the transient performance:
| Parameter | PI-MPC | SMC-MPC |
|---|---|---|
| Overshoot (%) | 11.5 | 0 |
| Settling Time | 50 ms | 0.4 ms |
5. Conclusion
The SMC-MPC strategy significantly enhances the performance of bidirectional DC-DC converters in energy storage battery systems:
- Robustness: Immune to parasitic mismatches and load disturbances.
- Precision: Reduces current imbalance to <2% in steady-state.
- Speed: Achieves sub-millisecond settling during transients.
Future work will explore scalability for higher-power applications and integration with advanced battery management systems.