High Performance Robust Predictive Control Of Dual Active Bridge Converters For Fuel Cell Vehicles

Faced with increasingly serious energy and environmental problems,the development of hydrogen energy has become a future trend.Fuel cell vehicle,as the typical application scenario for hydrogen energy utilization,is an important direction for the sustainable development of the automotive industry.Dual active bridge converter,is particularly suitable for fuel cell electrical systems,has received widespread attention due to the advantages of high power density,allowing bi-directional power transfer,and galvanic isolation capability.Based on this,we choose the dual active bridge converter as the core topology.Therefore,fast start-up and transient response capability,high stability and accurate output voltage have become typical requirements.As a result,this paper focuses on typical control requirements of a dual active bridge converter for fuel cell vehicles.Firstly,in order to improve the start-up performance of the dual active bridge converter,we analyze the DC bias in the start-up stage when the converter directly use a closed-loop control strategy,and propose a start-up bias eliminate method which can apply into the frame of phase-shift modulations.Furthermore,we explain the fundamental reason why single-phase-shift modulation can be hardly used for start-up optimization,and propose a fast start-up control method based on extended-phase-shift modulation without current bias.By considering the states in each operating mode of the extended-phase-shift modulation,the optimum phase-shift ratio is obtained for the given current stress constraint.The current of inductor can be always restricted within the allowed output current range,and the output power is maximized throughout start-up stage,increasing start-up speed of the converter as much as possible without DC bias.Simulation and experimental results show that the proposed method can improve the start-up stage performance.Secondly,in order to improve the dynamic response and output quality of the dual active bridge converter,we develop a voltage prediction model based on the power transfer relationship,propose a simplified continuous control set model predictive control method for industrial applications,based on the monotonic relationship between power and phase-shift.Then,we analyses the mechanism of the voltage bias problem and proposes a model predictive control method without voltage bias based on reference voltage compensation.In addition,the transient bias phenomenon when applying predictive control is also analyzed,and the effect of transient bias is reduced by combining the bias suppression method.Simulations and experiments demonstrate that the proposed method has excellent dynamic performance,shows no voltage bias and transient DC bias.Finally,in order to improve the robustness to system parameters and sensor noise of the model predictive control in the dual active bridge converter,we present a sensitivity analysis and propose a robust predictive control strategy based on Kalman filter,which achieved by reconfiguring the input variables.The proposed method not only improves the robustness of the model predictive control but eliminates the current sensor on the output side.Simulation and experimental results demonstrate that the proposed method improves the dynamic performance while reducing the sensitivity to system parameters and noise of model predictive control,and ensures the voltage quality of the converter.