基于事件触发的光储双向变换器有限控制集模型预测控制

Event-triggered finite control set model predictive control for bidirectional converter of PV and battery storage system

  • 摘要: 为了解决光储系统中Buck/Boost双向DC-DC变换器在采用有限控制集模型预测控制(Model Predictive Control with Finite Control Set,FCS-MPC)时时间成本较大的问题,研究了一种基于事件触发的有限控制集模型预测控制(Event-Triggered Finite Control Set Model Predictive Control,ETFCS-MPC)方法。FCS-MPC的策略在每个控制周期内求解在线优化问题,增加了大量的计算时间以及开关损耗。通过事件触发控制机制,一旦误差范数大于设定上限,系统便会自动触发FCS-MPC的运作,不仅减少了冗余的优化操作,而且确保了目标系统状态与其参考状态之间的误差被系统维持在预设上限内的性能水平,不会因事件触发控制机制对FCS-MPC的触发有所牺牲。在MATLAB/Simulink中搭建仿真模型,验证了该方法在保持满意调节性能的前提下,减少了计算量,降低了时间成本。

     

    Abstract: In order to solve the problem of high time cost faced by the Buck/Boost bidirectional DC-DC converter in the optical storage system when using the Model Predictive Control with Finite Control Set(FCS-MPC), this text discusses an Event-Triggered Finite Control Set Model Predictive Control(ET-FCS-MPC) way. Generally,the online optimization problem is cleared up in each control cycle,to achieve the control scheme of FCS-MPC.However, online optimization solution will increase a large amount of calculation time and switching losses. FCSMPC is activated by the system solely when the error norm surpasses a predefined threshold, utilizing eventtriggered control. This superfluous optimization operation can be reduced without sacrificing the performance of keeping the deviation between the target system’s state and its reference within a preset threshold. A simulation model was ideated in MATLAB/Simulink, which certificated that this method decreased a great deal of calculation and time cost while maintaining satisfactory adjustment performance.

     

/

返回文章
返回