基于自适应超螺旋滑模观测器的Buck变换器无模型预测控制

Model-free Predictive Control of Buck Converter Based on Adaptive Super-twisting Sliding Mode Observer

针对三相交错并联Buck变换器在外部干扰下影响系统鲁棒性和动态性能的问题,本文提出一种基于自适应超螺旋滑模观测器的无模型预测控制策略。首先,建立超局部模型代替原有的数学模型,设计自适应超螺旋滑模观测器估计超局部模型中的动态部分。然后,引入最小二乘法预测电流误差趋势,动态调整观测器增益矩阵。最后,构建离散方程,设计代价函数,实现无模型预测控制。实验结果验证了所提算法可以有效抑制因外部扰动产生的稳态误差,且相较于传统无模型预测控制具有更好的鲁棒性和动态性能。

Aimed at the problem that a three-phase interleaved Buck converter will affect the system robustness and dy⁃namic performance under external disturbances,a model-free predictive control strategy based on adaptive super-twisting sliding mode observer model-free predictive control(ASTSMO-MFPC)is proposed.First,an ultra-local model of Buck converter is established to replace the original mathematical model,and an ASTSMO is designed to estimate the dynam⁃ic part of the ultra-local model.Second,the least squares method is introduced to predict the current error trend,so as to dynamically adjust the observer gain matrix.Finally,a discrete equation is constructed,and a cost function is designed to achieve MFPC.Experimental results indicate that the proposed algorithm can effectively suppress the steady-state er⁃rors caused by external disturbances,and it has better robustness and dynamic performance than the traditional MFPC.