Modeling And Control Of Boost Converter Based On Mixed Logical Dynamical Model

Due to the small size,light weight and reliable performance of DC-DC converters,they are extensively used for industrial applications.Because of the different states of the switching devices,the working process of converters exhibits switching characteristic,which is a constituent of a typical hybrid system.Such system is non-linear in nature,but linear systems theory have been traditionally used for the modeling and analysis of the converters,which causes that many properties of the converter have not been fully understood.The methods using hybrid system theory to analyze and control the converter do not require any approximation,and directly reflect the nature of the working principle of the system.So hybrid system theory can solve some problems that traditional methods faced with in analyzing and control power electronic converters.In this paper we focus on Boost converter modeling and control using hybrid system theory.First,we establish a simplified mixed logical dynamical(MLD)model for Boost converter without approximate or linearization.Based on this model,we use model predictive control(MPC)method to control the system,and convert the control problem into a finite constraint optimal control problem,then get the optimal switching signal to regulate the output voltage by solving the problem online.Taking into account the real-time requirements of the controller,a explicit form of affine state feedback control law is derived offline and used as a look-up table to facilitate the real time implementation.The controller guarantees the optimal performance according to constrained items in the cost function of MPC.These features are impossible for the conventional controller designed using the small signal linearized model and the linear system theory.To deal with the degradation caused by the uncertain parameters,a PI compensator is proposed to adjust the current reference to reduce steady state error of the output voltage,and the PI parameters are optimized using Genetic Algorithm.Under parameters perturbation,our proposed method achieves very nice performance in the sense of no steady state error,fast recovey time,and consistent transient response mode.The proposed method has been validated by both simulation and prototype experiment using MC9S12DG128 micro-controller.Comparison results of both simulation and experiment show the effectiveness and superiority of our proposed scheme.