Dual-mode Predictive Control Of DC/DC Converter Based On State Estimation

DC/DC converter is the main component of switching power supply and the core of power conversion,which has the advantages of low cost,high accuracy,good stability and wide range of applications.Robust control studies for DC/DC converters have been reported extensively in the literature.In recent years,with the rapid development of electronic technology and control theory,the multi-constrained optimal control problem for DC/DC converters has become a research hotspot.However,it is noteworthy that most of the current studies use model predictive control methods such as minimum-maximization and explicit optimization for the constrained regulation of output voltage.And these methods generally suffer from the problems of excessive online computation or storage space occupation.Recently,a dual-mode predictive control method based on set theoretic has been proposed,which effectively balances the requirements of online computing burden,storage space,performance optimization and constraint satisfaction.Therefore,many scholars have extended this method to deal with the optimal control problems of different systems.Inspired by these studies,this thesis aims to apply the dual-mode predictive control method to solve the constrained optimal voltage regulation problem for different DC/DC converters.In addition,this paper further integrates the unknown input estimation theory to give an observer-based dual-mode predictive control scheme.The experiments are also given to confirm the feasibility of such method.The main work is as follows:(1)A state observer based on unknown input estimation is designed for the Buck converter.Specifically,the topology of the Buck converter is firstly analyzed,and a linear parameter varying(LPV)small signal model is developed by considering the uncertainty of the model parameters.Secondly,a reduced-order unknown input observer is constructed,and the system matrix parameterization method of the observer is given to make its design more general.Finally,the designed observer is applied to a Buck converter.(2)A dual-mode model predictive controller with unknown input estimation is designed for a Boost converter.Firstly,the topology of the Boost converter is analyzed,the LPV small-signal model is established,and the model is discretized for the subsequent implementation of the predictive control.Second,a dual-mode predictive control method based on polyhedral invariant sets is presented according to the control requirements of the system.The maximum controllable invariant set and the optimal control regulation algorithm are designed by defining suitable performance metrics and establishing ensemble constraints.(3)The Buck-Boost converter is used as a research object to build a real circuit to verify the effectiveness of the state observer and dual-mode predictive controller.First,the topology of the Buck-Boost converter is analyzed,and the LPV small-signal discrete model is established.Second,the state observer and dual-mode predictive controller are designed according to the design methods proposed in the previous two chapters.Finally,a Buck-Boost converter is built with actual electrical components,and an Arduino UNO card is used as the I/O interface with MATLAB Simulink to control the actual circuit.The effectiveness of the proposed unknown input estimator-based dual-mode predictive controller is finally validated.