Boundary Control Of Power Electronic Converters For DC Power Supply And Its Digital Implementation

In recent years,the DC power supply technology based on the power electronic converter has been applied in many fields,and the power quality problem of DC power system is increasingly prominent.To satisfy the power quality requirements,it is necessary to optimize the control strategy of power electronic converters to improve the dynamic performance.Boundary control is a kind of nonlinear control strategy.Compared with the traditional linear control,boundary control adopts the design method of state trajectory planning,which can provide a superior,or even close to the theoretical optimal dynamic performance,so it has been widely concerned.In this thesis,three typical power converters for DC power supply are considered,and the corresponding boundary control strategies are studied systematically and deeply.There are two limitations in the existing researches of boundary control.First,all the existing researches are based on the ideal mathematical model.The influence of non-ideal factors such as parametric variations lacks systematic analysis,and the robustness of boundary control lacks in-depth research.Second,from the perspective of digital implementation,the traditional boundary control is derived based on continuous-time system model,so it needs a continuous sampling of state variables,which means that an infinite sampling frequency is needed and the algorithm is not suitable for the digital controller with limited sampling frequency.With the wide application of digital control technology in the power electronics field,studying the boundary control adapted to the limited sampling frequency has an important theoretical and engineering significance.Based on these two problems,the research work of this thesis is as follows.1)A general expression of the high-order switching surface for Buck converter,which can provide a near time-optimal dynamic response,is derived.The influence of two kinds of non-ideal factors,parametric variations and parasitic parameters,on the sliding-mode region of boundary control is researched,and some points for attention when determining the coefficients of switching surface are proposed,to avoid the output voltage overshoot caused by the departure from the sliding-mode region and ensure the robustness of the boundary control.2)For Buck converter,by using the idea of traditional quasi-sliding-mode control and combining the discrete reaching law method,a PWM-based discrete-time second-order boundary control is proposed.The algorithm can be regarded as a discrete-time version of the existing boundary control with second-order switching surface.Furthermore,the sensitivity of the proposed control to parametric variations and parasitic parameters is discussed,and the current-limiting method of the inductor current is also discussed.Compared with the discrete-time first-order boundary control,the proposed discrete-time second-order boundary control has a better dynamic performance.Compared with the continuous-time version of this boundary control,the proposed control not only provides a dynamic response close to the time-optimal one,but also avoids the disadvantages of the traditional boundary control,such as variable switching frequency and high sampling frequency requirements.3)For Boost converter,considering that its circuit model is different from the buck converter,the bilinear system model of boost converter is transformed into a global linearized model by using the exact feedback linearization method.Based on the linearized model,a boundary control with second-order switching surface,which can provide near time-optimal dynamic performance,is derived.By using the idea of deriving the discrete-time boundary control of buck converter and combining the discrete reaching law method,a discrete-time second-order boundary control suitable for boost converter is proposed.Moreover,the influence of parametric variations and parasitic parameters is discussed,and the current-limiting control method is also discussed.Compared with the traditional linear control,the proposed discrete-time second-order boundary control has a significant superiority in dynamic performance.4)The idea of discrete-time boundary control is further extended to single-phase PWM rectifier with LCL filter.A double closed-loop control for single-phase PWM rectifier is proposed,which takes the discrete-time boundary control as the voltage outer loop and the improved deadbeat control as the current inner loop.For the current inner loop,the traditional deadbeat control,with the converter side current as the controlled variable,cannot realize the unit power factor on the grid side accurately.Therefore,an improved deadbeat algorithm with the weighted sum of the grid side current and the converter side current as the controlled variable is proposed,and the current reference value required to achieve the accurate unit power factor is given.For the voltage outer loop,the mathematical relationship between the peak value of half wave current and the voltage increment of DC side in the half wave period is derived.Based on that,a voltage outer loop control strategy derived by discrete reaching law method is proposed.The proposed double closed-loop control not only achieves the accurate unit power factor at the grid side,but also greatly improves the dynamic performance of the output voltage.In addition,the current overshoot on the grid side caused by the under-damped oscillation is discussed,and the method for eliminating this overshoot is given.For the above research work,the low-power experimental prototypes are built respectively.The correctness of the theoretical analyses and the effectiveness of the proposed control strategies are all verified by the experimental results.