Research On Model Predictive Control Of Three-level Inverters In Microgrids

In the background of achieving carbon peaking and carbon neutrality goals,the green-oriented transition of global energy is overwhelming.Developing and utilizing new energy is an essential strategic approach to realizing deep decarbonization in China,which has great significance in promoting energy innovation and high-quality development.As a fundamental component of the new power system,the microgrid is an important way for the efficient local consumption of distributed new energy generation,a powerful supplement and support for the power grid,and critical technology to facilitate the advancement of the new energy industry.The three-level inverter is the core interface equipment of microgrids due to the advantages of low switching loss,high conversion efficiency,and accessible expansion capacity,which can connect distributed new energy generation units to microgrids flexibly in grid-connected/islanded operation mode.The control performance of inverters directly affects the operation efficiency,stability and reliability of microgrids,which is significant to research.Among all control methods,model predictive control(MPC)is favored by academics and industry because of offering salient advantages,such as fast dynamic response,easy implementation,multi-objective optimization,and nonlinear constraints.However,there are still several crucial unsolved problems to the application of the three-level inverter MPC in microgrids,which seriously restrict the capability of new energy consumption.To this end,this paper takes the T-type three-level inverter as the research object to improve the control performance.From the perspectives of modeling and prediction,cost function,multi-objective constraints,vector selection,and optimization methods,this paper focuses on crucial control technologies such as CMV reduction and current distortion suppression,voltage quality improvement and observer design,voltage frequency adaptive regulation,circulating current suppression and control accuracy improvement.It satisfies the performance requirements of three-level that high reliability,high accuracy,adaptive regulation,low harmonics and less cost.The main contributions and innovations of this paper as follows:In order to solve the problems of high CMV and current distortion in the conventional MPC,the CMV-reduced double-vector MPC is proposed,which restricts the CMV within low magnitude by removing the high-CMV candidate vectors.In the double-vector MPC scheme,the evaluation process of the double vector is simplified by the cost function based on the midline distance and the vector preselection algorithm.Then,by optimizing the weightless factor strategy and the candidate set compensation method,the inverter suppresses the grid-current distortion while maintaining low CMV and neutral point(NP)balance,which improves the safety and reliability of microgrids.In order to solve the problems of heavily relying on sensors and poor voltage quality in the conventional model predictive voltage control method,an improved double-vector model predictive voltage control method with current sensorless is proposed to gain superior voltage performance.It simplified the mathematical model and control framework,and the inverter voltage increment is induced in the multi-objectives cost function,which optimizes the output voltage control performance.To estimate the capacitor current of the simplified discrete model,the strong tracking Kalman filter(STKF)algorithm-based observer is designed with output voltage feedback only,lessening hardware cost.In addition,the two-layer decision algorithm based on spherical decoding principle effectively reduces the complexity and computational burden of the double-vector selection process.In order to solve the problems of poor frequency dynamic response and weak power support for the abruptly varied load in islanded microgrids,a double-loop control strategy with cascaded MPC is proposed to improve frequency regulation,which includes inner-loop control(model predictive voltage control method)and outer-loop control.In the outer-loop,the MPC-based virtual synchronous generator(VSG)method is proposed to enhance dynamic frequency characteristics.Based on the dynamic load requirements analysis,the optimization objectives are constructed under different load switching scenarios,where the corresponding frequency response process speeded up or slowed down appropriately.Furthermore,based on the Lyapunov criterion,the stability of the proposed method is analyzed to guide the design of the critical parameters.In order to improve the performance in the current circulating suppression and current harmonics reduction of the conventional MPC method for paralleled three-level inverters system,a multivector MPC method is proposed to suppress the circulating current,where the control freedom is extended by constructing a virtual vector with zero-average CMV,enhancing the control capability.The small vector is introduced to the virtual vector to enhance NP balance capability without weight factors.Moreover,the multi-objective cost function is simplified as the univariate cost function since the candidate vector selection is optimized.Thus,it has the advantages of low computational burden,easy implementation,and no communication between each inverter.In summary,this paper proposed a series of MPC contributions in the three-level inverter,contributing to the reliable,efficient,high-quality operation and modular application for microgrids.It provides theoretical and technical support for promoting the development of microgrid technology,helps to implement the national energy conservation and emission reduction work,and helps China to promote deep decarbonization technology research.