Research On Model Predictive Current Control Of Open-Winding Five-Phase Permanent Magnet Synchronous Motor

Five-phase permanent magnet synchronous motor(PMSM)drives have been widely studied in many fields such as vehicles,ships and aerospace,due to the advantage of high power density,low torque pulsation and strong fault tolerance.Open-winding structure has lower voltage harmonic content,better fault tolerance and higher degree of freedom compared with the traditional motor stator winding star-connected structure.In addition,the dual inverter power supply can improve the DC bus voltage utilization and doesn’t change the structure of the motor,so combining it with multiphase motors will have more significant advantages in high-voltage and high-power applications.Model predictive control(MPC)is a high-performance control strategy with simple principles,fast dynamic response,and the ability to handle multivariable and complex constraints,which has been widely studied in the field of motor drives.In this paper,the open-winding five-phase PMSM(OWFP-PMSM)is taken as the research object,and the MPC algorithms are detailedly studied.Firstly,the mathematical models of OWFP-PMSM in natural and synchronous rotating coordinate systems is established.The mechanism of zero-sequence current generation in the common DC bus open-winding topology is analyzed.The space voltage vectors distribution characteristics of the single-inverter and double-inverter topologies are also analyzed.Secondly,the OWFP-PMSM prediction models of finite control set MPC and deadbeat MPC are derived.The impact of computational delay on the MPC algorithm is analyzed,and the corresponding delay compensation method is given.A simplified FCS-MPC algorithm based on the virtual voltage vector is studied in this paper.The third harmonic voltage is eliminated by constructing virtual voltage vectors.And a method to quickly screen the optimal voltage vector is given based on the DB-MPC principle,which reduces the number of candidate vectors and reduces the computational effort.A FCS-MPC algorithm based on double vectors synthesis is proposed for the problem that the single-vector FCS-MPC algorithm cannot suppress the zero-sequence current.The algorithm selects two optimal voltage vectors to synthesize according to a certain duty ratio,so as to eliminate the zero-sequence loop voltage and suppress the zero sequence current.Thirdly,a FCS-MPC algorithm based on optimal duty ratio is proposed in this paper.The algorithm adds on the basis of double vectors a voltage vector for regulating the output voltage amplitude.The duty ratio is introduced into the current prediction model to calculate the optimal duty ratio to minimize the evaluation function,which can suppress the zero sequence current and improve the control accuracy.All the above algorithms are complicated to implement,so a low complexity FCS-MPC algorithm based on single inverter vectors is proposed.The algorithm selects the optimal voltage vectors and calculates the duty ratio in two single inverter spaces.The complexity of algorithm implementation and the storage space occupied by algorithm program statements are reduced by modulating the inverter pulse signals separately.Finally,a low-power experimental platform of OWFP-PMSM drive system with DSP and FPGA as control chips is built.An experimental comparison study of the propoesd algorithms is conducted and the correctness and effectiveness of the proposed algorithms are verified.