Research On MC-PMSM System Based On RBF Neural Network

With the continuous exploit of rare earths,the continuous development of permanent magnetic materials,the continuous improvement of power electronic technology and motor control technology and the continuous optimization of vector control technology,permanent magnet synchronous motors are widely used in military,industrial and civil fields.Matrix converter stems from the advantages of its excellent input characteristics,output characteristics and omitted the direct current capacitors,favored in alternating current speed control system.Matrix converter permanent magnet synchronous motor system based on double space vector modulation has become the research focus of researchers all over the world.Firstly,the principle of vector control of the permanent magnet synchronous motors and mathematical derivation of coordinate transformation are introduced.The topology and working principle of the matrix converter are analyzed.The AC-AC matrix converter is equivalent to two parts:virtual rectifier and virtual inverter.A dual space vector modulation algorithm is deduced,nine-segment pulse width modulation strategy and semi-soften four-step commutation strategy are used for complete the control of the system.Secondly,due to the unbalanced input voltage on the network side,the input and output current waveforms are seriously distorted.This paper analyzes the relationship between the input voltage and the intermediate virtual DC voltage,and deduces the expression of the intermediate virtual DC voltage when the network voltage unbalanced input.A new modulation method based on the combination of virtual DC voltage regulation and dual space vector control strategy is presented.Through real-time monitoring of grid input voltage,The method of compensating the intermediate virtual DC voltage with preconditioning parameters is used to maintain the same and achieve good output characteristics.Thirdly,due to the input-output relationship of the system,the factors that influence the power factor of the network side to be overturned include the parameters of the input RLC filter and the output frequency,in this paper,a RBF neural network algorithm is proposed to predict the power factor of the matrix converter PMSM.The preset value is combined with the dual space vector modulation algorithm to achieve the goal of increasing the power factor of the network side.Fourthly,We Simulate system based on double space vector modulation technology by using MATLAB/Simulink software.The input voltage and current waveforms before and after regulation are compared,which proves the authenticity of the modulation method.Finally,ARM + FPGA is chosen as the control structure of matrix converter permanent magnet synchronous motor,the hardware circuit is designed and the software is programmed,and the experimental platform is set up.Through the data collection and waveform analysis of the real thing,it is proved that this method is reasonable.