Research On The Control Of Wireless Parallel-Connected Three-Phase Inverters In Microgrid

The parallel-connected inverter technology is an important way to expand inverter power supply range.There is no communication between each inverter module in the system of wireless parallel-connected inverters,which reduces the complexity of the system,and makes the reliability and redundancy higher.But the difference of circuit parameters between each parallel-connected inverter causes internal circulation and makes load power can not be shared during the operation of parallel system,therefore,aiming at sloving these problems,this article mainly works for the parallel-connected inverter technology in microgrid.First of all,the mathematical model of three-phase voltage source inverter is established in different coordinate system,then the inner loop control strategies of the parallel syatem are studied respectively.Meanwhile,the model predictive control method is introduced.According to the operation of two parallel inverters,the main reasons which caused circulation between parallel inverters are analyzed.Based on the illustration of the power transmission principle,the traditional droop control method of wireless parallel inverters is deduced,which is affected seriously by the line impedance.As a result,load power cannot be divided averagely,especially for the reactive power.Therefore,this paper proposes a Q-V'droop control strategy based on the changing rate of the voltage,and the output reactive power is changed with the output voltage amplitude until that it is not affected by line impedance,which achieves the average accuracy of the load power between parallel system.Secondly,the droop control and the model predictive control of the three-phase inverter are combined to optimize the parallel system performance.And the reference voltage signal of the MPC controler is provided by the improved droop controler.Then the value function is established,and the optimal switching combination of inverters is determined by calculating the minimum value of the value function.The proposed method avoids the PI controllers and PWM modulators,it can not only reflect the good dynamic characteristics of the model predictive control,but also realize the scalability and redundancy of parallel system.In addition,it can also meet the requirements of the tracking error and the minimum circulation,and the accurate distribution of the load power is achieved.Finally,the parallel system model of two inverters is built on the MTALAB/Simulink platform to verify the correctness and practicability of the control algorithm proposed in this paper.Firstly,the characteristic of load power sharing under the Q-V' droop control is tested in different line impedance and different inverter capacity,and then the Q-V' droop control and predictive control for the parallel system is checked by simulation.Experimental platform of the parallel system is established with the control chip DSPTMS320f2812 to verify the validity of the proposed control method.