Intelligent Photovoltaic Grid-connected Inverter Based On DSP

Grid-connected photovoltaic power generation is the typical application of solar energy,whose related technologies are the current research hotspot.In order to solve the shortcomings of photovoltaic technology,such as various faults in PV array under operation can not be effectively detected and the maximum power point tracking(MPPT)algorithm may have poor performance under non-uniform illumination,a fault diagnosis algorithm based on probabilistic neural networks(PNN)and a MPPT algorithm based on pollen propagation algorithm(FPA)are proposed in this thesis,and then an inverter prototype is developed by using TMS320F28335 as the control core.Firstly,the topology as well as key algorithms of PV grid-connected inverter are analyzed,in which the MPPT algorithm and the fault diagnosis method are mainly studied.Then,a simulation model is established to test the designed system,mainly focus on the correctness of parameters for the inverter and the effectiveness of the proposed algorithms.Finally,the hardware and software of the system is developed and tested,respectively.The system consists of hardware circuit,control software in DSP and applications for fault diagnosis on computer.Firstly,the hardware circuit is designed by adopting two-stage non-isolated BOOST topology.The auxiliary circuit include power supply circuit,drive circuit,and control circuit;Secondly,the control software in DSP consists of MPPT algorithm,grid connected control algorithm,islanding detection algorithm,and AD sampling module.The MPPT algorithm is based on FPA,whose control parameters are set properly according to a large number of simulation studies;The voltage and current double closed loop algorithm is used for the grid-connected control,the PI controller parameters are set by theoretical analysis and simulation;Active frequency drift method is applied for islanding detection,which can be implemented based on the grid-connected control algorithm.Finally,probabilistic neural networks is employed for faults diagnosis of PV array.The fault characteristics of PV array are analysed to determine the input variables,then a mass of data are collected for train and test the faults diagnosis model.In addition,the model is developed based on MATLAB.Based on the above mentioned theoretical analysis and simulation,the PV grid-connected inverter prototype is developed,According to the test,the system achieves the anticipated target.The output power of the inverter prototype reaches 1200W,the total efficiency is over 90%,the power factor of inverter is more than 0.99,and the THD is less than 4%.Moreover,the MPPT efficiency is above 99.5%,and the convergence time and the vibration output are significantly improved compared with the existing algorithms.For the islanding detection,the system can cut off the connection with the grid within 0.2 s.In addition,the accuracy of PV array fault diagnosis is above 95%.Consequently,this study can provide valuable reference for other researches on PV grid-connected inverter.