Research On The Maximum Power Point Tracking Algorithm For Photovoltaic System Of Multi-array

With the depletion of fossil energy,solar energy as an ideal renewable energy has important research value.Currently,photovoltaic power generation system as a generation system that can convert solar energy into electrical energy,is the most promising way to produce electricity.Maximum Power Point Tracking(MPPT)technology is a key technology in photovoltaic power generation system,and it plays an important role to improve the energy conversion efficiency of the system.However,under the complex and changeable external environment,PV array often worked under partial shaded conditions,in this case,the conventional MPPT algorithms are difficult to achieve good control effect.For this reason,in order to improve the efficiency of solar energy,taking multi-array photovoltaic system as the research object,the paper proposes a group MPPT control algorithmfor a variety of illumination state,which contains the parallel MPPT control and independent MPPT control.In order to solve the problem of partial shaded in photovoltaic system,the particle swarm optimization(PSO)algorithm often used in MPPT control of a single array,however,this method has a problem of slowing track speed and big system oscillation.To solve the problems,based on the similar output characteristics betweenthe different arrays,the paper proposes aparallel MPPT control methodbased on improved PSO algorithm.The method takes each array in the system as a particle in the PSO algorithm population.Based on the tracking characteristics of PSO algorithm,every arraycan obtain more effective information by using the interactive information from others,and to improving the speed of MPPT.The paper combined with the characteristics of the parallel MPPT control,and makes some adaptation improvement of the PSO algorithm's inertia weight,acceleration factor and other parameters,to reducing the oscillation of the control process in parallel MPPT.Parallel MPPT control also provide useful information for subsequent independent MPPT control of the arrays.For the array under uniform illumination,its sub controller takes personal best position corresponding of the array as the initial reference voltage directly,and use the modified variable step incremental conductance(MR-INC)forthe independent MPPT control;For the partial shaded arrays,its sub controller takes personal best position corresponding to the array as the global best position in the PSO algorithm,and use the improved PSO algorithm combined with MR-INC to restart the independent MPPT control.In order to improve the efficiency of independent MPPT control,it is necessary to identify the illumination state of the array accurately.Therefore,the paper presents anillumination state recognition algorithm based on SOM neural network.Byclustering of each array's individual optimal value at every sampling time in parallel MPPT control process,the algorithm obtains the illumination state of each array.The paper establisha simulation model of the system in MATLAB / Simulink,and take comparative experiment for MR-INC,PSO and the proposed group MPPT algorithm under four different illumination state.Experimental results show that compared with the other control methods,proposed MPPT algorithm of the papers can solve the multi–peak optimization problem under partial shaded conditioneffectively,improve the maximum power point tracking speed of system,and reduce the oscillation of the system produced in the process of tracking.