Research On Intelligent Fault Diagnosis Of Internal Transmission Line In Photovoltaic And Energy-storage Hybrid Microgrid

Detected accurately and excised rapidly the faults of transmission lines in micro grid has significant impact on shortening the time of failure,reducing the economic loss of users,and ensuring micro grid system safety and reliability.The intelligent fault diagnosis for internal transmission lines in microgrid is studied to improve the security and reliability of microgrid system.Wavelet packet analysis and neural network are applied to fault diagnosis of microgrid to complete the identification of the fault type and fault phase,and ensure the accuracy and reliability of the diagnosis results,and make a theoretical cushion to achieve intelligent fault diagnosis of internal transmission lines based on wavelet packet analysis and neural network in microgrid by hardware in the future.The specific work is as followed:First,the micro grid system which can relate to the grid stably is built on the Matlab/Simulink simulation platform.Owing to the wavelet packet analysis technique are used detect and display the transient non-stationary signals in normal signals,three-layer wavelet packet decomposition for three-phase output voltage signals of microgrid is proposed to detect whether the internal transmission line of microgrid has fault,and regard the wavelet packet energy value of the reconstructed voltage as the fault characteristic of the internal transmission lines in microgrid.And the different characteristics of curve of the wavelet packet voltage reconstruction signal and energy spectrum of fault phase and non-fault phase are pointed out.Secondly,the BP and RBF neural network are proposed and constructed for the fault diagnosis model of internal transmission line in microgrid based on the vector of wavelet packet energy,the performance of two fault diagnosis models are analyzed and compared.The results show that the mode of fault diagnosis of internal transmission line based on RBF is better than BP in training curve iteration speed and the correct diagnosis rate.Finally,because the model of fault diagnosis of internal transmission line based on BP neural network in microgrid has many disadvantages,such as slow convergence speed,long iterations,and cannot be effectively divided into two phases short-circuit fault and two phases grounding short-circuit fault.An adaptive genetic algorithm is proposed to optimize the initial weights and thresholds of the BP neural network fault diagnosis model to ensure the accuracy and reliability of the diagnosis results.The simulation results show that outcomes of the fault diagnosis model based on AGA-BPNN which has already made up the shortcomings and defects BP and GA-BPNN completes fault detection and the identification of the fault type and fault phase effectively.What's more,the results have high confidence level and its performance is also closed to RBF.