Research On Power Quality Disturbance Classification Method For AC/DC Hybrid Microgrid

With the rapid development of distributed energy,microgrids,as one of the technical means to solve the problem of large-scale access of distributed power sources to the grid,are an important form of future power systems,which have significant implications for the development of China’s energy strategy.Microgrid systems are usually composed of various new energy sources(such as wind power,solar power,etc.)and loads,and the mixed use of AC/DC hybrid microgrid is becoming increasingly common.The appearance of AC/DC hybrid microgrid systems improves the energy utilization efficiency,but as the penetration rate of new energy sources increases and a large number of power electronic components are connected,the frequency of power quality disturbances significantly increases.Therefore,it is of great theoretical and practical significance to conduct research on the classification of power quality disturbances in AC/DC hybrid microgrid systems.This paper analyzes and studies the formation mechanism and disturbance recognition methods of AC and DC power quality problems,and the main work is as follows:(1)A simulation model of an AC/DC hybrid microgrid was constructed,and the working principles and relevant characteristics of the photovoltaic power generation system and energy storage system within the hybrid microgrid were analyzed.Detailed analysis was conducted on the voltage fluctuation and harmonic mechanism caused by the photovoltaic power generation system interfacing with the microgrid,as well as the harmonic interaction principle between the microgrid and the main grid caused by load changes during microgrid operation.(2)The AC/DC power quality issues in the hybrid microgrid were analyzed.A known model for generating AC power quality disturbances with random elements was used to achieve simulation data similar to actual data.As the definition of DC power quality issues and national standards have not yet been unified,the DC voltage fluctuation,DC voltage ripple,DC voltage temporary rise,and DC voltage temporary drop under three typical operating conditions in the hybrid AC/DC microgrid were defined,and their formation mechanisms were analyzed.Using the simulation model of the AC/DC hybrid microgrid,the impacts of different operating conditions on the DC power quality in the hybrid microgrid were studied.(3)Aiming at the problems of artificial feature extraction for AC power quality disturbances,low accuracy of composite disturbance recognition,and long training time,a sparrow algorithm-optimized deep extreme learning machine(DELM)method for AC power quality disturbance classification is proposed.Firstly,the basic principles and characteristics of extreme learning machine(ELM)and autoencoder(AE)are explained,and these two methods are combined to construct an ELM-AE model and analyze its advantages in the classification field.Secondly,a deep ELM model is constructed by using the feature representation ability of ELM-AE.Then,the sparrow optimization algorithm(SSA)is combined with DELM to achieve the optimal input weight of DELM,thereby improving the robustness of the network.Finally,through experiments on mathematical model simulation data and working condition simulation data of AC power quality disturbances,the effectiveness of the proposed method is verified.(4)This study investigates the classification of direct current DC power quality disturbances in hybrid AC/DC microgrids and proposes a DC power quality disturbance classification method based on multidimensional feature analysis and decision tree.Six typical operating scenarios of hybrid AC/DC microgrids are analyzed for their DC power quality disturbance signals.Based on their time-domain and frequency-domain characteristics,as well as graphical features,five feature quantities are extracted as the basis for identification and classification of DC power quality disturbances.A CART decision tree classification method is constructed to achieve fast and accurate identification of DC power quality disturbance signals.The accuracy and effectiveness of the proposed method are verified through experimental simulation.