Research On Power Quality Situational Awareness In Power Distribution Station Area

As new technologies drive the development of power system,the electricity load tends to be more and more complex,and the fluctuation situation has strong uncertainty,leading to the normal production and life of high demand and high stability of power quality requirements are increasingly prominent.At the same time,the continuous development and application of power electronics technology,the popularity of new loads and power supply equipment,such as electric vehicle charging piles,new energy generation equipment,etc.also have higher and higher requirements for power quality.Therefore,the research and analysis of power quality in distribution station area can help power system operation and management personnel better understand power quality problems,take corresponding measures and technical means to solve power quality problems,and improve the stability,reliability and safety of power system.In this paper,situational awareness technology is used to solve the problem of multifactorial evaluation and prediction of power quality information in distribution station areas.Firstly,the characteristics of different types of distribution stations and the feasibility of applying the situational awareness model structure to distribution stations are described.Secondly,we establish a situational awareness framework model for power quality in the distribution area,analyze the model structure of the distribution area,and establish a fast simulation phase-volume method model based on the traditional continuous accurate simulation model for analyzing the real-time operation of the distribution area and quickly collecting the power quality information of the area;reflect the overall power quality of the area by the more serious end nodes of the distribution area.The particle swarm optimization algorithm is used to optimize the parameters of the long-short term memory network model to speed up the analysis of the situation and improve the prediction quality.Considering the influence of user behavior at the prediction point,we update the sampling data in real time by sliding window method,so that the input data can be updated in time to correct the prediction results.Finally,based on the evaluation and prediction of traditional single-indicator single monitoring results,the combination weighting method is proposed to make a comprehensive evaluation of power quality from two perspectives: the difference of multi-node power quality information collection results and the different evaluation tendencies among different power quality indicators,and through quantitative analysis,the situation value and situation level of power quality in the distribution station area at the corresponding moment can be reflected.The method of interval prediction by combining quantile and long-short term memory network is proposed in the prediction section,and the feasibility of the scheme is verified by comparing the interval prediction index and probability density function analysis with the calculation example.