Advanced Metering Infrastructure In The Power Distribution System And Data Application Methods Research

In the background of power big data,based on the three layer logic of data acquisition,data cleaning and data application to the distribution system,a theoretical framework and method of Engineering practice is proposed.Advanced Metering Infrastructure is a rounded network and system for electricity information measurement,which can read,storage and analysis to suit users,including the related hardware,communication facilities and software systems.An innovative network technology scheme based on secure and open IPv6 communication protocol deployment on AMI system is proposed.In order to combine the IPv6 protocol with the new smart meter network based on Wireless Sensor Networks,the research concerning about interconnection between smart meter network based on the 6LoWPAN and the IPv6 wired transmission network has carried on.It presents the entire sensor smart meter networks based on 6LoWPAN wireless as a virtual IPv6 network access point ideas,gives specific method wireless sensor network through a root node to communicate with IPv6 wired networks,this makes wireless sensor network based on 6LoWPAN can directly access the IPv6 wired network,and realizes interconnection at any point between the two networks,it provides a strong guarantee for the big data application in power distribution system.Due to the wide distribution,complex grid structure,poor operating environment and other objective conditions,the reliability of communication in power distribution and utilization system is difficult to guarantee.Distribution transformer as a key part in the power distribution and utilization system,there are a lot of anomalies and missing data in the data area of actual operation.The data in the actual operational distribution transformer are used as the study sample in this paper.The reasons of the abnormal data and the missing data are analyzed firstly,then use the algorithm of small sample data for research,a method based on Kernel Smoothing technology is proposed to solve abnormal data,a method based on Pearson correlation coefficients and the regression model is proposed to solve missing data.In order to verify the efficiency of the proposed method,Spark parallel computing architecture is built with 6 server computers.The identification of abnormal data and the recovery of missing data are carried out on the TB level distribution transformer data.The related results prove the practical value of the proposed method.The data collected by Advanced Metering Infrastructure is the basic data in the distribution system,with heterogeneous and large scale characteristics,showing the characteristics of big data.The use of smart meter data,using cluster analysis,correlation analysis and other statistical methods,combined with electrical professional knowledge,the value of data mining for the power industry can solve the practical engineering problems,creating benefits for all stakeholders.Based on the smart meter data,how to correct the relationship between the user and the distribution transformer is studied.It can save the cost of the asset management of distribution system,and an actual data example has been proposed to prove the practical value of this method.