| With the rapid development of energy meter technology, smart meters in the power system have been widely used. According to the statistics, before 2015, the smart meters coverage will reach more than 80%, and the amount of users will be more than 140 million in 2020 and even cover 100%. Compared with the traditional meters, power measurement level is increasing, it has good accuracy and accuracy of high precision and high advantages. However, in practice, there will still be some failures, not only will affect the correctness of the power meter,but also may endanger the equipment and personal safety. So the analysis in the causes of meter failure and prediction of meter failure is of great significance.In this paper, Bayesian network is used to predict and analyze the fault types of smart meters. Based on the Bayesian network, a classification and prediction model of smart meter fault types is proposed,and various kinds of smart meter faults are studied. We aimed at the fault factors to make predictions and analyses. This article completes the following work:First, we study Bayesian network-related knowledge and technology,and make a brief introduction to it.Secondly, introducing and comparing the two bayesian algorithm in this systemi,and analyzing their advantages and disadvantages, making an analysis of the actual problem, and finding related factors of building bayesian networks. Based on the large data platform, the design and implementation of the bayesian network build subsystem based on the large data platform with the Mapreduce programming framework.Finally through large data , using bayesian network constructs a subsystem,which can provide a bayesian network structure and have the conditional probability table to complete the forecast subsystem .In this paper, we propose a system based on the large data platform to deal with the fault data, and obtain a complete system of intelligent meter fault type prediction. |
PDF Full Text Request